STUDIA TRADITIONIS THEOLOGIAE

Explorations in Early and Medieval Theology

Theology continually engages with its past: the people, experience,

Scriptures, liturgy, learning, and customs of Christians. The past is pre-
served, rejected, modified; but the legacy steadily evolves as Christians
are never indifferent to history. Even when engaging the future, theol-
ogy looks backwards: the next generation’s training includes inheriting a
canon of Scripture, doctrine, and controversy; while adapting the past is
central in every confrontation with a modernity.

This is the dynamic realm of tradition, and this series’s focus. Whether
examining people, texts, or periods, its volumes are concerned with how
the past evolved in the past, and the interplay of theology, culture, and
tradition.
STUDIA TRADITIONIS THEOLOGIAE
Explorations in Early and Medieval Theology
26

Series Editor: Thomas O’Loughlin,

Professor of Historical Theology
in the University of Nottingham

EDITORIAL BOARD

Director
Prof. Thomas O'Loughlin

Board Members
Dr Andreas Andreopoulos, Dr Nicholas Baker-Brian,
Dr Augustine Casiday, Dr Mary B. Cunningham, Dr Juliette Day,
Dr Johannes Hoff, Dr Paul Middleton, Dr Simon Oliver,
Prof. Andrew Prescott, Dr Patricia Rumsey, Dr Jonathan Wooding,
Dr Holger Zellentin
 LATE ANTIQUE
CALENDRICAL THOUGHT
AND ITS RECEPTION IN
THE EARLY MIDDLE AGES

Proceedings of the 3rd International

Conference on the Science of
Computus in Ireland and Europe
Galway, 16–18 July, 2010

Edited by
Immo Warntjes
&
Dáibhí Ó Cróinín

H
F
Cover illustration: Tabula Peutingeriana ©  ÖNB Vienna: Cod. 324, Segm.
VIII + IX

© 2017, Brepols Publishers n.v., Turnhout, Belgium

All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording, or otherwise, without the prior permission
of the publisher.

D/2017/0095/215
ISBN 978-2-503-57709-8
e-ISBN 978-2-503-57710-4
DOI 10.1484/M.STT-EB.5.113899

Printed on acid-free paper

To the memory of Luciana Cuppo
 TABLE OF CONTENTS

Abbreviations ix
Foreword xi
IMMO WARNTJES, Introduction: state of research on late an-
tique and early medieval computus 1
ALDEN A. MOSSHAMMER, Towards a new edition of the Com-
putus of AD 243 43
JAN ZUIDHOEK, The initial year of De ratione paschali and the
relevance of its paschal dates 71
DANIEL MC CARTHY, The paschal cycle of St Patrick 94
LUCIANA CUPPO, Felix of Squillace and the Dionysiac compu-
tus II: Rome, Gaul, and the insular world 138
BRIGITTE ENGLISCH, Osterfest und Weltchronistik in den
westgotischen Reichen 182
DAVID HOWLETT, An addition to the Hiberno-Latin canon:
De ratione temporum 212
MARINA SMYTH, Once in four: the leap year in early medieval
thought 229
C. PHILIPP E. NOTHAFT, Chronologically confused: Claudius of
Turin and the date of Christ’s passion 265
LISA CHEN OBRIST, What a difference a day makes: the eighth
day of the week in book 10 of Hrabanus Maurus’ De rerum
naturis 293
DÁIBHÍ Ó CRÓINÍN, Archbishop James Ussher (1581–1656)
and the history of the Easter controversy 309
Bibliography 352
index 375


 ABBREVIATIONS

CCCM Corpus Christianorum, Continuatio Mediaevalis

CCSL Corpus Christianorum, Series Latina
CLA Codices Latini Antiquiores
CSEL Corpus Scriptorum Ecclesiasticorum Latinorum
GCS  Die griechischen christlichen Schriftsteller der ersten
Jahrhunderte
LM Lexikon des Mittelalters
MGH Monumenta Germaniae Historica
Auct. ant. Auctores antiquissimi
Epp. Epistolae (in Quart)
LL nat. Germ. Leges nationum Germanicarum
Poetae Poetae Latini medii aevi
SS rer. Merov. Scriptores rerum Merovingicarum
PG Patrologia Graeca
PL Patrologia Latina


 FOREWORD

This volume of essays represents the Proceedings from the third Inter-
national Conference on the Science of Computus in Ireland and Europe,
which took place in Galway, 16–18 July, 2010. It follows (belatedly) a
volume of papers from the previous Proceedings of the First Galway Con-
ference of 2006, that had the sub-title Computus and its cultural context
in the Latin West, AD 300–1200 (2010), which brought together papers
by ten of the leading scholars in the field, whose papers ranged from
the origins of the Anno Domini system of time-reckoning to the study
of computus in Ireland c.AD 1100. All those present at that pioneering
first event were agreed that a second, follow-up conference should be
organised, which duly took place (also in Galway), 18–20 July 2008. The
Proceedings of that volume too appeared subsequently, with the title, The
Easter Controversy of Late Antiquity and the Early Middle Ages (2011),
with eleven contributions ranging from the Easter-reckonings of Rome
and Milan in the late fourth century down to the Liber de astronomia of
the Irishman Dicuil c.AD 800.
It took longer than Immo or I had expected to produce the papers
of the third Galway Conference, but we hope that you will agree that
the wait was worth it! This volume brings together papers that range
in date from the Computus of AD 243 through the computistical writ-
ings of the ninth-century authors Claudius of Turin and Hrabanus Mau-
rus down to the great seventeenth-century scholar and chronographer,
Archbishop James Ussher (†1656). One of the contributions, by Leo-
franc Holford-Strevens on the Disputatio Chori, grew to monograph-
length and will be published separately in the Studia Traditionis Theo-
logiae series. As Conference organisers, Immo and I were delighted to


 Foreword

welcome back to Galway many of the colleagues who had been with us
for the first two gatherings, but also some younger scholars, such as Lisa
Chen and Philipp Nothaft. One of the very gratifying aspects of the Gal-
way Conferences, during those first years and in the years since, is the way
that younger scholars have joined our ranks, and the enthusiasm with
which the more established scholars have always welcomed them into
their ranks. Several of these younger colleagues have since gone on to
establish themselves among the first rank of scholars in the field.
The Third Conference attracted speakers from Canada, England, Ger-
many, Ireland, Italy, and the Netherlands. Sadly, however, Luciana Cup-
po, one of our founding-members and a regular speaker at our previous
gatherings, passed away before she could see the fruits of her present la-
bours in print. Luciana did, however, devote her final weeks to ensuring
that we received a clean copy of her paper; thanks to the generosity and
collaboration of her family, that paper is presented here in the form in
which she would have wanted to see it. We cherish fond recollections of
Luciana’s good-humoured and erudite contributions to all the Galway
Conferences, and we take this opportunity to mark our respect for her
profound scholarship by dedicating this collection to her memory.
A regular feature of the Galway Conferences has been the launch
of (at least) one book related to our activities. At the 2010 Conference
we had the pleasure of launching the Proceedings of the First Confer-
ence, Computus and its cultural context in the Latin West, AD 300–1200
(Studia Traditionis Theologiae 5), and Immo’s magnificent volume, The
Munich Computus: text and translation. Irish computistics between Isidore
of Seville and the Venerable Bede and its reception in Carolingian times
(Sudhoffs Archiv 59).
We look forward with anticipation to the Seventh Conference, which
will take place—in Galway!—in July 2018, when we hope to be able to
present these Conference Proceedings to our peers. Preparation is already
under way to publish the papers of the intervening Conferences in the
years to come.
We would like to thank all those whose encouragement and help
have made the Galway Computus Conferences to date such an outstand-
ing success. These include the authorities in the National University of
Ireland, Galway, which has hosted all of our events and provided gener-
ous funding for them; the Director and staff of the Moore Institute for
the Humanities, where the Conferences have always taken place; Hilary
Murphy in the Buildings Office, NUIG, for her unfailing kindness and
efficiency, and last—but certainly not least—Maura Walsh (Ó Cróinín),


 Foreword

who once again designed the wonderful Conference poster and pro-
gramme. The NUIG Medieval MA students were also a great help with
the day-to-day running of the Conference. We owe a particular debt to
George Janzen (University of Leipzig), who was an invaluable help with
the editorial preparation of the Proceedings.
To Immo, in particular, I owe special thanks for having once again
taken on, almost single-handedly, the heavy burden of seeing the Pro-
ceedings through the editorial process. If it has taken longer than we an-
ticipated to nurse the volume into print, that is no fault of his. I would
also like to thank the contributors to the volume, for their patience and
forbearance when it appeared that the volume might not see the light
of day. Immo and I hope that they—and the other readers of the pa-
pers—will feel that the volume was worth the wait. We are grateful also
to Prof. Thomas O’Loughlin (Nottingham), who has continued to offer
a home in his series of Studia Traditionis Theologiae, and to Dr Bart Jans-
sens, Brepols, who, by his steadfast support for the publication of these
volumes, has encouraged us to believe in the value of what we have been
doing.
It remains only to thank those institutions and repositories that
granted permission to reproduce images from manuscripts in their care:
the British Library, London; the Pierpont Morgan Library, New York;
the Archivo de la Catedral, León; the Bodleian Library, Oxford.
Finally, it should be noted that the contributions to this collection
(as in previous Proceedings of the Galway Computus Conference) are the
responsibility of the individual authors and do not necessarily represent
the views either of the editors or of the publishers.

Dáibhí Ó Cróinín,
Galway, 12 May 2017


 IMMO WARNTJES

INTRODUCTION:
STATE OF RESEARCH ON LATE
ANTIQUE AND EARLY MEDIEVAL
COMPUTUS

The essays in this volume emerged from papers presented at the 3rd Inter-
national Conference on the Science of Computus in Galway in July 2010.
The conference was established in 2006 and has since been held biannu-
ally. It may be fair to say that before this date, the late antique and medi-
eval science of computus was safely in the hands of only a few specialists.1
Computus is a decidedly Christian science of calendrical calcula-
tions, which evolved around the need to establish the date of Easter.
Since the third century, Easter was supposed to fall on the first Sunday
after the first full moon after the spring equinox, as is still the custom
today. Therefore, its calculation depended, from a late antique and medi-
eval perspective, on the mathematical modeling of the tropical year (the
Julian calendar) and of the lunar phases—the synodic lunar months, i.e.
the periods from one new moon to the next (84-year or 19-year lunar
cycle). From the seventh century onwards, computus evolved from its
core purpose of calculating Easter to become a more wide-ranging sub-
ject explaining the natural world as part of the quest of understanding
God’s creation.
In the 16th and 17th centuries, scholars like Scaliger, Petavius, and Bu-
cherius used computistical texts for their study of historical chronology

1
The following will also be published separately on <computus.lat>, and will be
updated on an annual basis. Writing a biographical essay is a difficult task, as it will never
be fully comprehensive and it reflects, probably more than other academic publications,
the interests and biases of the author. I am extremely grateful to Philipp Nothaft and
Dáibhí Ó Cróinín for discussing an early draft with me and for pointing out some pub-
lications I had initially neglected. As this essay is designed as a survey of past and present
publications on computus, it was agreed that, contrary to the procedure in the rest of this
volume, full bibliographical details are provided in the notes.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 1-42
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114732
 IMMO WARNTJES

and early church history.2 At the same time, James Ussher tried to recon-
struct the Easter controversy (i.e. the dispute about the theological and
scientific accuracy of conflicting lunar calendars) on this basis (as demon-
strated by Dáibhí Ó Cróinín in the present volume).3 The title of found-
ing father of the modern study of late antique and medieval computus,
however, must go to the early 18th-century Dutch scholar Johannes van
der Hagen, whose studies of the 1730s are still worth consulting.4
His lead was most prominently followed by Bruno Krusch. Initially,
Krusch wanted to work on the Easter reckoning of Victorius of Aquitaine
(published in AD 457) for his Ph.D. However, he soon realized that, in
order to do so, he first needed to come to grips with the pre-mid-fifth cen-
tury tradition.5 This study culminated in his celebrated Studien of 1880,
which contains the standard edition of numerous key texts (though many
of Krusch’s theories there expressed, like that of an older and newer Sup-
putatio Romana, have since been superseded). Shortly before his death in
1940, Krusch eventually produced new editions of the writings of Vic-
torius of Aquitaine and Dionysius Exiguus.6 In the meantime, the Irish
2
Scaliger, J.  J. (1583) De emendatione temporum, Paris, 2nd  ed. Geneva 1629;
Petavius, D. (1627) Opus de doctrina temporum, 2 vols, Paris; Bucherius, A. (1634) De
doctrina temporum. Commentarius in Victorium Aquitanum […] aliosque antiques can-
onum paschalium scriptores, Antwerp.
3
Ussher, J. (1639) Britannicarum ecclesiarum antiquitates, Dublin.
4
van der Hagen, J. (1733) Observationes in Prosperi Aquitani chronicon integrum
ejusque LXXXIV annorum cyclum, et in anonymi cyclum LXXXIV annorum a Muratorio
editum, nec non in anonymi laterculum paschalem centum annorum a Bucherio editum,
Amsterdam; van der Hagen, J. (1734) Observationes in veterum patrum et pontificium,
prologos et epistolas paschales, aliosque antiquos de ratione paschali scriptores, Amsterdam;
van der Hagen, J. (1736) Observationes in Heraclii imperatoris methodum paschalem;
ut et in Maximi monachi computum paschalem; nec non in anonymi chronicon paschale,
Amsterdam; van der Hagen, J. (1736) Dissertationes de cyclis paschalibus […], ut et de en-
neadecaeteridis […], nec non de computo solari, Amsterdam.
5
Krusch, B. (1880) Studien zur christlich-mittelalterlichen Chronologie: Der
84jährige Ostercyclus und seine Quellen, Leipzig, V; Krusch, B. (1938) ‘Studien zur christ-
lich-mittelalterlichen Chronologie: Die Entstehung unserer heutigen Zeitrechnung,’
Abhandlungen der Preußischen Akademie der Wissenschaften, Jahrgang 1937, phil.-hist.
Klasse, Nr. 8, Berlin, 5–6.
6
Krusch (1938); the 1892 edition of Victorius by Theodor Mommsen in MGH
Auct. ant. 9, 667–735, was based on Krusch’s notes. Of the other numerous important pub-
lications by Krusch, see especially Krusch, B. (1884) ‘Die Einführung des griechischen Pas-
chalritus im Abendlande,’ Neues Archiv der Gesellschaft für ältere deutsche Geschichtskunde 9,
99–169; Krusch, B. (1884) ‘Über eine Handschrift des Victurius,’ Neues Archiv der Gesell­
schaft für ältere deutsche Geschichtskunde 9, 269–81; Krusch, B. (1885) ‘Chronologisches aus
Handschriften,’ Neues Archiv der Gesellschaft für ältere deutsche Geschichtskunde 10, 81–94;
Krusch, B. (1910) ‘Das älteste fränkische Lehrbuch der dionysianischen Zeitrechnung,’ in
Mélanges offerts à M. Émile Chatelain, Paris, 232–42; Krusch, B. (1933) ‘Neue Bruchstücke


 STATE OF RESEARCH ON COMPUTUS

scholar Bartholomew Mac Carthy and the German classicist Eduard

Schwartz, in two seminal studies that still remain core readings on the sub-
ject (unfortunately both composed in somewhat inaccessible language),
corrected many of Krusch’s youthful and extremely bold reconstructions.7
Mac Carthy introduced computus into the study of medieval Ire-
land, tackling the intricacies of what he believed to have been the na-
tive tradition (the 84 (14)-year Easter reckoning now known as the
latercus). Another  true pioneer in the study of Irish computistics was
Mario Esposito. He certainly was not afraid of tackling difficult texts.
His first publication, of 1907, provides a study and transcript of Dicuil’s
Liber de astronomia, one of the most interesting texts of the Carolingian
age, which still awaits a full-scale study and critical edition.8 Esposito
followed this up by two more articles on this famous Irish scholar in
Carolingian Francia and his calendrical treatise, but also highlighted the
interrelation between computus and exegesis in the Irish tradition of the
seventh century by an analysis of De mirabilibus sacrae scripturae.9
If Esposito was an outsider in academia in Ireland, the same applied
to Krusch within Germany. He was respected for his editorial rigour
(but feared for his fierce temperament), and as a member of various
Akademien, but as an archivist not within the academic inner circle. In
German universities, computus received some attention from the phi-
lologists. Medieval Latin became a university subject in its own right,
Ludwig Traube, in 1902, the first full professor of the subject. His exten-
sive manuscript researches led him to a wonderful study of the transmis-
sion of Helperic’s De computo.10 His successor, Paul Lehmann, followed
this up in 1912 with the standard edition of the Computus of AD 562

der Zeitzer Ostertafel vom Jahre 447,’ Sitzungsberichte der Preußischen Akademie der Wis-
senschaften, Jahrgang 1933, philosophisch-historische Klasse, 981–97.
7
Mac Carthy, B. (1901) Annala Uladh: The Annals of Ulster, vol. 4: introduction
and sources, Dublin; Schwartz, E. (1905) ‘Christliche und jüdische Ostertafeln,’ Ab-
handlungen der Königlichen Gesellschaft der Wissenschaften zu Göttingen, philologisch-
historische Klasse, Band 8, Nr. 6, Berlin.
8
Esposito, M. (1907) ‘An unpublished astronomical treatise by the Irish monk
Dicuil,’ Proceedings of the Royal Irish Academy 26C, 378–446.
9
Esposito, M. (1914) ‘An Irish teacher at the Carolingian court: Dicuil,’ Studies:
an Irish quarterly review 3, 651–76; Esposito, M. (1920) ‘A ninth-century astronomical
treatise,’ Modern philology 18, 1–12; Esposito, M. (1919) ‘On the pseudo-Augustinian
treatise De mirabilibus sacrae scripturae written in Ireland in the year 655,’ Proceedings of
the Royal Irish Academy 35C, 189–207.
10
Traube, L. (1893) ‘Computus Helperici,’ Neues Archiv der Gesellschaft für ältere
deutsche Geschichtskunde 18, 73–105, repr. with editional notes in Traube (1909–20), iii
128–56.


 IMMO WARNTJES

attributed to Cassiodorus.11 Lehmann’s successor, Bernhard Bischoff, in

turn worked on ‘Ostertagstexte und Intervalltafeln’, published in 1940.12
During Krusch’s life-time, the late antique, early medieval Easter
controversy received considerable attention. Two scholars stand out in
this respect: In 1880, Louis Duchesne wrote an important article on the
alleged founding moment of Easter calculations, the Council of Nicea
of AD  325.13 In the very early 20th century, Joseph Schmid surveyed
the Easter controversy in two short monographs, which, together with
Krusch’s famous ‘Einführung des griechischen Paschalritus im Abend-
lande’ (‘The introduction of the Greek paschal reckoning in the Latin
West’), must form the basis of any new inquiry into this field.14
In the inter-war years, the Low Countries became one of the most
progressive regions for the study of the Middle Ages (the Belgian Henri
Pirenne and the Dutch scholar Johan Huizinga being only the most
prominent examples). Computistical research also profited from this
more general trend. André van de Vyver in Ghent made a name for him-
self in the second half of the 1920s initially as an expert on Boethius.
In the 1930s, however, he published two groundbreaking studies on
Cassiodorus and Abbo of Fleury, which remain seminal readings to the
present day.15 In the 1950s, he followed this up with a valuable study of
late antique Easter reckonings.16 Walter Émile van Wijk became privaat-
docent in de mathematische en technische tijdrekenkunde at the University
of Leiden in 1924. Unlike his predecessors in the field, Ludwig Ideler
and Friedrich Karl Ginzel, who produced monumental ‘handbooks’ on

11
Lehmann, P. (1912) ‘Cassiodorstudien,’ Philologus 71, 278–99, repr. in Lehmann
(1959­–62), ii 47–55.
12
Bischoff, B. (1940) ‘Ostertagtexte und Intervalltafeln,’ Historisches Jahrbuch 60,
549–80, repr. in Bischoff (1966–81), ii 192–227.
13
Duchesne, L. (1880) ‘La question de la pâque au concile de Nicée,’ Revue des
questiones historiques 28, 5–42. See also Daunoy, F. (1925) ‘La question pascale au con-
cile de Nicée,’ Echos d’Orient 24, 424–44.
14
Schmid, J. (1904) Die Osterfestberechnung auf den britischen Inseln vom Anfang
des vierten bis zum Ende des achten Jahrhunderts, Regensburg; Schmid, J. (1907) Die Os-
terfestberechnung in der abendländischen Kirche, Freiburg. See also Schmid, J. (1905) Die
Osterfestfrage auf dem ersten allgemeinen Konzil von Nicäa, Wien. Also worth consulting
is Grosjean, P. (1946) ‘Recherches sur les débuts de la controverse pascale chez les celts,’
Analecta Bollandiana 64, 200–45.
15
van de Vyver, A. (1931) ‘Cassiodore et son oeuvre,’ Speculum 6, 244–92; van de
Vyver, A. (1935) ‘Les oeuvres inédites d’Abbon de Fleury,’ Revue bénédictine 47, 125–69.
16
van de Vyver, A. (1957) ‘L’évolution du comput alexandrin et romain du IIIe au
V siècle,’ Revue d’histoire ecclésiastique 52, 5–25.
e


 STATE OF RESEARCH ON COMPUTUS

mathematical and technical chronology,17 van Wijk’s primary interest

was in twelfth- and thirteenth-century developments. His 1936 edition
of the Massa compoti of Alexander de Villa Dei, with its substantial in-
troduction, translation, and commentary, provides a good introduction
into the field.18 His 1951 edition of Reinher of Paderborn’s Computus
emendatus was as visionary as the 12th-century work itself.19
The 1930s saw the emigration of some of the brightest German
minds, principally to the US. It is worth remembering that this phenom-
enon was not only due to the rise of the Nazis in Germany. Academia in
general became more mobile and therefore global since the late 19th cen-
tury, and especially scholars of languages (German / English / Celtic)
often sought employment overseas. A good example is Heinrich Henel,
who became professor of German at Queen’s University in Kingston,
Canada in 1932. In 1934, Henel put Anglo-Saxon computistics on the
map, with his brilliant Studien zum altenglischen Computus (interest-
ingly written in German and published in Leipzig).20 This he followed
up with an edition of Ælfric’s De temporibus anni in 1942.21 Otto Neu-
gebauer, on the other hand, reacted to changed politics in Germany in
his decision to leave the country in 1933 (though his story is different to
the more traditional pattern of fleeing Nazi persecution). Neugebauer
taught history of mathematics in the famous mathematics department
of the University of Göttingen, many members of which were opposed
to the activities of the Nazi party. Still, few followed Neugebauer’s lead
and example. He was appalled by the Nazi policies that discriminated
against his Jewish students and colleagues, and refused the oath of loy-
alty to the new regime. Anticipating persecution, he moved first to Co-
penhagen. When academic publishing collapsed in Germany in 1938
(to which he had continued to contribute as editor of the Zentralblatt

17
Ideler, L. (1825–26) Handbuch der mathematischen und technischen Chronolo-
gie, 2 vols, Berlin; Ginzel, F. K. (1906–14) Handbuch der mathematischen und technis-
chen Chronologie, 3 vols, Leipzig. It is worth mentioning here also the studies of medieval
chronology by the Oxford archivist and scholar of diplomatics, Reginal Lane Poole, col-
lected in his 1934 Studies in chronology and history, Oxford.
18
van Wijk, W. E. (1936) Le nombre d’or: étude de chronologie technique suivie du
texte de la Massa compoti d’Alexandre de Villedieu (c. 1170–1250), La Haye.
19
van Wijk, W. E. (1951) ‘Le computu emendé de Reinherus de Paderborn (1171),
publié d’après le Ms V.P.L. 191-E de la Bibliothèque de l’Université de Leiden,’ Verhan-
delingen der Koninklijke Nederlandse Akademie van Wetenschappen, afd. letterkunde 59,
no. 3, Amsterdam.
20
Henel, H. (1934) Studien zum altenglischen Computus, Leipzig.
21
Henel, H. (1942) Ælfric’s De temporibus anni, London.


 IMMO WARNTJES

für Mathematik, published by Springer), he had to move again, this time

to the US, joining Brown University, where he stayed for the rest of his
distinguished career. His principal interest lay in Babylonian and Egyp-
tian astronomy and mathematics, but in the last few decades of his life
he got more and more interested in medieval science, both Arabic and
Latin. His monograph on Ethiopic computus contains some references
to the Alexandrian reckoning, and he also produced an article on the
Computus of AD 562.22 The Nazi destruction of the academic landscape
hit mathematics particularly hard, but also in other areas the brightest
minds, like Max Förster and Wilhelm Levison, had to emigrate to escape
almost certain death (Levison’s famous England and the Continent in the
eighth century derives from his 1943 Oxford lectures; Levison was one of
Krusch’s students and collaborators, and Krusch continued correspond-
ence with him right to the point of emigration).
A significant turn in the study of computus was marked by the publi-
cations of the great American scholar Charles W. Jones in the 1930s and
1940s. Before Jones, the principal interest in computus lay in the first
few centuries of Christianity. As a Bedan scholar (attracted to the field
by his teacher, the English-born Max Laistner), Jones shifted the focus
of research to the early medieval period. His grand tour of European
libraries was, unfortunately, cut short by the outbreak of Word War II
in 1939, but it had provided him with an unparalleled insight into the
manuscript transmission history of early medieval computistical texts,
which he published as Bedae pseudepigrapha in 1939.23 This preparatory
work was followed in turn by his masterful edition of Bede’s scientific
works in 1943, whose roughly 120-page introduction is still considered
the seminal account of the development of computistical thought from
the foundation of Christianity to the age of Bede in the eighth century.24
Jones had created an awareness that only through the systematic study
of the thousands of computistical manuscripts that survive from before
c.AD 1200 could one arrive at a thorough understanding of the late an-
tique and medieval science of computus. This lead was followed by Al-
fred Cordoliani, who, from the 1940s to the 1960s, mined the libraries

22
Neugebauer, O. (1979) Ethiopic Astronomy and Computus, Wien; Neugebauer,
O. (1982) ‘On the Computus paschalis of ‘Cassiodorus’,’ Centaurus 25, 292–302.
23
Jones, C. W. (1939) Bedae pseudepigrapha: scientific writings falsely attributed to
Bede, Ithaca.
24
Jones, C. W. (1943) Bedae opera de temporibus, Cambridge. See also the reviews
of this work by Heinrich Henel in American Historical Review 49 (1944), 694–96; Jour-
nal of English and Germanic Philology 43 (1944), 411–16.


 STATE OF RESEARCH ON COMPUTUS

of France, Switzerland, and Spain, from Mount St Michel to Tarragona,

in search of new discoveries. His method was more descriptive than ana-
lytical, however, and his interpretations not always well founded. But in
his more than 40 articles, he certainly identified key figures and themes
for future research, from Martin of Braga to Gerland the Computist,
from the Easter controversy to the re-calculation of the incarnation
era.25 More importantly, Cordoliani made the first (and, to the present
day, only) serious attempt at systematizing the thousands of short for-
mulae found in hundreds of manuscripts (something that had not been
achieved by Thorndike & Kibre’s famous Incipits of mediaeval scientific
writings in Latin of 1937).26 Cordoliani’s broad manuscript research was
complemented by Giles Meersseman and Edvige Adda’s detailed study
of a ninth-century Computus attributed to Pacificus of Verona.27
Whether triggered by Jones’s survey of the early development of com-
putistical thought in the Latin West, or simply following the established
focus, Byzantinists discovered Latin computus of the first centuries of
Christianity as a fruitful field of research in the 1950s and 1960s. Marcel
Richard worked extensively on Hippolytus and the underlying lunar cy-
cles.28 After a strong interest in chronology throughout his distinguished
career (culminating in La chronologie of 1958), Venance Grumel turned to
the early Easter controversy and the question of Anatolius’ paschal tract in
particular.29 During this time, theologians also showed considerable inter-
est in computus and its exegesis. George Ogg produced seminal studies on

25
For a full list of Cordoliani’s publications on computus, see the Appendix to this
introduction.
26
Cordoliani, A. (1960) ‘Contribution à la littérature du comput ecclésiastique au
moyen âge,’ Studi medievali, Ser. 3, 1, 107–37; Cordoliani, A. (1961) ‘Contribution à la
littérature du comput ecclésiastique au moyen âge,’ Studi medievali, Ser. 3, 2, 169–208.
Thorndike, L. and P. Kibre (1937) A catalogue of incipits of mediaeval scientific writings in
Latin, Cambridge.
27
Meersseman, G. G. and E. Adda (1966) Manuale di computo con ritmo mnemo-
tecnico dell’arcidiacono Pacifico di Verona, Padua.
28
Richard, M. (1950) ‘Comput et chronographie chez saint Hippolyte,’ Mélanges de
science religieuse 7, 237–68; Richard, M. (1951) ‘Comput et chronographie chez saint Hip-
polyte, 5. La durée de la vie du Christ,’ Mélanges de science religieuse 8, 19–50; Richard, M.
(1953) ‘Encore le problème d’Hippolyte,’ Mélanges de science religieuse 10, 13–52, 145–80;
Richard, M. (1966) ‘Notes sur le comput de cent-douze ans,’ Revue des études byzantines
24, 257–77; Richard, M. (1974) ‘Le comput pascal par octaétéris,’ Le muséon 87, 307–39.
29
Grumel, V. (1958) La chronologie, Paris; Grumel, V. (1960) ‘Le problème de la
date paschale aux IIIe et IVe siècles,’ Revue des études byzantines 18, 163–78; Grumel, V.
(1964) ‘La date de l’équinoxe vernal dans le canon pascal d’Anatole de Laodicée,’ in Mé-
langes Eugène Tisserant, 7 vols, Vatican, ii 217–40. Both the Easter controversy and Ana-
tolius have constantly been the subject of individual publications, too many to list here.


 IMMO WARNTJES

the Computus of AD 243, and later turned his attention to Hippolytus.30

In Germany, in more recent years, Wolfgang Huber and August Strobel
devoted important monographs to the early Easter question.31
From the late 1950s onwards, excellent work was done in Germany
on computistical chronology, or, as Joachim Wiesenbach would term
it, critical computists. Interestingly, this work exclusively derived from
exceptional Ph.D. theses of scholars who did not necessarily pursue an
academic career afterwards, which may explain why this area of research
remained occasional and never really entered the mainstream of Medi-
eval Studies. Anna-Dorothee von den Brincken submitted a thesis enti-
tled Studien zur lateinischen Weltchronistik bis in das Zeitalter Ottos von
Freising to the University of Münster in 1956, which was published the
following year.32 This she followed up with two important studies on
Heimo of Bamberg and Marianus Scottus in the early 1960s, but when
she became professor in Köln, her interest had moved more decidedly
towards cartography and medieval chronicles more broadly.33 Joachim
Wiesenbach identified the Liber decennalis of Sigebert of Gembloux,
which was considered lost by previous scholarship, in Rome, Biblioteca
Angelica, 1413.34 This discovery served as the foundation for his Ph.D.
at the University of Frankfurt submitted in 1979 and published in 1986,
30
Ogg, G. (1962) ‘Hippolytus and the introduction of the Christian era,’ Vigiliae
Christianae 16, 2–18; Ogg, G. (1954) ‘The tabella appended to the Pseudo-Cyprianic
De pascha computus in the Codex Remensis,’ Vigiliae Christianae 8, 134–44; Ogg, G.
(1955) The Pseudo-Cyprianic De pascha computus, London.
31
Huber, W. (1969) Passa und Ostern: Untersuchungen zur Osterfeier der al-
ten Kirche, Berlin; Strobel, A. (1977) Ursprung und Geschichte des frühchristlichen
Osterkalenders, Berlin; Strobel, A. (1984) Texte zur Geschichte des frühchristlichen
Osterkalenders, Münster. In this context, see also the later publication by Gerlach, K.
(1998) The antenicene pascha: a rhetorical history, Leuven.
32
von den Brincken, A.-D. (1957) Studien zur lateinischen Weltchronistik bis in das
Zeitalter Ottos von Freising, Düsseldorf.
33
von den Brincken, A.-D. (1960) ‘Die Welt- und Inkarnationsära bei Heimo
von St Jakob: Kritik an der christlichen Zeitrechnung durch Bamberger Komputisten
in der ersten Hälfte des 12. Jahrhunderts,’ Deutsches Archiv für Erforschung des Mittel­
alters 16, 155–94; von den Brincken, A.-D. (1961) ‘Marianus Scottus, unter besonderer
Berücksichtigung der nicht veröffentlichten Teile seiner Chronik,’ Deutsches Archiv für
Erforschung des Mittelalters 17, 191–238. See also von den Brincken, A.-D. (1982) ‘Ma­
rianus Scottus als Universalhistoriker iuxta veritatem Evangelii,’ in Heinz Löwe, Die Iren
und Europa im früheren Mittelalter, 2 vols, Stuttgart, ii 970–1009; von den Brincken, A.-
D. (1979) ‘Beobachtungen zum Aufkommen der retrospektiven Inkarnationsära,’ Archiv
für Diplomatik 25, 1–20.
34
Wiesenbach, J. (1977) ‘Der Liber decennalis in der Hs. Rom, Biblioteca Angelica
1413, als Werk Sigeberts von Gembloux,’ Deutsches Archiv für Erforschung des Mittel­
alters 33, 171–81.


 STATE OF RESEARCH ON COMPUTUS

a critical edition of the text with a solid introduction to the computis-

tical background.35 Another edition of a crucial text was prepared by
Hans Martin Weikmann for his 1984 Würzburg Ph.D., which focused
on Heimo of Bamberg’s De decursu temporum. It is now available as a
monograph of 2004.36
Inspired by Jones, Cyril Hart, in 1970, revived the study of late An-
glo-Saxon computus, following on where Henel left off with his study of
Aelfric and the anonymous tradition. Hart focused on Ramsey abbey,
its principal computist, Byrhtferth, and (as he believed) its key scientific
manuscript, Oxford, St. John’s College 17, eventually culminating in a
2003 Survey of the development of mathematical, medieval, and scientific
studies in England before the Norman conquest.37 Peter Baker contin-
ued this work with a preliminary study on Byrhtferth in Anglo-Saxon
England in 1982, and then the critical edition (together with Michael
Lapidge) of Byrhtferth’s Enchiridion in 1995.38 St John’s College 17 was
analysed in detailed by Faith Wallis for her 1985 Ph.D.,39 and she has
since published widely on the (principally English) computistical tradi-
tion. Most notably, she opened up the study of (not only Anglo-Saxon)
computistics to a wider audience by providing a translation of Bede’s
De temporum ratione, which she has followed up since, in collaboration
with Calvin Kendall, with translations of Bede’s De natura rerum and
De temporibus.40

Wiesenbach, J. (1986) Sigebert von Gembloux: Liber decennalis, Weimar.

35

Weikmann, H. M. (2004) Heimo von Bamberg: De decursu temporum, Hanno-

36

ver.
37
Hart, C. J. R. (1970) ‘The Ramsey computus,’ English Historical Review 85, 29–
44; Hart, C. J. R. (1970) ‘Byrhtferth and his manual,’ Medium aevum 41, 95–109; Hart,
C. J. R. (2003) Learning and culture in late Anglo-Saxon England and the influence of
Ramsey Abbey on the English monastic schools, vol. 2: a survey of the development of math-
ematical, medical, and scientific studies in England before the Norman conquest, Lewiston.
38
Baker, P.  S. (1982) ‘Byrhtferth’s Enchiridion and the computus in Oxford,
St  John’s College 17,’ Anglo-Saxon England 10, 22–37. Baker, P.  S. and M.  Lapidge
(1995) Byrhtferth’s Enchiridion, Oxford. For an older, pioneering study, see Kluge, F.
(1885) ‘Angelsächsische Excerpte aus Byrhtferth’s Hanboc oder Enchiridion,’ Anglia 8,
298–337.
39
See the excellent website which arose from her Ph.D. interest: http://digital.
library.mcgill.ca/ms-17/index.htm.
40
Wallis, F. (1999) Bede: The Reckoning of Time, Liverpool; Kendall, C. A. and
F. Wallis (2010) Bede: On the nature of things and On times, Liverpool. See also Wallis,
F. (1989) ‘The church, the world and the time: prolegomena to a history of the medieval
computus,’ in M.-C. Déprez-Masson, Normes et pouvoir à la fin du moyen âge, Montreal,
15–29; Wallis, F. (2015) ‘What a medieval diagram shows: a case study of computus,’
Studies in iconography 36, 1–40.


 IMMO WARNTJES

The renewed interest in Bede’s scientific texts may have been trig-
gered by Charles W. Jones’s second (re-)edition of the works in the Cor-
pus Christianorum Series Latina in 1975–80, this time including Bede’s
Chronicles, but lacking the substantial introduction and detailed com-
mentary of the 1943 edition.41 This publication also marked an impor-
tant turn towards the early medieval computistical tradition and its texts.
Jones’s student, Wesley Stevens, produced a critical edition of Hrabanus
Maurus’ De computo in Corpus Christianorum Continuatio Mediaevalis,
and has since worked extensively on Hrabanus’ pupil Walahfrid Strabo.42
In the 1980s, Dáibhí Ó Cróinín put early medieval Irish computistics
on the map. First, his discovery of the long-lost 84-year Easter table used
in the regiones Scottorum until AD 716, in Briton Wales until the later
eighth century, provided, for the first time, a solid basis for the study of
the early medieval Easter controversy, and it was pivotal in the recon-
struction of the chronological apparatus of the Irish annals.43 In effect,
this discovery provided, for the first time, a series of reliable dates for
the first three centuries of Irish history.44 Together with his wife, Maura
Walsh, he produced in 1988 an edition and translation of the key text for
the Irish side of the Easter controversy, Cummian’s letter of AD 632.45 In
the same volume, he edited the most sophisticated of the Irish compu-
tistical textbooks (De ratione conputandi), a model of its time, not only
rivaling but surpassing Bede’s rather wordy accounts on the subject.46

41
Jones, C. W. (1975–80) Bedae Venerabilis opera, pars VI, Corpus Christianorum
Series Latina CXXXIIIA–C, Turnhout.
42
Stevens, W.  M. (1979) ‘Rabani Mongontiacensis episcopi De computo,’ in
CCCM 44, 163–323.
43
Mc Carthy, D. P. and D. Ó Cróinín (1987–88) ‘The ‘lost’ Irish 84-year Easter
table rediscovered,’ Peritia 6–7, 227–42, repr. in Ó Cróinín (2003), 58–75; Mc Carthy,
D. P. (1993) ‘Easter principles and a fifth-century lunar cycle used in the British Isles,’
Journal for the History of Astronomy 24, 204–24; Mc Carthy, D. P. (1998) ‘The chronol-
ogy of the Irish annals,’ Proceedings of the Royal Irish Academy 98C, 203–55.
44
http://www.irish-annals.cs.tcd.ie/.
45
Walsh, M. and D. Ó Cróinín (1988) Cummian’s letter De controversia paschali,
together with a related Irish computistical tract, De ratione conputandi, Toronto.
46
The discovery was announced and its historical context discussed in Ó Cróinín,
D. (1982) ‘A seventh-century Irish computus from the circle of Cummianus,’ Proceedings
of the Royal Irish Academy 82C, 405–30. For Ó Cróinín’s other contributions to early
medieval computus, see the essays collected in his Early Irish history and chronology,
Dublin 2003. Ó Cróinín’s scholarship was complemented by the contemporary studies
by Kenneth Harrison: Harrison, K. P. (1973) ‘The Synod of Whitby and the beginning
of the Christian era in England,’ Yorkshire Archaeological Journal 45 (1973), 108–14;
Harrison, K. P. (1976) The framework of Anglo-Saxon history to A.D. 900, Cambridge;
Harrison, K. P. (1977–78) ‘Epacts in Irish chronicles,’ Studia Celtica 12–13, 17–32; Har-


 STATE OF RESEARCH ON COMPUTUS

In the late 1980s and early 1990s, Joan Gómez Pallarès, in various
articles that evolved out of his 1986 Barcelona Ph.D. thesis, aimed at
reconstructing Visigothic computistics, providing diplomatic editions
of the key tracts.47 Though not entirely successful in the end, this was the
first serious attempt at a comprehensive study of the computistical tradi-
tion of the Iberian peninsula from the seventh to the eleventh centuries.
The 1990s saw an increased interest in Carolingian computus. Two
Aachen-based scholars, Paul Butzer and Dietrich Lohrmann, utilised
the local dynamics in the former centre of the Carolingian Empire trig-
gered by the 1400th anniversary of the Admonitio generalis (AD 789) and
of the AD 790s, the formative decade in what is termed the Carolingian
Renaissance, to organize two international conferences on the wider con-
text of Carolingian science, a truly pioneering undertaking.48 Out of this
context emerged the only substantial study on Alcuin’s contribution to
the computistical endeavour in the Carolingian age by Kerstin Springs-
feld.49 The study of Carolingian diagrams and cosmology was taken up
anew by Barbara Obrist, culminating in her La cosmologie médiévale of
2004,50 which, in many ways, provides the cultural background to Bruce
Eastwood’s more technical pioneering work on astronomical diagrams
from the 1980s onwards. In 1998, Stephen McCluskey provided the first
survey of early medieval astronomy in his Astronomies and cultures in
early medieval Europe,51 while in 1994 Brigitte Englisch attempted to

rison, K. P. (1978) ‘Easter cycles and the equinox in the British Isles,’ Anglo-Saxon Eng-
land 7, 1–8; Harrison, K. P. (1979) ‘Luni-solar cycles: their accuracy and some types of
usage,’ in M. H. King and W. M. Stevens, Saints, scholars and heroes, 2 vols, Collegeville,
ii 65–78; Harrison, K. P. (1982) ‘Episodes in the history of Easter cycles in Ireland,’ in
D. Whitelock, R. McKitterick, and D. Dumville, Ireland in early medieval Europe, Cam-
bridge, 307–19; Harrison, K. P. (1984) ‘A letter from Rome to the Irish clergy, AD 640,’
Peritia 3, 222–29.
47
Gómez Pallarès, J. (1986) Estudios sobre el Computus Cottonianus, diss. Barce-
lona, www.tdx.cbuc.es/handle/10803/5248; his articles are collected in Gómez Pallarès,
J. (1999) Studia chronologica: estudios sobre manuscritos latinos de cómputo, Madrid.
48
Butzer, P. L. and D. Lohrmann (1993) Science in Western and Eastern civiliza-
tion in Carolingian times, Basel; Butzer, P. L. et al. (1998) Karl der Große und sein Nach-
wirken. 1200 Jahre Kultur und Wissenschaft in Europa, 2: Mathematical Arts, Turnhout.
49
Springsfeld, K. (2002) Alkuins Einfluß auf die Komputistik zur Zeit Karls des
Großen, Stuttgart.
50
Obrist, B. (2004) La cosmologie médiévale. I: Les fondements antiques: textes et
images, Firenze.
51
McCluskey, S.  C. (1998) Astronomies and cultures in early medieval Europe,
Cambridge.


 IMMO WARNTJES

define the place of computus within the late antique, early medieval con-
cept of the quadrivium.52
The turn of the millennium saw a brief resurgence of interest in the
incarnation era. George Declercq published a stimulating short volume
on Anno Domini, both in French and English, which he supplemented
with a more detailed study in Sacris Erudiri.53 The main focus of Anno
Domini lies on the Easter reckonings of Victorius of Aquitaine and Dio-
nysius Exiguus, and it therefore provides the context for Krusch’s 1938
editions of these works. The seventh International Medieval Congress
in Leeds in 2000 also took the turn of the millennium as an excuse to
address questions of time and eternity in the Middles Ages, which cul-
minated in a volume on the theme in 2003.54 Anna-Dorothee von den
Brincken’s introduction to historical chronology was also published ap-
propriately in 2000.55
The last in the list of scholars who promoted the study of computis-
tics before 2006 is Arno Borst. One of Borst’s trademarks was to work
on the history of the place he was living in. When he received the Ruf
to the newly founded University of Konstanz with its ambitious educa-
tional reform program, he immediately started to work on the monastic
culture of the Lake Constance region. He got particularly interested in
the Reichenau monk Hermannus Contractus (†1054), arguably one of
the most brilliant and wide-ranging intellectuals of the eleventh century.
From 1975, Borst pursued the project of writing a biography of Her-
mann. For this, he needed to understand Hermann’s oeuvre, which, to
Borst’s dismay, was not readily accessible. In particular, Hermann’s com-
putistica and astrolabica still remained largely unexplored in numerous
manuscripts, with no edition, for many texts not even a transcript avail-
able in print (Cordoliani had worked on Hermann’s computistica in the
early 1960s, Bergmann in the 1980s56).

52
Englisch, B. (1994) Die Artes liberales im frühen Mittelalter (5–9. Jh.), Stuttgart.
53
Declercq, G. (2000) Anno domini: les origins de l’ère chrétienne, Turnhout; in
English: Anno Domini: the origins of the Christian era, Turnhout 2000; Declercq, G.
(2002) ‘Dionysius Exiguus and the introduction of the Christian era,’ Sacris Erudiri 41,
165–246.
54
Jaritz, G. and G.  Moreno-Riaño (2003) Time and eternity: the medieval dis-
course, Turnhout.
55
von den Brincken, A.-D. (2000) Historische Chronologie des Abendlandes:
Kalenderreformen und Jahrtausendrechnungen – eine Einführung, Stuttgart.
56
Cordoliani, A. (1963) ‘Le computiste Hermann de Reichenau,’ Miscellanea sto­
rica ligure 3, 161–90; Bergmann, W. (1988) ‘Chronographie und Komputistik bei Her-


 STATE OF RESEARCH ON COMPUTUS

Therefore, Borst embarked on producing editions of Hermann’s texts

before tackling the broader goal of a biography. But Hermann’s compu-
tistical writings, in particular, proved difficult to edit, as the Reichenau
monk obviously based his knowledge on a centuries-old tradition, which
had not received adequate attention in modern scholarship. Borst tried
to fill that lacuna in 1990 with a sweeping overview of the history of
computus, which must have made him aware of all the work that still
needed to be done.57 Certainly, after this he concentrated on Hermann’s
potential sources. First, in 1994, he produced a full-scale study on the
reception of Pliny in the Middle Ages.58 Then he turned for a decade to
the early medieval calendar tradition, culminating in his monumental
edition of the so-called Reichskalender.59 Finally, he engaged with com-
putistical texts. His massive 2006 editions of 20 computistical texts from
Francia of the period AD 721–818 is a milestone in the modern study
of early medieval computus, and we were proud to honour this achieve-
ment by launching the 3-volume work at the First Galway Computus
Conference.60
From Jones to Borst, that is from the 1930s to the early 21st centu-
ry, the study of early medieval computus focused principally (with few
exceptions) on works by authors known by name. Borst, however, cre-
ated an awareness that for any proper understanding of this calendrical
science, it is essential to analyse the much more numerous anonymous
texts. The three Carolingian encyclopediae (Lib. ann. of AD 793, Lib.
comp. of AD 809/12, and Lib. calc. of AD 818) had previously received
some attention, but a critical edition of each of these was still wanting.61

mann von Reichenau,’ in D. Berg and H.-W. Goetz, Historiographia mediaevalis: Studien

zur Geschichtsschreibung und Quellenkunde des Mittelalters, Darmstadt, 103–17.
57
Borst, A. (1990) Computus: Zeit und Zahl in der Geschichte Europas, Berlin; it
was translated into English by Andrew Winnard as The ordering of time: from the ancient
computus to the modern computus, Chicago 1993; see also Borst, A. (1988) ‘Computus:
Zeit und Zahl im Mittelalter,’ Deutsches Archiv für Erforschung des Mittelalters 44, 1–82.
58
Borst, A. (1994) Das Buch der Naturgeschichte: Plinius und seine Leser im Zeit-
alter des Pergaments, Heidelberg.
59
Borst, A. (1998) Die karolingische Kalenderreform, Hannover; Borst, A. (2001)
Der karolingische Reichskalender und seine Überlieferung bis ins 12. Jahrhundert, 3 vols,
Hannover; Borst, A. (2004) Der Streit um den karolingischen Kalender, Hannover.
60
Borst, A. (2006) Schriften zur Komputistik im Frankenreich von 721 bis 818,
3 vols, Hannover.
61
E.g. Neuß, W. (1941) ‘Eine karolingische Kopie antiker Sternzeichenbilder im
Codex 3307 der Biblioteca Nacional zu Madrid,’ Zeitschrift des deutschen Vereins für
Kunstwissenschaft 8, 113–40; Boschen, L. (1972) Die Annales Prumiensis, Düsseldorf;
Springsfeld (2002); and numerous articles by Eastwood.


 IMMO WARNTJES

Other crucial texts (like Prol. Aquit. of AD 721, Dial. Burg. of AD 727,
Dial. Neustr. of AD 737, Cap. comp. of AD 809) had only been tran-
scribed or mentioned in passing in the literature, but remained sidelined
by a focus on Bede, Alcuin, and Hrabanus Maurus. However, other ex-
tremely important texts had remained hidden in the manuscripts (like
Dial. Langob. of c.AD 750, Quaest. Austr. of AD 764, Quaest. Langob. of
c.AD 780, Epist. Rat. of AD 809, Arg. Aquens. of AD 816), and part of
Borst’s achievement was to bring these texts to the attention of scholars.
To be sure, Borst’s corpus has its problems, especially in the representa-
tion of texts. But it still represents a landmark publication in the field of
computistical studies.
Schriften zur Komputistik im Frankenreich marked the culmination
and high-point of Borst’s impressive career. He was fully aware that
Schriften would be his last major publication. In fact, he postponed a
crucial medical operation in order to see the three volumes through the
press, as he feared that he would not be able to continue his normal,
highly intense level of work afterwards. Unfortunately, he died in the
following year. But the end of Borst’s scholarly endeavor heralded a new,
vibrant, and exceptionally productive interest in what had previously
been a neglected area of research. Whether the establishment of the
Galway Conference also contributed to this more general trend remains
for others to judge. Certainly, publications on late antique and early me-
dieval computistics have spiraled since 2006. In the following, I can only
name the most prominent publications, surveying the literature chrono-
logically according to content from AD 200 to 1200.
First place in such a survey must be given to Alden Mosshammer.
In 2008, Mosshammer produced his Easter Computus and the origins of
the Christian era, the first monograph in English on the beginnings of
the computistical tradition, c.AD 200–600.62 In effect, this updates (but
does not replace) Eduard Schwartz’s Christliche und jüdische Ostertafeln
published a century earlier. In subsequent years, Mosshammer has fol-
lowed up this survey with more detailed studies. In the present volume,
he outlines prolegomena to a new edition of the oldest computistical
text that has survived, the Computus of AD  243. Another and more
prominent focus of his is on the fifth century. This century is crucial for
our understanding of the history of computus, as it saw the biggest pro-
duction of new (and rivaling) Easter reckonings in the late antique and

62
Mosshammer, A. A. (2008) The Easter Computus and the origins of the Christian
era, Oxford.


 STATE OF RESEARCH ON COMPUTUS

medieval period. Mosshammer concentrated on the laterculus of Augus-

talis as outlined in the Carthaginian computus of AD 455 (preserved in
Lucca, Biblioteca Capitolare Feliniana, 490).63 Both the laterculus and
the Carthaginian computus as a whole are a great reminder of the active
role played by the Vandal kingdom in late antique culture, and especially
its scientific strand, which deserves a full-scale investigation in the future.
Besides the Easter reckonings described in the Computus Cartha­
ginensis (the laterculus of Augustalis, the circulus primus, and the circulus
secundus), the fifth century witnessed the creation of three more promi-
nent methods of calculating Easter: the latercus (the reckoning followed
in Britain and Ireland until the eighth century), probably invented by
Sulpicius Severus c.AD 410; the famous Zeitz table of AD 447; and the
532-year table established by Victorius of Aquitaine, which dominated
western Europe until the mid-eighth century.
Since the discovery of the only known copy of the latercus in Padua,
Biblioteca Antoniana, I 27, this system has been very much in the focus
of modern research. Most notably, Dan Mc  Carthy reconstructed the
chronological apparatus of the Irish annals on this basis, which culminat-
ed in his monograph of 2008, The Irish annals: their genesis, evolution and
history.64 In 2011, he rounded off a series of earlier publications on the
history and cultural impact of this Easter reckoning.65 In 2006, Caitlin
Corning reassessed the insular Easter controversy of the seventh century
on the basis of the reconstructed data of the latercus and contemporary
epistles and historiography (principally Bede).66 Surprisingly, however,
Corning did not draw on the substantial computistical corpus of the
seventh century, which would have refined, if not changed her picture.
In 2010, Leofranc Holford-Strevens demonstrated, how our improved
understanding of the latercus has helped to re-assess key episodes in the
insular Easter controversy, in this case the Synod of Whitby. He was also
instrumental in establishing technical details of this Easter reckoning,

63
Mosshammer, A.  A. (2011) ‘The Computus of 455 and the laterculus of Au-
gustalis, with an appendix on the fractional method of Agriustia,’ in Warntjes and Ó
Cróinín (2011), 21–47.
64
Mc  Carthy, D.  P. (2008) The Irish Annals: their genesis, evolution and history,
Dublin.
65
Mc Carthy, D. P. (2011) ‘On the arrival of the latercus in Ireland,’ in Warntjes
and Ó Cróinín (2011), 48–75. Of his earlier publications on this topic, see especially
Mc Carthy, D. P. (1994) ‘The origin of the latercus paschal cycle of the insular Celtic
churches,’ Cambrian Medieval Celtic Studies 28, 25–49.
66
Corning, C. (2006) The Celtic and Roman traditions: conflict and consensus in the
early medieval church, New York.


 IMMO WARNTJES

a process that began with Dan Mc Carthy’s 1993 reconstruction of the

data of the Padua latercus, followed by Holford Strevens’s analyses in his
1999 Oxford companion to the year (together with Bonnie Blackburn)
and in his seminal 2007 article.67 I then compared the Padua latercus data
with the latercus information in the Munich Computus in 2007.68
The Zeitz table regained considerable prominence in recent times
due to the rediscovery of some of its fragments in the Stiftsbibliothek
in Zeitz, which made the German newspapers in 2005.69 Among Ger-
man academics, however, this did not trigger any further curiosity.70 Dan
Mc Carthy presented an impressive re-evaluation of the Zeitz table at
the Fourth Galway conference in 2012 (to be published in the follow-
ing proceedings). It can only be hoped that this will lead to a large-scale
analysis of the reformation of the Roman 84-year reckoning (the Sup-
putatio Romana) in the fifth century, which appears to provide the wider
context for the latercus, the Zeitz table, and the tables described in the
Carthaginian Computus.
Victorius of Aquitaine still seems to suffer from Krusch’s scathing
verdict of 1938: ‘Der calculator scrupulosus [Victorius] war ein ganz
beschränker Kopf und außerdem nicht einmal ehrlich.’ (‘Victorius was
a very limited individual, and not even an honest one.’)71 Columbanus,
c.AD  600, was of a similar opinion, giving Victorius ‘bad press’ espe-
cially for his (diplomatic) use of alternative dates for Easter according
to what he believed were the Alexandrian and the Roman traditions.72
The Visigothic monk Leo, in AD 627, wrote along the same lines but

67
Mc Carthy (1993); Blackburn, B. and L. Holford-Strevens (1999) The Oxford
companion to the year, Oxford, 870–75; Holford-Strevens, L. (2008) ‘Paschal lunar cal-
endars up to Bede,’ Peritia 20, 165–208: 178–87.
68
Warntjes, I. (2007) ‘The Munich Computus and the 84 (14)-year Easter reckon-
ing,’ Proceedings of the Royal Irish Academy 107C, 31–85.
69
Frankfurter Allgemeine Zeitung, 4 November 2005.
70
At the time, and regularly since, the Zeitz fragments were and are on display
in Berlin or Zeitz (or both), and an exhibition booklet was produced: Overgaauw, E.
and F.-J. Steving (2005) Die Zeitzer Ostertafel aus dem Jahre 447, Petersberg. This, how-
ever, shows no interest either in the construction of the table itself or its scientific con-
text. The standard accounts remain Mommsen, T. (1863) ‘Zeitzer Ostertafel vom Jahre
447,’ Abhandlungen der Königlichen Akademie der Wissenschaften zu Berlin, Jahrgang
1862, philosophisch-historische Klasse, 537–66; MGH Auct. ant. 9, 501–09; Krusch, B.
(1933) ‘Neue Bruchstücke der Zeitzer Ostertafel vom Jahre 447,’ Sitzungsberichte der
Preußischen Akademie der Wissenschaften, Jahrgang 1933, philosophisch-historische Klasse,
981–97.
71
Krusch (1938), 15.
72
Columbanus, Epistolae 1–2 (ed. by Walker (1957), 2–22).


 STATE OF RESEARCH ON COMPUTUS

using less insulting language.73 This, however, should not lead us to the
conclusion that Victorius is not worth studying. Quite the contrary,
both Columbanus and Leo are important reminders of the relevance of
Victorius still in the seventh century, and Dan Mc Carthy’s contribution
in the present volume highlights the potential technical problems that
followers of this Easter reckoning faced.
One of the key factors in minimizing Victorius’ importance, no
doubt, was Bede’s insistence on ‘Roman unity’ when there was none.
The sixth to eighth centuries not only saw a conflict between ‘Celtic’
and ‘Roman’ tradition, as Bede, our only detailed witness, would want
to make us believe. Quite the contrary, ‘Romans’, in Ireland, Anglo-Sax-
on England, Visigothic Spain, Frankish Gaul, Italy, and Rome, debated
whether Victorius or the Alexandrian reckoning in the shape of Dio-
nysius Exiguus’ translation was to be followed. Masako Ohashi has for
years, since her 1999 Nagoya Ph.D. thesis, insisted on the centrality of
this issue, and more recent publications of 2015 by E. T. Dailey and my-
self emphasize this argument more forcefully.74 The Easter controversy
of the early Middle Ages needs to be rewritten on this basis.
Alden Mosshammer also leads the way in the reception of the Al-
exandrian reckoning in the Latin West. The Alexandrian system of cal-
culating Easter was formalized in the late third century. Until the mid-
seventh century, Rome followed different practices, first 84-year tables,
then Victorius of Aquitaine. The controversy between Rome and Alex-
andria is well documented especially for the fifth-century, but for the
sixth century comparable insight is lacking. The earliest reception of the
Alexandrian reckoning in the Latin West in the fifth and sixth centu-
ries is directly connected to a vaguely defined corpus of texts that bears

73
Krusch (1880), 298–302. Only one manuscript transmitting the full letter was
known to Krusch, Cologne, Dombibliothek 83-II, 184r–185v; a second copy has come
to light in Bremen, Universitätsbibliothek, msc 0046, 41r–44r. A new edition of this
central monument of the Iberian Easter controversy is currently prepared by José Carlos
Martín Iglesias as ‘La Epistola de computo paschali (CPL 2300) del monje León: nueva
edición y estudio de una obra probablemente hispano-visigoda’.
74
See especially Ohashi, M. (2005) ‘Theory and history: an interpretation of the
paschal controversy in Bede’s Historia ecclesiastica,’ in S. Lebecq, M. Perrin, and O. Szer-
winiack, Bède le Vénérable: entre tradition et postérité, Lille, 177–85; Ohashi, M. (2011)
‘The Easter table of Victorius of Aquitaine in early medieval England,’ in Warntjes and
Ó Cróinín (2011), 48–75; Dailey, E. T. (2015) ‘To choose one Easter from three: Os-
wiu’s decision and the Northumbrian synod of AD 664,’ Peritia 26, 47–64; Warntjes, I.
(2015) ‘Victorius vs Dionysius: the Irish Easter controversy of AD 689,’ in P. Moran and
I. Warntjes, Early medieval Ireland and Europe: chronology, contacts, scholarship, Turn-
hout, 33–97.


 IMMO WARNTJES

the name of the 17th-century Jesuit scholar Jacques Sirmond.75 Many of

these texts have mistakenly been labeled Irish forgeries of the seventh
century, but more recent research places them more securely in their pre-
dominantly fifth- and sixth-century context. Dan Mc Carthy and Aiden
Breen produced in 2003 a new edition and translation of De ratione
paschali ascribed to Anatolius of Laodicea, which they consider to be
a faithful Latin translation of Anatolius’ third-century original tract.76
Jan Zuidhoek analyses, in the present volume, the potential base year of
the Easter table incorporated in this text. Rick Graff reconsidered the
Disputatio Morini in the proceedings of the first Galway conference in
2010.77 Alden Mosshammer analysed the Prologus Cyrilli and its recen-
sions in Vigiliae Christianae in 2013, and a new edition of this text and
of the Prologus Theophili by him has just been published.78 Only critical
editions of all such texts will enable a solid understanding of the Easter
question in the transition period from Late Antiquity to the early Mid-
dle Ages.
Iberia and Italy were the first regions in the Latin West to embrace the
Alexandrian reckoning. The recensions of the Prologus Cyrilli analysed
by Mosshammer are transmitted principally in Visigothic manuscripts
or those with Visigothic connections. He has followed this up by an edi-
tion of three computistical texts from Visigothic Spain to be published
in the following Galway proceedings, and I will present an overview of
sixth- to eighth-century Visigothic computistics in the same volume.
The picture that emerges is one of discussion, debate, and conflict be-
tween the Victorian and the Alexandrian reckonings, which was solved
in favour of Alexandria by the AD 640s (though, in the present volume,
Brigitte Englisch suggests, rather controversially, that the Visigoths fol-

75
For the Sirmond corpus, see Jones, C. W. (1937) ‘The ‘lost’ Sirmond manuscript
of Bede’s Computus,’ English Historical Review 52, 205–19, repr. in C. W. Jones, Bede,
the schools and the computus, ed. by Wesley M. Stevens, Aldershot 1993, article X (omit-
ting the crucial final pages); Jones (1943), 105–13; Ó Cróinín, D. (1983) ‘The Irish
provenance of Bede’s computus,’ Peritia 2, 229–47, repr. in Ó Cróinín (2003), 173–90;
Ó Cróinín, D. (2003) ‘Bede’s Irish computus,’ in Ó Cróinín (2003), 201–12; Wallis
(1999), lxxii–lxxix; Springsfeld (2002), 64–80; Warntjes (2011).
76
Mc Carthy, D. P. and A. Breen (2003) The ante-Nicene Christian Pasch: De ra-
tione paschali – The Paschal tract of Anatolius, bishop of Laodicea, Dublin.
77
Graff, E. (2010) ‘The recensions of two Sirmond texts: Disputatio Morini and
De divisionibus temporum,’ in Warntjes and Ó Cróinín (2010), 112–42.
78
Mosshammer, A. A. (2013) ‘The Praefatio (Prologus) sancti Cyrilli de Paschate
and the 437-year (not 418!) paschal list attributed to Theophilus,’ Vigiliae Christianae
67, 49–78; Mosshammer, A. A. (2017) The Prologues on Easter of Theophilus of Alexan-
dria and [Cyril], Oxford.


 STATE OF RESEARCH ON COMPUTUS

lowed their own system based on actual lunar observation). Isidore’s cu-
rious Alexandrian Easter table remains one of the central avenues not
only into Visigothic computistics, but also into the transmission of the
Etymologiae as a whole.79 From a manuscript perspective, the mid-ninth
century Paris, Bibliothèque nationale de France, Lat. 609, certainly is our
key witness for the early period in Spain, and Alden Mosshammer is cur-
rently preparing an edition of its oldest, seventh-century layer. A new,
collaborative edition of the famous Antiphonary of León will shed more
light on the various phases of Visigothic computistics right up to the
eleventh century, and the Visigothic tradition will certainly be one of
the most fruitful areas of computistical studies in the future.
In Italy, the Alexandrian reckoning made an early appearance in a
letter of Ambrose of Milan written in AD 386. Its authenticity has long
been questioned, but Max Lejbowicz, in his analysis of the computis-
tical importance of the letter presented at the first Galway conference,
considers it authentic.80 The Alexandrian reckoning was popularized in
AD 525 by Dionysius Exiguus’ translation from Greek into Latin of the
table and accompanying explanatory material. The Alexandrian Easter
table has a 532-year cyclic structure. However, it was first designed as
a 95-year table, in which every fourth year (those designated bissextile)
had to be recalculated in order to update it to the immediately following
95-year period. Dionysius’ table of AD 532 to 626 followed from Cyril’s
of AD 437 to 531. The expiration of Dionysius’ table—and therefore
79
I have two articles on Isidore’s Easter table in preparation; the first will be pub-
lished in Revue d’histoire des textes, n.s. 13 (2018) as ‘The continuation of the Alexandri-
an Easter table in seventh-century Iberia and its transmission to ninth-century Francia
(Isidore, Etymologiae 6.17)’. For the reception of Isidore in seventh- and eighth-century
cosmology and computistics, see the studies of Marina Smyth: Smyth, M. (1987) ‘Isidore
of Seville and early Irish cosmography,’ Cambridge Medieval Celtic Studies 14 (1987),
69–102; Smyth, M. (1996) Understanding the universe in seventh-century Ireland,
Woodbridge; Smyth, M. (2015) ‘Isidorian texts in seventh-century Ireland,’ in A.T. Fear
and J. Wood, Isidore of Seville and his reception in the early Middle Ages: transmitting and
transforming knowledge, Amsterdam, 111–30.
80
Lejbowicz, M. (2008) ‘Une étape contournée dans l’unification des pratiques
computistes médiévales latines,’ Cahiers de recherches médiévales et humanistes 15,
277–305; Lejbowicz, M. (2010) ‘Les pâques baptismales d’Augustine d’Hippone: une
étape contournée dans l’unification des pratiques computistes latines,’ in Warntjes and Ó
Cróinín (2010), 1–39. See also his survey of the development and cultural importance
of computus up to AD  1200 in Lejbowicz, M. (2006) ‘Des tables pascales aux tables
astronomique et retour,’ Methodos 6, 1–67, as well as his earlier Lejbowicz, M. (1992)
‘Computus. Le nombre et le temps altimédiévaux,’ in B. Ribémont, Le temps, sa mesure
et sa perception au moyen âge, Caen, 151–96. For Ambrose’s paschal letter, Lejbowicz’s
study builds on Zelzer, M. (1978) ‘Zum Osterfestbrief des heiligen Ambrosius und zur
römischen Osterfestberechnung des 4. Jahrhunderts,’ Wiener Studien 91, 187–204.


 IMMO WARNTJES

its necessary continuation in AD  626—was a crucial moment in the

history of early medieval computus, especially since the papal curia had
not yet embraced the Alexandrian system. In Italy, this re-calibration of
Dionysius’ table is ascribed to a certain Felix, and Luciana Cuppo’s most
recent research focussed strongly on this key episode (both in the pre-
sent volume and in the previous conference proceedings).81 Essential for
this task of re-calculating the data of the Easter table was the set of nine
argumenta or calendrical algorithms which formed part of Dionysius’
explanatory material. Before the eighth century, these argumenta were
re-calibrated in AD  562 (in the circle of Cassiodorus), and expanded
and added to in AD 581, 625, 675, 689, and 695. This presents the only
traceable direct and continuous line of transmission of calendrical sci-
ence from Late antiquity into the early Middle Ages. I  have sketched
this development in the first proceedings of the Galway conference and
I hope to provide an edition with translation and commentary of all of
these texts shortly.82 In the eighth century, this genre of the computis-
tical formulary (defined as a collection of principally calendrical algo-
rithms) mushroomed and became one of the central features of Caro-
lingian s­ cience.
An equally important, third genre of computistical texts (after tables
and formularies) was invented in the second half of the seventh-century
in Ireland: the textbook. The process itself still needs to be studied in
detail, but the reception of Isidore’s writings and the quest for under-
standing the liturgical calendar and God’s creation as manifested in the
cosmos appear to have been the key triggers. Dáibhí Ó Cróinín led the
way here with his edition of De ratione conputandi in 1988. This was fol-
lowed by the edition of the Munich Computus in 2010, and a third full-
scale textbook, discovered in 2006, still awaits publication (the Compu-
tus Einsidlensis).83 With these three major texts at hand, it is possible to

81
Cuppo, L. (2011) ‘Felix of Squillace and the Dionysiac computus I: Bobbio
and Northern Italy (MS Ambrosiana H 150 inf.),’ in Warntjes and Ó Cróinín (2011),
110–36.
82
Warntjes, I. (2010) ‘The Argumenta of Dionysius Exiguus and their early recen-
sions,’ in Warntjes and Ó Cróinín (2010), 40–111; Warntjes, I. (2011) ‘The Computus
Cottonianus of AD 689: a computistical formulary written for Willibrord’s Frisian mis-
sion,’ in Warntjes and Ó Cróinín (2011), 173–212.
83
Warntjes, I. (2010a) The Munich Computus: text and translation. Irish computis-
tics between Isidore of Seville and the Venerable Bede and its reception in Carolingian times,
Stuttgart; Warntjes, I. (2005) ‘A newly discovered Irish computus: Computus Einsidlen-
sis,’ Peritia 19, 61–64; Bisagni, J. and I. Warntjes (2008) ‘The Early Old Irish material in
the newly discovered Computus Einsidlensis (c.AD 700),’ Ériu 58, 77–105.


 STATE OF RESEARCH ON COMPUTUS

define the Irish contribution to this calendrical science and to identify

further treatises that help to reconstruct the thought-world of the Irish
‘Golden Age’, c.AD  650–750. Particularly intriguing are the tract De
comparatione epactarum of AD 689 (which draws attention to the fact
that the systems ascribed to Victorius and Dionysius are irreconcilable),
the Victorian Prologue of AD  699 (paralleling the three major linear
time-lines, annus mundi, annus passionis, and AD), and the eclipse pre-
diction of AD 754, all discovered since 2006.84 But our improved knowl-
edge of early medieval Irish computistics also makes it possible to clearly
assess the Irish contribution to the Carolingian educational reform.85
Ultimately, this will provide the context for Dúngal and Dicuil, two of
the most important Irish peregrini of the Carolingian age. Werner Berg-
mann drew attention once more to the importance of Dicuil’s Liber de
astronomia in the proceedings of the first Galway conference,86 and a
critical edition of the text remains one of the main desiderata not only in
the study of early medieval computus, but also of the Carolingian intel-
lectual endeavour.
Scholarship on Bede’s computistica has flourished over the past dec-
ade principally due to the translations by Faith Wallis and the combined
efforts of Máirín MacCarron and Peter Darby, who, since 2011, have
held sessions on Bede annually at the major medievalists’ congresses in
Kalamazoo and Leeds (with a thorough eye on Bede’s contribution to
early medieval scientifica).87 There is an apparent shift in Bedan research,
systematically analyzing Bede’s oeuvre according to certain themes. Dar-
by applied this approach to themes relating to time, with a 2012 mono-
graph on Bede and the end of time, and a 2014 volume of collected essays
on Bede and the future (ed. together with Faith Wallis).88 This turn from
84
Warntjes (2010a), CLII–CLVIII, 322–26, 328; Warntjes, I. (2010) ‘A newly
discovered prologue of AD  699 to the Easter table of Victorius of Aquitaine in an
unknown Sirmond manuscript,’ Peritia 21, 255–84; Warntjes, I. (2013–14) ‘An Irish
eclipse prediction of AD 754: the earliest in the Latin West,’ Peritia 24–25, 108–15.
85
Cf.  Warntjes, I. (2013) ‘Seventh-century Ireland: the cradle of medieval sci-
ence?,’ in M. Kelly and C. Doherty, Music and the stars: mathematics in medieval Ireland,
Dublin, 44–72; Warntjes, I. (2016) ‘Computus as scientific thought in Ireland and the
early medieval West,’ in R. Flechner and S. Meeder, The Irish in early medieval Europe:
identity, culture and religion, New York, 158–78.
86
Bergmann, W. (2011) ‘Dicuils Osterfestalgorithmus im Liber de astronomia,’ in
Warntjes and Ó Cróinín (2011), 242–87.
87
See their website: bedenet.com.
88
Darby, P. (2012) Bede and the end of times, Farnham; Darby, P. and F.  Wallis
(2014) Bede and the future, Farnham, with essays by J. T Palmer and M.  MacCarron
on Bede’s computistical texts. See also von den Brincken, A.-D. (2006) ‘Jahrtausend­


 IMMO WARNTJES

focusing on individual works to the entire œuvre of an author is very

laudable, but from an historian’s point of view, obviously still quite lim-
ited. Bede, like any other author, was the product of his time. It was only
towards the end of the eighth century that he became standard reading,
as Joshua Westgard’s recent work on the manuscript transmission of Be-
de’s texts has highlighted again.89 Throughout the eighth century, he was
only one of many computists, many anonymous, but with equally inter-
esting (and, in some cases, more advanced) ideas. MacCarron analyzed
the Irish background to Bede’s computistical texts in 2015,90 and more
such studies are needed in the future, focusing on Bede’s background
and his immediate, eighth-century impact.91
This brings us right into the Carolingian age. Since the publication
of Schriften, Borst’s methodology and main thesis has been under scru-
tiny. The late Hartmut Hoffmann questioned the corpus itself, arguing
that the sermon Arn Serm. (the eleventh text of the volume) should
not, strictly speaking, be considered computistical.92 Most other critical
voices questioned Borst’s theory of the study (and teaching) of compu-
tus being a heavily centralized project at the heart of the Carolingian
educational reform, personified by Alcuin and Charlemagne.93 Accord-

rechnung, christliche Ära und Eschatologie bei Beda Venerabilis,’ Sitzungberichte der
geisteswissenschaftlichen Klasse der Akademie gemeinnütziger Wissenschaften zu Erfurt 5,
11–26.
89
Westgard, J. A. (2010) ‘Bede and the continent in the Carolingian age and be-
yond,’ in S. De Gregorio, The Cambridge companion to Bede, Cambridge, 201–15. See
also Faith Wallis’s ‘Bede and science’ in the same volume, pp. 113–26.
90
MacCarron, M. (2015) ‘Bede, Irish computistica and annus mundi,’ Early Medi-
eval Europe 23, 290–307.
91
The seventh- and eighth-century context is missing or is only briefly referred to
in the studies of Francisca Plaza Picón and Antonio Gonzáles Marrero: del Mar Plaza
Picón, F. and J. A. González Marrero (2004) ‘El vocabulario del cómputo en el De tem-
poribus liber de Beda,’ Minerva 17, 125–37; del Mar Plaza Picón, F. and J. A. González
Marrero (2006) ‘De computo uel loquela digitorum: Beda y el cómputo digital,’ Faventia
28, 115–23; del Mar Plaza Picón, F. and J. A. González Marrero (2006) ‘Un acercami-
ento a los tratados del cómputo de Beda,’ Fortunatae 117–25; del Mar Plaza Picón, F. and
J. A. González Marrero (2011) ‘La Epistola ad Wicthedum: un apéndice del De tempo-
rum ratione de Beda,’ in J. M. Gázquez, Ó. de la Cruz Palma, and C. Ferrero Hernández,
Estudios de latín medieval hispánico, Firenze, 579–88. Worth consulting is the Fribourg
Ph.D. thesis by R.-P. Pillonel-Wyrsch, published in 2004, which is a commentary to De
temporum ratione: Le calcul de la date de pâques au moyen âge: analyse et commentaires sur
de temporum ratione de Bède, Fribourg.
92
Hoffmann, H. (2012) ‘Abisag calefaciente oder Der karolingische Traktat De sole
et luna,’ Deutsches Archiv für Erforschung des Mittelalters 68, 445–78.
93
Cf. Dobcheva, I. (2013) ‘The umbrella of Carolingian computus,’ in M. J. Mu-
ñoz Jiménez, P.  Cañizares Ferriz, and C.  Martin, La compilación del saber en la edad


 STATE OF RESEARCH ON COMPUTUS

ing to Borst, a canon of computistical knowledge was set, first in the

key monastic centres, then in the Aachen palace school, from where it
radiated out to the more peripheral educational institutions (monastic
and cathedral schools). This approach is reflected in the editions prop-
er: Especially for the Cologne formulary of AD  760/92 (Lect. comp.)
and the three Carolingian encyclopediae of AD 793, 809/12, 817 (Lib.
ann., Lib. comp., Lib. calc.), Borst listed and used numerous manuscripts
which contain only very few of the passages / algorithms of these texts.
This creates the misleading impression that those passages were directly
copied from these, allegedly normative encyclopediae. Quite the contra-
ry, those concepts and algorithms circulated widely in Western Europe,
they were common knowledge and therefore did not need the push by a
centralized authority. More importantly, Borst stripped individual pas-
sages of their immediate context, and therewith clouded the fact that
they belonged to separate cohesive texts in their own right, like the
Computus of AD 757, the Computus Rhenanus of AD 776, the unfin-
ished Computus of AD 789, etc.
These texts will be properly introduced into scholarship in a sepa-
rate volume on Carolingian computistics arising out of the last Galway
conference. Dating clauses are a good indicator of the vibrancy of calen-
drical science in the early Middle Ages, but the vast number of them
preserved in hundreds of manuscripts of especially the ninth and tenth
centuries still needs to be examined and systematized (which, no doubt,
can only be achieve as a big project with considerable (wo)manpower).
Ninth-century computistics have therefore been approached from other
perspectives.
The historians of science have focused specifically on diagrams.
Most prominent in this area is the scholarship of Bruce Eastwood, who,
since the 1980s, has analysed the planetary diagrams occurring in the
Carolingian commentaries to Pliny, Macrobius, Martianus Capella, and
Calcidius. His collected essays appeared in the Variorum series in 2003,
and in the following year he published, together with Gerd Grasshoff,
a slim but influential volume which neatly outlines the manuscript evi-

media, Porto, 211–30; Palmer, J. T. (2011) ‘Computus after the paschal controversy of
AD 740,’ in Warntjes and Ó Cróinín (2011), 213–41; Palmer, J. T. (2011) ‘Calculat-
ing time and the end of time in the Carolingian world, c. 740–820,’ English Historical
Review 126, 1307–31; Warntjes, I. (2012) ‘Köln als naturwissenschaftliches Zentrum in
der Karolingerzeit: die frühmittelalterliche Kölner Schule und der Beginn der fränkis-
chen Komputistik,’ in H. Finger and H. Horst, Mittelalterliche Handschriften der Kölner
Dombibliothek, Viertes Symposion, Köln, 41–96.


 IMMO WARNTJES

dence.94 In 2007, i.e. in the period since 2006 that primarily interests us
here, Eastwood produced a wonderfully detailed monograph in Order-
ing the heavens, effectively harvesting the fruits of 30 years of research.95
The book may not make for light reading for the faint-hearted, but it
provides Carolingian scholars with an indispensable guide to planetary
theory. In 2011, both Eastwood and Stephen McCluskey, the doyens
of the study of early medieval astronomy, contributed to a fine collec-
tion of essays on the Carolingian reception of Martianus Capella (ed. by
Mariken Teeuwen and Sinéad O’Sullivan).96
The second approach to this material is an art historical one. The
Carolingian period continues to fascinate scholars and the interested
public alike also because of its lavish depictions of constellation figures
and celestial phenomena. The pioneering study here probably is Fritz
Saxl’s 1915 Verzeichnis astrologischer und mythologischer illustrierter
Handschriften des lateinischen Mittelalters. More recently, in 2003, Bi-
anca Kühnel provided a sweeping survey in The end of times in the order
of things,97 and, in the last few years of his productive life, since 2005,
the Cologne art historian Anton von Euw turned his interest to com-
putistical manuscripts.98 Chief among the early medieval representation
of constellation figures is the so-called Aratus Latinus, to which Floren-

94
Eastwood, B. S. (2002) The revival of planetary astronomy in Carolingian and
post-Carolingian Europe, Aldershot; Eastwood, B. S. and G. Graßhoff (2004) Planetary
diagrams for Roman astronomy in medieval Europe ca. 800–1500, Philadelphia.
95
Eastwood, B. S. (2007) Ordering the heavens: Roman astronomy and cosmology
in the Carolingian Renaissance, Leiden.
96
Eastwood, B.  S. (2011) ‘The power of diagrams: the place of the anonymous
commentary in the development of Carolingian astronomy and cosmology,’ in M. Teeu-
wen and S.  O’Sullivan, Carolingian scholarship and Martianus Capella: ninth-century
commentary traditions on De nuptiis in context, Turnhout, 193–220; McCluskey, S. C.
(2011) ‘Martianus and the traditions of early medieval astronomies,’ in ibidem, 221–44.
97
Kühnel, B. (2003) The end of time in the order of things: science and eschatology in
early medieval art, Regensburg.
98
Early studies, which presumably started his interest in astronomical depic-
tions, are his descriptions of Cologne, Dombibliothek, 83-II and 103 in Plotzek,
J. M. and U. Surmann (1998) Glaube und Wissen im Mittelalter, Die Kölner Dombib-
liothek, München, 129–56. See von Euw, A. (2005) ‘Astronomie und Zeitrechnung im
Karolingerreich,’ in H. Finger, Mittelalterliche Handschriften der Kölner Dombibliothek:
Erstes Symposium, Köln, 21–64; von Euw, A. (2006) ‘Artes liberales und artes technicae
im Spiegel der antiken, früh- und hochmittelalterlichen Handschriftenüberlieferung,’ in
C. Stiegemann and M. Wemhoff, Canossa 1077: Erschütterung der Welt, 2 vols, München,
i 544–54; von Euw, A. (2010) ‘Alkuin als Lehrer der Komputistik und Rhetorik Karls
des Großen im Spiegel der St Galler Handschriften,’ in E. Tremp and K. Schmuki, Alkuin
von York und die geistige Grundlegung Europas, St Gallen, 251–62.


 STATE OF RESEARCH ON COMPUTUS

tine Mütherich drew renewed attention in the late 1970s.99 It survives in

various recensions in numerous manuscripts, and Ivana Dobcheva is cur-
rently preparing a full-scale study of these for her Ph.D. Dieter Blume,
Mechthild Haffner, and Wolfgang Metzger have supplied, in 2012, a
massive catalogue of 68 manuscripts containing depictions of the 42
constellations.100 Hot off the press is Eric Ramírez-Weaver’s A saving sci-
ence, which analyses the pictorial program in the Libri Computi of 809 in
its political, ecclesiastical, intellectual, and artistic context.101
The third approach to the ninth century lies in detailed analyses of
selected manuscripts, particularly those attributed to famous Carolin-
gian intellectuals. Most prominent among these (beside the art histori-
cal study by Ramírez-Weaver on the famous codex Madrid, Biblioteca
Nacional, 3307 dedicated to Drogo of Metz) are the monograph-length
studies on Walahfrid Strabo’s Vademecum (St  Gall, Stiftsbibliothek,
878) and related manuscripts. Wesley Stevens has worked on the subject
since the 1980s, and his Rhetoric and reckoning in the ninth century: The
vademecum of Walahfrid Strabo is due to appear this year.102 The idea of
connecting the various parts of the manuscript to the biography of its
compiler is an excellent one, though the details will surely prove contro-
versial. In 2014, Richard Corradini submitted a monumental Habilita-
tionsschrift to the University of Vienna, entitled Karolingische Gelehr-
samkeit und Zeitforschung im Kompendium des Walahfrid Strabo, which
will also appear shortly in two volumes.
Very recently, Jacopo Bisagni in Galway has drawn attention to the
potential of studying the transmission of computistical ideas from Ire-
land through Brittany to the western Carolingian Empire.103 Method-
ologically, the focus lies principally on Breton and Loire valley manu-

99
Mütherich, F. (1989) ‘Die Bilder des Leidener Aratus,’ in B.  Bischoff et  al.,
Aratea: Kommentar zum Aratus des Germanicus Ms.  Voss. Lat.  Q. 79, Bibliotheek der
Rijksuniversiteit Leiden, Luzern, 31–68, repr. in F. Mütherich, Studies in Carolingian
manuscript illumination, London 2004, 147–265.
100
Blume, D., M. Haffner, and W. Metzger (2012) Sternbilder des Mittelalters: der
gemalte Himmel zwischen Wissenschaft und Phantasie, Bd. 1: 800–1200, 2 vols, Berlin.
Unfortunately, this 2-volume work is neither up to date nor affordable for the interested
individual.
101
Ramírez-Weaver, E. M. (2017) A saving science: capturing the heavens in Carolin-
gian manuscripts, University Park.
102
Stevens, W. M. (2017) Rhetoric and reckoning in the ninth century: the vademe-
cum of Walahfrid Strabo, Turnhout.
103
As a starting point, see Bisagni, J. (2014) ‘A new citation from a work of Colum-
banus in BNF Lat. 6400B,’ Peritia 24–25, 116–22.


 IMMO WARNTJES

scripts showing unambiguous Breton features (Old Breton glosses). This

project is in its infancy, but it will eventually shed very bright new light
on eighth-century Breton culture. As a side product, this research draws
attention to neglected texts crucial for the understanding of the scien-
tific strand of the Carolingian educational reform movement.104 It needs
to be remembered that Borst’s corpus breaks off at AD 818. It is far from
complete up to this point, however, and for the remaining eight decades
of the ninth century the manuscripts need to be mined systematically
for cohesive texts. Two impressive cases to illustrate this are the ency-
clopediae preserved in Laon, Bibliothèque municipale, 422 and Vatican,
Biblioteca Apostolica, Reg. lat. 123.105 Both appear to have been pro-
duced in the mid-ninth century in the Loire valley, and they are prime
witnesses of how the Carolingian scientific endeavour progressed from
the celebrated three encyclopedias of AD 793, 809/12, 818 published
by Borst to these hardly known, but much more structured and more
sophisticated texts. Certainly, a study of these and, no doubt, other com-
parable collections will shed much more light on the Carolingian inven-
tion of the scientific encyclopedia and its development in the slowly but
surely disintegrating Empire. This also points to what I would consider
the biggest desideratum in the study of the mid- to late ninth century:
detailed studies of key (calendrical) scientific centres or Wissenschafts-
landschaften, like St Gall and Fleury on the one hand, the Lake Constanz
region and the Loire valley on the other.
If the second half of the ninth century is massively understudied, this
applies even more so to the tenth century, at least to its first seven dec-
ades. One of the most promising ways of approaching this century will
be through the Liber de computo of Helperic, which was composed in
AD 900/03, but then constantly updated right into the eleventh century
and beyond. The transmission of this text and the manuscript contexts
in which it appears (Helperic’s computus is preserved in over 80 manu-

104
Besides the two massive works discussed in the following, see also especially the
neglected scientific encyclopedia in Cologne, Dombibliothek, 83-II, 15r–36v, which is
currently also being studied by Bisagni.
105
On both, Bisagni has articles in preparation. For discussion of some of the texts
and diagrams in those two manuscripts, see Obrist, B. (2002) ‘Les manuscrits du De
cursu stellarum de Grégoire de Tours et le MS Laon, Bibliothèque municipale 422,’ Scrip-
torium 56, 335–44; Castiñeiras González, M.  A. (1998) ‘Diagramas y esquemas cos-
mográficos en dos misceláneas de cómputo y astronomía de la Abadía de Santa María de
Ripoll (ss. XI–XII),’ Compostellanum 43, 593–646; Martínez Gázquez, J. and J. Gómez
Pallarès (1992) ‘La Epistola de ciclo paschali del monje Oliba de Ripoll,’ Mittellateinis-
ches Jahrbuch 27, 103–40.


 STATE OF RESEARCH ON COMPUTUS

scripts), will shed substantial light on the computitistical tradition of

the two centuries after its original composition. This fundamental text
still awaits a proper introduction into scholarship by means of a modern
critical edition, which hopefully will reverse the negligence of the past
40 years which saw no publication on this crucial textbook.106
One of the updates of Helperic’s text was produced by Abbo of
Fleury in AD 978. More importantly, Abbo designed his own computus,
a new edition of which will be published by Alfred Lohr shortly. Abbo’s
computus, best preserved in Berlin, Staatsbibliothek, Phillipps 1833, sig-
nals a major change in the presentation of computistical material. When
computus turned into a major subject in monastic education in the sev-
enth century, it was almost exclusively text-based (with the exception of
the Easter tables proper, of course). In the Carolingian era, some of the
algorithms were pressed into table format, while diagrams helped to ex-
plain more complex concepts. Abbo’s computus, then, was almost exclu-
sively based on tables and diagrams, the textual element was minimized.
Whether Abbo can be credited with this revolution in the presentation
of computistical material can only be determined by a systematic analy-
sis of tenth-century manuscripts. Certainly, the tenth century has much
more to offer than the current neglect of this period suggests, and Nadja
Germann’s 2006 study of Abbo’s scientific mindset, together with the
excellent volume on Abbo edited by Barbara Obrist two years earlier, are
only a first step towards a better understanding of this crucial century.107
Abbo certainly set a trend by questioning the incarnation era. One
of the reasons why the Easter controversy lasted well into Charlemagne’s
reign was the fact that the Alexandrian reckoning in Dionysius’ transla-
tion was intrinsically connected to AD. AD marked one of the columns

106
Helperic’s Liber de computo was edited by Bernard Pez in 1721 (repr. in PL 137,
17–48), on the basis of one, late witness (of AD 1090!). A list of manuscripts can be found
in Jullien, M.-H. (2010) Clavis scriptorum Latinorum medii aevi: auctores Galliae 735–
987, vol. 3, Turnhout, 421–29. Key studies are: Traube, L. (1893) ‘Computus Helperici,’
Neues Archiv der Gesellschaft für ältere deutsche Geschichtskunde 18, 73–105, repr. with ad-
ditional notes in Traube (1909–20), iii 128–56; McGurk, P. (1974) ‘Computus Helperici:
its transmission in England in the eleventh and twelfth century,’ Medium Aevum 43, 1–5.
107
Germann, N. (2010) De temporum ratione: quadrivium und Gotteserkenntnis am
Beispiel Abbos von Fleury und Hermanns von Reichenau, Leiden. Obrist, B. (2004) Ab-
bon de Fleury: philosophie, science et comput autour de l’an mil, Paris. See also the articles
by Germann (‘A la recherche de la structure du temps: Abbon de Fleury et le comput’)
and Obrist (‘Abbon de Fleury: cosmologie, comput, philosophie’) in Dufour-Malbezin,
A. (2008) Abbon, un abbé de l’an mil, Turnhout. For an earlier study, see Engelen, E.-M.
(1993) Zeit, Zahl und Bild: Studien zur Verbindung von Philosophie und Wissenschaft bei
Abbo von Fleury, Berlin.


 IMMO WARNTJES

of Dionysius’ Easter table, and this column provided its most conten-
tious datum. Dionysius’ original table covered the years AD 532–626,
but there was enough expertise in the seventh and eighth centuries not
only to extend that table, but also to calculate its data back to the time
of Christ. The Gospels suggest that Christ died in either the 31st or
34th year of his life.108 The Dionysiac reckoning provided lunar data for
both years which could not be reconciled with the explicit information
found in the Gospels. This made the Dionysiac reckoning theologically
problematic at best. There are interesting seventh-century attempts to
solve this obvious problem, and Bede tried to dismiss it by a blunt ref-
erence to authority. The question re-emerged, however, in Abbo’s days.
Between the AD 990s and 1135, at least eight attempts at recalculating
the incarnation era prove the seriousness of the issue. These have been
thoroughly discussed by Peter Verbist in his 2003 Ph.D. thesis, which
appeared in English translation (with financial help from the Arno-
Borst-Foundation) in 2010 (following in the footsteps of van de Vyver,
von den Brinken, and Wiesenbach, discussed above).109 Philipp Nothaft
corrected Verbist’s account of Marianus Scottus’ recalculation in a semi-
nal article of 2013, and he proved that Marianus’ ideas were still being
discussed in the early 14th century.110 For the eighth and last of these re-
calculations of the incarnation era by Heimo of Bamberg, Verbist based
his study on the later, longer of two versions of the text, and worked
directly from the manuscript. As mentioned above, between Verbist’s
Ph.D. of 2003 and the English translation of 2009, Hans Martin Weik-
mann published a critical edition of both versions in parallel, which will
have to form the basis of all subsequent studies. In fact, the text discov-
ered by Luciana Cuppo and published in the present volume, appears to
represent a ninth attempt at recalculating the incarnation, rather than
belonging to the seventh century as she argues.
Abbo also set a trend in Anglo-Saxon England. Little is presently
known about the Anglo-Saxon computistical tradition between Bede
and the Benedictine reform movement, and this certainly would make
108
For the wider context, see Nothaft, C. P. E. (2012) Dating the Passion: the life of
Jesus and the emergence of scientific chronology (200–1600), Leiden.
109
Verbist, P. (2010) Duelling with the past: medieval authors and the problem of the
Christian era (c. 990–1135), Turnhout.
110
Nothaft, C. P. E. (2013) ‘An eleventh-century chronologer at work: Marianus
Scottus and the quest for the missing twenty-two years,’ Speculum 88, 457–82; Nothaft,
C. P. E. (2012) ‘Nicholas Trevet and the chronology of the crucifixion,’ The Mediaeval
Journal 2 (2012), 37–51. See also Baran-Kozłowski, W. (2009) Kronika świata, Mariana
Szkota: studium źródłoznawcze, Poznań.


 STATE OF RESEARCH ON COMPUTUS

for an excellent study for more enterprising students or scholars. Tra-

ditionally, most of what has been published on late Anglo-Saxon com-
putus follows the practice inherent in English Studies of focusing on
prominent authors, here two scholars influenced by Abbo, Ælfric and
Byrhtferth. A new edition of Ælfric’s De temporibus anni was published
by Martin Blake in 2009.111 In 2011, Ælfric was the theme of the XII
Seminario avanzato in filologia germanica, and Giuseppe de Bonis pre-
pared an article on Ælfric’s computus for the subsequent proceedings.112
Byrhtferth has been more prominently in the spotlight of recent scholar-
ship. This is in part due to rekindled interest in the glosses to Bede’s com-
putistica ascribed to his name, which John Contreni discussed in Galway
in 2012.113 Even more attention has been given to Byrhtferth’s Enchiri­
dion, especially its diagrams. Philippa Semper in 2004 and Peter Baker in
the 2005 Festschrift for Michael Lapidge reopened discussion,114 which
was continued by Francesca Chiusaroli and Cristina Raffaghello in Ita-
ly.115 The last Galway conference in 2016 saw a strong panel on Byrht-
ferth with papers by Rebecca Stephenson and Sabine Rauch.116

111
Blake, M. (2009) Ælfric’s De temporibus anni, London.
112
De Bonis, G. D. (2012) ‘Il De temporibus anni di Ælfric: una rielaborazione con-
tenutistica e sintattica dell’ opera De temporibus di Beda,’ in V.  Dolcetti Corazza and
R. Gendre, Lettura di Ælfric, Alessandria, 347–88. For Ælfric, see also Chardonnens,
L. S. (2013) ‘Ælfric and the authorship of the Old English De diebus malis,’ in C. Gili­
berto and L. Teresi, Limits to learning: the transfer of encyclopaedic knowledge in the early
Middle Ages, Leuven, 123–53.
113
Contreni, J.  J. (2011–12) ‘‘Old orthodoxies die hard’: Herwagen’s Bridferti
Ramesiensis Glossae,’ Peritia 22–23, 15–52. Contreni’s recent scholarship on the subject
is a reaction to earlier articles by Lapidge and Gorman: Lapidge, M. (2007) ‘Byrhtferth
of Ramsey and the Glossae Bridferti in Bedam,’ Journal of Medieval Latin 17, 384–400;
Gorman, M. (1996) ‘The glosses on Bede’s De temporum ratione attributed to Byrhtferth
of Ramsey,’ Anglo-Saxon England 25, 209–32, repr. in M. Gorman, Biblical commentar-
ies from the early Middle Ages, Firenze 2002, 175–98.
114
Semper, P. (2004) ‘Doctrine and diagrams: maintaining the order of the world
in Byrhtferth’s Enchiridion,’ in P. R. Cavill, Christian tradition in Anglo-Saxon England:
approaches to current scholarship and teaching, Woodbridge, 121–37. Baker, P. S. (2005)
‘More diagrams by Byrhtferth of Ramsey,’ in K. O’Brien O’Keeffe and A. Orchard, Latin
learning and English lore, 2 vols, Toronto, ii 53–73.
115
Chiusaroli, F. (2005) ‘Costituzione e impiego del lessico tecnico nell’Enchiridion
di Byrhtferth: l’ambito dell’astronomia,’ in D. Gottschall, Testi cosmografici, geografici e
odeporici del medioevo germanico, Louvain-la-Neuve, 1–40; Raffaghello, C. (2009) ‘Il
computo del tempo in ambiente anglosassone: l’Enchiridion de Byrhtferth di Ramsey,’
in L. Vezzosi, La letteratura tecnico-scientifica nel medioevo germanico: Fachliteratur e Ge-
brauchstexte, Allessandria, 213–36.
116
See also Stephenson, R. (2012) ‘Byrhtferth’s Enchiridion: the effectiveness of
hermeneutic Latin,’ in E. Tyler, Conceptualizing multilingualism in England, 800–1250,


 IMMO WARNTJES

More problematic is the question of how to approach the vast num-

ber of English computistical manuscripts of c.AD 950–1150. One route
taken in recent years is through a series of texts that are classified as prog-
nostics by modern scholars. Prognostics occur, more often than not, in
computistical manuscripts because of the calendrical element in many of
these texts. From the ninth to the eleventh centuries, prognostics appear
to have developed into a genre in its own right. The lead here was taken
by Roy Liuzza’s seminal article of 2001, which he followed up by a mono-
graph on London, British Library, Cotton Tiberius A III in 2010.117 Sán-
dor Chardonnens produced a detailed overview of the texts and manu-
script tradition in 2007, which he in turn followed up with numerous
articles.118 Marilina Cesario produced a series of articles on weather prog-
nostics.119 This massive increase in interest in late Anglo-Saxon prognos-
tics led to a panel on the subject at the 2012 Galway conference which
brought the three aforementioned scholars into dialogue. Prognostics

Turnhout, 121–44; Stephenson, R. (2016) ‘Saint who? Building monastic identity

through computistical inquiry in Byrhtferth’s Vita  S. Ecgwini,’ in R.  Stephenson and
E. Thornbury, Latinity and identity in Anglo-Saxon literature, Toronto, 118–37.
117
Liuzza, R. (2003) ‘Anglo-Saxon prognostics in context: a survey and handlist of
manuscripts,’ Anglo-Saxon England 30, 181–230; Liuzza, R. (2010) Anglo-Saxon prog-
nostics: studies and texts from London, British Library, MS Cotton Tiberius A.iii, Wood-
bridge.
118
Chardonnens, L. S. (2007) Anglo-Saxon prognostics, 900–1100: study and texts,
Leiden. Of his subsequent articles, see, e.g., Chardonnens, L. S. (2007) ‘Context, lan-
guage, date and origin of Anglo-Saxon prognostics,’ in R. H. Bremmer, Jr and K. Dek-
kers, Foundations of learning: the transfer of encyclopaedic knowledge in the early Middle
Ages, Groningen, 317–40; Chardonnens, L. S. (2007) ‘London, British Library, Harley
3271: the composition and structure of an eleventh-century Anglo-Saxon miscellany,’
in P.  Lendinara, L.  Lazzari, and M.  A. D’Aronco, Form and content of instructions in
Anglo-Saxon England in the light of contemporary manuscript evidence, Turnhout, 3–34;
Chardonnens, L.  S. (2010) ‘Appropriating prognostics in late Anglo-Saxon England:
a preliminary source study,’ in R. H. Bremmer, Jr and K. Dekker, Practice in learning:
the transfer of encyclopaedic knowledge in the early Middle Ages, Leuven, 203–55; Char-
donnens, L.  S. (2011) ‘Norm and practice of divination and prognostication in late
Anglo-Saxon England,’ in L. Sturlese and K. Nauer, Mantik, Schicksal und Freiheit im
Mittelalter, Wien, 51–64; Chardonnens, L. S. (2016) ‘Beyond long lines and column:
experiments in the visual structure of knowledge in Anglo-Saxon manuscripts,’ R. H.
Bremmer, Jr and K. Dekker, Fruits of learning: the transfer of encyclopaedic knowledge in
the early Middle Ages, Leuven, 35–74.
119
See, e.g., Cesario, M. (2009) ‘La Revelatio Esdrae nella tradizione latina e an-
glosassone,’ in L. Vezzosi, La letteratura tecnico-sceintifica nel medioevo germanico: Fach-
literatur e Gebrauchstexte, Allessandria, 57–84; Cesario, M. (2012) ‘Weather prognostics
in Anglo-Saxon England,’ English Studies 93, 391–426; Cesario, M. (2015) ‘An English
source for a Latin text? Wind prognostication in Oxford, Bodleian MSS Hatton 115
and Ashmole 354,’ Studies in Philology 112, 213–33.


 STATE OF RESEARCH ON COMPUTUS

provide an interesting inroad into the English computistical tradition of

the high Middle Ages, but a comprehensive focus on the computistica,
especially those composed in Latin, still remains a major desideratum.
Thorough studies of individual manuscripts as undertaken, e.g., by Faith
Wallis will be one of the main avenues to open up this material to the
wider scholarly community, and her website on Oxford, St. John’s Col-
lege, 17 (launched in 2007) is an excellent starting point for anybody
interested in English calendrical science of the high Middle Ages.120
The Benedictine reform movement brings us to the eleventh century,
to Germany, and back to Arno Borst and Hermann of Reichenau. As
outlined above, since 1975 Borst had toyed with the idea of writing a
biography of Hermann of Reichenau, one of the most prominent intel-
lectuals of that century. In order to do that, he started preparing first
editions of Hermann’s key scientific texts. The project stalled in 1990,
however, for various reasons, one of them being that more groundwork
was needed on Hermann’s sources, which led to Schriften, Borst’s last
publication. The basis for an appreciation of Hermann’s scientific writ-
ings had been laid, but, sadly, Borst was not given the time to finalize the
project he had been working on for four decades.
With Borst’s death in 2007, his microfilm copies went to the Monu-
menta Germaniae Historica in Munich, while his remaining papers went
to the Universitätsbibliothek in Konstanz (his books were transferred to
the Stiftsbibliothek in Zeitz). Among the Konstanz papers was the 1990
draft of his edition of Hermann’s scientifica, which came to the atten-
tion of the Monumenta Germaniae Historica. It was immediately under-
stood that Borst’s edition of Hermann’s computistica was much further
advanced than his work on the astrolabica. Therefore, it was agreed to
publish one volume on Hermann’s computistica on the basis of Borst’s
manuscript first, which will hopefully be followed by a thorough study
of Hermann’s astrolabica.121
Hermann’s interest in the subject was sparked by Notker of St Gall
(†AD 1022), whose Quatuor questionibus compoti Hermann considered
inadequate. In order to understand the context, Borst decided to include
Notker’s text in the volume.122 Notker was principally known for his

120
http://digital.library.mcgill.ca/ms-17/index.htm.
121
Borst, A. and I. Warntjes (2018) Hermann der Lahme: Schriften zur Zeitrech-
nung, Wiesbaden. See also the study by Germann (2006).
122
In this, Borst followed the pioneering study by Meier, G. (1887) ‘Die sieben
freien Künste im Mittelalter 2,’ Jahresbericht über die Lehr- und Erziehungs-Anstalt des
Benediktiner-Stiftes Maria-Einsiedeln der Studienjahre 1886/87, Einsiedeln, 3–36.


 IMMO WARNTJES

writings in Old High German, which led to his epithet ‘the German’.
Incidentally, he also translated his own computus from Latin into Ger-
man, as a wonderful discovery by Norbert Kruse in the Fürstlich Quadt
zu Wykradt und Isny’schen Archiv proves, published in 2003.123 One of
the questions that Notker posed, on the basis of an eighth-century tract,
was about the exact length of the mean synodic lunar month (the period
from one new moon to the next). The mathematical exercise was a divi-
sion of 6939 ¾ days by 235 lunar months (of the 19-year lunar cycle).
Notker did not consider it worth the effort to follow the calculation
through to the smallest fractions. Hermann, however, did just this (in his
Epistola), and then created a new 19-year cycle on the basis of his precise
value (Abbreviatio) and even tried to use it for the calculation of eclipses
(Prognostica). Alongside Notker’s Questiones and Hermann’s three works,
Borst also included commentaries on Hermann’s texts in this volume. In
total, some 30 manuscripts from the eleventh and twelfth centuries are
thoroughly analysed in this study, which therefore provides a good (but
certainly not exhaustive) insight into the study of computus in high me-
dieval southern Germany. Similar studies for other regions of Western
Europe in the eleventh and twelfth centuries are still major desiderata.
Borst’s ultimate goal of producing a biography of Hermann was final-
ly achieved in a collective effort through a conference held in Weingar-
ten in 2013. Its proceedings of 2016 show a strong focus on Hermann’s
scientific (often mistaken as ‘quadrivial’) works.124 Besides a discussion
of Hermann’s astrolabica (by David Juste) and his computistica (by my-
self ), the volume includes an introduction to Hermann’s abacus tracts
(by Martin Hellmann) and his version of the Rithmomachia (by Menso
Folkerts). It is a forceful reminder that computists, from the seventh
to the eleventh century, were not scholars of a single discipline. What
changed were the other areas they studied. In the seventh century, com-
putists quite certainly were also trained grammarians and accomplished
exegetes. From the tenth or eleventh century onwards, however, compu-
tists were actively embracing the new scientific knowledge entering the
Latin West from the Arabic world; accordingly, they would also have
called themselves experts on the astrolabe or abacists.125 The interrela-

123
Kruse, N. (2003) ‘Eine neue Schrift Notkers des Deutschen: der althoch-
deutsche Computus,’ Sprachwissenschaft 28, 123–55.
124
Heinzer, F. and T. Zotz (2016) Hermann der Lahme: Reichenauer Mönch und
Universalgelehrter des 11. Jahrhunderts, Stuttgart.
125
See, e.g., the case of Abbo: Burnett, C. (2006) ‘Abbon de Fleury, abaci doctor,’ in
B. Obrist, Abbon de Fleury: philosophie, science et computu autour de l’an mil, Paris, 129–39.


 STATE OF RESEARCH ON COMPUTUS

tion of these subjects still remains to be studied in detail, for the seventh
century as much as for the eleventh.
Hermann’s ideas principally spread from his monastery of Reichenau
in Lake Constance in two directions: north-eastwards into Bavaria and
up the Rhine as far as Liège. One of the key computists influenced by
Hermann was a certain Gerland, who worked in Lotharingia. His com-
putus (of c.AD 1060–93) has been studied principally for his re-dating
of the incarnation era, a chronological problem that Abbo of Fleury ap-
pears to have popularized.126 But his text contains many more important
ideas, which can only be appreciated since Alfred Lohr’s editio princeps
of 2013.127 Gerland applied Hermann’s exact value of the mean synod-
ic lunar month more rigorously to the 19-year lunar cycle. The prob-
lem was that the 19-year cycle did not consist of an integer number of
days (6939 ¾). This made the equal distribution of the fractions of the
mean synodic lunar month difficult, as they could only be assigned to
full calendar days, not their quarters. The solution was a 76-year period
(4×19  years=4×6939  ¾  days=27,759  days). In one of the eleventh-
century redactions of Hermann’s computistica, his results were already
applied to a 76-year period. Gerland followed this lead, and, like, Her-
mann, used a solar eclipse (of 29 September 1093) as the basis for his
calculations.
Lotharingia towards the end of the eleventh century and at the begin-
ning of the twelfth appears to have been a hotbed of progressive compu-
tists. Shortly after Gerland, Walcher, also from Lotharingia, added more
accuracy to Gerland’s model by using the lunar eclipse of 18  October
1092 as the basis of his calculations of all 940 conjunctions in 76 years
(in a text now labelled De lunationibus by its recent editor). Unlike Ger-
land, who worked from an eclipse record rounded to the hour, Walcher
was able to draw on his own astrolabe calculations to gain a precision of
the event unparalleled in the Latin West. Alas, Walcher soon realized
that his calculations did not match the observable realities of subsequent
years, as solar and lunar eclipses occurred a few hours after Walcher’s
estimated time.
Walcher had the benefit of being able to draw on two research tradi-
tions. The ideas underlying his De lunationibus he surely acquired in his
native Lotharingia. But he soon moved to England, where he became

See above pp. 27–28.

126

Lohr, A. (2013) Der Computus Gerlandi: Edition, Übersetzung und Erläuterung,

127

Stuttgart.


 IMMO WARNTJES

prior of Great Malvern. The West Midlands, at this time, developed into
a centre of the early reception of Arabic science in Western Europe, per-
sonified by Abelard of Bath. Walcher came into dialogue with Petrus
Alfonsi, a Jew trained in Arabic Iberia, who provided some insight into
why Walcher’s calculations did not match the observable reality: the mo-
tions of sun and moon are not uniform, and the lunar node (the points
when the moon’s orbit crosses the ecliptic) are not fixed, they have their
own motion, in the opposite direction to the planets (De dracone).
Walcher deserves the credit of heralding the potential of applying
Arabic science to computistics. His pioneering work used to be known
only to the initiated few among modern scholars, and it had never been
studied either in detail or in full. It is only thanks to the masterful editio
princeps published by Philipp Nothaft this year that the scholarly com-
munity can fully appreciate Walcher’s ingenious thought-world.128
Philipp Nothaft’s interest in the introduction of Arabic science into
computus was partially triggered by the study of manuscripts trans-
mitting Hermann’s work. Hermann paved the way for the division of
computus into vulgaris or ecclesiasticus, and naturalis. Vulgaris was the
traditional design for the ordering of the liturgical calendar, while na­
turalis attempted to find solid mathematical models for observable ce-
lestial phenomena. Hermann produced the first attempt at naturalis in
the second part of his Abbreviatio and in his Prognostica. However, as
Hermann’s theories did not work, this opened up the search for more
accurate models. In the manuscripts, Hermann’s study of the computus
naturalis was soon replaced, in south-eastern Germany, by discussions of
the Arabic lunar calendar and its relation to the traditional system. A pi-
oneering article in this respect is Nothaft’s ‘The application of Arabic
science in twelfth-century computistics’ of 2013, which puts a spotlight
on the diocese of Salzburg.129 In a second publication of 2015, Nothaft
opens the same discussion for twelfth-century England by editing a hith-
erto unknown Collatio compoti Romani et Arabici.130

128
Nothaft, C. P. E. (2017) Walcher of Malvern, De lunationibus and De dracone:
study, edition, translation, and commentary, Turnhout.
129
Nothaft, C.  P.  E. (2013) ‘The reception and application of Arabic science in
twelfth-century computistics: new evidence from Bavaria,’ Journal for the History of As-
tronomy 44, 35–60.
130
Nothaft, C. P. E. (2015) ‘Roman vs. Arabic computistics in twelfth-century Eng-
land: a newly discovered source (Collatio compoti Romani et Arabici),’ Early Science and
Medicine 20, 187–208.


 STATE OF RESEARCH ON COMPUTUS

The discussion of foreign lunar calendars as alternatives to the al-

most a millennium old Christian tradition reached its first height in
the AD 1170s. At least four authors compiled in this decade substantial
treatises on computus and the relation of the traditional Christian sys-
tem to the newly introduced Jewish and Arabic lunar calendars (as well
as additional astronomical information from newly translated Greek
authorities). The first is Reinher of Paderborn in Westphalia, who ve-
hemently advocated the supremacy of Jewish calculation (presumably
influenced by intellectual developments in neighbouring Lotharingia).
His Computus emendatus of AD 1170–71 was first transcribed in print
by Walter van Wijk in 1951, and Werner Herold produced a new edi-
tion in 2011, which, however, falls short of modern critical standards.131
In 2015, Alfred Lohr produced a definitive text for Corpus Christiano-
rum, and, admirably, this volume also contains editiones principes of the
following two computists, thereby bringing to a successful fruition the
pioneering (though sometimes misguided) studies by the Dublin-based
scholar, Jennifer Moreton.132 If Reinher was the reformer, a certain Mag-
ister Cunestabulus, writing in AD 1175, presumably in Canterbury ac-
cording to Lohr, was the traditionalist (and a new study of this text has
just been published by Nothaft133). Cunestabulus was provoked by Ger-
land’s redating of the incarnation era, which he refuted extensively in his
final chapter (39). In the three chapters before that (36–38), he cleverly
employed Greek, Arabic, and Jewish knowledge to explain the differ-
ence between traditional computus and observable reality in the date
of the equinoxes, solstices, and new moons. The rest of the work is more
or less a textbook along conventional lines. The third author, Roger of

131
van Wijk, W. E. (1951) Le comput emendé de Reinherus de Paderborn (1171),
Amsterdam; Herold, W. (2011) Reinher von Paderborn – Computus emendatus: die
verbesserte Osterfestberechnung von 1171, Paderborn; see also Herold, W. (2005) ‘Der
Computus emendatus des Reinher von Paderborn,’ Beiträge zur Geschichte des Bistums
Regensburg 39, 39–47.
132
Lohr, A. (2015) Opera de computo saeculi duodecimi: Reinheri Paderbornensis
Computus emendatus, Magistri Cunestabuli Computus, Rogeri Herefordensis Computus,
Turnhout (CCCM 272). Moreton, J. (1995) ‘Before Grosseteste: Roger of Hereford
and the calendar reform in twelfth-century England,’ Isis 86, 562–86; Moreton, J. (1999)
‘The Compotus of Constabularius (1175): a preliminary study,’ in J. Biard, Langage, sci-
ences, philosophie au XIIème siècle, Paris, 61–82. For Moreton’s contribution to the study
of medieval computus, see P. Nothaft, ‘On Jennifer Moreton’s edition of the Computus
ecclesiasticus,’ https://durhamgrossetesteproject.files.wordpress.com/2015/10/nothaft_
preface-to-moreton_computus-ecclesiasticus_online-edition.pdf.
133
Nothaft, C. P. E. (2017) ‘A reluctant innovator: Graeco-Arabic astronomy in the
Computus of Magister Cunestabulus (1175),’ Early Science and Medicine 22, 24–54.


 IMMO WARNTJES

Here­ford, writing in AD 1176, was then the conciliator, outlining the

details of the old system (books 1–3), Gerland’s new approach (here up-
dated; book 4), and finally contrasting the computus vulgaris and natu-
ralis with the Hebrew lunar calendar and Arabic astronomy (book 5).
A fourth text of the AD 1170s, a computus by a certain Peter, has only
very recently been (re-)introduced into scholarship by Philipp Nothaft
and awaits detailed analysis.134
Twelfth-century computus therefore provides a wealth of informa-
tion concerning the practical impact of the so-called Renaissance of the
twelfth century.135 Even more so, it opens an interesting window into the
cultural difficulties, the prejudices and biases of those responsible for the
introduction of new knowledge that ran counter to established church
doctrine. There has been a considerable turn over the past two decades
within Medieval Studies towards comparative history between different
cultures (especially between Western Europe and the Arab world) and
towards what is often termed cultural transfer (principally of ideas). It
must appear surprising that computus has not entered the discussion yet,
as there is hardly any other field in which the appropriation of Greek
and Arabic knowledge directly questions Christian authority. Compu-
tus has constantly been marginalized within Medieval Studies, for vari-
ous reasons, but particularly because of its perceived complexity and an
underestimation of its cultural impact. Sometimes, the more time-con-
suming engagement with complex topics yields the more fruitful results,
and it can only be hoped that the current tendency of research being
determined by arbitrary metrics, public impact, and projects designed
to please funding bodies rather than simply pushing the boundaries of
knowledge, will not create further obstacles in the future.
Nothaft and Lohr have raised twelfth-century computistics, almost
single-handedly, to a new level. Their research shows the massive po-
tential of manuscript studies, which will certainly lead to new and im-
portant discoveries. On the other hand, it demonstrates the existence
of exciting Wissenschaftslandschaften (regions of specialized learning)
in south-eastern Germany, Lotharingia, and the West Midlands in Eng-
land. In this, they have only started the academic discourse, and more
systematic studies of all computistical manuscripts of these regions, and

Nothaft (2017), 28; earlier Haskins (1924), 86.

134

For the scientific strand of the Renaissance of the twelfth century, see especially
135

the pioneering studies by Charles Homer Haskins: his detailed analyses collected in
Studies in the history of mediaeval science (Cambridge 1924) and his sweeping classic The
Renaissance of the twelfth century (Cambridge 1927).


 STATE OF RESEARCH ON COMPUTUS

even more so of the so-far neglected France, will be very rewarding areas
of research in the future. It certainly is crucially important to understand
the role of computistics in the adoption of Arabic science. In the one-
and-a-half centuries from Hermann to Roger of Hereford, computus
became thoroughly divided into two separate strands: the traditional
computus ecclesiasticus which was soon taught as the first basic scientif-
ic instructions in the newly emerging universities, while the computus
naturalis had fulfilled its purpose by providing inroads into the critical
examination of Arabic and Greek science which became the basis of the
scientific discourse of subsequent centuries. C.AD 1200 therefore marks
a major watershed in the study of the medieval computistical traditions,
and I therefore end my survey here. I can do this with a good conscience,
knowing that Philipp Nothaft’s seminal volume on the pre-history of the
Gregorian calendar reform (Scandalous error) has been submitted to the
publisher and will appear shortly (and is sure to replace Kaltenbrunner’s
classic136). This will provide an excellent overview of the continued story
of the calendrical endeavour from the late twelfth to the sixteenth centu-
ries. It certainly is very heartening to see that computus has made it into
the most recent surveys of medieval mathematics, which is a welcome
present acknowledgement of its past importance.137

***

Among the six conferences on Computus so far held in Galway since

2006, the third, which forms the basis for the present volume, marked
a notable watershed. The initial idea was to bring together all scholars
worldwide who worked on the technical details of late antique and early
medieval computus. The field was small enough, the time-range covered
substantial, the source material overwhelming, so that the pattern estab-
lished was to meet regularly to learn from each other as much as possible
in a short period of time and to create a platform for communication.
The third conference still followed this original spirit, with a strong fo-
cus on the scientific theory behind late antique and early medieval calen-
drical concepts, as this book demonstrates. One of the contributions to

136
Kaltenbrunner, F. (1876) ‘Die Vorgeschichte der Gregorianischen Kalender-
reform,’ Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften, philosophisch-
historische Classe 82, 289–414.
137
Hein, W. (2010) Die Mathematik im Mittelalter: von Abakus bis Zahlenkampf-
spiel, Darmstadt, 87–102; Katz, V. J. et al. (2016) Sourcebook in the mathematics of medi-
eval Europe and North Africa, Princeton, 29–36.


 IMMO WARNTJES

the third conference, Leofranc Holford-Streven’s edition of the Dispu-

tatio Chori et Praetextati, grew to monograph length in the process of
preparation for print, so that it was decided to publish this seminal study
as a separate book.
The first five essays in this volume principally deal with Easter reck-
onings and the tables preserving the relevant data. Alden Mosshammer
starts off with a discussion of the oldest computistical text that has sur-
vived, the famous Computus of AD 243. Mosshammer, who has been
working on an edition of this important text, here explains its Roman
contexts and the misrepresentation of its 112-year Easter cycle in the
available editions. Jan Zuidhoek analyzes the Easter table in Anatolius
Latinus’ De ratione paschali. He concludes that the data provided there
suggest AD 271 as the base year of the table. Dan Mc Carthy examines
the reference to St Patrick in Cummian’s letter of AD 632 and its con-
text. In his view, the southern Irish as represented by Cummian adopted
a Victorian 532-year Easter table with lunar limits 15–21 devised by Pat-
rick, which provides additional evidence for the late chronology of the
saint’s floruit. Luciana Cuppo continues the discussion of the introduc-
tion of Easter tables into the insular world in the seventh century. First
she focuses on three manuscripts which she believes represent Victorian
reactions to the gradual spread of the Dionysiac reckoning. The second
part of her article argues that part of the famous codex Oxford, Bodleian
Library, Digby 63 contains a Dionysiac dossier that was, in her opinion,
probably sent to Anglo-Saxon England by Pope Vitalian (AD 672–76).
From Britain to Iberia: Brigitte Englisch (in a more speculative article)
assumes that the Christian communities in Spain followed an Easter
reckoning that was based, at least in parts, on actual observations.
The second half of this volume deals with texts of various genres and
their computistical ideas. David Howlett discusses a didactic poem on
computistics which he places in seventh-century Ireland. Marina Smyth
examines the early medieval perception of one of the most though-pro-
voking feature of the Julian calendar, the bissextile (leap-year) day. She
surveys its understood length (three vs six hours per year) and the vari-
ous theological, allegorical, etymological, and pseudo-scientific expla-
nations of this calendrical device that circulated from seventh-century
Ireland to eighth-century Francia. Philipp Nothaft reconstructs the
‘computistical chronology’ of Claudius of Turin, who calculated the Ju-
lian calendar date and weekday data for key biblical events. The most
important one of these obviously was the passion of Christ, for which
Claudius implied, according to Nothaft’s reconstruction (contra Borst),


 STATE OF RESEARCH ON COMPUTUS

the impossible 21 March AD 34. Lisa Chen Obrist turns attention to

Claudius’s contemporary Hrabanus Maurus, proving that in one strand
of the manuscript transmission of his De rerum naturis he made use of
Isidore of Seville’s De ecclesiasticis officiis.
This volume therefore principally deals with the technical features of
lunar and solar calendars designed in late antiquity and their application
in the early Middle Ages. In the final article, Dáibhí Ó Cróinín rounds
off the volume by turning to the history of scholarship on late antique
and early medieval computus. The interests of 16th- and 17th-century
scholars in the texts and tables discussed in the present volume lay in
establishing biblical chronology and in understanding one of the most
central debates within the early Church, the Easter controversy. James
Ussher, working in the first half of the 17th century, was a key figure in
this emerging scholarship, and Ó Cróinín unearthed in the Bodleian Li-
brary a list of late antique and early medieval texts related to the Easter
question in chronological order, compiled by Ussher. This list demon-
strates the rigour of 17th-century scholarship and serves as a reminder of
the many texts written before AD 700 which still await a definite study.
The third Galway conference was the last with such a strong focus on
technical details. It was felt that, after establishing the technicalities, it
was time to allow more room for the scientific, cultural, political, theo-
logical, and manuscript contexts of computus. To facilitate this and to
attract more colleagues from neighbouring disciplines, special themes
were developed for each of the subsequent conferences. The first of these,
in 2012, was on early medieval prognostics, attracting scholars like Da-
vid Juste, Roy Liuzza, Sándor Chardonnens, Faith Wallis, Marilina Ce-
sario, and Richard Landes. The dialogue started there was very inspiring,
and it certainly served one of the purposes of introducing these themes,
namely creating a continued interest in the topics discussed for future
conferences. Prognostics has since formed a central part of the Galway
conference. In 2014, a platform was created to discuss the extremely im-
pressive advances in high medieval computus, which I have highlighted
in some detail above. In 2016, the impact of Borst’s Schriften on the
advances in Carolingian computistics was reviewed ten years after the
publication of his monumental work. For the first time, a call for papers
was issued, which received an overwhelming response. In fact, the con-
ference had to be extended from its traditional 1 ½ to 2 ½ days, which
stretched the organizational limits of an academic culture that does not
have a tradition of undergraduate and postgraduate Hilfskräfte; without
the voluntary help of Galway MA students, the tremendous organiza-


 IMMO WARNTJES

tional skills of Maura Walsh (Ó Cróinín), and Dáibhí’s exceptional abi­

lity of securing the basic funding necessary for such an event, this would
not have been possible. It is hoped that the next Galway conference will
have ‘computus and the vernacular’ as its special theme.
The introduction of special themes also changes the direction of
publication. The 2016 Carolingian theme was so popular that a separate
volume on Carolingian computistics is envisaged for the near future.
The proceedings of the 2012 and 2014 conferences will be published
in one volume. In the future, to keep the publications as focused as pos-
sible, a volume on the special theme will be produced. At the same time,
the Galway Conference will continue to invite papers on topics other
than the special theme, to keep its character as a platform for the latest
research in late antique and early medieval computistics; publication of
these papers in key journals will be strongly encouraged to spread the
awareness of the massive progress that has been made in this field.
It is very much hoped that the present volume, the two previous ones,
and those published in the future will push the boundaries of knowledge
in this field, and will attract new recruits to the vibrant research culture
of late antique and early medieval computus and its scientific, cultural,
political, theological, and manuscript context.

Appendix: Alfred Cordoliani’s contribution to the study

of computus138
‘Les traités de comput ecclésiastique de 525 à 990,’ Positions des thèses de l’école
des chartes (1942).
‘Études de comput I: notes sur le manuscrit latin 7418 de la Bibliothèque Na-
tionale,’ Bibliothèque de l’école des chartes 103 (1942), 61–65.
‘Études de comput II: un texte espagnol de comput du VIIe (?) siècle,’ Biblio-
thèque de l’école des chartes 103 (1942), 65–68.
‘Les traités du comput du haut moyen âge (526–1003),’ Archivum latinitatis
medii aevi 17 (1943), 51–72.
‘Une encyclopédie carolingienne de comput: les  Sententiae in laude compoti,’
Bibliothèque de l’école des chartes 104 (1943), 237–43.
‘Notes sur un auteur peu connu: Gerland de Besançon (avant 1100–après
1148),’ Revue du moyen âge latin 1 (1945), 411–19.
‘Les computistes insulaires et les écrits pseudo-Alexandrins,’ Bibliothèque de
l’école des chartes 106 (1945–46), 5–34.

This list was compiled by Philipp Nothaft and myself.

138


 STATE OF RESEARCH ON COMPUTUS

‘Le comput de Gerland de Besançon,’ Revue du moyen âge latin 2 (1946), 309–13.
‘A propos de premier chapitre de De temporum ratione de Bède,’ Le moyen âge
54 (1948), 209–23.
‘Abbon de Fleury, Hériger de Lobbes et Gerland de Besançon sur l’ère de
l’incarnation de Denys le Petit,’ Revue d’histoire ecclésiastique 44
(1949), 463–87.
‘La Logica de Gerland de Besançon,’ Revue du moyen âge latin 5 (1949), 43–47.
‘Un manuscrits de comput et d’astronomie des XIIe-XIVe siècles: le ms. 467 de
l’Université de Glasgow,’ Scriptorium 3 (1949), 69–79.
‘A propos du manuscrit 467 de l’Université de Glasgow,’ Scriptorium 4 (1950), 115.
‘Les manuscrits de comput écclésiastique conservés dans les bibliothèques
d’Aragon,’ Universidad Zaragoza 27 (1950), 592–616.
‘Los manuscritos de cómputo eclesiástico en las bibliotecas de Barcelona,’ Ana-
lecta sacra tarraconensia 23 (1950), 1–28.
‘Inventaire des manuscrits de comput ecclésiastique conservés dans les biblio-
thèques de Catalogne,’ Hispania sacra 4 (1951), 359–84.
‘Los textos y figuras de cómputo en los códices Emilianense y Vigiliano y el
Tratado del cómputo de Rodríguez Campomanes,’ Revista bibliográ-
fica y documental 5 (1951), 117–52.
‘Manuscrits de comput ecclésiastique de l’Escorial,’ La Ciuidad de Dios 163
(1951), 277–317.
‘Un manuscrits de comput ecclésiastique mal connu de la Bibliothèque Nation-
ale de Madrid,’ Revista de archivos, bibliotecas y museos 57 (1951), 5–35.
‘Les manuscrits de comput ecclésiastique de la Bibliothèque Capitulaire de
Tolède,’ Revista de archivos, bibliotecas y museos 58 (1952), 323–52.
‘Inventaire des manuscrits de comput ecclésiastique conservés dans les biblio-
thèques de Catalogne,’ Hispania sacra 5 (1952), 121–64.
‘La connaissance du comput ecclésiastique au moyen âge dans les abbayes de
l’ancienne province de Normandie du VIIIe au XIIIe siècle,’ Bulletin
philologique et historique du Comité des travaux historiques et scienti-
fiques (1953–54), 359–76.
‘Manuscrits bernardins des bibliothèques des anciens monastères d’Utrecht,’ in
Mélanges saint Bernard: XXIVe Congrès de l’Association bourguignonne
des Sociétés savantes, Dijon, 1953 (Dijon, 1954), 399–407.
‘Les textes et figures de comput de l’Antiphonaire de León,’ Archivos Leonenses
8 (1954), 258–87.
‘Inventaire des manuscrits de comput ecclésiastique conservés dans les biblio-
thèques de Madrid,’ Hispania sacra 7 (1954), 111–43.
‘Les manuscrits de comput ecclésiastique des bibliothèques du Levant,’ Anales
Universidad de Valencia 28 (1954–55), 2–22.
‘Les manuscrits de comput ecclésiastique de l’Abbaye de Saint Gall du VIIIe au
XIIe siècle,’ Zeitschrift für schweizerische Kirchengeschichte 49 (1955),
161–200.


 IMMO WARNTJES

‘L’évolution du comput ecclésiastique à Saint Gall du VIIe au XIe siècle,’

Zeitschrift für schweizerische Kirchengeschichte 49 (1955), 288–323.
‘Un autre manuscrit de comput ecclésiastique mal connu de la Bibliothèque Na-
tionale de Madrid,’ Revista de archivos, bibliotecas y museos 61 (1955),
435–86.
‘Les manuscrits de comput ecclésiastique des bibliothèques de Madrid,’ Hispa-
nia sacra 8 (1955), 177–208.
‘Textes de comput espagnol du VIe siècle: encore le problème des traités de com-
put de Martín de Braga,’ Revista de archivos, bibliotecas y museos 62
(1956), 685–97.
‘Textos de cómputo espagñol del siglo VI: el Prologus Cyrilli,’ Hispania sacra
9 (1956), 127–39.
‘Les plus anciens manuscrits de comput ecclésiastique de la bibliothèque de Berne,’
Zeitschrift für schweizerische Kirchengeschichte 51 (1957), 101–12.
‘Un manuscrit de comput ecclésiastique du fonds de Jumièges,’ in Jumièges
Congrès scientifique du XIIIe centenaire, Rouen, 10–12 juin 1954, 2 vols
(Rouen, 1955), ii 691–702.
‘Les manuscrits de la bibliothèque de Berne provenant de l’abbaye de Fleury au
XIe siècle le comput d’Abbon,’ Zeitschrift für schweizerische Kirchenge-
schichte 52 (1958), 135–50.
‘Un manuscrit de comput intéressant: Schaffhouse, Ministerialbibliothek, 61,’
Scriptorium 12 (1958), 247–53.
‘Textes de comput espangnol du VIIe siècle: le Computus Cottonianus,’ Hispania
sacra 11 (1958), 125–36.
‘Le comput de Dicuil,’ Cahiers de civilisation médiévale 3 (1960), 325–37.
‘Contribution à la littérature du comput ecclésiastique au moyen âge,’ Studi me-
dievali, Ser. 3, 1 (1960), 107–37.
‘Contribution à la littérature du comput ecclésiastique au moyen âge,’ Studi me-
dievali, Ser. 3, 2 (1961), 169–208.
‘La table paschale de Périgueux,’ Cahiers de civilisation médiévale 4 (1961), 56–60.
‘Comput, chronologie, calendriers,’ in C.  Samaran, L’histoire et ses méthodes
(Paris, 1961), 37–51.
‘Les manuscrits de comput des bibliothèque d’Utrecht,’ Scriptorium 15 (1961),
76–85.
‘Le computiste Hermann de Reichenau,’ Miscellanea storica ligure 3  (1963),
165–90.
‘L’activité computistique de Robert, évêque de Hereford,’ in P. Gallais and Y.-
J. Riou, Mélanges offerts à Réné Crozet, 2 vols (Poitiers, 1966), i 333–40.
‘Le comput ecclésiastique de l’abbaye du Mont-Cassin au Xe siècle,’ Anuario de
estudios medievales 3 (1966), 65–89.
‘Les manuscrits de comput de l’abbaye du Mont-Saint-Michel,’ Sacris Erudiri
17 (1966), 55–65.


 ALDEN A. MOSSHAMMER

TOWARDS A NEW EDITION OF THE

COMPUTUS OF AD 243

Abstract
The Pseudo-Cyprianic De pascha computus, composed in the year AD 243, is
the earliest extant paschal computus (that attributed to Hippolytus being a list
of dates, rather than a computistical text). This paper discusses the manuscripts
and editions, summarizes the contents, and finally focuses on the reconstruc-
tion of the 16-year cycle and 112-year paschal period appended to the now lost
Codex Remensis.
Keywords
Pseudo-Cyprian; Computus of AD  243; 112-year paschal cycle; Hippolytan
cycle.

Introduction
The Computus of AD 243 is the earliest extant Christian computisti-
cal text.1 It first came to the attention of scholars in the 1640s. While
resident in London during the last fifteen years of his life, James Ussher
discovered among the manuscripts in Robert Cotton’s library an anony-
mous text under the title Expositio Bissexti. Ussher mentions the work
in his Chronologia Sacra, which was published posthumously from his
manuscripts by Thomas Barlow in 1660. Ussher remarks that, according
to a formula at the end, the text was written in the fifth year of Gordian,

1
The 112-year cycle preserved in an inscription at Rome and attributed to Hip-
polytus dates from AD 222, but this is a list of dates, not a computistical text. On that
list, see Mosshammmer (2008), 116–25.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 43-70
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114733
 ALDEN A. MOSSHAMMER

the consulship of Arianus and Papus, which is the year 243 of the Chris-
tian era.2
About the same time, the Jesuit scholar Jacques Sirmond discovered
a similar text in a manuscript at Reims. That manuscript is now lost,
probably destroyed in the fire of 1774.3 Sirmond sent a copy to Gilles
Bouchier (alias Aegidius Bucherius) in Paris in August of 1648. Accord-
ing to Bouchier, the copy he received carried the name of Cyprian. He
thought the attribution genuine and reported that the astronomer Gott­
fried Wendelin agreed.4

The editions
John Fell, dean of Christ Church and bishop of Oxford, decided to in-
clude the tract in an appendix to his edition of the works of Cyprian,
published at Oxford in 1682. In his preface, Fell defended Cyprianic au-
thorship and identified the work with the Chronicam ualde utilem men-
tioned in an anonymous Life of Cyprian falsely attributed to Paul the
Deacon, who was a contemporary of Charlemagne.5 Even if this ‘very
useful chronicle’ is a reference to the Computus of AD 243, we would
know only as much as can be inferred from the Reims manuscript—that
the name of Cyprian had been attached to this work by the end of the
ninth century.
Fell entrusted the editorial work to the Oxford mathematician John
Wallis. Wallis worked from Ussher’s copy of the Cotton manuscript,
which he sent to Thomas Gale for collation with the original. Gale re-
turned the text with some notes of his own. There are nevertheless a
number of errors in the readings that Wallis attributes, either explicitly
or by silence, to the manuscript.6
2
James Ussher in Barlow (1660), 29: In Paschali tractatu, quem ineditum habeo,
anno quinto Gordiani, Ariano et Papo consulibus, id est, anno aerae Christianae 243. The
form Ariano is based on the reading of the Cotton manuscript. The correct spelling is
Arrianus, as in the Reims manuscript.
3
Mattei (2001), 35.
4
Bouchier (1655), 202 col. 1 (book 6, chapter 9, paragraph 1). I thank Leofranc
Holford-Strevens for consulting a copy of Bouchier’s report for me in Oxford.
5
Fell (1700), 3.
6
For example, in Fell (1700), 211 (line 8 in the first full paragraph = p. 252.2
in Hartel’s CSEL edition, below n. 10), Wallis reads iniunxisse, with no note, whereas
the reading of the Cotton manuscript is inluncxisse and the Vatican copy of the Reims
manuscript (discussed below) has illuxisse.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

From Jean Mabillon Wallis obtained a copy of the text from the
Reims manuscript. Wallis found that the text was the same as that con-
tained in the Cotton manuscript, but in a fuller and more reliable form.
The Reims text also included a set of tables illustrating the 16-year cycle
and 112-year period that the author describes.7
In a discussion of the life and works of Cyprian published in 1797,
Gottfried Lumper argued on stylistic grounds that the work could not
possibly have come from Cyprian’s pen.8 Lumper also pointed out, as
had others before him, that it is unlikely that Cyprian was a Christian
as early as AD 243. Most scholars since that time have agreed, and the
work is therefore often referred to as the Pseudo-Cyprianic De pascha
computus.
Wallis’s edition was reprinted in the Patrologia Latina in 1844.9
There are numerous typographical errors in that edition, and the edi-
tors also disturbed the format and arrangement of the appended tables.
Wallis’s notes appear erroneously and inexplicably under the title Notae
Henrici Dodwelli.
Wilhelm von Hartel included the treatise among the dubia in the
third volume of his edition of the works of Cyprian published in the
Corpus Scriptorum Ecclesiasticorum Latinorum in 1871.10 Hartel was un-
able to find the Reims manuscript or any new witnesses to the text. He
therefore asked his readers to be content with a redaction of Wallis’s edi-
tion, in which Hartel adopted a few readings rejected by Wallis.11 Un-
fortunately, Hartel seems to have used the edition from the PL reprint
of 1844, instead of from the Oxford original or the Paris reprint. This
conclusion follows from such erroneous readings as sexta decennitas feria
quarta at p. 255, lines 14–15, in agreement with PL 4, 951, where Wal-
lis, following both the Reims text and the logic of the description, reads
sexta decennitas feria tertia. Hartel also reprinted and further disturbed
the arrangement of the tables as they appear in the Patrologia Latina.
Indeed, Hartel’s version of these tables is all but unintelligible.

7
This and the previous paragraph are derived from Wallis’s Monitio ad lectorem in
Fell (1700), 208.
8
Lumper (1795), 375, repr. in PL 4, 816.
9
PL 4, 937–74.
10
CSEL 3,3, 248–71.
11
Wilhelm von Hartel in CSEL 3,3, lxii, lxv.


 ALDEN A. MOSSHAMMER

The manuscripts
Vatican, Biblioteca Apostolica, Reg. lat. 324
Of the two witnesses, the better text is that included in the now lost
manuscript discovered at Reims. As remarked above, Wallis used a copy
made for him by Mabillon. In 1896, Johann Ernst brought to the at-
tention of scholars the existence of another seventeenth-century copy
in Vatican, Biblioteca Apostolica, Reg. lat. 324.12 The manuscript con-
sists of 25 leaves in two distinct parts.13 The first part includes the same
Pseudo-Cyprianic works listed in a description of the Reims manuscript
extant in Paris, Bibliothèque nationale de France, Lat. 11777, 244v, fol-
lowed by the spurious Letter of Cornelius to Cyprian, apparently copied
from a different (now lost) manuscript at Reims.14
The titles of De rebaptismate and De resurrectione are in the same
hand as the text. The titles of De pascha and the Epistula Cornelii are
in a different, but approximately contemporaneous hand. At the bot-
tom of folio 20v, from the hand of the second writer, is a letter of Pope
Alexander III to William of Sens, concerning the feast of St Thomas of
Canterbury. This text too came from that second Codex Remensis.
The last five pages of the manuscript are from a different hand entire-
ly and on a different type of paper. On folios 21–25 is a copy, in French,
of the Peace of Vervins between Henry III of France and Philipp II of
Spain dated 2 May 1598.
There are numerous corrections entered into the text of the three
Pseudo-Cyprianic treatises.15 Some of these corrections seem to come
immediately from the pen of the original copyist, making good his own
errors. Other corrections come from the hand of the same writer who
added the titles of De pascha and the Epistula Cornelii. They appear
sometimes in the margin, sometimes supra lineam. That writer also sup-
plies some omissions in the text and was therefore working by collation
with the manuscript. The identity of these writers cannot be determined.
They may have been the same person, possibly Sirmond or Mabillon,
working at different times. The Vatican copy of De pascha is different
from what Mabillon sent to Wallis, as several variant readings attest.16

12
Ernst (1896); Ernst (1898).
13
Wilmart (1937–45), ii 224–25; see also Petitmengin (1974), 28–29.
14
Petitmengin (1975), 126, 131.
15
Mattei (2001), 35.
16
For examples, see below on the text of the appended tables.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

London, British Library, Cotton Caligula A XV

The only other direct witness to the text is London, British Library,
Cotton Caligula A XV, 97v–105v. The manuscript consists of 153 leaves
containing a miscellany of material, ranging in date from the eighth to
the eleventh century.17 Folios 73r–117v contain a late eighth-century
copy of a collection of computistical texts, including (fols 73r–80r) the
so-called Cotton Computus of AD 689.18 On folio 107r there is a calen-
dar giving the dates of new and full moons for every month of a year des-
ignated as AD 743 from the Incarnation, an eleventh indictional year,
the first year of King Childeric, with Easter on 14 April, moon 15.19 This
collection of material was therefore probably made somewhere in Gaul,
during the reign of Childeric III (AD 743–51).
The text in question begins in the middle of folio 97v, with the title
Incipit Expositio Bissexti. That title is entirely inappropriate to the text.
While the author mentions bissextile days and years, there is no discus-
sion of the leap-year intercalation. The title perhaps originally belonged
to the following text in the collection, which occupies folios  106r–v.
That text is untitled, but begins certum est quod byssexti diem et aetatem
lunae deputandam est in luna Feb.20
The Expositio Bissexti of Cotton Caligula A XV is a heavily redacted
version of the text to which the lost Reims manuscript attests.21 The au-
thor of the original text argues (251.11) that the Sunday of Creation was
25 March and that the earliest permissible date for the paschal full moon
is 17  March (252.19–22). The redactor of the Cotton copy changes
both of those dates to 22 March. In the Reims copy, the author criticizes
(251.11–15) certain predecessors for reckoning the first day of the new
month between 15 March and 13 April. By that phrase, the computist
actually means the full moon of the first month. The redactor substitutes
the standard Alexandrian dates for the paschal new moon, as ranging
from 8 March to 5 April.22 Where the Reims text refers (253.18–23) to

17
For descriptions, see Planta (1802), 45–46; CLA 2, 19 (no. 183); Gómez Pal-
larès (1986), 24–32.
18
See Gómez Pallarès (1986) and Warntjes (2011).
19
On the date, see Palmer (2011), 218–19, and Warntjes (2011), 174–76.
20
For a transcription, see Jones (1943), 372.
21
In what follows, citations are by page and line from Wilhelm von Hartel’s CSEL
edition.
22
For these Alexandrian limits, see Dionysius Exiguus, Epistola ad Petronium
(ed. by Krusch (1938), 65).


 ALDEN A. MOSSHAMMER

the creation of the sun and the moon or its anniversary on 28 March,
the redactor either substitutes the more standard 25 March or omits the
datum.23 The Reims text says (256.18–20) that Jesus shared a paschal
meal with his disciples on 8 April and suffered on 9 April. The redactor
substitutes the traditional Roman dates of 24 and 25  March.24 Friday
9 April corresponds to AD 28 as the year of the Passion, which is con-
sistent with the interval of 215 years from the Passion to the consulship
of Arrianus and Papus in the dating formula at the end of the Cotton
manuscript (216.18–20).25 The Reims manuscript changed the interval
to 220 and moved the present year to a position in the 16-year cycle that
corresponds to AD  248, instead of 243 (268.25). This is the only in-
stance where the Cotton manuscript preserves a significantly better text.
Sometimes the redactor omits whole sections of the text in order
to suppress arguments with which he disagrees. Thus, in Chapter  4,
where the author explains how the predecessors erroneously arrived at
15 March, the redactor omits the passage (251.12–13). In Chapter 7, the
redactor omits a section (254.7–9,12–20) explaining how the Hebrews
added an embolismic month to avoid a full moon earlier than 17 March.
At the end of Chapter 18 and the beginning of 19, the redactor omits
the computist’s unusual date for the Nativity on 28 March (266.8–11).
In two places, where the redactor omits what he regards as an unor-
thodox passage, he interpolates into the text phrases borrowed from the
last paragraph of the Disputatio Morini Alexandrini episcopi de ratione

23
For other texts with 22 March as the date of creation and 25 March for the sun
and the moon, see the note on Munich Computus 44, ll. 18–36 (ed. by Warntjes (2010a),
146).
24
For 25  March 29 as the date of the Passion see Nothaft (2012), 38–56, who
argues (p. 50) that the date in Tertullian, Aduersus Iudaeos 8.18 (ed. by Emil Kroyman
in CCSL 2, 1363) is a later interpolation and (p.  51) that the tradition derives from
Hippolytus. The date is attested also in the Chronograph of AD 354, ed. by Theodor
Mommsen in MGH Auct. ant. 9, 57. For a convenient list of other texts attesting to this
tradition see Warntjes (2010a), 151 (note on Munich Computus 44, ll. 55–80).
25
The traditional date for the Passion was the consulship of the two Gemini
(C. Fufius and L. Rubellius), in the year AD 29. In some versions of the Roman consular
list, that year corresponds to AD 28, instead of AD 29. The best-known example is the
Cursus paschalis of Victorius, ed. by Krusch (1938), 16–52, which begins (p. 26) in the
year of the Passion in the consulship of the two Gemini, with 1 January on a Thursday
and 28 March on a Sunday—the calendrical data for the year AD 28. He finished his
work, as he says in the prefatory letter (Victorius, Prologus 7 (ed. by Krusch (1938), 23))
in the consulship of Constantinus and Rufus, AD 457, which he numbers (p. 48) as the
year 430 from the Passion. The reasons for the choice of the year 28 in the Computus of
AD 243 are different. See below, near n. 50.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

paschali, often referred to as Pseudo-Morinus.26 That last paragraph is

not included in all of the manuscripts, but it does appear in the copy
of Pseudo-Morinus preserved in the Cotton collection (fol. 83v). The
relevant text is as follows:

Post haec enim omnia breuiter dicam quod numquam factus est pascha apud
Iudaeos ante XII K. Apr., quae luna nata in VIII id. Marti, XIIII est in XI
[sic!] kal. Aprelium. Obserua igitur cursum lunarem iuxta regulam I Greco-
rum more Aegyptiorum et non secundum aepactas id est adiectiones lunares.

‘After this, I will briefly explain that Passover was never done among the
Jews before 12 Kalends April (21 March), because the moon that is new
on 8 Ides March (8 March) is the 14th on 11 Kalends April (22 March).27
Therefore observe the course of the moon according to the first rule of
the Greeks in the manner of the Egyptians, and not according to the
epacts, which are lunar additions.’

At the beginning of Chapter 6 in the Cottonian version of the Com-

putus of AD 243, where the Reims text had Propter hoc ergo (252.13),
the redactor substitutes Post haec enim omnia breuiter dicam quod. The
redactor then omitted the section where the computist explains how the
Hebrews arrived at 17 March as the date of the full moon in the first year
after the Creation, and wrote instead: numquam fecerunt Pascha ante
XII kal. Apr. quae luna nata in VIII id. Mart. XIIII est in XI kal. Apr.
(‘They never did Passover before 12 Kalends April, 21 March, because
the moon that is new on 8 Ides March, 8 March, is the 14th on 11 Kal-
ends April, 22 March’).
At the beginning of Chapter 7, where the computist says (253.25)
that the Hebrews were divinely instructed circa cursum lunae, the redac-
tor wrote: circa cursum lunarem iuxta regulam primam Graecorum more
Aegyptiorum et non secundum aepactas (‘about the lunar course accord-
ing to the first rule of the Greeks in the manner of the Egyptians and not
according to the epacts’).
When Wallis sent his copy of the Cottonian text to Gale for colla-
tion with the manuscript, Gale drew his attention to the parallel with
Pseudo-Morinus in Chapter 6. Wallis included that report in his notes.28

26
See Graff (2011), 125–33.
27
Correctly, 12 Kalends April, 21 March. For reconstruction of the text based on
the witness of several manuscripts, see Graff (2011), 142.
28
Fell (1700), 211 n. 1.


 ALDEN A. MOSSHAMMER

Either Gale did not note or Wallis did not report the parallel in Chap-
ter 7. Wallis included the phrase iuxta regulam primam Graecorum more
Aegyptiorum et non secundum epactas in the text.
The inclusion of this phrase in Wallis’s text, as in the Patrologia Latina
reprint and in the edition of Hartel, has misled several scholars to state
that the Computist of AD 243 rejected calculation by the epacts.29 If this
phrase were genuine, it would be the earliest known reference to epacts
of the moon in any Greek or Latin text.30 The Computist of AD 243 uses
what looks like a system of epacts in his description of the 8-year cycle
(253.18–23). He cannot therefore have rejected such a method.
The redactor of the Cottonian version of the Computus of AD 243
may have belonged to the same seventh-century context from which
what few references we have to the Disputatio Morini derive.31 It is un-
likely that the person who collected these texts undertook this redaction
himself.

Origins
The original author of this tract must remain unknown. The consu-
lar date in the year of Arrianus and Papus (AD 243) is secure and, as
Lumper argued, excludes Cyprian as author, who is unlikely in any case
to have written so barbarous a text.32 Since the year AD 243 does not
correspond to a first year of the author’s cycle, it is likely to have been the
date of composition.33 The author (216.18–20) calculated an interval
of 215 years from that date to the Passion, which yields AD 28 as the
year of the Passion. That agrees with the author’s calendar date on Friday
9 April (256.20).
Jerome attributed a work De pascha to the Roman schismatic Nova-
tian.34 Adolf von Harnack raised the possibility that the Computus of

29
Schwartz (1905), 39; Fotheringham (1922), 53; Mosshammer (2008), 126.
30
The Greek word is first attested in the later fourth century in Epiphanius’ de-
scription of the 8-year cycle (Panarion, ed. by Holl (1915–33), iii 246) and in Theon’s
Lesser Commentary on the Handy Tables (ed.  by Tihon (1978), 256), where Theon
explains how to calculate the epacts for any year of a 19-year cycle. The Latin word is not
attested before Dionysius Exiguus, except in a fragmentary letter of Cyril to Leo (PL 54,
603), which is of doubtful authenticity and uncertain date.
31
Graff (2010), 125–33, and Warntjes (2011), 183, who suggests that this collec-
tion of texts comes from the circle of Willibrord.
32
See above, n. 8.
33
For the author’s cycle, see the discussion below under the heading Pinax.
34
Jerome, De viris illustribus 70 (ed. by Bernoulli (1895), 41).


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

AD 243 might be that work.35 August Strobel accepted that identifica-

tion in a book published in 1977, but in the commentary on a transla-
tion of the text that he published in 1984 stated only that there are influ-
ences of Novatianist and Montanist theology in the tract.36 A work De
pascha need not have been a computus. Novatian was a far better writer
than the computist, and whether his literary activity can be dated to as
early as AD 243 is doubtful.37
An African provenance for the work is suggested by the similarities
between the Old Latin text of the Hebrew Bible used by the author and
that used by Cyprian, in the few instances where comparison is possi-
ble.38 Apart from the manuscripts themselves, the only witnesses to the
Computus of AD  243 are African. The Carthaginian Computist of
AD 455 quotes from a brief portion of the text.39 Quintus Julius Hilari-
anus, author of a treatise De die paschae, subscribed on 5 March AD 397,
seems to have known the Computus of AD  243. Hilarianus says that
some people call the full moon the first day of the month, and this must
be a reference to the peculiar usage of the Computist of AD 243.40

Contents
The author advocates the use of a 16-year luni-solar cycle with a 112-year
paschal period.41 Eusebius attributes to an author named Hippolytus a
work De pascha with a ‘Canon’ of dates for a 16-year cycle ‘to the first
year of Alexander Severus’, AD 222.42 Among the other works that Euse-
bius attributes to this Hippolytus are commentaries on the six days of
creation and a book Against All Heresies. Whether this Hippolytus is
the same person whom Eusebius had just mentioned as a prolific writer

35
von Harnack (1893–1904), i 653.
36
Strobel (1977), 174; Strobel (1984), 65–67.
37
On Novatian, see Papandrea (2008).
38
Monceaux (1901–23), i 121–22; ii 98.
39
Computus Carthaginiensis 1.6; 2.4–5 (ed. by Krusch (1880), 282–83, 287).
40
Quintus Julius Hilarianus, Expositum de die paschae et mensis 2 (ed. by Chris-
topher Pfaff (Paris 1712), repr. in PL 13, 1107): alii in orbem plenam, qualis Dei jussu
mundo orta fuerat, primam appellabant.
41
For discussion, see also de Rossi (1857), lxxxi–lxxxii; Mac Carthy (1901), xl–
xliv; Schwartz (1905), 36–40; Ogg (1955). Krusch (1880), 189–92, provides a recon-
struction of the 112-year list as it must have appeared in the Codex Remensis. He men-
tions the work in his narrative (esp. p. 20 n. 5), but there is no continuous discussion.
42
Eusebius, Historia ecclesiastica 6.22 (ed. by Schwartz and Mommsen (1903–09),
ii 568–69).


 ALDEN A. MOSSHAMMER

and the bishop of an unknown, but apparently eastern city, or the same
that the Chronograph of AD 354 says was exiled from Rome to Sardinia
in AD 235, is a much-debated problem.43
Eusebius does not mention a 112-year list in connection with the 16-
year cycle of Hippolytus. About 1550, excavations at Rome unearthed
just such a list. The headless statue of a figure seated on a chair was found
near the church of San Lorenzo fuori le Mura. On one side of the chair
there is inscribed in Greek characters a 16-year list of dates in the Roman
calendar, with a letter from α to ζ indicating the weekday from 1 to 7 on
which that date occurs. There follow six more columns with only the
letters from α to ζ indicating the weekday for the full moon in each of
six iterations, for a total of 112 years. Within that matrix there are short
notes showing to what year in the 112-year period the several observa-
tions of Pascha mentioned in the Bible correspond. On the other side of
the chair is a 112-year list of Easter Sundays in seven parallel columns of
16 years each. At the top of the first list is an inscription stating that the
table begins in the first year of Alexander Severus, which was an embo-
lismic year with the 14th day of the moon on Saturday, the Ides of April
(13  April AD  222). The list of dates for Moon 14 repeats after eight
years, so that this system is based on doubling an 8-year cycle.44 The earli-
est date for Moon 14 is 18 March, and the latest is 13 April.45 By inspec-
tion of the 112-year list of Sundays, one can see that what we know from
later sources as the distinctively Roman rules are already in effect. Easter
can be observed no earlier than the 16th day of the moon, no later than
the 22nd day of the moon, with an outer limit of 21 April.46

43
Eusebius, Historia ecclesiastica 6.20 (Schwartz and Mommsen (1903–09), ii
566–67); Chronograph of 354 (MGH Auct. ant. 9, 72–75). For a brief summary of this
‘Hippolytan’ problem, see Mosshammer (2008), 118–21, and Burgess and Kulikowski
(2013) 366–71, with full references to the earlier literature. The most comprehensive
discussion is that of Brent (1995).
44
An 8-year cycle is based on the correspondence between eight Julian years
(365.25×8=1922 days) and 99 lunar months, with 96 ordinary months alternating be-
tween 29 and 30 days (29.5×96=2832 days) and seven embolismic months of 30 days
each (3×30=90). The cycle conveniently returns the same calendar date for the 14th
moon every eight years, but with an error of about 1 ½ days with respect to the observ-
able moon, three days after a 16-year cycle.
45
For further discussion, see de Rossi (1857), lxxix–lxxx; Mac Carthy (1901),
xxxii–xl; Schwartz (1905), 29–35; Mosshammer (2008), 122–25. On the statue, see
Guarducci (1977). Photographs of the inscription are among the plates in Brent (1995),
at the front of the volume. The text is ed. by Scaliger (1595), 1–3; Fabricius (1716–18),
i 37–41 (repr. in PG 10, 875–84).
46
For these ‘Latin’ rules, see Victorius, Prologus 4–5 (Krusch (1938), 19–21).


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

The name of Hippolytus does not appear on this statue, nor is the list
of writings on the back of the chair the same as the list attributed to him
by Eusebius. The 16-year cycle may or may not be the same list that Euse-
bius attributes to Hippolytus. Unless there was a very early corruption in
the text, Eusebius said that the list of Hippolytus ended in the first year
of Alexander, while that on the statue began in that year.47
Neither the Roman exemplar nor the Computus of AD 243 is likely
to have been composed entirely independently of the 16-year cycle that
Eusebius attributes to Hippolytus. The author’s cycle is not, however,
simply a continuation of the Roman list that began in AD 222, nor of
the Hippolytan list that either began or ended in that year.
While the inscription at Rome began with a paschal full moon
on 13  April, which was correct for the year AD  222, the Computus
of AD 243 begins twenty years later with 1 April, correct for the year
AD 242. It is thus a recalibrated version of the Hippolytan list, which
in the Roman exemplar would produce 29 March in the year AD 242. It
also has different lunar limits. The Roman list restricts the 14th day of the
moon to the period between 18 March and 13 April, while the Compu-
tist of AD 243 uses the limits of 17 March and 12 April.
The contents of the tract can be divided into three parts. In the first
(250.1–253.23), the author argues from the facts of Creation that what
he calls the first day of the new month must be 17 March. The Sunday
of Creation was 25  March. The moon was created on the fourth day
of the week. It was a full moon rising in its 15th day in the evening of
28 March. Therefore, in the following year, which is the first year when
there was actually a 14th day of the moon in the first month, that moon
corresponded to the daytime hours of 17 March. The author does not
explain why that date should also be the earliest date for Passover. The
reason presumably is that the first year of Creation cannot have been

47
Most scholars (see n. 43) have assumed that the list of Hippolytus began in the
same year as the list on the statue. Recently, however, Burgess and Kulikowski (2013),
367–68, have challenged that consensus, pointing out that Eusebius says, in Histo-
ria ecclesiastica 6.22 (Schwartz and Mommsen (1903–09), ii 568), ἐπὶ τὸ πρῶτον ἔτος
αὐτοκράτορος Ἀλεξάνδρου τοὺς χρόνους, which means ‘to the first year’, not ‘from the first
year’. Burgess and Kulikowski do not refer to the ancient translations. Rufinus (Schwartz
and Mommsen (1903–09), ii 569) translated intra primum Alexandri imperatoris an-
num concludit. The Armenian version (German translation by Preuschen (1902), 29)
has, ‘setze er auch die Zeitrechnung auseinander und den Kanon von Zehn (sic!) Jahren
und (zwar) mit den ersten zehn (sic!) Jahren des Kaisers Alexandros beendete er die
ganze Berechnung’. Book 6 is missing in a lacuna in the Syriac text (ed. by Nestle (1901)).
Perhaps the text originally read (or Eusebius meant) ἀπὸ τοῦ πρώτου ἔτους. Burgess and
Kulikowski are not the first to notice this discrepancy. See Brent (1995), 308–09.


 ALDEN A. MOSSHAMMER

embolismic. Therefore, the 14th day of the first month in the following
year must be 17 March, not 16 April. The author also claims that the
11 ¼-day differential between the lunar and solar year arises from the
11 ¼ hours during which the sun preceded the moon on that first day
of their creation. The difference is 11 ¼ hours, not 12. The moon rose
at the beginning of the night, he says (253.10–15), but not after sunset,
because it had to appear on the same ‘day’ (i.e., during daylight).
In the second part of the text (253.24–258.2), the author describes a
16-year cycle, with the year of the Exodus as its base-date, with a Passo-
ver moon on Monday, 12 April. Intervals are counted from the Exodus,
which was in effect the year 0. The cycle itself begins the next year with
moon 14 on 1 April. The author says that the Hebrews discovered the
8-year cycle and then doubled it because of the sequence of weekdays. He
refers to the fact that after a 16-year interval, any given date in the Julian
calendar will recede by one weekday. After seven repetitions, 112 years,
12 April returns to Monday. Unfortunately, because of the error in a 16-
year cycle, 12 April will not be moon 14, but moon 22.48
The third and longest portion of the text (258.3–269.5) offers a
chronological summary of the years from the Exodus to the Passion.
The author calculates the date of every Passover observance mentioned
in the Bible. That which took place in the 18th year of Josiah (2 Kings
23:23), for example, was 970 years after the Exodus, on 12 Kalends April
(21 March), weekday #2 (Monday).49 He enjoys finding typological as-
sociations between the week of Creation and events in biblical history.
Thus (266.3–8), he dates the birth of Christ to 28 March and chooses the
year 1548 from the Exodus, corresponding to 4 BC, because 28 March
was a Wednesday in that year. To that he adds 31 years and dates the Pas-
sion to 1579, AD 28, choosing that year, rather than the traditional AD
29, because in his pinax the year 1579 has Moon 14 on a Thursday, to
accommodate a Passover meal on that weekday.50 He also reflects on the
mystical significance of the various numerals he generates, commenting
especially (257.8–10, 268.10) on the 318 servants of Abraham (Genesis
14:14) as being represented in Greek by the numeral τιη, which consists

48
Over the period of 112 years, an error of –22 days with respect to the observed
moon will accrue, which is the equivalent of +8.
49
The author does not show the calculation, which is as follows: The interval 970
is divisible by 112 with a remainder of 74, which is divisible four times by 16 with a
remainder of 10. The 10th year of a cycle always has Passover on 12 Kalends April. In the
fifth column of the pinax, the 10th year has feria 2.
50
See Nothaft (2012), 52–54.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

of tau, the symbol of the cross, plus iota-eta, the first two letters of the
name Jesus (Ίησοῦς).51

The pinax and 112-year list of the Codex Remensis

I am working on a new edition of the text for publication in the Corpus
Christianorum Series Latina. Here I will focus on the set of tables in-
cluded in the lost Reims manuscript, which implement the 16-year cycle
and 112-year period that the author describes. A redactor has introduced
a few changes. Figure 1 shows the 16-year cycle and Figure 2 the 112-year
period as the author describes them in the text. In reconstructing what
the author refers to as his pinax, I have followed Wallis’s lead and used
the format in the Roman exemplar of the Hippolytan cycle, with seven
parallel columns for the weekdays on which the paschal full moon oc-
curs through the 112-year period.52 The 16-year cycle is actually a pair
of 8-year cycles, and the list of Sundays repeats therefore after 56 years.
The base-date is the year corresponding to the Exodus, while the
16-year cycle begins in the following year. The author says in Chapter 7
that the Hebrews first observed the Pasch on 12 April (254.2), in the
next year on 1 April (254.10), and that the second year has moon 14 on
21 March (254.11). At the end of Chapter 18 (266.6) and again at the
beginning of Chapter 23 (268.21), he says the year of the Exodus is the
first line of the pinax (a primo pinacis uersu). We must therefore distin-
guish between the 17 lines of the pinax and the 16 years of the cycle.
In constructing the 112-year list of Easter dates, I  have followed
­Krusch and Strobel in assuming that the author adopted the rule ap-
parent in the Roman exemplar of the Hippolytan tables and explicit in
descriptions of the later Roman 84-year cycle, that Easter Sunday may
not precede the 16th day of the moon.53 My reconstruction in Figures 1
and 2 therefore does not differ from either of theirs.

51
Later the numeral became significant also as the traditional number of bishops
assembled at the Council of Nicaea; so, for example, Dionysius Exiguus in the Epistola
ad Petronium (ed. by Krusch (1938), 20). The earliest (c.AD 359) reference to the ‘318
fathers’ is from Hilary of Poitiers, De Synodis 86 (PL 10, 538), who also connects the
number with the story of Abraham. See Aubineau (1966).
52
Fell (1700), 213.
53
Krusch (1880), 189–91; Strobel (1984), 57–59; for the rule, see Victorius of
Aquitaine, Prologus 4 (Krusch (1938), 19).


 Year Age of Moon Type Passover Weekday I II III IIII V VI VII
on 28 March (Luna XIV)
Exodus 30 Prid Id Apr 2 1 7 6 5 4 3
12 April
I 11 Common Kl Apr 6 5 4 3 2 1 7
1 April
II 22 Common XII Kl Apr 3 2 1 7 6 5 4
21 March
III 4 Embolismic VI Id Apr 2 1 7 6 5 4 3
Bissextile 8 April
IIII 15 Common V Kl Apr 6 5 4 3 2 1 7


28 March
V 26 Common XVI Kl Apr 3 2 1 7 6 5 4
17 March
ALDEN A. MOSSHAMMER

VI 7 Embolismic Non Apr 2 1 7 6 5 4 3

5 April
VII 19 Common IX Kl Apr 6 5 4 3 2 1 7
Bissextile 24 March
VIII 30 Embolismic Prid Id Apr 5 4 3 2 1 7 6
12 April
 Year Age of Moon Type Passover Weekday I II III IIII V VI VII
on 28 March (Luna XIV)
VIIII 11 Common Kl Apr 2 1 7 6 5 4 3
1 April
X 22 Common XII Kl Apr 6 5 4 3 2 1 7
21 March
XI 4 Embolismic VI Id Apr 5 4 3 2 1 7 6
Bissextile 8 April
XII 15 Common V Kl Apr 2 1 7 6 5 4 3
28 March
XIII 26 Common XVI Kl Apr 6 5 4 3 2 1 7


17 March
XIV 7 Embolismic Non Apr 5 4 3 2 1 7 6
5 April
XV 19 Common IX Kl Apr 2 1 7 6 5 4 3
Bissextile 24 March
XVI 30 Embolismic Prid Id Apr 1 7 6 5 4 3 2
12 April
TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

Figure 1 The pinax and 16-year cycle as reconstructed from the author’s description.
 Luna XIV Easter Sunday, Cycle II Cycle III Cycle IIII Cycle V Cycle VI Cycle VII
Cycle I
Kl Apr 1. III Non Apr 17. Prid Non 33. Non Apr 49. VIII Id Apr 65. VII Id Apr 81. VI Id Apr 97. V Id Apr
1 April 3 April Apr 5 April 6 April 7 April 8 April 9 April
4 April
XII Kl Apr 2. VII Kl Apr 18. VI Kl Apr 34. V Kl Apr 50. IV Kl Apr 66. X Kl Apr 82. IX Kl Apr 98. VII Kl Apr
21 March 26 March 27 March 28 March 29 March 23 March 24 March 25 March
VI Id Apr 3. XVIII Kl 19. XVII Kl 35. XVI Kl 51. IV Id Apr 67. III Id Apr 83. Prid Id Apr 99. Id Apr
8 April Mai Mai Mai 10 April 11 April 12 April 13 April
14 April 15 April 16 April
V Kl Apr 4. III Kl Apr 20. Prid Kl Apr 36. Kl Apr 52. IV Non 68. III Non 84. Prid Non 100. Non Apr


28 March 30 March 31 March 1 April Apr Apr Apr 5 April
2 April 3 April 4 April
XVI Kl Apr 5. XI Kl Apr 21. X Kl Apr 37. IX Kl Apr 53. VIII Kl 69. XIV Kl 85. XIII Kl 101. XII Kl
ALDEN A. MOSSHAMMER

17 March 22 March 23 March 24 March Apr Apr Apr Apr

25 March 19 March 20 March 21 March
Non Apr 6. III Id Apr 22. Prid Id Apr 38. Id Apr 54. VII Id Apr 70. VI Id Apr 86. V Id Apr 102. IV Id Apr
5 April 11 April 12 April 13 April 7 April 8 April 9 April 10 April
IX Kl Apr 7. VII Kl Apr 23. VI Kl Apr 39. V Kl Apr 55. IV Kl Apr 71. III Kl Apr 87. Prid Kl Apr 103. Kl Apr
24 March 26 March 27 March 28 March 29 March 30 March 31 March 1 April
Prid Id Apr 8. XVII Kl Mai 24. XVI Kl 40. XV Kl Mai 56. XIV Kl 72. XIII Kl 88. XII Kl Mai 104. XVIII Kl
12 April 15 April Mai 17 April Mai Mai 20 April Mai
16 April 18 April 19 April 14 April
 Luna XIV Easter Sunday, Cycle II Cycle III Cycle IIII Cycle V Cycle VI Cycle VII
Cycle I
Kl Apr 9. VII Id Apr 25. VI Id Apr 41. V Id Apr 57. III Non 73. Prid Non 89. Non Apr 105. VIII Id
1 April 7 April 8 April 9 April Apr Apr 5 April Apr
3 April 4 April 6 April
XII Kl Apr 10. X Kl Apr 26. IX Kl Apr 42. VIII Kl 58. VII Kl Apr 74. VI Kl Apr 90. V Kl Apr 106. IV Kl Apr
21 March 23 March 24 March Apr 26 March 27 March 28 March 29 March
25 March
VI Id Apr 11. III Id Apr 27. Prid Id Apr 43. Id Apr 59. XVIII Kl 75. XVII Kl 91. XVI Kl 107. IV Id Apr
8 April 11 April 12 April 13 April Mai Mai Mai 10 April
14 April 15 April 16 April
V Kl Apr 12. III Non 28. Prid Non 44. Non Apr 60. III Kl Apr 76. Prid Kl Apr 92. Kl Apr 108. IV Non
28 March Apr Apr 5 April 30 March 31 March 1 April Apr


3 April 4 April 2 April
XVI Kl Apr 13. XIV Kl 29. XIII Kl 45. XII Kl Apr 61. XI Kl Apr 77. X Kl Apr 93. IX Kl Apr 109. VIII Kl
17 March Apr Apr 21 March 22 March 23 March 24 March Apr
19 March 20 March 25 March
Non Apr 14. VI Id Apr 30. V Id Apr 46. IV Id Apr 62. III Id Apr 78. Prid Id Apr 94. Id Apr 110. VII Id
5 April 8 April 9 April 10 April 11 April 12 April 13 April Apr
7 April
TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

IX Kl Apr 15. III Kl Apr 31. Prid Kl Apr 47. Kl Apr 63. VII Kl Apr 79. VI Kl Apr 95. V Kl Apr 111. IV Kl Apr
24 March 30 March 31 March 1 April 26 March 27 March 28 March 29 March
Prid Id Apr 16. XIII Kl 32. XII Kl Mai 48. XVIII Kl 64. XVII Kl 80. XVI Kl 96. XV Kl Mai 112. XIV Kl
12 April Mai 20 April Mai Mai Mai 17 April Mai
19 April 14 April 15 April 16 April 18 April

Figure 2  112-year list of Easter Sundays.

 ALDEN A. MOSSHAMMER

This original arrangement was disturbed by a redactor, then distorted

by the copyist of the lost Reims manuscript, slightly disarranged in the
transcript that Mabillon sent to Wallis and in the copy at the Vatican,
misunderstood by the editors of the reprint in the Patrologia Latina
and finally rendered unintelligible in Hartel’s edition. The Vatican copy
shows that Wallis faithfully reproduced the tables as they appeared in
the copy of the Reims manuscript that Mabillon sent him. I use an image
from Wallis’s edition to represent the basic arrangement, as it is in the
public domain, more easily legible than Vatican, Biblioteca Apostolica,
Reg. lat. 324, and with fewer errors.
Plate 1 shows the tables as they appeared in the 1700 reprint of the
1682 edition.54 The tables in the manuscript began with columns on the
left for the age of the moon and the weekday (the feria) on the Kalends
of March. That information is useful, but in the text the author makes
no reference to such data. It is probably a later addition.55 The pinax on
the right, headed ‘Exodus’, has 16 lines, instead of 17, with the result that
this table makes the 16-year cycle begin with 12 April in the year cor-
responding to the Exodus, instead of with 1 April in the following year.
This is the work of a redactor. Either the redactor did not understand
why there were 17 lines in the exemplar and omitted the last one or he
made a deliberate decision to recalibrate the cycle with the Exodus as
year 1.
The redactor also entered a note at the eighth line that the year is
numbered ∞DCCXVIII (1718).56 This is a scribal error for ∞DCCXC-
VIIII (1799), which is the numeral that appeared at the end of the nar-
rative text in the Reims manuscript. There (268.18) the interval from
the Exodus to the fifth year of Gordian, the consulship of Arrianus and
Papus, is ∞DCCXCVIIII (1799), and the writer refers to the eighth line
of the table. In the Cotton manuscript, the year is 1794, and is said to
correspond to the third line of the first set of 16 years.
That the Cotton manuscript is correct has already been noted above.
The intervals given (268.20) from the Passion to the present year show
54
The page also shows on the lower right Wallis’s reconstruction of the Hippol-
ytan table. It is not the pinax of the Computus of AD 243, which Wallis reconstructs on
p. 213.
55
Krusch (1880), 162, says: ‘Luna und ferie des 1. März finden wir übrigens schon
in dem sogenannten Computus Cypriani notirt’; but the table that includes this column
is clearly the work of a redactor and not necessarily a product of the third century.
56
The earlier form of the Roman numeral for 1000 was CIƆ, which evolved into a
symbol resembling a figure-8 turned on its side. I have used the unicode symbol for the
‘infinity’ sign to represent it. See Capelli (1962), 44.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

that 1794 is the correct interval from the Exodus. The consulship of Ar-
rianus and Papus was AD 243. The author dates the Passion to Friday
8 April in the year 1579. Friday 8 April is calendrically correct for the
year AD 28. In the Cotton manuscript the interval from the Passion to
the present year is 215. In the Reims copy the interval is 220. The Cot-
ton copy refers to the third line of the table with the paschal full moon
on Tuesday, 12 Kalends April (21 March), which is correct for the year
AD 243. The Reims copy makes it the eighth line, with Passover on Fri-
day, 9 Kalends April (24 March), correct for AD 248.57
The decision to include a table for the luna and feria of 1  March
forced a change in the arrangement of the columns for the feriae in the
pinax. There was not room for an additional six columns of feriae on
the right side of the page. The list of feriae is arranged instead in four
columns of 24 lines each and a fifth column of 11 lines, for a total of 107
lines, instead of six columns of 16 lines each. The fifth column appears
on the left below the first column.
To understand the ferial columns as they appeared in the Reims
manuscript, the reader must compare Plate 1 with Figure 1. Consistent
with his own 16-line version of the pinax, the redactor began with feria
ii in the list of 14th moons and therefore with feria i in the first of the
now displaced ferial columns. Although there are 107 lines, a horizontal
space across the whole width, and another space within the second col-
umn, reduce the total number of entries to 102. There would have been
96 feriae included in the original six columns. A copyist erroneously re-
peated the last six entries at the end of the fifth column. There is only one
mistake in Wallis’s version as compared with what appears in the Vatican
manuscript. In the second column, after the horizontal space, the second
numeral in Wallis’s copy was 5, instead of 6. The Vatican copy correctly
has 6.
There follows a list of calendar dates. They are all Sundays, as indi-
cated by the word Domini or the abbreviation dom, for dies domini, the
Lord’s Day. The entry for year eight is missing, so that the list has 111
entries, instead of 112. The list begins erroneously with XVI Kl. Mart.
instead of XIV Kl. Mai.

57
Krusch (1880), 20 n. 5, thought that the change in the dating formula in the
Reims manuscript derived from a copy or redaction made in the year AD 248. George
Ogg (1955), 40, argued that a scribal error in the numerals led a copyist to make all of
these changes.


 ALDEN A. MOSSHAMMER

Plate 1  Fell (1700), 220–21.

At the end of the list is Wallis’s note, as follows:58

Sed non fui sollicitus de his vel ordinandis, vel restituendis: Quoniam
quicquid sani hic contineatur, id omne haberi potest in Pinace quem (ad
Auctoris mentem) ante restitueram et Notis inserueram, quam hanc con-
spexerim tabellam. Qua nihil aliud continetur, quam confusa materia, et
male digesta, et depravata pro istiusmodi pinace: neque operae pretium est
corrigere, vel ordinare, vel etiam divinare quo casu in hanc confusionem
devenerit. Si cui id libet, per me licet.

‘I have not troubled to order or reconstruct these lists. Whatever here

makes sense can be found in the pinax that I reconstructed according
to the author’s intent and the notes I have inserted rather than in what
I observed in this table. In it there is nothing but confused material, a
poorly understood and corrupted version of that pinax. It would not

58
Fell (1700), 221; the translation is mine.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

be worth the effort to try to correct or rearrange it, much less to try to
figure out how it came to its present condition. If that task appeals to
anyone, he has my permission to try.’

George Ogg attempted to reconstruct the tables as he thought they

must have appeared in the lost Reims manuscript.59 Unfortunately, Ogg
worked from Hartel’s 1871 CSEL edition. As we saw previously, Hartel
in turn used the reprint of Wallis’s edition published in the Patrologia
Latina. The editors of the PL altered the format to fit the requirements
of a page divided into two columns. Hartel seems not to have under-
stood the format and rearranged its parts beyond recognition.60 Ogg
correctly identified some of the elements of Hartel’s tables, but because
of his failure to consult Wallis’s edition his attempt to explain how the
tables came to be so disarranged is unnecessarily complex.
The arrangement is less chaotic than Wallis realized. The Vatican copy
has the same arrangement of four columns of feriae, with a fifth column
appended to the first. In the Vatican copy, however, a page break occurs
after line 17, instead of a horizontal space after line 16 (see Figure 4b,
line  17). Where Wallis’s copy had the incorrect feria 5 (see Figure  4b,
year 55), the Vatican has the correct numeral 6. Both manuscripts share
five other mistakes compared with the original matrix, and those errors
must therefore have already appeared in the lost Remensis (see Figure 4b,
years 66, 89, 97, 104, 105). At the end of the list of feriae, Wallis’s copy
had an extra six entries. The Vatican copy has three more extra entries,
but a corrector has marked them for deletion.
The list of Easter Sundays had few errors, and the order of the entries
is less disturbed than it appeared to Wallis. The list of Sundays in the
Vatican copy has a slightly different arrangement compared with Wallis’s
copy. In particular, what was the fifth column in Wallis’s arrangement
appears in the Vatican copy as if it were an extension of the third column.
Figure 3 compares the list of Sundays as it appears in Wallis’s edition
with that of the Vatican copy. The numerals represent the position of
each date in a correctly reconstructed version (Figure 2). By numbering
the list, one can easily see how the original arrangement came into its
present format. Wallis’s version is almost correct. The year 8 is missing.
The entries for years 11/34 and 12/33 have been reversed. The entries for
years 79 to 83 and 108 to 112 have been shifted to the left.

Ogg (1954).
59

CSEL 3,3, 269–71.

60


 ALDEN A. MOSSHAMMER

Wallis’s Copy Vat. Reg. lat. 324

13 11 34 84 12 39 61
14 35 57 85 13 40 62
15 36 58 86 14 41 63
16 37 59 87 15 42 64
17 38 60 88 16 43 65
18 39 61 89 17 44 66
19 40 62 90 18 45 67
20 41 63 91 19 46 68
21 42 64 92 20 47 69
22 43 65 93 21 48 70
23 44 66 94 22 49 71
1 24 45 67 95 20 50 72
2 25 46 68 96 21 51 73
3 26 47 69 97 22 52 74
4 27 48 70 98 23 53 75
5 28 49 71 99 24 54 76
6 29 50 72 100 1 25 55 77
7 30 51 73 101 2 26 56 78
9 31 52 74 102 3 27 79 108
10 32 53 75 103 4 28 80 109
12 33 54 76 104 5 29 81 110
55 77 105 6 30 82 111
56 78 106 7 31 83 112
79 108 107 9 32 84
80 109 10 33 85
81 110 11 34 86
82 111 35 57 87
83 112 36 58 88
37 59 89
38 60 90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
Figure 3  Numbered list of Easter Sundays. Numerals in bold type highlight
the displacements. For the dates associated with each numeral, see Figure 2.

 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

In the Vatican version, the year  8 is also missing. The years  11/34
correctly follows the years 10/33. The years 20 to 22 have been errone-
ously repeated. The years 35 to 38 and 57 to 60 appear at the bottom of
the first two columns, instead of the top of columns three and four. The
years 79 to 83 and 108 to 112 are shifted to the left as in Wallis’s copy, but
what appeared as a fifth column in Wallis’s version is written as if it were
an extension to the third column.
Although the Vatican version is more seriously disarranged than what
was in the transcript that Mabillon sent to Wallis, having the two copies
of the Codex Remensis for comparison helps us to reconstruct what must
have been the arrangement in that lost manuscript. Figures 4a–e show
that reconstruction.
Figure 4a shows the pinax and 4b the list of feriae. Figures 4c–e have
the end of the list of feriae followed by the list of Sundays, written two
columns to the page. A space before the last entry in the first two col-
umns (Figure 4c, line 24) led Mabillon to treat them as the beginning
of new columns (Figure 3, line 1), when he was transcribing the Reims
manuscript for Wallis. A  space after the first four entries in columns
three and four (Figure  4d) led the copyist of the Vatican manuscript
to see those four entries as continuations of the first two columns (see
Figure 3), rather than as new columns following in sequence after the
second column. Both copyists (Mabillon and the unknown writer of
Vatican, Biblioteca Apostolica, Reg. lat. 324) read the first five entries on
the third page of Sundays (Figure 4e) as continuations of columns three
and four, instead of four and five. The list of feriae was originally written
23 lines to the page. The list of Sundays was written 23 lines to the page,
except that on the last page the copyist managed to fit 30 lines.


 ALDEN A. MOSSHAMMER

Kl. Mart. Luna et Exodus

Feria
Luna II feria ii Embolismus Prid Id Apr 1 feria ii
Luna XIII feria iii Comm Kl Apr 2 feria vi
Luna XXIII feria iiii Comm XII Kl Apr 3 feria iii
(recte XXIIII)
Luna VI feria vi Bissextus et <VI Idus Apr> 4 <feria> vi
Embolismus
Luna XVII feria vii Comm V Kl Apr 5 feria vi
Luna XXIIX feria i Comm XIII Kl Apr 6 feria iii
(recte XVI Kl Apr)
Luna IX feria ii Embolismus Non Apr 7 feria ii
Luna XXI feria iiii Bissextus VIII Kl Apr 8 feria vi
Comm (recte VIIII Kl
Apr)
∞dcxviii
Luna II feria v Embolismus Prid Id Apr 9 feria v
Luna XIII feria vi Comm Kl Apr 10 feria ii
Luna XXVI feria vii Comm XII Kl Apr 11 feria vi
(recte XXIIII)
Luna VI feria vi Bissextus et VI Idus Apr 12 feria v
(recte ii) Embolismus
Passio
Luna XVII feria iii Comm V Kl Apr 13 feria ii
Luna XXVIII feria iv Comm XVI Kl Apr 14 feria vi
Luna VIIII feria v Embolismus Non Apr 15 feria v
Luna XXI feria vii Bissextus VIIII Kl Apr 16 feria ii
Comm
Figure 4a  Reconstruction of first page of tables in the Codex Remensis Deper-
ditus.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

1 17 feria i 40 feria iiii 62 feria iii 85 feria i

2 18 feria v 41 feria iii 63 feria ii 86 feria v
3 19 feria ii 42 feria vii 64 feria vi 87 feria iiii
4 20 feria i 43 feria iiii 65 feria v 88 feria i
5 21 feria v (blank both 66 feria iii 89 feria v (recte vii)
copies) (recte ii)
6 22 feria ii 44 feria iii 67 feria vi 90 feria iiii
7 23 feria i 45 feria vii 68 feria v 91 feria i
8 24 feria v 46 feria iiii 69 feria ii 92 feria vii
9 25 feria iiii 47 feria iii 70 feria vi 93 feria iiii
10 26 feria i 48 feria vii 71 feria v 94 feria i
11 27 feria v 49 feria vi 72 feria ii 95 feria vii
12 28 feria iiii 50 feria iii 73 feria i 96 feria iiii
13 29 feria i 51 feria vii 74 feria v 97 feria i
(recte feria iii)
14 30 feria v 52 feria vi 75 feria ii 98 feria vii
15 31 feria iiii 53 feria iii 76 feria i 99 feria iiii
16 32 feria i 54 feria vii 77 feria v 100 feria iii
17 Blank Space W
18 33 feria vii 55 feria vi 78 feria ii 101 feria vii
(feria v Wallis)
Page Break V
19 34 feria iv 56 feria iii 79 feria i 102 feria iiii
20 35 feria i 57 feria ii 80 feria v 103 feria iii
21 36 feria vii 58 feria vi 81 feria iiii 104 feria iiii
(recte feria vii)
22 37 feria iv 59 feria iii 82 feria i 105 feria iiii
(recte feria vi)
23 38 feria i 60 feria ii 83 feria v 106 feria iii
24 39 feria vii 61 feria vi 84 feria iiii 107 feria vii
Figure 4b  Reconstruction of second page of tables in the Codex Remensis Dep-
erditus.


 ALDEN A. MOSSHAMMER

1 108 feria vi 12 III Id Apr Dom

2 109 feria iii 13 III Non Apr Dom
3 110 feria vii 14 XIIII Kl Apr Dom
4 111 feria vi 15 VI Id Apr Dom
5 112 feria iii 16 III Kl Apr Dom
6 feria vii 17 XIII Kl Mai Dom
7 feria vi 18 Prid Non Apr Dom
8 feria iii 19 VI Kl Apr Dom
9 feria vii 20 XVII Kl Mai Dom
10 feria vi 21 Prid Kl Apr Dom
11 feria iii 22 X Kl Apr Dom
12 23 Prid Id Apr Dom
13 24 VI Kl Apr Dom
14 1 XVI Kl Mart Dom 25 XVI Kl Mai Dom
(recte XIV Kl Mai)
15 2 III Non Apr Dom 26 VI Id Apr Dom
16 3 VII Kl Apr Dom 27 VIIII Kl Apr Dom
17 4 XIIX Kl Mai Dom 28 Prid Id Apr Dom
18 5 III Kl Apr Dom 29 Prid Non Apr Dom
19 6 XI Kl Apr Dom 30 XIII Kl Apr Dom
20 7 III Id Apr Dom 31 V Id Apr Dom
21 9 XVII Kl Mai Dom 32 Prid Kl Apr Dom
22 10 XVII Id Apr Dom 33 XII Kl Mai Dom
(recte VII Id Apr Dom)
23
24 11 X Kl Apr Dom 34 Non Apr Dom
Figure 4c  Reconstruction of third page of tables in the Codex Remensis Dep-
erditus.


 TOWARDS A NEW EDITION OF THE COMPUTUS OF AD 243

1 35 V Kl Apr Dom 57 XIII Kl Mai Dom

2 36 XVI Kl Mai Dom 58 III Non Apr Dom
3 37 Kl <Apr> Dom 59 VII Kl Apr Dom
(Kl Dom WV) (VII Id. Apr W)
4 38 VIIII Kl Apr Dom 60 XVII Kl Mai Dom
(recte XVIII Kl Mai)
5
6 39 Id Apr Dom 61 III Kl Apr Dom
7 40 V Kl Apr Dom 62 XI Kl Apr Dom
8 41 XV Kl Mai Dom 63 III Id Apr Dom
9 42 V Id Apr Dom 64 VII Kl Apr Dom
10 43 VIII Kl Apr Dom 65 XVII Kl Apr Dom
(recte XVII Kl Mai Dom)
11 44 Id Apr Dom 66 VII Id Apr Dom
12 45 Non Apr Dom 67 X Kl Apr Dom
13 46 XII Kl Apr Dom 68 III Id Apr Dom
14 47 IIII Id Apr Dom 69 III Non Apr Dom
15 48 Kl Apr Dom 70 XIIII Kl Apr Dom
16 49 XVIII Kl Apr Dom 71 VI Id Apr Dom
(recte XVIII Kl Mai)
17 50 VIII Id Apr Dom 72 III Kl Apr Dom
18 51 IIII Kl Apr Dom 73 XIII Kl Mai Dom
19 52 IIII Id Apr Dom 74 Prid Non Apr Dom
20 53 IIII Non Apr Dom 75 VI Kl Apr Dom
21 54 VIII Kl Apr Dom 76 XVIII Kl Apr Dom
(recte XVII Kl Mai Dom)
22 55 VII Id Apr Dom 77 Prid Kl Apr Dom
23 56 IIII Kl Apr Dom 78 X Kl Apr Dom
24
Figure 4d  Reconstruction of fourth page of tables in the Codex Remensis Dep-
erditus.


 ALDEN A. MOSSHAMMER

1 79 Prid Id Apr Dom 108 IIII Id Apr Dom

2 80 VI Kl Apr Dom 109 IIII Non Apr Dom
3 81 XVI Kl Apr Dom 110 VIII Kl Apr Dom
(recte XVI Kl Mai)
4 82 VI Id Apr Dom 111 VII Id Apr Dom
(V Id Apr W)
5 83 VIIII Kl Apr Dom 112 IIII Kl Apr Dom
6
7 84 Prid Id Apr Dom
8 85 Prid Non Apr Dom
9 86 XIII Kl Apr Dom
10 87 V Id Apr Dom
11 88 Prid Kl Apr Dom
12 89 XII Kl Apr Dom
(recte XII Kl Mai)
13 90 Non Apr Dom
14 91 V Kl Apr Dom
15 92 XVI Kl Apr Dom
(recte XVI Kl Mai)
16 93 Kl Apr Dom
17 94 VIIII Kl Apr Dom
18 95 Id Apr Dom
19 96 V Kl Apr Dom
20 97 XV Kl Mai Dom
21 98 V Id Apr Dom
22 99 VIII Kl Apr Dom
23 100 Id Apr Dom
24 101 Non Apr Dom
25 102 XII Kl Apr Dom
26 103 IIII Id Apr Dom
27 104 Kl Apr Dom
28 105 XVIII Kl Apr Dom
(recte XVIII Kl Mai)
29 106 VIII Id Apr Dom
30 107 IIII Kl Apr Dom
Figure 4e  Reconstruction of fifth page of tables in the Codex Remensis Dep-
erditus.


 JAN ZUIDHOEK

THE INITIAL YEAR OF DE RATIONE

PASCHALI AND THE RELEVANCE OF
ITS PASCHAL DATES

Abstract
According to Mc Carthy and Breen, the Latin text De ratione paschali (DRP) is
a translation of the original Greek paschal tract of Anatolius, bishop of Laodi-
cea (†c.AD  282), containing his famous 19-year paschal cycle. The 19-year
periodic sequence of epacts of this paschal cycle has a Metonic structure. This
implies that there must have been a Metonic sequence of dates of paschal full
moons (PFM) on the basis of which Anatolius constructed his paschal cycle.
This particular Metonically structured predecessor of the classical Alexandrian
cycle must have been constructed by Alexandrian computists somewhere be-
tween the years AD 250 and 270. Thus the sequence of dates of the Anatolian
PFM marks the connection between the proto-Alexandrian cycle and the se-
quence of paschal dates of DRP.
Subsequently, somewhere in the first quarter of the fourth century, the church
of Alexandria replaced its Metonic sequence of dates of paschal full moons then
in use, possibly the proto-Alexandrian cycle, with the classical Alexandrian cy-
cle. It will be demonstrated that this was done by advancing the date of the
March equinox by one day and the computistical lunar months by two days.
Keywords
Anatolius, Annianus, Bede, Theodor Zahn, Bruno Krusch; Christian era, Ni-
san, Jewish calendar, Julian calendar, Metonic sequence, Alexandrian calendar,
classical Alexandrian PFM, proto-Alexandrian cycle, Six Millennium Catalog,
lunisolar conjunction, Metonic adaptation, Passover, March equinox, synodic
period, Anatolian PFM, Anatolian paschal Sunday, Anatolian calendar, Anato-
lian paschal data; Alexandria, Jerusalem, Nicaea; De ratione paschali.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 71-93
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114734
 JAN ZUIDHOEK

Introduction
According to Daniel Mc  Carthy and Aidan Breen, the Latin text De
ratione paschali (DRP)1 is a fourth-century accurate translation of the
original Greek paschal tract of Anatolius,2 a native of third-century Al-
exandria, who was a famous computist and bishop of Laodicea (on the
coast of Syria), around the seventies of the third century.3 This implies
that the paschal table that forms part of DRP is nothing less than Ana-
tolius’ lost 19-year paschal cycle. This agrees with the view of Theodor
Zahn that the Greek original of DRP could very well have been written
around the year AD 270.4 This text was of great importance in the early
Middle Ages to scholars of the Christian paschal calendar;5 however, it
still remains unclear whether the paschal table of DRP was part of the
original work. For Bruno Krusch, writing in the year 1880, the text was
no more than a forgery.6 On the other hand, according to Mc Carthy
and Breen, DRP represents a consistent whole,7 which was consulted
by early medieval scholars of the Christian Pasch. The very core of DRP
is formed by the 19-year periodic sequence of paschal dates, and it is by
locating this sequence of dates in the Christian era that the relevance of
these dates can be established.
Paschal tables were used since the third century, most of them pro-
vided with a periodic sequence of dates of the paschal full moon (PFM)
for consecutive calendar years. These periodic sequences of dates of PFM
were originally intended as substitutes for dates of the observed four-
teenth day of the Jewish month Nisan. It was not before the middle of
the fourth century that the Jewish calendar began gradually to become
fixed.8

1
Mc Carthy and Breen (2003), 25. I would like to express my gratitude to both
Dáibhí Ó Cróinín and Daniel Mc Carthy for their encouragement and support in de-
veloping my ideas about the development of the Alexandrian computus in the second
half of the third century and the first quarter of the fourth; to Dáibhí for his constructive
editorial patience; to Dan for the continuously inspiring e-mail correspondence on these
ideas since July 2009.
2
Mc Carthy and Breen (2003), 142.
3
Mc Carthy and Breen (2003), 18.
4
Zahn (1884), 196.
5
Mc Carthy and Breen (2003), 24.
6
Krusch (1880), 311.
7
Mc Carthy and Breen (2003), 142.
8
Mosshammer (2008), 88.


 THE INITIAL YEAR OF DE RATIONE PASCHALI

In this article various periodic sequences of dates of PFM falling be-

tween 21 March and 20 April (inclusive) of 19 consecutive Julian calen-
dar years will be considered. In these sequences each following date can
be obtained by advancing its immediate predecessor by 10, 11, or 12 days
modulo 30 days, such that, over each period of 19 years, the total number
of advanced days amounts to 210 (e.g., 4×10+10×11+5×12=210 days).
Among such 19-year periodic sequences of dates, particularly those in
which each following date can be obtained by advancing its immedi-
ate predecessor by 11 days modulo 30 days, or, once every 19 years, by
12 days modulo 30 days (i.e. 18×11+1×12=210 days), are of especial
interest, because they reflect in the most natural way the phenomenon of
the 19-year lunar cycle. That is, a time interval of 19 solar years consists
on average of nearly as many days as 235  synodic lunar months. Tak-
ing the solar year to be the tropical year consisting of 365.2422 days,9
and the synodic lunar month to consist of 29.53059  days,10 we find
19×365.2422≈6939.602  days and 235×29.53059≈6939.689  days re-
spectively. Thus in about 6940 days the sun and moon return to the same
relative positions. This astronomical synchronism was known as early
as the fifth century before Christ in Mesopotamia, as well as in Greece,
where the Athenian astronomer Meton rediscovered it. Hence, such se-
quences of dates, as well as their particular structure, are called Metonic.
Metonic sequences of dates can be separated into two types: the first
type is characterized by 11 ordinary advances of 11 days, 1 saltus advance
of 12 days, and 7 ordinary regressions of 19 days; the second type is char-
acterized by 12 ordinary advances of 11 days, 6 ordinary regressions of
19 days, and 1 saltus regression of 18 days. For example, it is easy to verify
that four out of the five 19-year periodic sequences of dates of PFM in
Figure 6, being Metonic, are Metonic sequences of dates of the first type.
We note that two of them have not 21 March but 23 March as their earli-
est permissible date, and, as a consequence, not 18 April but 20 April as
their latest permissible date.
The historically most important example of a Metonic sequence
of dates is the classical Alexandrian cycle, which is of the first type. Its
initial year is considered to be the year AD 285, the first year of con-
sulship of the Emperor Diocletian,11 and its saltus is inserted between
the years AD 303 and 304 modulo 19, i.e., at the end of its nineteenth

9
Smart (1958), 146.
10
Smart (1958), 169.
11
Mosshammer (2008), 26–27.


 JAN ZUIDHOEK

year AD 303 modulo 19.12 This Metonic sequence of Alexandrian cal-

endar dates of the classical Alexandrian PFM, which was constructed
by Alexandrian computists somewhere in the first quarter of the fourth
century, say around the year AD 310,13 forms the backbone of the classi-
cal Alexandrian paschal cycle.14 That is the great 532-year Easter cycle,15
which was invented, and drawn up, by the Alexandrian monk Annianus
around the year AD 400.16 Its 532-year structure, unknown to Dionysius
Exiguus, was reconstructed by insular computists (most famously Bede)
in the late seventh-, early eighth-century.17 The initial year of Annianus’
Alexandrian calendar version is AD 285;18 and the initial year of Bede’s
Julian calendar version is AD 532.19
Although constructed before the council of Nicaea in AD 325, the
classical Alexandrian cycle cannot have been the very earliest example of
a Metonic sequence of dates of PFM, because the Metonic sequence of
epacts of the first type is a structural feature of the framework of DRP.20
Consequently it must also have formed a structural feature of the Greek
original of DRP, and therefore must have originally derived from a third-
century proto-Alexandrian cycle, i.e., the Metonic sequence of dates of
the proto-Alexandrian PFM, from which Anatolius started in order to
construct his 19-year paschal cycle. Because 19-year periodic sequences
of dates of PFM were not constructed before the middle of the third
century,21 it must have been around the year AD 260 that this particular
Metonic sequence of dates of PFM was constructed by computists of the
church of Alexandria, among them possibly Anatolius before his epis-
copal consecration.22 From the dates of this proto-Alexandrian PFM it
was relatively straighforward to calculate the corresponding dates of the

12
Mosshammer (2008), 149.
13
Declercq (2000), 66.
14
Neugebauer (1979), 59–63.
15
Wallis (1999), 392–404.
16
Declercq (2000), 30.
17
Declercq (2000), 156–58.
18
Neugebauer (1979), 59; Wallis (1999), 392–404.
19
Wallis (1999), 392.
20
Mc Carthy and Breen (2003), 68.
21
Lejbowicz (2006), 45.
22
Declercq (2000), 65–66.


 THE INITIAL YEAR OF DE RATIONE PASCHALI

paschal Sunday, namely by applying the third-century Alexandrian prin-

ciple ‘paschal Sunday is the first Sunday after the paschal full moon’.23
Unfortunately, no ancient or medieval manuscript has survived con-
taining a third-century Metonic sequence of dates of PFM; in fact, we
know next to nothing with certainty about the proto-Alexandrian cy-
cle. However, by using NASA’s Six Millennium Catalog of Phases of the
Moon24 we can calculate Metonically structured good approximations
to it. Thus, with the help of the Six Millenium Catalog we can endeavour
to reconstruct the Julian calendar equivalent of the proto-Alexandrian
cycle. These Metonically structured approximations, with the 19-year
periodic sequence of dates of the Anatolian PFM, will enable us to pre-
cisely locate the sequence of paschal dates of DRP in the Christian era.
To summarize: in this article I wish to work out how the following
three fundamental 19-year periodic sequences of Julian calendar dates of
PFM relate to each other:

1. The proto-Alexandrian cycle, i.e., the unknown Metonic se-

quence of dates of the PFM constructed by Alexandrian compu-
tists around the year AD 260.
2. The sequence of dates of the PFM constructed by Anatolius
around the year AD 270, deduced from the sequence of paschal
dates of DRP.
3. The classical Alexandrian cycle, i.e., the important Metonic se-
quence of dates of the classical Alexandrian PFM constructed by
Alexandrian computists around the year AD 310.

Constructing Metonically structured approximations

to the proto-Alexandrian cycle
To reconstruct the proto-Alexandrian cycle, it is necessary to first con-
struct the best possible Metonically structured approximations to the
spring equinoctal full moons (luna 14) as observed from Alexandria. To
do this, we need to imagine the way in which, around the year AD 260,
the third-century Alexandrian computists endeavoured to construct
their first Metonic sequence of dates of PFM. They must first have ex-
perimented with various sequences of successive ‘dates of the fourteenth

Lejbowicz (2006), 48.

23

Espenak (2007).
24


 JAN ZUIDHOEK

day of Nisan’, remembering that, at that time, dates of the fourteenth day
of Nisan were generally not exactly calculable. To be more precise, vari-
ous sequences of successive dates had to be substituted for the observed
fourteenth day of Nisan. I take these dates as the fourteenth day of an
abstract computistical month, Nisan*, to be defined hereafter in this sec-
tion. To obtain these dates of the fourteenth day of the computistical
month Nisan*, of course via their dates of the first day of the computisti-
cal month Nisan*, the computists in question needed tables of then rela-
tively recent dates and times of lunisolar conjunction; such tables they
could have derived from the schematic lunar calendar as adapted to the
Alexandrian calendar by Alexandria’s Jewish community of the time.25
These dates of new moons must have belonged to a sufficient time inter-
val, encompassing at least twice 19 calendar years. Let us take the inter-
val I* consisting of the period of years AD 220 to 260.
It is likely that, in order to obtain their dates of the fourteenth day of
their computistical month Nisan*, these Alexandrian computists made
use of the old rule concerning the beginning of Nisan, i.e. the rule that
the first day of Nisan usually began with the second sunset in Jerusalem
after the new moon of Nisan, each individual new moon of Nisan actu-
ally being the lunisolar conjunction just preceding Nisan. The rule that
Nisan usually began with the second sunset in Jerusalem after lunisolar
conjunction is an obvious consequence of the Babylonian criterion for
first new-moon visibility, namely the rule that around the beginning of
spring every new moon will be visible for the first time, weather permit-
ting, at the beginning of the evening, between 24  and 48  hours after
lunisolar conjunction.26
Applying their rule concerning the beginning of Nisan means that, for
each lunisolar conjunction which heralded the beginning of the month
Nisan, the date of the first day of this month Nisan can be estimated
quite accurately by simply adding 2 or 3 days to the local Jerusalem date
(from midnight to midnight) of this lunisolar conjunction, depending
on whether the local Jerusalem time of this lunisolar conjunction fell
before or after 18:00, respectively. However, in order to reconstruct the
proto-Alexandrian PFM cycle, we have not only to take into account
the Babylonian 24 hours rule, but also the Jewish religious principle that
Pesach had to be celebrated as early as possible in spring, which meant
since the mid-third century that the fourteenth day of Nisan had to fall

25
Neugebauer (1979), 8.
26
Bruin (1977), 333.


 THE INITIAL YEAR OF DE RATIONE PASCHALI

on or immediately after the spring equinox.27 It is because of this Jewish

religious principle that, in the second half of the third century and in
the earlier fourth century, Alexandrian computists reckoned with the
March equinox. At least from about the year AD 253, their date for the
March equinox was the Ptolemaic date of 22 March.28 This implies that
the earliest possible date of the proto-Alexandrian PFM must have been
either 22 or 23 March. Incidentally, in the second half of the third cen-
tury, the astronomical date of the March equinox was, in reality, either
21 March or 20 March. We may safely assume that it is only since about
the year AD 303 that the church of Alexandria considered 21 March to
be the date of the March equinox.29
Reconstructing the proto-Alexandrian cycle, as well as reconstruct-
ing the classical Alexandrian cycle, appears to be possible because of the
fact that, at least during the years AD 253 and 325, Alexandrian compu-
tists respected the rule of the equinox. On the other hand, estimating the
dates of the fourteenth day of the real Jewish month Nisan during these
same years is a precarious undertaking, because, at the time, their Jewish
contemporaries in actual practice ignored the rule of the equinox, and
in consequence began their month Nisan, and celebrated their paschal
feast, in fact a month too early many a time.30
Normally, in order to obtain estimated dates of the fourteenth day
of Nisan, we would apply the rule concerning the beginning of Nisan
to the local Jerusalem dates and times of lunisolar conjunction. How-
ever, in order to reconstruct the proto-Alexandrian cycle, we must apply
an adapted version of the rule concerning the beginning of Nisan, in
which the computistical month Nisan* differs from Nisan in that, firstly,
its fourteenth day must obey the rule of the March equinox falling on
22 March; secondly, its first day must begin with the second sunset in
Alexandria after its new moon. In practice, this means that from each
individual local Alexandrian date and time of new moon of the compu-
tistical month Nisan* belonging to I*, an approximate (from midnight
to midnight) date of the daylight part of the first (from sunset to sunset)
day of Nisan* can be obtained by simply adding 2 or 3 days to the date
in question, depending on whether the time in question falls before or
after 18:00, respectively.

27
Stern (2001), 50.
28
Lejbowicz (2006), 48.
29
Lejbowicz (2006), 48.
30
Stern (2001), 66–67.


 JAN ZUIDHOEK

The local Alexandrian dates and times of the new moon of the com-
putistical month Nisan* that we need for the determination of the ap-
proximate dates of the first day of Nisan* can be obtained from the ta-
ble of local Greenwich dates and times of lunisolar conjunction in the
NASA Six Millennium Catalog. Local Greenwich dates and times can
be converted to local Alexandria dates and times by adding 1:59, this
being the time difference between the geographical longitudes of Green-
wich and Alexandria. After having thus obtained approximate dates of
all first days and then of all fourteenth days of the computistical month
Nisan* belonging to the time interval I*, we are in a position to construct
the best possible Metonically structured approximation(s) to the proto-
Alexandrian cycle. To do this, Alexandrian third-century computists
must have been able to distinguish good Metonically structured approx-
imations to sequences of their dates of the fourteenth day of Nisan from
poor ones, and able to determine which good ones were the very best.
In order to be able to appropriately constrain the Julian calendar
position of the computistical month Nisan* over the time interval I*,
we must establish suitable lower and upper limit dates, with a difference
of about 29.5  days, between which limits to analyze local Alexandria
times of lunisolar conjunction. These must determine a first day of the
month Nisan*, such as to guarantee that the corresponding dates of the
fourteenth day of Nisan* will be not earlier than 22  March, and not
later than 20 April (a period of 30 days, inclusively). This is achieved by
taking our lower limit as 6 March 18:00, and our upper-limit as 5 April
6:00, because adding 3+13 days to 6 March gives 22 March, and adding
2+13 days to 5 April gives 20 April.
The structure of Figure 1 suggests the way in which third-century Al-
exandrian computists must have succeeded in constructing their Meton-
ic sequence of dates of the proto-Alexandrian PFM. Hence the column
Conjunction* shows the estimated local Alexandria times of new moon
of the computistical month Nisan*, each one obtained by adding 1:59
to the corresponding local Greenwich times of lunisolar conjunction as
reckoned by the NASA Six Millennium Catalog.31 These conjunctions are
chosen so that the corresponding local Alexandria times all fall between
6 March 18:00 and 5 April 6:00. Column 1 Nisan* shows our estimated
dates of the first day of the computistical month of Nisan*, each of them
obtained from the corresponding time in Conjunction* by applying the
Nisan* version of the rule concerning the beginning of Nisan. Column

Espenak (2007).
31


 THE INITIAL YEAR OF DE RATIONE PASCHALI

Year Conjunction* 1 Nisan* 14 Nisan* Adv ProtoAlexPFM*

221 11 March 17:01 13 March 26 March – 26 March
222 30 March 14:57 1 April 14 April −19 14 April
223 19 March 19:58 22 March 4 April +10 3 April
224 7 March 20:06 10 March 23 March +12 23 March
225 26 March 13:13 28 March 10 April −18 11 April
226 15 March 19:16 18 March 31 March +10 31 March
227 3 April 17:35 5 April 18 April −18 19 April
228 23 March 9:06 25 March 7 April +11 8 April
229 13 March 1:49 15 March 28 March +10 28 March
230 1 April 2:02 3 April 16 April −19 16 April
231 21 March 12:21 23 March 5 April +11 5 April
232 9 March 15:21 11 March 24 March +12 24/25 March
233 28 March 7:53 30 March 12 April −19 12 April
234 17 March 9:27 19 March 1 April +11 1 April
235 5 April 4:55 7 April 20 April −19 20 April
236 24 March 16:42 26 March 8 April +12 9 April
237 14 March 8:54 16 March 29 March +10 29 March
238 2 April 10:08 4 April 17 April −19 17 April
239 23 March 0:47 25 March 7 April +10 6 April
240 11 March 8:57 13 March 26 March +12 26 March
241 30 March 3:00 1 April 14 April −19 14 April
242 19 March 3:12 21 March 3 April +11 3 April
243 8 March 6:12 10 March 23 March +11 23 March
244 26 March 2:50 28 March 10 April −18 11 April
245 15 March 16:09 17 March 30 March +11 31 March
246 3 April 16:58 5 April 18 April −19 19 April
247 24 March 9:45 26 March 8 April +10 8 April
248 12 March 22:50 15 March 28 March +11 28 March
249 31 March 19:38 3 April 16 April −19 16 April
250 20 March 22:26 23 March 5 April +11 5 April
251 9 March 22:37 12 March 25 March +11 24/25 March
252 27 March 16:41 29 March 11 April −17 12 April
253 17 March 1:18 19 March 1 April +10 1 April
254 5 April 0:27 7 April 20 April −19 20 April
255 25 March 16:54 27 March 9 April +11 9 April
256 14 March 8:59 16 March 29 March +11 29 March
257 2 April 8:22 4 April 17 April −19 17 April
258 22 March 16:13 24 March 6 April +11 6 April
259 11 March 17:29 13 March 26 March +11 26 March

Figure 1  Best Metonic approximations to the proto-Alexandrian PFM cycle

All years are AD and all dates Julian calendar dates; the column Conjunction*
refers to the estimated local Alexandrian times of the new moon of the com-
putistical month Nisan* as obtained from the NASA Six Millenium Catalog;
1 Nisan* refers to the estimated dates of the first day of Nisan* on the basis of
Conjunction*; 14 Nisan* refers to the estimated dates of the fourteenth day of
the computistical month Nisan*; Adv refers to the advance in days from the
preceeding 14 Nisan*; ProtoAlexPFM* contains the two equally best Metoni-
cally structured approximations to the sequence of dates of 14 Nisan*.


 JAN ZUIDHOEK

14 Nisan* shows our estimated dates of the fourteenth day of the com-
putistical month Nisan*, each of them obtained by adding 13 days to the
corresponding date in column 1 Nisan*. Note that the sequence of dates
in column 14 Nisan* is not Metonic, not even 19-year periodic.
It is by systematically examining the advance or regression in 14 Ni-
san* from year to year given in the column Adv of Figure 1 that we obtain
the best two Metonically structured approximations to that sequence
of dates, as given in protoAlexPFM* of Figure 1. The first of these two
sequences, which includes 24 March at AD 232 and 251 and which I
will refer to as MSA1, is a Metonic sequence of the first type and has its
saltus advance between the years AD 231–32 and 250–51. The second
sequence, which includes 25  March at AD  232 and 251 and which I
will refer to as MSA2, is of the second type and has its saltus regression
between the years AD 232–33 and 251–52.
Next we turn our attention to the column Adv of Figure  1 where
positive values show the advances in 14 Nisan* ranging from +10 to
+12 days, and negative values show the regressions in 14 Nisan* rang-
ing from −17 to −19 days. Examination of these shows that in most in-
stances a non-standard Metonic advance of +10 or +12, or regression of
−17 or −18, is followed within two or three years by another, balancing,
advance or regression. For example, at AD 223–24 the advances of +10
and +12 may both be adjusted to the standard advance of +11 by ad-
vancing the date of 14 Nisan* at AD 223 by one day. We may concisely
list all such instances as follows:

Years Adv 14 Nisan* Adjustment Resultant Adv

223–24 +10; +12 Advance 223 one day +11; +11
225–26 −18; +10 Retard 225 one day −19; +11
227–29 −18; +11; +10 Retard 227–28 one day −19; +11; +11
236–37 +12; +10 Retard 236 one day +11; +11
239–40 +10; +12 Advance 239 one day +11; +11
244–47 −18; +11; −19; +10 Retard 244–46 one day −19; +11; −19; +11
252–53 −17; +10 Retard 252 one day −18; +11

These ten one-day adjustments leave only AD 232 with its advance
of +12, and AD 252 with its regression of −18, as the only non-standard
values, and consequently these identify the two possible locations for the
saltus. If we advance 25 March AD 251 by one day we obtain MSA1, or
if we retard 24 March AD 232 by one day we obtain MSA2.


 THE INITIAL YEAR OF DE RATIONE PASCHALI

Each of these choices to locate the saltus implies just one further ad-
justment to the 14 Nisan* sequence, and so they both represent equally
good Metonic approximations, and consequently both dates are given
under ProtoAlexPFM*. Thus the minimum number of adjustments re-
quired to obtain a Metonic sequence from 14 Nisan* is eleven, and these
are shown in ProtoAlexPFM*. This results in two possible best Metonic
sequences, MSA1 with date 24 March, and MSA2 with date 25 March.
In more mathematical terms: in order to obtain the best Metonically
structured approximation(s) to the sequence of dates in column 14Ni-
san*, divide the period AD 222–59 (because the dates of 14Nisan* are
the same in AD 221 and 259) into two parts of 19 years each, AD 222–
40 and 241–59. Compare the dates of 14Nisan* of the first 19 years with
the second. The maximum of consecutive years in which the two 19-
year periods agree is 3, and this happens only once, in AD 229–31 vs
AD 248–50 (28 March, 16 April, 5 April). This three-year sequence is
Metonically structured, with a regression of 19 days between 28 March
and 16 April, and an advance of 11 days between 16 April and 5 April.
Because of the already existing Metonic structure of the two advances
(–19 and +11) between these three years, it is best to insert the saltus
in any of the remaining 17 advances of the years AD 231–48 (and its
19-year equivalents). This leads to 17 possible Metonically structured
approximations, depending on the place of the saltus, defined as MSAx,
with the saltus inserted at the end of the year AD (230+x)modulo19
(e.g., MSA9 refers to the saltus being inserted at the end of the year
AD 239modulo19).
For each integer x with 0<x<18, the margin between any given date
of MSAx and the corresponding date in the column 14Nisan* in Fig-
ure 1 is 0, 1, 2, or 29 days. Let T(x) be the sum of all 38 margins between
MSAx and 14Nisan* for the period AD 222–59 (with 0<x<18). Since
the two advances AD  229–31 and the two advances AD  248–50 are
Metonic, the margin is 0 for the six years covered by these two periods;
this leaves 38–6=32 margins to be considered (for each 0<x<18). MSAx
is a best approximation to the sequence of dates in column 14Nisan*,
if and only if T(x) is minimal. It turns out that T is a monotonic non-
decreasing function, with T(1)=T(2)=11<T(x)<T(17)=93, for each
integer x with 2<x<17.
Thus, MSA1 and MSA2 (combined in the column protoAlexPFM*
of Figure 1) are the two best Metonically structured approximations to
the sequence of dates in column 14Nisan. The only difference between
MSA1 and MSA2 is that the former has 24 March, the latter 25 March


 JAN ZUIDHOEK

in AD 232modulo19. MSA1 is of the first type, MSA2 of the second. At

this point of the investigation, MSA1 and MSA2 appear to be our best
two candidates for the proto-Alexandrian cycle, but it will be demon-
strated below that only MSA1 can be the proto-Alexandrian cycle.
Qualitively, it can be seen that the real moon exhibits intermittent
accelerations and decelerations, but these balance out within a few years,
so that the average advance or regression corresponds to the standard
Metonic advance of 11 days or regression of 19 days. Then, at approxi-
mate 19-year intervals an acceleration occurs which does not balance
out, and this marks the location of the saltus needed in the Metonic se-
quence. These features of the moon’s average annual advance or regres-
sion narrowly constrain the best possible Metonic sequences.

Defining the sequence of dates of Anatolian PFM

At first sight, the paschal dates of DRP (each of them accompanied by
an integer lunar age between 14 and 20 inclusive) seem to be Julian cal-
endar dates. However, according to Daniel Mc Carthy and Aidan Breen,
they are calendar dates belonging to a special variant of the Julian calen-
dar ingeniously invented by Anatolius;32 this Anatolian version differs
from the Julian calendar in the length of its solar year, 365 2/19 compared
to the Julian 365 ¼ days (leading to a difference of approximately one
day in seven years). Consequently, I call them dates of the Anatolian pas-
chal Sunday, because, within the framework of this Anatolian calendar
they all did indeed fall on a Sunday.33 However, during a certain rela-
tively short time they must have been not only Anatolian but also Julian
calendar dates.34 Therefore, it is of interest to examine the dates which
we obtain by considering the original paschal dates of DRP simply as if
they were not Anatolian but pure Julian calendar dates. In which case it
is preferable to speak of dates of the Anatolian paschal day, because with-
in the framework of the Julian calendar they did not all fall on a Sunday.
Starting from these dates of the Anatolian paschal day, including
their respective lunar age, we next focus on the sequence of Julian calen-
dar dates that I refer to as dates of the Anatolian PFM. These dates can
be obtained from the former by advancing each to a date having lunar

32
Mc Carthy and Breen (2003), 100.
33
Mc Carthy and Breen (2003), 68.
34
Mc Carthy and Breen (2003), 100.


 THE INITIAL YEAR OF DE RATIONE PASCHALI

age 14. Hence, in Figure 2 we show the Metonic sequence of epacts of

the first type that was used for the construction of the sequence of pas-
chal dates of DRP.35 Next, the dates of the Anatolian paschal day, with
their respective lunar ages,36 followed by dates of the Anatolian PFM,
each with lunar age 14.

Epact (1 Jan) AnatPday luna AnatPFM luna adapAnat

1 16 April 18 12 April 14 12 April
12 1 April 14 1 April 14 1 April
23 21 April 16 19 April 14 20 April
4 13 April 18 9 April 14 9 April
15 29 March 14 29 March 14 29 March
26 18 April 16 16 April 14 17 April
7 9 April 17 6 April 14 6 April
18 1 April 20 26 March 14 26 March
29 14 April 14 14 April 14 14 April
10 6 April 17 3 April 14 3 April
21 29 March 20 23 March 14 23 March
2 11 April 14 11 April 14 11 April
13 3 April 17 31 March 14 31 March
24 23 April 19 18 April 14 19 April
5 8 April 14 8 April 14 8 April
16 31 March 17 28 March 14 28 March
27 19 April 18 15 April 14 16 April
8 11 April 20 5 April 14 5 April
20 27 March 17 24 March 14 24 March

Figure 2  The sequence of dates of the Anatolian PFM

All dates are Julian calendar dates and all six sequences are 19-year periodic.
The column Epact refers to the Metonic sequence of epacts of DRP; AnatPday
refers to the Anatolian paschal day, followed by the respective lunar age; Anat-
PFM refers to the dates of the Anatolian PFM, followed by its lunar age, 14;
adapAnat contains the Metonic adaptation of the Anatolian PFM.

Because of the proximity in place and time of their origins, it is an obvi-

ous question whether, and if so to what extent, the sequences of dates
of the Anatolian PFM and the proto-Alexandrian PFM relate to each
other. While the sequence of dates of the Anatolian PFM is not Me-
tonic, it is straightforward to obtain the best Metonic approximation by
simply retarding by one day the dates for epacts (of 1 January) 23, 26, 24,
and 27, and this is included in Figure 2 as adaptAnat. We observe that

35
Mc Carthy and Breen (2003), 68.
36
Mc Carthy and Breen (2003), 68.


 JAN ZUIDHOEK

the position of the saltus in this Metonic adaptation agrees with the one
in the Metonic sequence of epacts that underlies the structure of DRP.
In spite of their entirely different sources, the 19-year periodic se-
quences of dates in column ProtoAlexPFM* of Figure  1 and column
adapAnat of Figure  2 show a striking correspondence, which suggests
the hypothesis that the Anatolian PFM could be just a Metonic adapta-
tion of the sequence of dates of the proto-Alexandrian cycle. In order to
examine this suggestion in a constructive and systematic way, we need
to compare the two sequences of dates of PFM in question. However,
the paschal table of Anatolius has no known initial year. On the other
hand, both of them are 19-year periodic, and it is therefore appropriate
to compare the sequence of dates of the Anatolian PFM with MSA1 and
MSA2 in all 19 possible alignments.

Determining the initial year of DRP

In this section, I locate the paschal dates of DRP in the Christian era.
This will be done by comparing the sequence of dates of the Anatolian
PFM with MSA1 and MSA2 as substitutes for the unknown proto-Al-
exandrian cycle. Needless to say, the way in which the sequence of dates
of the Anatolian PFM has been obtained is completely independent of
the way in which MSA1 and MSA2 were obtained.
Until recently, there was difficulty with regard to locating the pas-
chal dates of DRP in the Christian era,37 but in view of the closeness
in place and time of the origins of the sequences of dates of the proto-
Alexandrian and the Anatolian PFM, there is likely to be a connection
between these two sequences of dates of PFM. Because we do not know
the real sequence of dates of the proto-Alexandrian PFM exactly, I will
endeavour to solve this problem by means of comparing the sequence
of dates of the Anatolian PFM with MSA1 and MSA2. Thus I compare
AnatPFM of Figure 2 with protoAlexPFM* of Figure 1, in each of the 19
possible alignments, since both columns have a period of 19 years.
I begin by considering the year AD 260, which was the initial year of
DRP suggested by Daniel Mc Carthy and Aidan Breen,38 and so align the
first Anatolian PFM of 12 April with the protoAlexPFM* of 14 April.
Consequently there is a two-day difference and these values, and those for

Mc Carthy and Breen (2003), 100.

37

Mc Carthy and Breen (2003), 101.

38


 THE INITIAL YEAR OF DE RATIONE PASCHALI

the subsequent years are shown in Figure 3. We see that in only 5 or 6 in-
stances out of 19 does the date of the Anatolian PFM differ by 0 or 1 day
from that of proto-Alexandrian PFM. However, if the first paschal date of
DRP had indeed belonged to any year AD 260modulo19, then we would
have expected that in all, or almost all, 19 instances out of 19 the absolute
difference in question would have been 0 or 1 day. This implies that, in
all probability, the initial year of DRP is not any year AD 260modulo19.

Year protoAlexPFM* AnatPFM Diff

260 14 April 12 April 2
261 3 April 1 April 2
262 23 March 19 April 27
263 11 April 9 April 2
264 31 March 29 March 2
265 19 April 16 April 3
266 8 April 6 April 2
267 28 March 26 March 2
268 16 April 14 April 2
269 5 April 3 April 2
270 24/25 March 23 March 1/2
271 12 April 11 April 1
272 1 April 31 March 1
273 20 April 18 April 2
274 9 April 8 April 1
275 29 March 28 March 1
276 17 April 15 April 2
277 6 April 5 April 1
278 26 March 24 March 2

Figure 3  Testing AD 260 as the initial year of DRP

All years are AD, all dates Julian calendar dates, and all three intended se-
quences of dates 19-year periodic. The column protoAlexPFM* contains the
two Metonically structured approximations to the proto-Alexandrian cycle;
AnatPFM refers to the dates of the Anatolian PFM aligned to the year AD 260
(modulo19); Diff shows the absolute difference in days between the proto-Al-
exandrian PFM date and the corresponding date of the Anatolian PFM.

For the case that the first paschal date of DRP belonged to some year
AD 263 modulo 19 (an alternative initial year discussed by Mc Carthy
and Breen39), we establish in just the same manner that in only 10 or
11 instances is the absolute difference 0 or 1 day. Therefore, in all prob-
ability, the initial year of DRP is not any year 263modulo19.

39
Mc Carthy and Breen (2003), 101.


 JAN ZUIDHOEK

However, comparing the sequence of dates of the Anatolian PFM

with MSA1 and MSA2 for the case where the first paschal date of DRP
is aligned to the year AD 271 (another alternative initial year discussed
by Mc Carthy and Breen40) it turns out that in all 19 instances the ab-
solute difference in question is 0 or 1 day. Thus the initial year of DRP
seems to be a year AD 271modulo19.

Year protoAlexPFM* AnatPFM Diff

271 12 April 12 April 0
272 1 April 1 April 0
273 20 April 19 April 1
274 9 April 9 April 0
275 29 March 29 March 0
276 17 April 16 April 1
277 6 April 6 April 0
278 26 March 26 March 0
279 14 April 14 April 0
280 3 April 3 April 0
281 23 March 23 March 0
282 11 April 11 April 0
283 31 March 31 March 0
284 19 April 18 April 1
285 8 April 8 April 0
286 28 March 28 March 0
287 16 April 15 April 1
288 5 April 5 April 0
289 24/25 March 24 March 0/1

Figure 4  Testing AD 271 as the initial year of DRP

All years are AD, all dates Julian calendar dates, and all three sequences of dates
19-year periodic. The column protoAlexPFM* contains the two Metonically
structured approximations to the proto-Alexandrian cycle; AnatPFM refers
to the dates of the Anatolian PFM aligned to the year AD 271 (modulo19);
Diff shows the absolute differences in days between the approximate date of the
proto-Alexandrian PFM and the corresponding date of the Anatolian PFM.

There are 16 remaining possible alignments, in each of which it turns out

that in 0 instances out of 19 does the date of the Anatolian PFM differ by
0 or 1 day from the corresponding date of the proto-Alexandrian PFM.
This then implies that the year AD 271 must be considered the initial
year of DRP, the year AD 252 being too early, and the year AD 290 too
late for the floruit of Anatolius.

Mc Carthy and Breen (2003), 101.

40


 THE INITIAL YEAR OF DE RATIONE PASCHALI

Completing the reconstruction of the proto-Alexandrian cycle

From Figure 4 it follows not only that the year AD 271 must be the ini-
tial year of DRP, but also that one of our two Metonically structured
approximations to the proto-Alexandrian cycle, namely MSA1, not only
differs by just one day in only 4 out of the 19 years with the sequence of
dates of Anatolian PFM, but is also identical with its Metonic adapta-
tion. As a matter of fact, the Metonic structure of MSA1 reflects the
structure of the sequence of epacts of DRP, exactly as the Metonic struc-
ture of the classical Alexandrian cycle reflects the structure of the se-
quence of epacts of Dionysius Exiguus’ paschal table. Therefore, MSA1
is the most likely candidate for consideration as the proto-Alexandrian
cycle itself, given the definition of the latter as the particular Metonic se-
quence of dates of PFM from which, around the year AD 270, the bish-
op of Laodicea constructed his 19-year paschal cycle. This completes our
reconstruction of the proto-Alexandrian cycle.
The sequence of dates of the Anatolian PFM marks the connection
between the proto-Alexandrian cycle and the sequence of paschal dates
of DRP. This underlines the relevance of all these three 19-year periodic
sequences of dates.

Comparing reconstructions of the Anatolian

19-year paschal cycle
The results of the two previous sections enable us to work out when the
Anatolian paschal day was a Julian Sunday, and on which Julian calendar
dates between the years AD 260 and 290 the Anatolian paschal Sunday
must have fallen. This is shown in Figure 5, which relates the sequences of
dates of the Anatolian PFM, the Anatolian paschal day, and the Anatolian
paschal Sunday to each other. In addition, in order to facilitate comparison
of these three sequences of dates with their predecessors, the sequences of
dates of the proto-Alexandrian PFM and of the related proto-Alexandri-
an paschal Sunday (obtained as the first Sunday after the date of the proto-
Alexandrian PFM) are also included in this table. The column AnatPday
shows that only in eight instances did the Anatolian paschal day fall on a
Julian Sunday, namely in the years AD 264 up to and including 271. This
concentration of Anatolian paschal days falling on Julian Sunday immedi-
ately before the year that Anatolius’ paschal cycle was constructed, is quite
remarkable. Similarly, it is precisely in this seventh decade of the third cen-


 JAN ZUIDHOEK

tury that the dates of the proto-Alexandrian and the Anatolian paschal
Sunday frequently coincided, which fact, for its part, may be considered
a strong confirmation of our supposition that the Greek original of DRP
must have been constructed around the year AD 270.

Year Proto­ AnatPFM AnatPday AnatPSun ProtoAlexP­

Alex­PFM Sun
261 3 April 3 April Sa. 6 April Su. 7 April Su. 7 April
262 23 March 23 March Sa. 29 March Su. 30 March Su. 30 March
263 11 April 11 April Sa. 11 April Su. 12 April Su. 12 April
264 31 March 31 March Su. 3 April Su. 3 April Su. 3 April
265 19 April 18 April Su. 23 April Su. 23 April Su. 23 April
266 8 April 8 April Su. 8 April Su. 8 April Su. 15 April
267 28 March 28 March Su. 31 March Su. 31 March Su. 31 March
268 16 April 15 April Su. 19 April Su. 19 April Su. 19 April
269 5 April 5 April Su. 11 April Su. 11 April Su. 11 April
270 24 March 24 March Su. 27 March Su. 27 March Su. 27 March
271 12 April 12 April Su. 16 April Su. 16 April Su. 16 April
272 1 April 1 April Mo. 1 April Su. 31 March Su. 7 April
273 20 April 19 April Mo. 21 April Su. 20 April Su. 27 April
274 9 April 9 April Mo. 13 April Su. 12 April Su. 12 April
275 29 March 29 March Mo. 29 March Su. 28 March Su. 4 April
276 17 April 16 April Tu. 18 April Su. 16 April Su. 23 April
277 6 April 6 April Mo. 9 April Su. 8 April Su. 8 April
278 26 March 26 March Mo. 1 April Su. 31 March Su. 31 March
279 14 April 14 April Mo. 14 April Su. 13 April Su. 20 April
280 3 April 3 April Tu. 6 April Su. 4 April Su. 4 April
281 23 March 23 March Tu. 29 March Su. 27 March Su. 27 March
282 11 April 11 April Tu. 11 April Su. 9 April Su. 16 April
283 31 March 31 March Tu. 3 April Su. 1 April Su. 1 April
284 19 April 18 April We. 23 April Su. 20 April Su. 20 April
285 8 April 8 April We. 8 April Su. 5 April Su. 12 April
286 28 March 28 March We. 31 March Su. 28 March Su. 4 April
287 16 April 15 April Tu. 19 April Su. 17 April Su. 17 April
288 5 April 5 April We. 11 April Su. 8 April Su. 8 April
289 24 March 24 March We. 27 March Su. 24 March Su. 31 March

Figure 5  The weekdays of Anatolian paschal dates

All years are AD and all dates Julian calendar dates. The column protoAl-
exPFM contains the Metonic sequence of dates of the proto-Alexandrian PFM
(MSA1); AnatPFM contains the 19-year periodic sequence of dates of the Ana-
tolian PFM; AnatPday contains the 19-year periodic sequence of dates of the
Anatolian paschal day; AnatPSun refers to the dates of the Anatolian paschal
Sunday, based on AnatPday; protoAlexPSun refers to the dates of the proto-
Alexandrian paschal Sunday, based on protoAlexPFM.


 THE INITIAL YEAR OF DE RATIONE PASCHALI

From the results of the two previous sections we are also able to com-
pare the crucial sequence of dates of the Anatolian PFM, which has no
Metonic structure, to other Metonic sequences of dates. For example,
the proto-Alexandrian cycle, the reconstructions of the 19-year cycle
of Anatolius by Eduard Schwartz,41 and by Alden Mosshammer,42 and
the classical Alexandrian cycle. Figure 6 compares all these interesting
19-year periodic sequences of dates of PFM over the time interval of
the years AD 270 to 290. Although the year AD 285 is considered to
be its initial year, the classical Alexandrian cycle did not really exist in
the third century, nevertheless it is of interest to see how it relates to its
third-century predecessors.

Year protoAlexPFM AnatPFM AnatSchw AnatMoss AlexPFM

271 12 April 12 April 11 April 10 April 10 April
272 1 April 1 April 31 March 30 March 30 March
273 20 April 19 April 19 April 18 April 18 April
274 9 April 9 April 8 April 7 April 7 April
275 29 March 29 March 28 March 27 March 27 March
276 17 April 16 April 16 April 15 April 15 April
277 6 April 6 April 4 April 4 April 4 April
278 26 March 26 March 24 March 24 March 24 March
279 14 April 14 April 12 April 12 April 12 April
280 3 April 3 April 1 April 1 April 1 April
281 23 March 23 March 20 April 21 March 21 March
282 11 April 11 April 9 April 9 April 9 April
283 31 March 31 March 29 March 29 March 29 March
284 19 April 18 April 17 April 17 April 17 April
285 8 April 8 April 6 April 6 April 5 April
286 28 March 28 March 26 March 26 March 25 March
287 16 April 15 April 14 April 14 April 13 April
288 5 April 5 April 3 April 2 April 2 April
289 24 March 24 March 23 March 22 March 22 March

Figure 6  Comparing 19-year PFM sequences

All years are AD, all dates Julian calendar dates, all sequences of dates 19-year
periodic and all, except AnatPFM, Metonic. The column protoAlexPFM con-
tains the sequence of dates of the proto-Alexandrian PFM (MSA1); AnatPFM
contains the sequence of dates of the Anatolian PFM; AnatSchw contains the
19-year cycle of Anatolius as reconstructed by Schwartz; AnatMoss contains
the 19-year cycle of Anatolius as reconstructed by Mosshammer; AlexPFM
contains the classical Alexandrian cycle.

41
Mosshammer (2008), 149.
42
Mosshammer (2008), 161.


 JAN ZUIDHOEK

Considering Figure 6, it is important to note how little the sequences of

dates of the proto-Alexandrian PFM and the Anatolian PFM differ. The
most obvious explanation for this small difference is that it must have
been via the sequence of dates of the Anatolian PFM that the sequence
of paschal dates of DRP was developed, starting from the Metonic se-
quence of dates of the proto-Alexandrian PFM, by advancing just four
dates by one day. A second agreement between these two sequences of
dates of PFM should be mentioned, namely that AnatPFM has not 21 or
22  but 23  March as its earliest permissible date. We also note that in
both AnatPFM and protoAlexPFM, the immediate successor of 1 April
is not 21  March, which strongly supports our assumption that, in the
second half of the third century, the church of Alexandria considered
not 21 March, but 22 March as the date of the March equinox.
It is also of interest to compare the positions of the saltus in the various
sequences of dates of Figure 6. The location of the saltus of the Anatolian
PFM is inserted at the end of the eighteenth year of the epacts. Moreover,
the position of this saltus agrees with the position of the saltus in its Me-
tonic adaptation, see Figure 2. Therefore, the position of the saltus in the
Anatolian PFM agrees also with the position of the saltus in the proto-
Alexandrian PFM. But the positions of the saltus in the reconstructions
by Schwartz and Mosshammer, and the classical Alexandrian cycles, all
differ from that of the proto-Alexandrian and the Anatolian PFM, see
Figure 6. Incidentally, all four of these Metonic 19-year sequences of dates
of PFM in Figure 6, are of the first type. Moreover, comparing protoAl-
exPFM, AnatSchw, and AnatMoss with column AnatPFM makes clear
that the reconstructions of Anatolius by Schwartz and by Mosshammer
differ to such an extent from the sequence of dates of the Anatolian PFM
that they, contrary to the sequence of dates of the proto-Alexandrian
PFM, cannot possibly represent Anatolius’ paschal cycle.
On the other hand, the remarkably simple difference between pro-
toAlexPFM and AlexPFM of Figure 6 of mostly 2 days, but sometimes
3  days, requires some explanation. It is the much simpler relationship
between protoAlexPFM and AlexPFM, rather than between AnatPFM
and AlexPFM that makes clear that there is no evidence for any evolu-
tion from the proto-Alexandrian cycle to the classical Alexandrian cycle
via the sequence of dates of the Anatolian PFM. The classical Alexan-
drian cycle must have been constructed on the basis of new principles,
not by manipulating the dates of the Anatolian PFM.
We may wonder which motives the church of Alexandria could have
had around the year AD 310 in opting for such a drastic change of its


 THE INITIAL YEAR OF DE RATIONE PASCHALI

dates of PFM, such that most of them were abruptly advanced by 2 days,
some of them even by 3 days. On reflection, the church of Alexandria
then had really just two motives for such a drastic recalculation of its
dates of PFM. First, of course, her change of view with regard to the
date of the March equinox, from 22 March to 21 March. Secondly, its
change of view on the extent to which it wanted to take into account the
Jewish principle of observing the crescent of the new moon. This latter
change was part of the overall development whereby, around the turn of
the third century, many churches, among them the churches of Rome
and Alexandria, became more and more sceptical about Jewish princi-
ples, and tried to distance themselves as much as possible from them. For
instance, the bishops who assembled in Nicaea in the year AD 325 were
‘united in their rejection of Jewish practices’.43
I conclude that, around the year AD 310, the church of Alexandria
decided to construct a new Metonic sequence of dates of PFM, in or-
der to accommodate the new date of the March equinox, and to leave
as little time as possible for catching sight of the crescent of the new
moon after lunisolar conjunction. It is obvious that it realized the lat-
ter by simply defining the first day of its new month Nisan as the lo-
cal Alexandrian ‘from sunset to sunset’ day of the moment of its new
moon, which is equivalent to defining the beginning of the first day of
the month Nisan as the moment of the last sunset in Alexandria before
its new moon, instead of defining it as the moment of the second sunset
in Alexandria after its new moon. This amounts to advancing its month
Nisan by 2 days, and implies that this new month of Nisan began just
2 days before first crescent visibility, somewhere between 0 and 24 hours
before its new moon.
Incidentally, in Antiquity lunar months beginning before first cres-
cent visibility were not unusual. Ancient Egyptian lunar months—their
first day being the first day of old moon crescent invisibility at sunrise—
began on average more than one day before first crescent visibility, be-
cause in ancient Egypt the first day of new moon crescent visibility at
sunset was, in 70 percent of the cases, 1 day, in the remaining 30 per-
cent of the cases, 2 days after their first day.44 Even ancient Greek lunar
months began usually before first crescent visibility, because they began
just 1 day later than ancient Egyptian lunar months.45 Around the thir-

43
Declercq (2000), 52.
44
Parker (1950), 13.
45
Pritchett (1982), 266.


 JAN ZUIDHOEK

ties of the fourth century lunar months of the special calendar used by
the Jewish community of Antioch began mostly 1 day before first cres-
cent visibility.46

Summary
In an endeavour to reconstruct the real but unknown proto-Alexandrian
PFM cycle, I made use of NASA’s Six Millennium Catalog in construct-
ing the two best Metonically structured approximations to it. Then by
comparing these Metonic sequences of dates with the sequence of dates
of the Anatolian PFM deduced from the sequence of paschal dates of
DRP I located these dates in the Christian era commencing in the year
AD 271.
As a matter of fact, my first approximation, MSA1, of the first type,
not only differs by one day in just 4 out of 19 dates with the dates of the
Anatolian PFM, but is also its Metonic adaptation. Therefore, MSA1
is our best representative of the proto-Alexandrian cycle. We conclude
that the sequence of dates of the Anatolian PFM can be considered to
be the causal link from the proto-Alexandrian cycle to the sequence of
paschal dates of DRP, which underlines the relevance of these three 19-
year periodic sequences of dates.
The fact that the year AD 271 is the initial year of DRP has other
remarkable consequences, one being that it is only in the years from
AD 264 up to and including 271 that the Anatolian paschal day was a
Julian Sunday. Further, the Metonic sequence of dates of PFM referred
to as the 19-year cycle of Anatolius as reconstructed by Schwartz, and
the one referred to as the 19-year cycle of Anatolius as reconstructed by
Mosshammer, contrary to our proto-Alexandrian cycle, cannot possibly
have underlain Anatolius’ paschal cycle.
Finally, the remarkably simple difference of mostly 2 days, or 3 days,
between the proto-Alexandrian and the classical Alexandrian cycle can
be qualitatively explained by realizing that the latter is nothing other
than the Metonic sequence of dates of PFM with which the church of
Alexandria around the year AD  310 replaced her proto-Alexandrian
PFM then in use. This must have been the result, not of adapting the
sequence of dates of the Anatolian PFM, but of a completely new defini-
tion of her Metonic sequence of dates of PFM. This was done by advanc-

Stern (2001), 130–32; Mosshammer (2008), 185.

46


 THE INITIAL YEAR OF DE RATIONE PASCHALI

ing her date of the March equinox by 1 day, and defining the first day of
her new month Nisan as the local Alexandria ‘from sunset to sunset’ day
of its new moon, thus apparently advancing her dates of PFM by 2 days
(mostly) or 3 days. This qualitative account for the simple difference in
question can be supplemented with a quantitative one by performing a
complete reconstruction of the classical Alexandrian cycle, which I will
leave for a different occasion.


 DANIEL MC CARTHY

THE PASCHAL CYCLE OF ST PATRICK

Abstract
Notwithstanding the substantial corpus of early medieval references to St Patrick
and his works, the only account we have of a paschal cycle associated with him is
that provided by Cummian in his letter addressed to Ségéne of Iona and Béccán
the hermit composed in c.AD 633. In this letter, Cummian identified himself and
his community with Patrick, but he furnished only limited technical details for
both Patrick’s cycle and the cycle he indicated that he and his community had re-
cently adopted. However, critical examination of Cummian’s account shows that
Patrick had adapted the 532-year paschal cycle compiled by Victorius of Aquit-
aine in AD 457, and that this was the cycle that Cummian’s community and other
influential southern Irish churches resolved to adopt at the synod of Mag Léne in
c.AD 630. Consequently, Cummian’s account of Patrick’s cycle, the earliest attest-
ed reference to him, holds significant implications for both the chronology of Pat-
rick’s mission to Ireland, and for the expansion of his cult in the seventh century.
Keywords
Patrick, Cummian, Ségéne; Iona, Mag Léne; paschal cycles, paschal controversy,
Alexandrian Pasch, early Roman Pasch, latercus Pasch; early Irish Christianity.

Introduction
In c.AD 633 Cummian, a scholar-monk located in southern Ireland ad-
dressed a letter concerning the celebration of the Pasch to Ségéne, abbot
of Iona c.AD 624–52, and to a hermit, Béccán, and to their sages (sapi­
entibus). The circumstances of the letter have been well-documented by
modern scholarship: in AD 628–29 Pope Honorius wrote to the Irish
churches criticizing their divergence in the matter of celebration of the

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 94-137
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114735
 THE PASCHAL CYCLE OF ST PATRICK

Pasch, and Cummian’s community agreed to conform; Cummian him-

self delayed for a year while he studied the question, and then with the
support of the principals of at least five influential churches he initiated
a synod at Mag Léne; here it was unanimously agreed that ‘they would
celebrate Easter with the Universal Church’; shortly thereafter, however, a
solitary dissenter precipitated division, and consequently a delegation was
sent to Rome to observe the practice there, and upon its return reported
that the different nations had all observed one Pasch; Cummian then
wrote his letter in c.AD 633 to Ségéne and Béccán, evidently in response
to earlier criticism from them.1 A copy of Cummian’s letter has survived
in a single manuscript, London, British Library, Cotton Vitellius A XII,
79r–83r, written in a twelfth-century English Carolingian minuscule,
and this was discovered and first published by Archbishop James Ussher
in 1632, whose edition was reprinted in Paris under the patronage of An-
toine Vion in 1665, by Charles Elrington in 1846, and by Jacques-Paul
Migne in 1863. In 1988, Maura Walsh and Dáibhí Ó Cróinín published a
new edition from the manuscript collated with Ussher’s edition, to which
they added a parallel translation, and an invaluable introduction and
commentary. In 1995, David Howlett published a detailed analysis of the
structure of the letter demonstrating the sophisticated character of Cum-
mian’s composition, and most recently Immo Warntjes published an ex-
tensive literature survey and a re-analysis of the chronology of the letter.2
In the opening sentences of his letter, Cummian indicated both the
circumstances that had prompted his writing and his modus operandi:3

Uerba excusationis meae in faciem sanctitatis uestrae proferre procaciter

audeo, sed excusatum me habere uos ut patres cupio, testem Deum inuo­
cans in animam meam, quod non contemtus uestri gratia nec fastu moralis
sapientiae, cum caeterorum despectu, sollempnitatem festi paschalis cum
caeteris sapientibus suscepi. Ego enim primo anno quo cyclus quingentorum

1
Kenney (1929), 220–21 (circumstances and chronology); Walsh and Ó Cróinín
(1988), 3–7 (circumstances and date of letter), 7–19 (identities of Cummian, Ségéne
and Béccán), 93 (citation); Ó Cróinín (1995), 152–54 (circumstances).
2
Ussher (1632), 24–35 (edition); Vion (1665), 17–23 (reprint of Ussher); El-
rington (1847–64), iv 432–44 (reprint of Ussher); PL 87, 969–78 (Migne’s reprint of
Ussher); Walsh and Ó Cróinín, (1988), 3–54 (introduction), 51–53 (manuscript), 56–
97 (edition, translation, and commentary); Howlett (1995a), 91–102 (structural analy-
sis); Warntjes (2010a), LXXXI–LXXXII n. 219 (literature survey and date of letter);
Warntjes (2015), 41–47 (identifies AD 633 and 643 as possible dates for the letter, but
concludes that the earlier ‘still has to be preferred on the evidence known at present’).
3
Cummian, De controversia paschali ll.  6–17 (ed. and trans.  by Walsh and Ó
Cróinín (1988), 56–59).


 DANIEL MC CARTHY

.xxx. duorum annorum a nostris celebrari orsus est non suscepi, sed silui; nec
laudare nec uituperare ausus utpote Aebreos, Gregos, Latinos—quas linguas
ut Ieronimus ait, in crucis suae titulo Christus consecrauit—superare mini­
me in scientia me credens. Deinde apostolum interrogans dicentem: Omnia
probate, quod bonum est tenete, antequam gustarem non fastidiui. Hinc
per annum secretus sanctuarium Dei ingressus, hoc est scripturam sanctam,
ut u‹a›lui inuolui, deinde historias, postremo cyclos quos inuenire potui.

‘I dare to offer the words of my excuse boldly in the face of your holiness
but I hope that you as fathers forgive me, calling upon God as witness
on my soul that it is not out of scorn for you nor puffed up with pride of
moral wisdom and the contempt of others that I have undertaken, along
with other sages, the solemnity of the Paschal feast. For in the first year in
which the cycle of 532 years began to be celebrated by our party I did not
accept it, but remained silent; I dared neither to praise nor to condemn
it, since I believe that I by no means surpass in knowledge the Hebrews,
Greeks and Latins, whose languages, as Jerome says, Christ consecrated
in the inscription of His cross. Therefore, asking the Apostle who says:
Test everything, hold fast what is good, before I tried it, I did not disdain
it. Hence, having cloistered myself for a year and having entered the sanc-
tuary of God (that is sacred Scripture) I studied as much as I was able,
then I examined the histories, and finally the cycles which I could find.’

Thus, while his community had recently changed their paschal tradition
to one employing a cycle of 532 years, Cummian more cautiously spent a
year studying, with other sages, first Scripture, then histories, and finally
the paschal cycles he was able to find. The subsequent arrangement of his
letter closely reflects these three categories.4 His scriptural section is de-
voted to establishing the limits for the age of the paschal moon approved
by scriptural and patristic authorities, while his historical section is al-
most entirely concerned with exhortations on the necessity for church
unity and universality, and this theme remains prominent through his
discussion of paschal cycles and his conclusion. While Cummian no-
where explicitly indicated what paschal tradition his community had
previously followed, his remark that ‘our elders, however, whom you hold
as a cloak for your rejection, kept simply and faithfully, without blame of
any contradiction or animosity, that which they knew to be best in their
day, and prescribed for their posterity thus, according to the Apostle’ ac-

4
Cummian, De controversia paschali ll. 1–17 (introduction), 18–85 (Scripture),
86–203 (histories), 204–57 (cycles), 258–304 (conclusions) (Walsh and Ó Cróinín
(1988), 56–97).


 THE PASCHAL CYCLE OF ST PATRICK

knowledges that his community shared a common ecclesiastical heritage

with Ségéne and Béccán.5 Further, his statement that the delegation sent
to Rome had witnessed Pasch celebrated on a date ‘in which we differed
by a whole month’, infers that they too had hitherto been observing the
same paschal table.6 These references, and indeed the authority of his ad-
dress to Ségéne and Béccán, imply that Cummian’s community had pre-
viously followed the 84-year latercus brought to Ireland in c.AD 425, and
taken subsequently by St Columba to Iona in c.AD 562, and followed by
his community there up until their paschal reform of AD 716.7
The letter therefore provides an extensive account of Cummian’s
views on the celebration of the Pasch, including in particular some com-
putistical details of the paschal cycles known to him. As part of this he
enumerated ten ‘cycles of different computations’ (cyclos computationum
diuersarum), and the first of these furnishes the only known reference to
a paschal cycle associated with St Patrick. It is the purpose of this article
to examine Cummian’s account of Patrick’s cycle and his own paschal
preferences in order to determine the characteristics of Patrick’s cycle
and the cycle adopted by Cummian and his community.
Regarding nomenclature, in this paper I shall refer to the practice of
celebrating Pasch on the Sunday falling on luna 15–­­­2­1 after the spring
equinox on 21 March between the termini of 22 March and 25 April in-
clusive as the ‘Alexandrian Pasch’, or the ‘Alexandrian tradition’. This pas-
chal tradition is often misrepresented as ‘Roman’ in western literature be-
cause the Roman church adopted the Alexandrian tradition in the form of
the cycle of Dionysius Exiguus in the seventh century, and eventually suc-
ceeded in persuading all the other western churches to adopt that cycle.8

5
Cummian, De controversia paschali ll. 114–18 (Walsh and Ó Cróinín (1988),
74–75): Seniores uero nostri, quos in <ue>lamine ‹repulsionis habetis, quod optimum in
diebus suis esse nouerunt simpli›citer et fi<deli>ter sine <culpa> contradictionis ullius <et
animositatis obs>eruauerunt et suis poster<is sic> mandauerunt, iuxta <apostolum […] >.
Passages between angle brackets ‘<…>’ identify restorations by the editors from Ussher’s
edition of passages illegible in the manuscript; cf. Walsh and Ó Cróinín (1988), 54.
6
Cummian, De controversia paschali l. 283 (Walsh and Ó Cróinín, (1988), 94–
95): in quo mense integro disiuncti sumus; Warntjes (2010a), LXXXI n. 219, and Warn-
tjes (2015), 43–49 (chronological re-analysis based upon collating Victorian, Dionysiac,
and latercus paschal dates).
7
Mc Carthy (2011), 69 (date of latercus arrival); Mc Carthy (1999–2005), s.a.
562 (Columba’s navigatio to Iona), s.a. 716 (Iona paschal reform); Mc Carthy (1993),
218–19 (latercus table); Blackburn and Holford-Strevens (1999), 870–75 (latercus pas-
chal dates and moons for AD 354–773).
8
Corning (2006), 4–8 (‘Roman tradition’ referencing the Alexandrian Pasch;
cf. her title); Hughes (1966), 115 (‘Roman Easter’ referencing the Dionysiac cycle).


 DANIEL MC CARTHY

However, before adopting the Alexandrian tradition the Roman church

had celebrated Pasch on the Sunday falling on luna 16–22 between the
termini of 22 March and 21 April inclusive. This tradition dated from at
least the time of Hippolytus (†c.AD 236), a priest and teacher of Rome,
who compiled a 112-year paschal cycle with these principles.9 These prin-
ciples were maintained in the fourth century by the 84-year Supputatio
Romana, and subsequently their lunar limits were accommodated in the
532-year paschal table compiled by Victorius of Aquitaine in AD 457,
and I shall refer to this tradition as the ‘earlier Roman Pasch’, or ‘earlier
Roman tradition’.10 Finally, I shall refer to the 84-year paschal table com-
piled c.AD 410 by Sulpicius Severus in southern Gaul which celebrated
Pasch on the Sunday falling on luna 14–20 between the termini 26 March
and 23 April inclusive, as the ‘latercus Pasch’, or ‘latercus tradition’.11
It is worth emphasizing that, while these various paschal traditions
are principally distinguished here just in terms of the age limits of their
paschal moon, these age limits derive from theological analyses of Scrip-
ture and the pronouncements of early church fathers. It was the theologi-
cal differences between these analyses that generated great passion in the
­medieval paschal controversies, and this passion may be readily observed
in Cummian’s letter.12 If the medieval paschal controversy is character-
ized as merely concerning the calculation of the date of Easter, then the
crux of the controversy is missed. Finally, for simplicity of presentation,
I shall, like Cummian, refer to all paschal tables as ‘paschal cycles’, regard-
less of whether their tabulated paschal data constitute a cycle or not.

9
Mac Carthy (1901), xxxi–xl; Schwartz (1905), 29–40; Blackburn and Holford-
Strevens (1999), 805–06; Mosshammer (2008), 116–29 (Hippolytus’ cycle).
10
Jones (1943), 27 (Supputatio Romana); Krusch (1880), 32–84 (Supputatio Ro­
mana—Krusch’s hypothesis of two versions of this Supputatio, an older and younger,
was shown by Schwartz (1905), 40–58, to be mistaken; see Warntjes (2010a), XXXVI
n. 76); Blackburn and Holford-Strevens (1999), 792–93, 807–08 (Roman paschal his-
tory); Mosshammer (2008), 204–13 (‘Paschal calculations at Rome’).
11
Mc Carthy (1993), 218 (celebration after 25 March), 218–19 (luna 14–20 and
the 84-year paschal table); Mc  Carthy (1994), 38–44 (latercus compiled by Sulpicius
Severus in c.AD 410).
12
Mc Carthy (1993), 204, and Warntjes (2015), 79–82 (theological significance
of luna 14–16); Cummian, De controversia paschali ll.  47–51 (Walsh and Ó Cróinín
(1988), 60–61): ‘Lex loquitur, apostolus probat; hoc superest, ut contradictor abiciatur ut
peruicax. Manifestum est enim transitum post pascha fuisse. Sanguis igitur salutem presti­
tit, non transitus, quia, ut transitus non noceret, obsistit sanguis.’ Hoc timui et me perculit
(‘‘The Law speaks, the Apostle proves, this remains: that the contradictor be cast out
as a liar. For it is clear that the passing over was after the Pasch. Therefore the blood
gave safety, not the passing over, because the blood prevents the passing over from doing
harm.’ I feared this and it disturbed me.’)


 THE PASCHAL CYCLE OF ST PATRICK

Plate 1  London, British Library, Cotton Vitellius A XII, 81v, where Cummi-
an’s account of the ten paschal cycles commencing Postremo ad cyclos starts on
line 20, and concludes with in aequinoctio in the middle of line 41. The upper
and lower margins of the folio were damaged in the Cottonian fire of 1730.


 DANIEL MC CARTHY

Cummian’s account of Patrick’s paschal cycle

I commence with Walsh and Ó Cróinín’s text and translation of Cum-
mian’s introduction to and enumeration of the ten cycles, the Latin of
which is reproduced in lines 20–41 of Plate 1 from the manuscript Lon-
don, British Library, Cotton Vitellius A XII, 81v:13

Postremo ad cyclos computationum diuersarum quid unaquaeque lingua

de cursu solis et lunae sentiret, conuersus totus, licet diuerse alium in die,
alium in luna, alium in mense, alium in bissexto, alium in epacta, alium
in aucmento lunari, quod uos saltum dicitis, inueni cyclos contra hunc,
quem <u>os tenetis, esse contrarios. Primum illum quem sanctus Patricius
papa noster tulit et fecit, in quo luna a .xiiii. usque in .xxi. regulariter, et
equinoctium a .xii. Kl. Aprilis obseruatur. Secundo Anatolium, quem uos
extollitis, qui dicit ‘ad ueram paschae rationem numquam peruenire’ eos
qui cyclum ‘.lxxxiiii. annorum’ obseruant. Tertio Theophilum. Quarto Di­
onisium. Quinto Cyrillum. Sexto Morinum. Septimo Agustinum. Octauo
Uictorium. Nono Pacomium monachum Aegipti cenobiorum fundatorem
cui ab angelo ratio paschae dictata est. Decimo trecentorum decem et octo
episcoporum decennouennalem cyclum, ‘qui Grece enneacedeciterida’ dici­
tur, in quo Kalendae Ianuarii lunaeque eiusdem diei et initia primi men­
sis ipsiusque .xiiii.mae14 lunae recto iure, ac si quodam clarissimo tramite,
‘ignorantiae relictis tenebris,’ studiosis quibusque cunctis temporibus sunt
adnotatae, quibus paschalis sollennitas probabiliter inueniri potest. Hunc
inueni ualde huic, cuius auctorem locum tempus incertum habemus, esse
contrarium in kalendis, in bissexto, in epacta, in .xiiii. luna, in primo
mense, in aequinoctio.

‘Finally I thoroughly examined the cycles of different computations

to see what each language thinks about the course of the sun and the
moon, and I found cycles that are in disagreement with the one which
you hold, although diversely one in the day, another in the moon, an-
other in the month, another in the bissextile, another in the epact, and
another in the lunar augment (which you call the saltus). The first is
that which holy Patrick, our bishop, brought and followed, in which
the moon is regularly observed from the fourteenth to the twenty-first,
and the equinox from March 21st. Secondly, I found Anatolius (whom
you extol) who says that those who observe a ‘cycle of eighty-four years

13
Cummian, De controversia paschali ll. 204–23 (Walsh and Ó Cróinín (1988),
82–87).
14
Walsh and Ó Cróinín (1988), 86 l. 218: the edition reads .xiiiii.mae lunae, but
examination of the manuscript shows the reading to be .xiiiimæ in accordance with the
translation ‘fourteenth moon’; cf. l. 36 of Plate 1).


 THE PASCHAL CYCLE OF ST PATRICK

can never arrive’ at the correct reckoning of Easter. Thirdly Theophilus;

fourth Dionysius; fifth Cyril; sixth Morinus; seventh Augustine; eighth
Victorius; ninth the monk Pacomius, founder of the monasteries of
Egypt, to whom the reckoning of Easter was dictated by an angel; tenth
the nineteen-year cycle of the 318 bishops ‘which is called enneacedeci­
terida in Greek’ in which the Kalends of January and the fourteenth
moons of that month have been correctly noted, as if by a most clear
path, ‘leaving aside the shadows of ignorance’, for studious men for all
times, by which the feast of Easter can with certainty be found. Indeed,
I found this cycle contrary in the kalends, the bissextile, the epact, the
fourteenth moon, the First Month and the equinox to that one whose
author, place and time we are uncertain of.’

Here Cummian began by making the taunting statement that the cycles
he found all disagreed with the cycle insisted upon by Ségéne and Béc-
cán. He then enumerated ten different cycles, supplying for each a named
authority, and concluded with the affirmation that the tenth cycle, that
of the 318 bishops at Nicaea, conflicted in virtually every computistical
criterion with ‘that one whose author, place and time we are uncertain
of ’.15 In this way Cummian bracketed his ten authorized cycles with two
disparaging references to the latercus maintained by Ségéne and Béccán,
the authorship of which he asserted to be uncertain.
Turning to his first named cycle, that of Patrick, here Cummian as-
sociated the equinox with the twelfth kalends of April, that is 21 March,
and stated that the moon was regularly observed from luna 14 to luna
21.16 He thus suggested that in Patrick’s cycle luna 14–21 fell systemati-
cally after the equinox, and, while he did not explicitly indicate on which
lunar ages the Pasch should be celebrated, his statement of these limits
is in accordance with orthodox Alexandrian doctrine, as Walsh and Ó
Cróinín pointed out.17 This doctrine, while repeating the luna 14–21
specification of Exodus 12:18, insisted that for Christians, should luna

15
Cummian, De controversia paschali ll. 93–94 (Walsh and Ó Cróinín (1988), 70),
identifies the synod of the 318 bishops with Nicaea: Nicena etiam sinodus trecentorum .x.
et .viii. episcoporum.
16
See text and transalation at n. 13, where Cummian’s expression equinoctium a
.xii. Kl. Aprilis is curious, since the equinox was considered to fall ‘on’ 21 March rather
than ‘from’ that date. However, Leofranc Holford-Strevens considers that the second
‘a’ has been ‘induced by the previous ‘a’ by false parallelism’, and consequently that the
translation would better read ‘the equinox on 21 March’ (personal communication, 29
February 2012).
17
Walsh and Ó Cróinín (1988), 30 (Alexandrian doctrine); cf. Jones (1943), 30
(rejection of celebration on luna 14 by Theophilus, bishop of Alexandria).


 DANIEL MC CARTHY

14 fall on Sunday, then Pasch must be deferred to the following Sun-

day on luna 21, so that in Alexandrian practice only luna 15–21 were
celebrated. We shall see that examination of Cummian’s discussion of
Scripture supports this interpretation.
While Cummian supplied no account of the length of the lunar cy-
cle employed by Patrick, it is the case that the lunar cycles employed by
seven of his nine other authorities are readily established to have been
of length 19  years. Namely, Anatolius, Dionysius, Theophilus, Cyril,
Victorius, and Pacomius were all considered in medieval times to have
compiled paschal tables employing a 19-year lunar cycle, while Cum-
mian himself expressly attributed a 19-year cycle to the 318 bishops at
Nicaea.18 On the other hand, while no paschal cycle is known to be as-
sociated with the name Morinus, the tract attributed to him, Disputa­
tio Morini, emphatically endorsed the Dionysiac version of the 19-year
cycle.19 Similarly, in the case of Augustine, while there is no evidence
that he compiled a paschal cycle, there can be no doubt that Cummian
considered him an advocate of the Alexandrian paschal tradition. For he
approvingly cited Augustine by name three times, two of which vigor-
ously repudiate celebration on luna 14. These are:20

Quod uenerandae memoriae sanctus Agustinus pulchre explanat: ‘Dies

festus,’ inquiens, ‘pascha est, iuxta euangelistam Lucam dicentem: Appro­
pinquabat autem dies festus azimorum in quo necesse erat occidi pascha.
Pascha occiditur et uiuificatur. Pars autem diei festi, .xiiii. luna, non totus
dies, in qua seducimur; sed in parte et sabbato otioso, et neomenia buci­
nata, quae sunt umbra, non corpus, Christi. Umbra occiditur, ueritas ui­
uificatur.’

18
Mc Carthy and Breen (2003), 51 (Anatolius’ cycle); Krusch (1938), 69–74 (Di-
onysius’ cycle); Jones (1943), 31 (Theophilus’ cycle); Jones (1943), 47–49 (Cyril’s cycle);
Krusch (1938), 27–52 (Victorius’ cycle); Jones (1943), 33, 82 (Pacomius’ and Nonae
Aprilis); Jones (1943), 17–25, especially 20–22 (table attributed to Nicaea).
19
Disputatio Morini ll. 61–67 (ed. by Graff (2010), 142): Obserua igitur cursum
lunarem iuxta regulam Graecorum more Aegypiorum […] iuxta computacionem Eusebii,
qui primus conscripsit circum X et VIIII annorum, Athanasii, Theophili, Cyrilli Dionysii­
que exigui, usque dum scripsit Victorius Hilaro papae urbis Romae episcipo; cf.  Graff
(2010), 132 for a translation, and 133: ‘the message is clear; adopt the lunar calendar of
Dionysius and discard the lunar data as given in Victorius’.
20
Cummian, De controversia paschali ll. 123–29, 183–86 (Walsh and Ó Cróinín
(1988), 74–75, 80–81); the citations are from Augustine’s De haeresibus and Enarratio
in Psalmo; the source of Cummian’s citation attributed to Augustine at lines 124–29 is
unidentified.


 THE PASCHAL CYCLE OF ST PATRICK

‘Holy Augustine of venerable memory explains this beautifully, saying:

‘The feast day is the Pasch, as the evangelist Luke says: Now the feast of
unleavened bread approached when it was necessary to kill the Pasch.
The Pasch is killed and is brought to life. It is a part of the feast day, the
fourteenth moon, not the whole feast, in which we are led astray, but
in the part, and in the restful Sabbath and in the heralded new moon,
which are but a shadow, not the substance of Christ. The shadow is
killed, the truth is brought to life.’’

Occurrit michi item Agustinus sanctus, ubi contra totas hereses sub anath­
ematis titulo disputat dicens, ‘Teserescedecadite’—id est quartanae deci­
mae—esse hereses eorum qui pas<cha ‘.x>iiii. luna cum Iudaeis faciunt.’

‘Again holy Augustine comes to me, where he disputes against all here-
sies under the heading of anathema, saying that ‘Teserescedecadite’ (that
is ‘of the fourteenth’) is the heresy of those who ‘hold’ Easter ‘on the
fourteenth moon with the Jews’.’

Furthermore, Augustine undoubtedly was a proponent of the Alexan-

drian paschal tradition, for in his Epistola ad Ianuarium he clearly ap-
proved the Alexandrian lunar limits, writing ‘that which is called Pasch
would be observed from the fourteenth moon […] up to the twenty first
moon to the precise number of seven’, and this letter was known in Ire-
land, as Jones pointed out.21 There can be no doubt, therefore, that Au-
gustine did indeed endorse the Alexandrian lunar limits of luna 15–21,
suggesting thereby that he favoured their 19-year lunar cycle. In sum-
mary, seven of Cummian’s authorities compiled paschal cycles employ-
ing 19-year lunar cycles, and two commended them.
I submit that the consistency of Cummian’s grouping of these nine
authors, together with the warmth and the emphasis of his own com-
mendation of the 19-year cycle that he attributed to the 318 bishops of
Nicaea, require that Patrick’s cycle also employed a 19-year lunar cycle.
Moreover, the antithetical way in which Cummian cited Anatolius’ De
ratione paschali in order to dismiss comprehensively all 84-year cycles,
including, of course, the latercus that Ségéne and Béccán wished to
maintain, makes it absolutely certain that Patrick’s cycle could not have
21
Augustine, Epistola ad Ianuarium (ed. by Alois Goldbacher in CSEL 34, 179–
80): quod pascha nominatur a quarta decima luna uoluit obseruari […] usque ad uicesi­
mam uero et primam propter ipsum numerum septenarium; much of the letter, including
this citation, was transcribed by Eugipius, Excerpta ex operibus S. Augustini 118 (ed. by
Pius Knöll in CSEL 9,1, 432); Jones (1943), 97 n. 1 (Epistola ad Ianuariam known in
Ireland).


 DANIEL MC CARTHY

been an 84-year cycle. All of these considerations imply that Patrick em-
ployed a 19-year lunar cycle, as Jones also concluded.22
Finally, regarding the origin of Patrick’s cycle, Cummian referred to
this as illum quem sanctus Patricius papa noster tulit et fecit, which pas-
sage Walsh and Ó Cróinín translated as ‘that which holy Patrick, our
bishop, brought and followed’.23 However, while fecit can certainly ac-
commodate their translation of ‘followed’, this interpretation makes
Patrick simply an agent of transport, leaving the authority of the cycle
‘followed’ by him as unidentified, and this is in complete contrast to
Cummian’s account of the other nine cycles, for each of which he has
named the primary authority. Indeed, for six of these the authority’s
name is the only information given by Cummian, namely, Theophilus,
Dionysius, Cyril, Morinus, Augustine, and Victorius. I conclude, there-
fore, that fecit should be understood rather in its primary sense of ‘made’,
so that Cummian here assigned authority for this cycle to Patrick, in
the same way that he does for each of the other cycles. This implies then
that Patrick had adapted his cycle from an earlier, yet to be identified
cycle, which he had brought with him to Ireland, and we shall see fur-
ther evidence in support of this interpretation. In summary, Cummian’s
account makes Patrick the author of an adaptation of an earlier paschal
cycle which employed a 19-year lunar cycle, and Patrick scheduled Pasch
on Sunday falling on luna 15–21 inclusive, and also required that luna
14–21 inclusive should fall after the equinox on 21 March.

The paschal cycle adopted by Cummian and his community

I next consider the question of which paschal cycle Cummian and his
community had come recently to favour, beginning with the observa-
tion that Cummian’s comprehensive dismissal of all 84-year cycles and
his enumeration of ten authorities for the 19-year cycle clearly requires
that their change had been to a 19-year cycle. Regarding which of these
22
Jones (1943), 86: ‘His [Patrick’s] table was based on the decennovenal cycle, as
Cummian makes clear’.
23
Cummian, De controversia paschali ll. 208–09 (Walsh and Ó Cróinín (1988),
84–85); a similar translation was given by de Paor (1993), 152: ‘that which the holy
Patrick, our papa, brought to us and practised’. On the other hand, Mac Carthy (1901),
cxxxvii: ‘1) that brought and composed by Patrick, their pope’; and similarly Meyer at
Zimmer (1902), 81: ‘that first cycle which our holy father Patrick brought and com-
posed’; Zimmer (1901), 229, had cited only the Latin: Primum illum […] XII kal. Apr.
observatur.


 THE PASCHAL CYCLE OF ST PATRICK

ten authorities had been adopted by them, we have seen that in Cum-
mian’s introduction he wrote of ‘the first year in which the cycle of 532
began to be celebrated by our party’. Now of these ten cycles only one,
that compiled by Victorius in AD 457, is known to have been of length
532 years. However, if the Victorian was the cycle adopted by the com-
munity it is remarkably incongruous that Cummian placed it eighth in
his list of cycles, and he made no reference whatsoever to its lunar limits,
for this was a matter of utmost concern to him, as we shall see. If, on the
other hand, we consider which of the ten cycles Cummian presented
most prominently, then there can be no question but that it was Patrick’s
cycle. For this was the cycle that he placed first, and with which he as-
sociated himself and his community by the words Patricius papa noster,
and the only authority to whom he applied the title sanctus. It was also
the only cycle for which Cummian provided lunar limits, a matter of
prime concern to him, so could then Patrick’s paschal cycle have been
of length 532 years? The length of his cycle is not known, but we have
seen above that Patrick adapted his cycle from one that he had brought
to Ireland, so it is appropriate therefore to consider what is known of the
circumstances of Patrick’s mission to Ireland.
Both of the earliest Lives of Patrick, those by Tírechán and Muirchú,
recount that, following his escape from servitude in Ireland aged about
22, Patrick spent an extended period in Gaul, after which he returned to
Ireland on a mission that occupied the remainder of his life. While Tíre-
chán gives no chronology for this mission, he dates his obit to Annus
Passionis 433, AD 460; on the other hand, Muirchú locates the start of
his mission immediately after the death of Palladius following the failure
of his brief mission in AD 431, and he dates his obit to Annus Passionis
436, AD 463.24 Consequently, for well over a thousand years it has been
believed that Patrick’s mission to Ireland commenced in AD 432. How-
ever, in 1942 and 1946, Thomas O’Rahilly proposed c.AD 462 as the
year of the commencement of Patrick’s mission to Ireland. Subsequently,
James Carney suggested the year c.AD 456, and in 1993 Liam de Paor
effectively reiterated O’Rahilly’s proposal as AD  460–61.25 Most re-
24
Tírechán, Collectanea (ed. by Bieler (1979), 126–27): Patricks’s obit at AP 433;
Muirchú, Vita Patricii (ed. by Bieler (1979), 62–63, 70–75): Patrick’s obit at AP 436,
and chronology of his mission. Roman ccccxxxiii and ccccxxxui readily interchange as a
consequence of scribal misreading of u as ii, and vice-versa. This ambiguity of three years
also appears in Tírechán’s ambivalent continuation of Loíguire’s reign after Patrick’s
death: Duobus uel quinque annis regnavit Loiguire post mortem Patricii.
25
O’Rahilly (1942), 36: ‘When did the later Patrick come to Ireland as mission-
ary […] we may accept 461 or 462 as the probable date of Patrick’s arrival’; O’Rahilly


 DANIEL MC CARTHY

cently, having reviewed the annalistic and other chronological accounts

of Patrick, I concluded that his mission commenced in c.AD 458, and
he died in AD  491.26 Thus these re-evaluations of the chronology of
Patrick’s mission all place it in close temporal proximity to Victorius’
compilation of his paschal table in AD 457, and this is in full accordance
with the implications of Cummian’s account. Furthermore, in Patrick’s
own account in the Confessio of embarkation on his mission he attrib-
utes the inspiration for this to a vision in which ‘I saw a man coming,
as it were from Ireland. His name was Victoricus’, and this name ‘Vic-
toricus’ means literally ‘like Victor’.27 Taking these details together with
Cummian’s account, I conclude that the cycle that Patrick brought was
none other than the 532-year cycle of Victorius, and it was this that he
adapted, with the consequence that his paschal cycle was also of length
532 years. This then was the cycle that Cummian and his community
had adopted and, thus, they also observed the equinox on 21  March
and celebrated Pasch within the lunar limits luna 15–21. Hence, once
it is understood that Cummian and his community had adopted Pat-
rick’s cycle, then Cummian’s letter provides independent evidence that
Patrick’s mission to Ireland must postdate Victorius’ compilation in
AD 457, and, as a corollary, that Tírechán and Muirchú misrepresent
Patrick’s chronology.
I turn next to consider how the foregoing deductions compare with
Cummian’s own statements regarding the paschal lunar limits. In total
Cummian makes 25 references to lunar ages relating to the celebration
of Pasch, and these together with his associated discussion provide the

(1946), 399: ‘Patricius  […] whose missionary work in Ireland did not begin until af-
ter […] 461’; Carney (1958), 24–37 (Patrick’s mission in 457); Carney (1961), 26: ‘April
5 […] 456’; de Paor (1993), 90: ‘But it was as a bishop that he [Patrick] came to Ireland.
This, by the reckoning above, would have been about 460–61’.
26
Mc Carthy (2008), 146: ‘c. 416 natus, 432 captivity, c. 438 release, c. 458 mission,
491 obit aged about seventy-five’. For his obit in AD 491, see now Mc Carthy (2017),
181–85, 188, 190, 192. I am aware, of course, that another school of thought has placed the
commencement of Patrick’s floruit in the fourth century, but this required that the annalis-
tic obits for him be completely set aside. For example, Esposito (1956–57), 144: ‘There is
to-day a tendency to regard the Patrick problem as principally a chronological puzzle to be
solved by more or less ingenious juggling with the data furnished by the so-called ‘Ulster
Chronicle’.’ Instead, Esposito (1956–57), 147, merely offered the speculation: ‘His floruit
may be conjectured as extending from c. 350 to c. 430’. The presence in the fifth- and sixth-
century annals of phenomenological entries which can be independently verified to be ac-
curately dated shows that their chronology cannot be summarily dismissed; cf. Mc Carthy
(2008), 159–62; Ludlow (2010), i 281, ii 2–3; and Ludlow (2013), 2–3.
27
Patrick, Confessio (ed. by Bieler (1952), 71; trans. by de Paor (1993), 100): uidi
in uisu noctis uirum uenientem quasi de Hiberione, cui nomen Victoricus.


 THE PASCHAL CYCLE OF ST PATRICK

substance of his discussion of Scripture.28 He commenced this discus-

sion by citing all of the lunar references from Exodus 12:2–18, conclud-
ing these as follows:29

Item in Exodo: .vii. diebus comedetis azima usque ad diem .xxi.; et hoc in
tractatibus diligenter inuestigaui quid sentirent de .xxi. eruditissimi uiri,
quod Ieronimus pulcherrime explanat dicendo: ‘Pascha immolat populus
et alias celebrat festiuitates, omnis eius sollennitas die finitur octauo.’ Prima
dies in azimis .xu., dies octaua .xxi.

‘Also in Exodus: For seven days you shall eat unleavened bread until the
twenty-first day; I  diligently investigated in treatises what the most
learned men might think about the twenty-first, which Jerome explains
most beautifully when he says: ‘The people sacrifice the Pasch, and cel-
ebrate other festivities. Their whole solemnity is finished on the eighth
day’; the first day of unleavened bread is the fifteenth [moon], the eighth
day [of the solemnity] is the twenty-first [moon].’

Note here that, having cited Exodus 12:15,18 for a seven-day celebra-
tion continuing until luna 21, Cummian immediately emphasized his
own concern to ‘diligently’ investigate luna 21 in treatises written by
‘the most learned men’, and then approvingly cited Jerome’s In Aggaeum
to identify that the solemnity is ‘finished on the eighth day’. Cummian
then cryptically stated the consequence of these citations, namely that
the lunar limits of the days of unleavened bread are luna 15–21. Fol-
lowing this Cummian turned to the Liber Quaestionum, from which he
took two citations, of which the first is:30

Item in Libro Questionum: ‘A .xiiii. luna, quae nobis secundum legem

prima est, rationem paschae obseruemus. Omnia enim plena Deus insti­
tuit: ideo a .xiiii. luna usque in .xxi. his .vii. diebus pascha nobis cele<brare
concessum e>st’.
28
For Cummian’s lunar references, see De controversia paschali ll. 20, 22, 26, 32 (luna
10, 14, 14, and 21 from Exodus 12:2–17), 33, 36 (luna 21 and 15–21 by Cummian), 36,
38, 40–41 (luna 14, 14–21, and 13–15 from Liber Quaestionum), 65, 71, 72, 76, 78, 82, 83
(luna 14, 15, 16, 16, 16, 14–20, 15–21, and 16–22 from the Gospels), 126 (luna 14 from
Augustine), 181 (luna 14 from Jerome), 185 (luna 14 from Augustine), 209 (luna 14–21
from Patrick’s cycle), 218 (luna 14 from the Nicaean cycle), 222 (luna 14 ‘contrary’ to that
of the latercus), 225, 228 (luna 14, twice from Cyril); Walsh and Ó Cróinín (1988), 58–88.
29
Cummian, De controversia paschali ll.  32–36 (Walsh and Ó Cróinín (1988),
58–61).
30
Cummian, De controversia paschali ll.  36–39 (Walsh and Ó Cróinín (1988),
60–61).


 DANIEL MC CARTHY

‘Again in the Book of Questions: ‘We may observe the computation of

Easter from the fourteenth moon, which is the first for us according
to the Law. For God made all things whole, therefore it was granted to
us to celebrate the Pasch during these seven days, from the fourteenth
moon until the twenty-first’.’

While this citation ambivalently admits celebration on luna 14–21,

its restriction of the celebration to seven days, when taken with Cum-
mian’s previous acceptance of luna 21, restricts his paschal limits to luna
15–21.
Cummian next reviewed the lunar dates of the Passion, Sepulture,
and Resurrection, assigning them to luna 14, 15, and 16 respectively, and
then concluded with the statement:31

Unde orie<ntalis tota aecclesia> tres ebdomadas tribus sacratissimis so­

lennitatibus Domini <nostri Iesu> Christi uenerabiliter, id est passioni,
sepulturae, resurrectioni, deputau<it>: passioni a .xiiii. in .xx.; sepultu­
rae a .xv. in .xxi.; resurrectioni a <.x>vi in .xxii. lunam, septimanam
pro reuerentia Dominici diei consecrans. Quia si .xiiii. luna resurrectioni
deputetur, ut uos facitis, .xiii. in sepultura et .xii. in passione, prepostero
ordine, fiet.

‘Hence the entire church of the East has reverently allotted three weeks
for the three most holy feasts of Our Lord Jesus Christ, that is [a week] to
the Passion, [a week] to the Sepulture, and [a week] to the Resurrection:
from the 14th to the 20th moon for the Passion, from the 15th to the 21st for
the Sepulture and from the 16th to the 22nd for the Resurrection, conse-
crating a week for reverence of the Lord’s Day. For if the fourteenth moon
were assigned to the Resurrection, as you do, then the thirteenth would
fall on the Sepulture and the twelfth on the Passion, in an inverted order.’

This passage may be interpreted to imply that Cummian repudiated cel-

ebration of Pasch on both luna 14 and 15 and hence that he favoured
the earlier Roman limits of luna 16–22, and indeed this was the inter-
pretation made by Bartholomew Mac Carthy, Joseph Schmid, Thomas
Charles-Edwards, Immo Warntjes, and Caitlin Corning.32 There are,
however, a number of serious obstacles to this interpretation:

31
Cummian, De controversia paschali ll.  80–85 (Walsh and Ó Cróinín (1988),
66–69).
32
See n. 45.


 THE PASCHAL CYCLE OF ST PATRICK

1. Cummian himself only uses the Eastern church’s allocation of

weeks in order to dismiss Ségéne and Béccán’s celebration of Pas-
ch on luna 14.
2. To approve luna 16–22 would require that Cummian also ex-
cluded luna 15, but nowhere in his letter does he condemn this;
rather in several places he explicitly includes it in limits that he
extols.
3. Celebration on luna 16–22 was a tradition of the earlier Roman
church, and had Cummian been commending that he could not
have repeatedly represented that this provided the basis for a uni-
versal Pasch and paschal unity. In particular, he could not have
claimed, as he does, that he was observing the first canon of Arles
concerning Pasch, that ‘we should keep it on one day at one time
throughout the entire world’, nor that the delegates to Rome had
stated: ‘this Easter is celebrated throughout the whole world’. In
the seventh century, such assertions could only plausibly refer-
ence the Alexandrian paschal tradition.33
4. All of the Patristic authorities that Cummian names explicitly and
whose work he cites in relation to the age of the paschal moon,
namely Cyril, Jerome, and Augustine, can be shown to have en-
dorsed the Alexandrian paschal tradition. These authorities, from
churches distributed around the Mediterranean basin, can indeed
be considered to represent ‘the entire world’, and hence a ‘paschal
unity’.

Thus the only conclusion that may be drawn from this passage that is
consistent with Cummian’s other declarations is that he intended it to
reject Ségéne and Béccán’s celebration of Pasch on luna 14, as he explic-
itly states in the concluding sentence of the citation.
Cummian concluded his discussion of lunar limits with citations
from Jerome, Augustine, and Cyril as follows:34

Item Ieronimus: ‘Moyses,’ inquit, ‘moriens plangitur; Iesus’ absque lacrimis

‘in monte’ sepelitur. In lege morientes sub .xiiii. luna Dominum crucifigen­
tes plangam, sed cum Christo resurgentes in euangelio suscipiam.

33
Cummian, De controversia paschali ll. 99–100, 284–85 (Walsh and Ó Cróinín
(1988), 72–73, 94–95): ‘ut uno die et uno tempore per totum orbem terrarum a nobis
conseruetur’;‘Per totum orbem terrarum hoc pascha, ut scimus, celebratur’.
34
Cummian, De controversia paschali ll. 180–83 ( Jerome), 183–86 (Augustine),
223–29 (Cyril); Walsh and Ó Cróinín (1988), 80–81, 86–89.


 DANIEL MC CARTHY

‘Again Jerome: ‘Moses dying,’ he says, ‘is mourned; Jesus’ is buried ‘on
the Mount’ without tears. I would mourn for those who die in the Law,
crucifying the Lord under the fourteenth moon, but I would receive
those rising again with Christ in the Gospel.’

Occurrit michi item Agustinus sanctus, ubi contra totas hereses sub
anathematis titulo disputat dicens, ‘Teserescedecadite’—id est quartanae
decimae—esse hereses eorum qui pa<scha ‘.x>iiii. luna cum Iudaeis faci­
unt’.

‘Again holy Augustine comes to me, where he disputes against all her-
esies under the heading of anathema, saying that ‘Teserescedecadite’
(that is ‘of the fourteenth’) is the heresy of those who ‘hold Easter on the
fourteenth moon with the Jews’.’

‘Scrutamini<que’ ut> Cyrillus ait, ‘quod ordinauit sinodus Nicena lu<nas

quartas decimas omnium annorum per d>ece<n>nouenalem cyclum’—
quem <Uictorius, per uicesimas> et oc<tau>as uic<es> cum kalendis
.dxxxii. et bissext<is .cxxxiii., i>n id i<psum>, unde ortus est, redire
fecit—‘ut non fallam<ur in> luna primi mensis et celebre<mus> pascha
in sequenti Dominico, et non faciemus in luna .xiiii. cu<m> Iudaeis et
hereticis, qui dicuntur thesserescedecadite.’

‘‘Note,’ as Cyril says, ‘that the Council of Nicaea ordained the four-
teenth moons of all years through the nineteen-year cycle’ (which Vic-
torius made return to where it started through twenty-eight turns with
532 kalends and 133 bissextiles) ‘so that we may not be deceived in the
moon of the First Month and so that we should celebrate Easter on the
following Sunday, and not keep it on the fourteenth moon with the Jews
and heretics who are called Thesserescedecadite.’’

Hence all three citations emphatically reject celebration on luna 14,

with Augustine and Cyril characterising it as heretical, and Cyril un-
equivocally implying that celebration be restricted to luna 15–21. The
only conclusion that may be drawn that is congruent with all these cita-
tions is that Cummian repudiated celebration of Pasch on luna 14 and
favoured celebration according to the Alexandrian tradition on luna
15–21. Consequently this conclusion is indeed in accordance with the
lunar limits that Cummian attributed to Patrick’s cycle. Regarding the
lunar cycle followed by Cummian, as well as his repudiation of the 84-
year cycle and his enthusiastic endorsement of the 19-year cycle of the
318 bishops of Nicaea, he also approvingly cited Cyril in support of this,


 THE PASCHAL CYCLE OF ST PATRICK

as may be seen in the citation just given. Thus, it is absolutely certain that
Cummian wished to advocate a 19-year lunar cycle.
Finally, regarding the length of Cummian’s paschal cycle, we have
already seen in the opening lines of his letter the statement that his com-
munity had begun to celebrate using a cycle of 532  years. Moreover,
Cummian parenthetically interjected into his citation of Cyril’s exal-
tation of the 19-year cycle the words ‘which Victorius made return to
where it started through twenty-eight turns with 532 kalends and 133
bissextiles’.35 With this interjection Cummian thus accurately identified
Victorius’ significant contribution to the application of the 19-year lu-
nar cycle to paschal computation, namely that 28 repetitions resulted
in a return to the starting ferial and lunar incidence. The implication
is therefore unmistakeable that the computistical structure of the cycle
followed by Cummian and his community, and hence that of Patrick,
derived from Victorius’ 532-year cycle.
In summary, Cummian’s paschal cycle was of length 532-years, and
he celebrated on luna 15–21 inclusive after equinox on 21 March using
a 19-year lunar cycle, and each one of these parameters is congruent with
his having adopted the cycle that Patrick had adapted from Victorius’
532-year cycle. This conclusion then requires that we examine the struc-
ture of Victorius’ cycle to see whether such an adaptation could plausi-
bly have been made by Patrick.

Victorius’ paschal cycle

Following the exchange between Pope Leo and Bishop Paschasinus of
Lilybaeum regarding the Pasch of AD 455, Hilarius, archdeacon to Leo,
requested Victorius of Aquitaine to examine the causes of differences
in paschal computation. Victorius responded with a 532-year table in
which he systematically tabulated for each year the ferial and epact of the
Kalends of January, followed by the Julian date of Pasch and the age of
the paschal moon. Because the Alexandrian initial lunar limit of luna 15
conflicted with the final lunar limit of luna 22 of the earlier Roman pas-
chal tradition, Victorius stated in his prologue to the table that he would
supply both dates and leave it to the pope to decide which he considered

35
Cummian, De controversia paschali ll. 225–26 (Walsh and Ó Cróinín (1988),
88–89): quem <Uictorius, per uicesimas> et oc<tau>as uic<es> cum kalendis .dxxxii. et
bissext<is .cxxxiii.,i>n id i<psum>, unde ortus est, redire fecit.


 DANIEL MC CARTHY

appropriate. While the occurrence of these double-dates in the surviving

manuscripts of the paschal table is variable, as Warntjes has documented,
Victorius’ statement in his prologue leaves no room for doubt as to his
intention to provide them.36
In Figure 1 I display concisely all of the essential computistical fea-
tures of the 532 years of Victorius’ paschal cycle by first tabulating all the
possible paschal dates across the base of the diagram, and then by listing
vertically the nineteen Victorian lunar epacts (luna 1, luna 3, …, luna
30). The heavy vertical bars before 22 March and after 24 April represent
the paschal termini evidently assumed by Victorius, in that all of his tab-
ulated paschal dates fall in the inclusive interval 22 March to 24 April.37
To explain the construction of Figure 1 I begin by considering a year
commencing with epact luna 1 on the Kalends of January, i.e. 1 January, in
which case the first occurrence of luna 15 after the equinox on 21 March
will fall during the April lunation on 14 April. Then luna 16–22 will fol-
low on 15–21 April, and these eight days are registered by the eight boxes
inscribed above these dates. Next, if the year is common (365 days) and
1 January falls on feria 3 (Tuesday), then 14 April will fall on a Sunday,
and so according to Alexandrian lunar limits Pasch will be celebrated on
this day with luna 15; hence this relationship is registered by the feria ‘3’
inscribed in this box. However, if 1 January falls one day earlier on feria 2,
then the Sunday must fall one day later on 15 April, and Pasch will be cel-
ebrated in both traditions on luna 16, and hence the feria ‘2’ is inscribed
in this box. In a similar fashion the five subsequent boxes receive the ferial
numbers ‘1’, ‘7’, ‘6’, ‘5’, and ‘4’ respectively. Lastly the eighth box, associated
with luna 22, receives the feria ‘3’ in order to accommodate the lunar
limit of the earlier Roman paschal tradition. Thus the two boxes with an
inscribed ‘3’ represent Victorius’ double-dates.
Regarding the bissextile years of 366 days, because of the additional
day in February, the feria of 1 January must be first incremented by one
in order to index the correct date and moon of the Pasch in the diagram.

36
Victorius’ prologue and paschal table are edited by Theodor Mommsen in MGH
Auct. ant. 9, 677–735, and Krusch (1938), 16–52. Regarding the luna 15/22 conflict,
Victorius, Prologus 11 (Krusch (1938), 26), wrote: Illud praeterea insinuari non desititi
propter diversorum paschalium conditores, ubi in hoc eodem cyclo dies paschae gemina de­
signatione positus invenitur, id est, ubi luna XV. die dominica et post septem dies vicensima
secunda conscribitur, non meo iudicio aliquid definitum, sed pro ecclesiarum pace apostoloci
pontificis electioni servatum. Warntjes (2010a), LXXXIV n. 228 surveys double-dates in
the surviving manuscripts; for a technical account see Holford-Strevens (2008), 192–96.
37
24 April is also the latest paschal date discussed in Victorius, Prologus 12 (Kr-
usch (1938), 26; Victorius’ discussion of epact luna 27, feria 7).


 THE PASCHAL CYCLE OF ST PATRICK

L L
15 22
Epact ↓ ↓ Epact
30 2 1 7 6 5 4 3 2 30

28 L L 7 6 5 4 3 2 1 7 28
27 16 22 6 5 4 3 2 1 7 6 27
↓ ↓ ↑ ↑
25 5 4 3 2 1 7 6 L L L 25
22 15 22
23 4 3 2 1 7 6 5 4 ↓ 23
22 3 2 1 7 6 5 4 3 22
↑
20 L 1 7 6 5 4 3 2 1 20
19 15 7 6 5 4 3 2 1 7 19

17 5 4 3 2 1 7 6 5 L 17
16 4 3 2 1 7 6 5 4 22 16
↓
14 2 1 7 6 5 4 3 2 14
↑
12 L 7 6 5 4 3 2 1 7 12
11 15 6 5 4 3 2 1 7 6 L 11
22
9 4 3 2 1 7 6 5 4 ↓ 9
8 3 2 1 7 6 5 4 3 8
↑
6 L 1 7 6 5 4 3 2 1 6
15
4 6 5 4 3 2 1 7 6 L 4
3 5 4 3 2 1 7 6 5 22 3
↓
1 L 15 → 3 2 1 7 6 5 4 3 1
Month M M M M M M M M M M A A A A A A A A A A A A A A A A A A A A A A A A A Month
Day 22 23 24 25 26 27 28 29 30 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Day
Julian 11 10 9 8 7 6 5 4 3 2 kl 4 3 2 no 8 7 6 5 4 3 2 id 18 17 16 15 14 13 12 11 10 9 8 7 Julian

Figure 1  A tabulation showing for each Victorian epact luna 1, 3, …, 30 and

each feria 1–7 of common Julian years, the date and lunar age of the celebration
of the Pasch in the range luna 15–22 in accordance with the paschal table of
Victorius (after Krusch (1937), 27–52).

For example, the Pasch of a bissextile year with epact luna 1 and feria 3 on
1 January will be indexed by the box with feria 3+1=4, and so it falls on
20 April with luna 21 in both the Alexandrian and earlier Roman tradi-
tions. These eight boxes and their inscribed ferial data therefore compre-
hensively, but concisely, represent all 28 instances of the Kalends epact
luna 1 and ferial data, and the corresponding paschal date and moon
given by Victorius in his table. In this way considerable compression of
the data and a more intuitive representation of the inter-relationships
between them is achieved, and this is especially useful when considering
the situations that arise near the dates of the paschal termini.
Next, when we consider epacts greater than luna 1, we find that as
the age of the moon on the Kalends of January increases by one, in con-
sequence the Victorian eight-day paschal period must advance by one
day. Thus for the next Victorian epact, luna 3, the moon on 1 January
is two days older so that the eight boxes representing luna 15–22 must
advance by two days and so are positioned above 12–19 April. The f­ erial


 DANIEL MC CARTHY

considerations are independent of the epact and so the feria ‘5’, ‘4’, ‘3’, ‘2’,
‘1’, ‘7’, ‘6’, ‘5’ are inscribed respectively in these eight boxes. These prin-
ciples may be applied repeatedly for epacts up to and including luna 23,
for which the Victorian eight-day paschal period is 23–30 March.
However, when he reached epact luna 25, Victorius here tabulated
only the earlier Roman paschal tradition of luna 16–22 falling on 22–
28 March; he could not include luna 15 because this falls on the equi-
nox and so both it and luna 14 fall before his earlier paschal terminus.
Thus with this choice Victorius scheduled Pasch during the April luna-
tion (defined as the lunar month ending in the Julian month of April)
whose luna 14 fell on 20 March, and thus before the equinox. While this
date and luna 14 were acceptable to the earlier Roman paschal tradition,
Victorius’ choice represented a serious violation by him of Alexandrian
paschal principles. In consequence the 28 entries in Victorius’ paschal
table for epact luna 25 made no provision for celebration according to
the Alexandrian tradition.38 This conflict arose because Victorius’ lunar
cycle includes epact luna 25, whereas the Alexandrian lunar cycle avoids
this particular epact.
Finally, considering the remaining epacts luna 27, 28, and 30, we find
that for these Victorius scheduled Pasch on the May lunation whose
luna 14 clearly falls long after 21 March. Here we can see that, while all
his luna 15–21 dates fall within his 24 April terminus, for epact luna
27 and feria 6 the Roman paschal luna 22 date falling on 25  April is
outside this terminus, and so it fails to satisfy his own paschal termini.39
Indeed for epacts luna 27–30 all eight of Victorius’ paschal dates fall-
ing on 22–25  April lie outside the earlier Roman paschal terminus of
21  April.40 These conflicts are an inevitable consequence of Victorius’
use of a 19-year cycle which combines each epact with all seven ferial

38
The solitary paschal date of 21  April (xi kal. Mai, luna xvi) given by Krusch
(1938), 51, at AD 547 (AP 520) for epact luna 25 and feria 3 occurs only in MS N,
which, being a continuation from AD 464 (AP 437), is unlikely to be the work of Victo-
rius; cf. Krusch (1938), 11, 48 (MS N). MS N, otherwise known as Paschale Campanum,
is edited separately by Theodor Mommsen in MGH Auct. ant. 9, 744–50.
39
For the three common years with epact luna 27 and feria 6, AD 45, 387, and 482,
and the bissextile year with epact luna 27 and feria 5, AD 140, all MSS of the Victorian
table give the Alexandrian Pasch on xiiii kal Mai (18 April), luna 15; some MSS give the
Roman luna 22 as a double date, but erroneously locate this on viii kal. Mai (24 April)
rather than vii kal. Mai (25 April); cf. Krusch (1938), 27, 32, 44, 49. The occurrence of
epact luna 27 and feria 6 in AD 577 resulted in a major conflict recorded by Gregory of
Tours, Historia Francorum 5.17; cf. Krusch (1884), 127–28.
40
Jones (1943), 26 (Roman paschal termini: ‘Easter must occur on one of the
31 days from March 22 to April 21’); Holford-Strevens (2005), 45–46.


 THE PASCHAL CYCLE OF ST PATRICK

data, whereas both Hippolytus’ 112-year cycle and the 84-year Suppu­
tatio Romana were able to avoid those epactal and ferial combinations
that yielded paschal dates after 21 April. Reviewing this we see that for
epact luna 25 Victorius failed to provide any paschal dates fulfilling the
Alexandrian paschal principles, while for epacts luna 27–30 eight of
his paschal dates did not fulfil the earlier Roman paschal principles. In
view of these shortcomings the remark of St Columbanus in his letter of
c.AD 600 to Pope Gregory (‘For you must know that Victorius has not
been accepted by our teachers, by the ancient Irish philosophers, by the
mathematicians most skilled in reckoning chronology, but has earned
ridicule or indulgence rather than authority’) seems well justified.41

Patrick’s adaptation of Victorius’ paschal cycle

I turn next to consider whether Victorius’ cycle could be adapted to
conform to Cummian’s account of Patrick’s cycle, and first recall that he
stated explicitly that the moon was ‘regularly observed from the four-
teenth to the twenty-first, and the equinox from March 21st’.42 Thus it
is clear that some emendation of the cycle must be made in the case of
epact luna 25 for which Victorius has the paschal luna 14 falling on
20 March, and the most straightforward way in which this may be made
to agree with Cummian’s account is by moving the celebration of Pasch
for this epact from the April to the May lunation. This emendation will
indeed place the paschal luna 15–21 on 20–26 April, and so regularly
after the equinox on 21 March. Further, Cummian’s account of Patrick’s
lunar limits, and the limits adopted by him and his community, allowed
celebration only on luna 15–21, so that the second emendation required
is the deletion of all of Victorius’ luna 22 dates. Thus in Figure  2, for
epact luna 25 the paschal luna 15–21 is moved to the May lunation, and
all luna 22 dates have been deleted.
This reconstruction accords with all of the details of Patrick’s cycle
given by Cummian, and also with the details that he affirms about the

41
Columbanus, Epistola 1.4 (ed. and trans.  by Walker (1957), 6–7; following
consultation with David Howlett, I  have emended Walker’s translation of Hibernicis
antiquis philosophis as ‘the former scholars of Ireland’ to ‘the ancient philosophers of
Ireland’): Scias namque nostris magistris et Hibernicis antiquis philosophis et sapientissimis
componendi calculi computariis Victorium non fuisse receptum, sed magis risu vel venia dig­
num quam auctoritate.
42
See the citation at n. 13.


 DANIEL MC CARTHY

L L
15 21
Epact ↓ ↓ Epact
30 2 1 7 6 5 4 3 30
28 7 6 5 4 3 2 1 28
27 6 5 4 3 2 1 7 27
L L
25 15 21 4 3 2 1 7 6 5 25
↓ ↓ ↑ ↑
23 4 3 2 1 7 6 5 L L 23
22 3 2 1 7 6 5 4 L 15 21 22
21
20 1 7 6 5 4 3 2 ↓ 20
19 7 6 5 4 3 2 1 19
↑
17 L 5 4 3 2 1 7 6 L 17
16 15 4 3 2 1 7 6 5 21 16
↓
14 2 1 7 6 5 4 3 14
↑
12 L 7 6 5 4 3 2 1 12
11 15 6 5 4 3 2 1 7 L 11
21
9 4 3 2 1 7 6 5 ↓ 9
8 3 2 1 7 6 5 4 8
↑
6 L 1 7 6 5 4 3 2 6
15
4 6 5 4 3 2 1 7 L 4
3 5 4 3 2 1 7 6 21 3
↓
1 L 15 → 3 2 1 7 6 5 4 1
Month M M M M M M M M M M A A A A A A A A A A A A A A A A A A A A A A A A A A
Day 22 23 24 25 26 27 28 29 30 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Julian 11 10 9 8 7 6 5 4 3 2 kl 4 3 2 no 8 7 6 5 4 3 2 id 18 17 16 15 14 13 12 11 10 9 8 7 6

Figure 2  A tabulation showing a reconstruction of Patrick’s adaptation of Vic-

torius’ paschal cycle in accordance with Cummian’s account, showing all the
luna 15–21 paschal dates and luna 14 after March 21, and their relationship to
the feria of Julian common years.

cycle adopted by himself and his community. In this way it can be seen
that it was relatively straightforward to adapt Victorius’ cycle to con-
form with the Alexandrian principles of celebration of luna 15–21 after
21 March, but with the additional constraint that the paschal luna 14
should always fall after 21 March, rather than be allowed to fall on that
date. I propose therefore that this represents the cycle that Patrick adapt-
ed from Victorius’ cycle, that is tulit et fecit, and that this was the paschal
cycle adopted by Cummian’s community by c.AD 630. The paschal ter­
mini of this cycle are 23 March to 26 April, and thus are one day in ar-
rears of the Alexandrian paschal termini. These one-day retardations are
an inescapable consequence of Patrick requiring that the paschal luna
14 should fall after the equinox, whereas the Alexandrian principles al-
lowed paschal luna 14 to fall on the equinox.43 Regarding Cummian’s at-

Dionysius, Cyclus (Krusch (1938), 69–74; paschal luna 14 of each sixteenth year
43

of the Dionysiac 19-year cycle falls on xii kal. Apr. (21 March)).


 THE PASCHAL CYCLE OF ST PATRICK

titude to paschal termini we note his complete silence on the subject; the
only Julian date he cites is that of the equinox in his account of Patrick’s
cycle, and it seems quite likely that this silence is a consequence of the
incongruence of Patrick’s paschal termini to the Alexandrian termini.
It is appropriate at this point to review the conclusions of earlier
scholarship regarding the paschal cycles of Cummian and Patrick. Here,
on the one hand Cummian’s references to a 532-year cycle inevitably
prompted identifications of Victorius’ cycle, while on the other his re-
peated rejection of Pasch on luna 14 and endorsement of luna 21 en-
couraged identification of the Alexandrian tradition. Thus Bruno Kr-
usch, James Kenney, Bonnie Blackburn and Leofranc Holford-Strevens,
and Masako Ohashi simply identified Cummian’s cycle as ‘Victorian’,
without specifying any lunar limits.44 As mentioned already, Bartho-
lomew Mac Carthy, Joseph Schmid, Thomas Charles-Edwards, Immo
Warntjes, and Caitlin Corning considered that it was specifically the
luna 16–22 limits of Victorius’ cycle that were adopted.45 On the other
hand John Ernest Oulton, Charles Jones, and Kathleen Hughes all con-
sidered that Cummian and his community had adopted the Dionysiac

44
Krusch (1884), 152: ‘Es ist ganz klar, dass Cummian in dem Paschale des Vic-
turius den vom Concil von Nicaea eingeführten Cyclus zu besitzen glaubte, dessen An-
nahme die Curie in allen Schreiben empfahl’; Kenney (1966), 221: ‘The greatest part of
the epistle is taken up with arguments in favor of the Victorian paschal system’; Black-
burn and Holford-Strevens (1999), 795: ‘The southerners thereupon adopted Victorius’
tables’; Ohashi (2010), 196–97: ‘As the letter of Cummian  […] reveals, the table [of
Victorius] was also accepted by the southern Irish in the early 630s’.
45
Mac Carthy (1901), cxlv: ‘As to the Paschal system advocated so strenuously by
Cummian, no ambiguity can arise. A cycle of 532 (arranged by Victorius in 133 bissex-
tiles), with Easter from the 16th to the 22nd of the moon, exclusively designates the Victo-
rian’; Schmid (1904), 26: ‘wenn er aber sagt, dass die orientalische Kirche die Erinnerung
an das Leiden des Herrn von luna 14–20, an die Grabesruhe von luna 15–21 und an die
Auferstehung von luna 16–22 ansetze, so ist klar, dass er das Paschale des Vikturius vor
sich hatte und nicht jenes des Dionysius; denn nach ersterem wurde Ostern von luna
16–22, nach letzterern dagegen von luna 15–21 gefeiert’; Charles-Edwards (2000), 402:
‘Cummian […] attacked the Celtic paschal reckoning and argued in favour of that of
Victorius of Aquitaine [citing the translation of lines 66–68]. Cummian is here draw-
ing on an earlier source to defend the Victorian lunar limits (16–22) as against those
adhered to by the abbot of Iona (14–20)’; Warntjes (2010a), LXXXVIII n. 238: ‘Cum-
mian’s letter gives unambiguous evidence that it was the Victorian reckoning which was
adopted by the southern Irish clergy in AD 632, since Cummian speaks of a cycle of
532 years and lunar limits for Easter Sunday of 16 to 22, which are Victorian character-
istics’; Corning (2006), 87: ‘What can be said with certainty is that Cummian advocated
a nineteen-year cycle with a lunar range of 16–20 [sic], in other words, the Victorian
cycle’; Warntjes (2015), 49: ‘Cummian is very specific about its 532-year structure and
lunar limits 16 to 22 for Easter Sunday’.


 DANIEL MC CARTHY

cycle.46 Earlier James Ussher had described the cycle adopted by Cum-
mian as ‘the great cycle of five hundred and thirty-two years, and the
Roman use of celebrating the time of Easter’; Ussher identified the ‘Ro-
man use’ as celebration ‘betwixt the fifteenth and twenty-first day of the
moon’, suggesting that he too had the Dionysiac cycle in mind.47 Finally,
Maura Walsh and Dáibhí Ó Cróinín, and Faith Wallis proposed a per-
plexing and ambivalent assortment of either unspecified adaptations of
Victorius’ cycle, or hybrid Victorian-Dionysiac compilations.48
Fewer scholars offered an opinion on the question of Patrick’s cycle,
and those that did either rejected it outright, or cast doubt on Cum-
mian’s account of it. Johannes van der Hagen, following Ussher, rejected
Cummian’s account as mistaken. Heinrich Zimmer considered that
Cummian anachronistically attributed the introduction of the Dionysi-
ac cycle to Patrick and dismissed it as the first appearance of the ‘Patrick-
legende’. Charles Plummer effectively rejected Cummian’s account by
characterising Patrick’s mission as ‘supposed’, while Mac Carthy scorned

46
Oulton (1957), 130–31: ‘In the third decade of the seventh century certain in-
dividuals (or communities) in Celtic circles had begun to conform to Roman usage in
regard to the date of Easter  […] The limits between which Easter must fall were  […]
for the Romans, as to-day, March 22 – April 25’; Jones (1943), 90: ‘The direct result of
the synod [of Mag Léne] was […] the adoption of the Alexandrian reckoning’; Hughes
(1966), 107: ‘Cummian did not succeed in persuading Ségéne of Iona to join the Roman
observance. The churches of southern Ireland seem to have come over in the 630s’—
Hughes identified ‘Roman observance’ with the cycle endorsed by Bede, the Dionysiac.
47
Ussher (1631), 336, 339 (citations).
48
Walsh and Ó Cróinín (1988), 6, 18–19, 22, 28, 43, 46: ‘628/629 […] Some of
the southern Irish immediately accept the Roman reckoning’; ‘this [Victorius’ cycle] (or
an adaptation of it) was indeed the cycle intended by Cummian’; ‘(by Cummian’s tes-
timony) the southern Irish churches […] decided to adopt the same [Victorian] cycle
or an adaptation of it’; ‘The southern Irish churches had begun to observe Easter in ac-
cordance with the cycle of Victorius (or a version thereof ) […] The Irish in the south
had adopted (or adapted) the Victorian cycle  […] Certainly it [Victorius’ cycle] was
the cycle which the southern Irish delegation to Rome seem to have regarded as the
one being observed by the universal church’; ‘Cummian does not say why the Victorian
tables were adopted in southern Ireland; he simply defends their adoption by reference
to the delegation to Rome which reported that the (Victorian?) Easter date for that year
was the one observed by the universal Church’; ‘An alternative explanation might be to
suggest that the southern Irish combined the data of Victorian and Dionysiac tables’;
Wallis (1999), lx, lix, lxiii: ‘the southern Irish in 633 adopted ‘Alexandrian reckoning’
generically, and seem to have used Victorius’ or Dionysius’ tables interchangeably, or
perhaps even a fusion of the two. In fact, the Letter of Cummian seems to indicate that
the Victorian system, or a hybrid Victorian-Dionysian system, was the preferred one […]
When the envoys returned to report that the Alexandrian reckoning was observed eve-
rywhere but in the Western island, the southern Irish converted to the new system […]
The southern Irish followed the Roman computus by 631’.


 THE PASCHAL CYCLE OF ST PATRICK

Cummian’s account as ‘false’. Richard Hanson and Edward Thompson

likewise disparaged it simply as an invention by Cummian, while John
Bury’s tentative suggestion that Patrick had brought the 84-year Suppu­
tatio Romana completely ignored Cummian’s outright rejection of all
such cycles. Jones, however, considered that Patrick had introduced to
Ireland the Alexandrian tradition as followed in Milan, but that use of
his cycle did not survive to the seventh century, and similarly Walsh and
Ó Cróinín inferred that Patrick had introduced Alexandrian reckoning
but denied that Cummian was prepared to follow it. Earlier Francis John
Byrne, while expressing respect for Cummian’s erudition, considered
that his account of Patrick’s cycle could not be accepted.49 Thus there
has been no consensus whatsoever on the tradition of the paschal cycle

49
van der Hagen (1733), 342: ‘Usserius hanc Cummiani de S. Patricii Cyclo opin-
ionem refutat, […]. Ex quo Userii ratiocinio patet Cummianum hac in re erasse’; Zim-
mer (1901), 229: ‘Er [Cummian] sagt in der Aufzählung der verschiedenen Ostercyklen
Primum illum […] observatur, schrieb also klar die Einführung des in Rom selbst erst im
6. ­Jahrh. eingeführten Dionys-Cyklus in Irland dem Patrick zu […] Damit ist die Patrick-
legende bei ihrem ersten Auftreten charakterisiert’; cf. Zimmer (1901), 208, and Zimmer
(1902), 81; Plummer (1896), ii 26: ‘[Patrick is mentioned] earlier still in Cummian’s letter
on the Easter question. But even this is still two hundred years later than his supposed mis-
sion’; Mac Carthy (1901), cxl: ‘the description of the Patrician cycle is demonstrably false’;
Hanson (1968), 104 n. 4: ‘Cummian in his letter to Segene would, of course, ascribe to
Patrick as traditional author any Paschal cycle of which he approved and whose immediate
origin he could not trace’; Thompson (1985), 159: ‘Cummian is not known to have been
aware of anything about Patrick except that he was a venerated but almost forgotten figure,
that his name would give authority to a new Easter cycle’; Bury (1905), 283 (‘the 84 Paschal
cycle’), 373 (‘Supputatio’); Jones (1943), 86, 95 n. 1: ‘I have showed that northern Italy, at
least, was using Alexandrian tables with early Easter-limits March 22 and lunar limits […]
lunae xiv–xxi. Patrick therefore introduced into Ireland in the first part of the fifth century
the Paschal usage of Milan […] Patrick would claim for his Easter-usage the authority of Al-
exandria […] His table was based on the decennovenal cycle, as Cummian makes clear […]
it is clear that no one used it [Patrick’s cycle] by the beginning of the seventh century’;
Walsh and Ó Cróinín (1988), 31–32:‘Patrick seems, on Cummian’s evidence, to have in-
troduced the Alexandrian reckoning into Ireland […] Why Cummian was not prepared
to follow the Patrician cycle is another question’. Earlier, Ó Cróinín (1986), 277–78, pro-
posed that Cummian’s ‘Patrick’ actually referred to Palladius: ‘Traces of Palladius’ Easter
table are to be found in […] Cummian’s Paschal Letter […] Cummian refers to a cycle that
passed under Patrick’s name’; this was repeated in Ó Cróinín (1995), 22–23: ‘one text […]
is an Easter table, which can be associated with Palladius. The Palladian Easter table shows
that the first continental missionaries brought with them the doctrines that were current
in the Gallican and north Italian churches in the late fourth century’. However, he entered
a caveat regarding this hypothesis in Ó Cróinín (2000), 209 n. 14: ‘I published what I took
then to be a fragment of Palladius’s Easter table. That interpretation of the new evidence
now requires revision’; then in Ó Cróinín (2003), 34, he withdrew the hypothesis: ‘The
proposed connection with Palladius, therefore, cannot stand either’; Byrne (1967), 173,
175: ‘The epistle is long and highly erudite […] the paschal computation he [Patrick] is
alleged to have brought is not accepted as valid’. See also n. 79.


 DANIEL MC CARTHY

adopted by Cummian and his community, and no serious endeavour to

analyse the significance of his conspicuous acclaim for, and knowledge
of, the details of Patrick’s cycle, or his understanding of the computisti-
cal structure underlying Victorius’ 532-year cycle. Rather, there has been
a widespread reflex to dismiss his account of Patrick’s cycle as either inac-
curate or irrelevant. These dismissals all neglect to consider the fact that
Cummian, in writing to Ségéne, was addressing the principal of one of
the foremost ecclesiastical institutions in these islands. For Cummian
to have proffered an inaccurate or extravagant account of Patrick’s cy-
cle would have been to invite certain scorn and ridicule from the Iona
scholars. The dismissals are, in effect, anachronistically imposing mod-
ern scepticism regarding Patrick on to a seventh-century scholar.

Patrick as a computist
The foregoing adaptation represented a significant computistical exercise
requiring the modification of paschal dates and moons in at least 28 years
of Victorius’ paschal table. But what evidence do we have to show that
Patrick possessed such expertise? In their edition of Cummian’s letter,
Walsh and Ó Cróinín drew attention to a hitherto unnoticed reference
to Patrick in the computistical compilation known as the Sirmond manu-
script, namely Oxford, Bodleian Library, Bodley 309. In 1937, Jones had
demonstrated that this eleventh-century manuscript is the best surviving
representative of the computus used by Bede for his two works on time
composed in the earlier eighth century.50 Then, in 1983, Ó Cróinín pub-
lished the evidence showing that the exemplar of this computus had been
assembled in southern Ireland in c.AD 658, and in 2003 he published an
edition of this Sirmond reference to Patrick from folio 106r as follows:51

Sed cum tres menses uernum habeat tempus, horum trium medius est, qui
initium mundo dedit. Neque solum mensis medius, sed etiam dies men­

50
Jones (1937), 204–19 (Sirmond MS the best representative of Bede’s computus).
51
Ó Cróinín (1983), 246 (‘the collection received its definite form in southern
Ireland c.AD 658’); Ó Cróinín (2003a), 206 (citation); earlier, Ó Cróinín (1986), 277–
78, had mistakenly associated this reference to Patrick with the text that follows hoc ius
ostendit, and this mistake was repeated in Walsh and Ó Cróinín (1988), 29, and sub-
sequently again by Dumville (1993), 87; the mistake was acknowledged in Ó Cróinín
(2003), 34, and remedied in Ó Cróinín (2003a), 206. I am grateful to Immo Warntjes
who drew to my attention that the opening sentences of this Sirmond passage are cited
from the Tractatus Adthanasi; cf. Krusch (1880), 332.


 THE PASCHAL CYCLE OF ST PATRICK

sium medius est. Ex .v. Idus enim Febroarii (ueris inchoatio) in .v. Idus
Martii unus mensis est; ex .v. Idus autem Martii in .viii. Kl. Aprilis .xv.
dies sunt, id est medietas mensis. Ita unus et dimedius mensis subsequitur,
hoc est in .v. Idus Maias. Ex .v. Idus Maias in .v. Idus Augusti, secundum
hanc rationem, tempus aestatis est. Ex .v. Idus Augusti in .v. Idus Nouim­
bris, autumni tempus est; item ex .v. Idus Nouimbris licet dissimiliter in
.v. Idus Februarii, hiemale tempus est. Patricius, in prologo suo, secundum
rationem Anatolii, hoc ius ostendit.

‘But as spring time has three months, it is the middle of these three that
gave a beginning to the world. Nor is it just the middle month, but in-
deed it is the middle day of the months. From the fifth ides of February
(the start of spring) to the fifth ides of March is one month; from fifth
ides of March to the eighth kalends of April are fifteen days, which is
half of a month. So one month follows and a half, this is to the fifth ides
of May. From the fifth ides of May to the fifth ides of August, according
to this reckoning, is the time of summer. From the fifth ides of August to
the fifth ides of November is the time of autumn; likewise from the fifth
ides of November, although different, to the fifth ides of February is the
time of winter. Patrick, in his prologue, demonstrated this rule accord-
ing to the reckoning of Anatolius.’

The anonymous author of this discussion first cites a passage from the
Tractatus Adthanasi which states that spring extends from the fifth ides
of February to the eighth kalends of April, and from thence for another
one and a half months to the fifth ides of May. The author then employs
this reckoning to systematically place the boundaries between the sea-
sons of spring, summer, autumn, and winter on the fifth ides of May,
August, November, and February respectively. His concluding sentence
then represents Patrick as the author of a prologue to an unidentified
work, and states that he had demonstrated this principle of placing the
seasonal boundaries by employing the reckoning of Anatolius. Now, in
the final section of De ratione paschali, Anatolius did indeed discuss the
four seasons, and he placed their mid-points on the eighth kalends of
April, July, October, and January. So this Sirmond account of the sea-
sonal boundaries, which likewise employs the eighth kalends of April as
the median date between the fifth ides February and the fifth ides May,
is fully in accordance with Anatolius’ De ratione paschali, which states:52

52
Anatolius, De ratione paschali 14 (ed. and trans. by Mc Carthy and Breen (2003),
53, 70); for a discussion, see Mc Carthy and Breen (2003), 112–13.


 DANIEL MC CARTHY

sed ita unumquodque tempus inchoandum est ut a prima die ueris tempus
aequinoctium diuidat et aestatis viii Kal. Iulii et autumni viii Kal. Octo­
bris et hiemis viii Kal. Ianuarii similiter diuidat.

‘But each season is to begin in such a way that, from the first day of
Spring, the equinox [sc. viii Kal. Apr.] divides time, and similarly viii
Kalends of July divides the Summer, viii Kalends of October the Au-
tumn and viii Kalends of January the Winter.’

Thus this brief Sirmond account represents Patrick as making compu-

tistical deductions based upon Anatolius’ De ratione paschali, a work
which we know to have circulated in Gaul in the earlier fifth century,
because it was the primary source used by Sulpicius Severus in the com-
pilation of the latercus. It is clear from this, therefore, both that Patrick
was a capable computist, and that he was familiar with an important
fifth-century paschal tract that circulated in Gaul. Moreover, since this
computistical demonstration appeared in a prologue by Patrick, it car-
ries the further implication that his ensuing compilation was a paschal
table, as Walsh and Ó Cróinín likewise concluded.53 This inference is, of
course, implicit in Cummian’s tulit et fecit, for Patrick could hardly have
made a convincing adaptation of Victorius’ paschal cycle without first
presenting an explanation of his reasons for doing this.
The second item of evidence occurs in a Carolingian computus pub-
lished by Arno Borst in 2006 and entitled by him as Kölner Lehrbuch
von 805, and it reads as follows:54

De augmento lunae, in quod et in quibus locis sit? Id in quo loco speciali

ponitur?
Decem. In duobus <iuxta> libros Patricii et Laterci. In sex locis, in XII.
Kalendas Ianuarii, in Idus Decembris, in Nonas Decembris, in XI. Kalen­
das Decembris, in VI. Kalendas Ianuarii iuxta Anatolium.

53
Walsh and Ó Cróinín (1988), 31: ‘there can be no doubt that the author of the
Sirmond text thought of this work as the Prologue to an Easter table of Patrick’.
54
Comp. Col. 4.8 (ed. by Borst (2006), 918); the edition reads quinque locis where the
sole manuscript has sex locis, which I have restored on the advice of Immo Warntjes. The
five dates imply that the epacts on 1 January of the following years were, respectively, luna
11, 19, 27, 11/12, and 5, and these do not reconcile with any such years of the latercus or
Anatolius’ De ratione paschali. Furthermore, the increment of eight days between the first,
second, third, and fifth epacts would imply a saltus inserted at an interval of seventeen years
(16×11+1×12=188=6×30+8), an interval not known from any lunar cycle. Consequently,
nothing can be inferred from these dates regarding either the latercus or Patrick’s cycle.


 THE PASCHAL CYCLE OF ST PATRICK

‘Regarding the increment of the moon [saltus], to what purpose and

in what places may it be? In what special place is it put? Ten. At two
<according> to the books of Patrick and the Latercus. In six places, at
the twelfth kalends of January, at the ides of December, at the nones of
December, at the eleventh kalends of December, at the sixth kalends of
January according to Anatolius.’

This text thus first poses questions concerning the placing of the incre-
ment of the moon (or the saltus) and then in its response refers to the
books of Patrick, the latercus, and Anatolius. It is evident from this refer-
ence, therefore, that in c.AD 805 a book was associated with Patrick in
which the technicality of the augmentum lunae was discussed, and such
a discussion almost certainly had been in the context of a paschal cycle.
It seems most likely, therefore, that the Kölner Lehrbuch’s liber Patricii,
the Sirmond’s prologus Patricii, and Cummian’s cyclus Patricii all refer to
the same paschal work.
Turning from computistical to hagiographical works, I next consider
the earliest Vitae Patricii, namely those of Muirchú and Tírechán, which
were attributed by their editor, Ludwig Bieler, to the last four decades
of the seventh century.55 As already stated above, both of these authors
date Patrick’s obit using the Annus Passionis, Victorius’ systematic enu-
meration of the years from the first year of his paschal table in which
he considered that the Crucifixion had occurred. This would indeed
be the appropriate chronological framework for Patrick’s obit when he
had adapted his paschal table from that of Victorius.56 Moreover, in a
protracted, highly dramatic narrative, occupying almost one third of
his Vita, Muirchú represented that Patrick had brought the ‘first Easter’
to Ireland, which, in chronological terms is untrue, since the evidence
points to the latercus having reached Ireland by c.AD 425.57 However, if
Muirchú held with Cummian that ‘those who observe a ‘cycle of eighty-
four years can never arrive’ at the correct reckoning of Easter’, and that
those who kept Pasch on luna 14 were ‘heretics’, then he could maintain
that Patrick’s cycle observing Alexandrian lunar limits and employing
a 19-year lunar cycle was the first ‘correct reckoning’ of Easter in Ire-
55
Bieler (1979), 2 (date of Muirchú), 42 (relative date of Tírechán).
56
Muirchú, Vita Patricii (Bieler (1979), 62–63); Tírechán, Collectanea (Bieler
(1979), 126–27).
57
Muirchú, Vita Patricii (Bieler (1979), 82–99; Muirchú’s account of Patrick’s ‘first
Easter’ in Ireland occupies almost one third of his Vita); Tírechán, Collectanea (Bieler
(1979),130–33; Tírechán’s summary account of the same event); Mc Carthy (2011), 69
(date of latercus in Ireland).


 DANIEL MC CARTHY

land.58 Indeed the reiteration by Muirchú of the otherwise redundant

‘first’ suggests this to have been the case. His account of this ‘first Easter’
commences as follows:59

Adpropinquauit autem pasca in diebus illis, quod pasca primum Deo in

nostra Aegipto huius insolae uelut quondam in genesseon celebratum est, et
in[uen]ierunt consilium, ubi hoc primum pasca in gentibus ad quas missit
illum Deus celebrarent.

‘In those days Easter was approaching, the first Easter to be offered to
God in the Egypt of this our island as it once was (offered), as we read in
Genesis, in Gessen; and they took counsel where they should celebrate
this first Easter among the pagans to whom God had sent him.’

Muirchú’s reference here to Gessen is especially relevant since the land

of Gessen was the place in Egypt where, according to Genesis 45:10–
47:27, the sons of Israel settled at the invitation of Joseph. Consequently,
it was in Gessen that God prescribed to Moses the rite of the Pasch, as is
recounted in Exodus 12:2,18:

mensis iste vobis principium mensium, primus erit in mensibus anni […]

primo mense quartadecima die mensis ad vesperam comedetis azyma usque
ad diem vicesimam primam eiusdem mensis ad vesperam.

‘This month shall be for you the beginning of months; it shall be the
first month of the year for you […] In the first month, on the fourteenth
day of the month at evening, you shall eat unleavened bread, and so until
the twenty-first day of the month at evening.’

In this way Muirchú’s reference to the ‘first Easter […] in Gessen’ implic-

itly invoked biblical authority for the lunar limits luna 14–21, the same
limits ascribed by Cummian to Patrick. This is, of course, the basis for
the Alexandrian paschal tradition, and Muirchú’s reference to ‘Egypt’
in ‘the first Easter to be offered to God in the Egypt of this our island’,
likewise implicitly identifies Patrick’s paschal tradition as Alexandrian.
What should be noted is that, notwithstanding his protracted account,
Muirchú gave absolutely no technical details of the paschal cycle used

58
Cummian, De controversia paschali ll. 211–12 (Walsh and Ó Cróinín (1988),
84–87): ‘ad ueram paschae rationem numquam peruenire’ eos qui cyclum ‘.lxxxiiii. an­
norum’ obseruant.
59
Muirchú, Vita Patricii (Bieler (1979), 83).


 THE PASCHAL CYCLE OF ST PATRICK

by Patrick, for to have done so would have revealed that it derived from
that of Victorius compiled in AD 457, and hence that Patrick’s mission
could not immediately follow that of Palladius in AD  431. In short,
both these Vitae associate Patrick’s obit with Victorius’ Annus Passionis,
and Muirchú implicitly attributes to Patrick the introduction to Ireland
of the Alexandrian paschal tradition.
In summary, the evidence, though fragmentary, indicates the surviv-
al, at least to the early ninth century, of a paschal work attributed to Pat-
rick, and this, together with Cummian’s account, suggests that Patrick
was indeed a competent computist who was actively involved with the
two important paschal tracts known in fifth-century Gaul, Anatolius’
De ratione paschali and Victorius’ paschal cycle. Furthermore, this con-
clusion of Patrick’s competence as a computist is in accordance with the
embedded numerical phenomena that David Howlett has extensively
demonstrated in his edition of Patrick’s Confessio.60

The use of Patrick’s cycle

Cummian’s account of the synod of Mag Léne is the only explicit de-
scription we have of a group of Irish churches agreeing to reform their
paschal tradition, so it is worth examining it in some detail. It reads:61

Anno igitur, ut predixi, emenso, iuxta Deuteronomium, interrogaui pa­

tres meos ut annuntiarent michi, maiores meos ut dicerent michi, suc­
cessores uidelicet nostrorum patrum priorum Ailbei episcopi, Quera<ni
C>oloniensis, Brendini, Nessani, Lugidi <quid sentirent de excommuni­
catione nostra, a supradictis sedibus apostolicis facta. At> illi congregati in
unum, alius per se, <aliu>s per <legatum suum uice suo> missum, in Cam­
po Lene sancxerunt et dixerunt: ‘Dec<essores> nostri <mandauerunt> per
idoneos testes, alios uiuentes, alios in pace dormientes, ut <mel>iora et po­
tiora probata a fonte baptismi nostri et sapientiae et successoribus apostolo­
rum Domini delata, sine scrupulo humiliter sumeremus.’ Post in commone
surrexerunt, et super hoc, orationem, ut moris est nobis, celebrauerunt ut
pascha cum uniuersali aecclesia in futuro anno celebrarent.

Howlett (1994), 94–115 (‘Confessio – Analysis and commentary’).

60

Cummian, De controversia paschali ll. 259–70 (Walsh and Ó Cróinín (1988),

61

90–93). Another meeting to discuss Easter, approximately contemporaneous with that

of Mag Léne, is described in Vita Sancti Munnu, where Laserian, abbot of Leighlin, and
Munnu, abbot of Taghmon, defended new and old paschal traditions respectively, but
reached no agreement; see Warntjes (2015), 45–46.


 DANIEL MC CARTHY

‘Therefore after a full year (as I said above), in accordance with Deu-
teronomy, I asked my fathers to make known to me, my elders (that is
to say, the successors of our first fathers: of Bishop Ailbe, of Ciaran of
Clonmacnois, of Brendan, of Nessan, and of Lugid) to tell me what they
thought about our excommunication by the aforementioned Apostolic
Sees. Having gathered in Mag Léne, some in person others through rep-
resentatives sent in their place, they enacted and said: ‘Our predeces-
sors enjoined, through capable witnesses (some living, some resting in
peace), that we should adopt humbly without doubt better and more
valid proofs proffered by the font of our baptism and our wisdom and
by the successors of the Lord’s Apostles.’ Then they arose in unison and
after this, as is our custom, they performed a prayer, that they would
celebrate Easter with the Universal Church the next year.’

Thus Cummian, having completed his year of study of Scripture, histo-

ry, and paschal cycles, approached the successors of the church found-
ing fathers, listing five by name, seeking their views on ‘our excommuni-
cation by the aforementioned Apostolic Sees’, that is the Sees of Rome,
Jerusalem, Antioch, and Alexandria.62 Now, as Walsh and Ó Cróinín
pointed out, this was not an excommunication formally issued by these
Apostolic Sees, but rather an excommunication implicit in a canon re-
corded by Dionysius Exiguus in both the prologue to his paschal cycle,
and his Regulae expositae apud Antiochiam in Encaenis.63 This is iden-
tified by Cummian’s cross reference to ‘the aforementioned Apostolic
Sees’, where his citation of the canon corresponds to those of Diony-
sius. From this it appears that the initiative that led to the synod was
taken by Cummian himself, by confronting the principals of at least five
churches with the Dionysiac canon of excommunication. Cummian’s
action in confronting these principals with this canon underscores on
the one hand the depth of his respect for, and knowledge of, Dionysius’
paschal work, and on the other hand, that none of these churches was
then following the Alexandrian paschal tradition. That they were most
likely all following the latercus tradition follows from the facts that this
was the tradition hitherto followed by Cummian’s community, and
that he wrote inclusively of ‘our excommunication’ (excommunicatione
nostra), and he emphatically and repeatedly repudiated celebration on
luna 14.

62
Cummian, De controversia paschali ll.  92–93 (names of the Apostolic Sees),
262–63 (excommunication); Walsh and Ó Cróinín (1988), 70–71, 90–91.
63
Walsh and Ó Cróinín (1988), 70–71 n. 90 (identification of Dionysius’ canon).


 THE PASCHAL CYCLE OF ST PATRICK

Regarding the synod itself, Cummian does not indicate either who,
or how many attended, but it does seem implicit that the principals of
the foundations of Emly, Clonmacnoise, Clonfert Molua, Mungret,
and Clonfert Brendan attended or sent representatives.64 Since these
five foundations were all prominent churches it also seems likely that
the number at the synod considerably exceeded this five. Having ex-
plained how he instigated the synod, Cummian proceeds straight to the
outcome, namely the unanimous agreement that ‘they would celebrate
Easter with the Universal Church next year’ (ut pascha cum uniuersali
aecclesia in futuro anno celebrarent). While he expressed this in terms of
a universal unity, it should be recalled that his account of the paschal
cycles requires that their decision was actually to adopt the Alexandrian
paschal tradition in the form of Patrick’s cycle.
This is the only account that we have of Irish churches deciding to
celebrate Pasch according to Patrick’s cycle, but it surely was a most
significant event in Irish ecclesiastical history. For the principals of the
five influential foundations of Emly, Clonmacnoise, Clonfert Molua,
Mungret, and Clonfert Brendan to agree unanimously to abandon two
centuries of the latercus tradition and instead to adopt Patrick’s cycle
had major doctrinal implications for Irish Christianity, not the least of
which was for Patrick’s status as a cult figure and as an icon of orthodoxy.
This then raises the question of the usage of Patrick’s cycle both before
and after the synod of Mag Léne.
From Cummian’s description that Patrick tulit et fecit, it appears cer-
tain that he established his cycle in his own church foundations. It seems
equally certain that the cycle survived in use in at least some of these
foundations until c.AD 630, for it is very difficult to imagine the assem-
bly at Mag Léne electing to resuscitate a completely defunct paschal cycle
to represent a universal unity. Indeed, the fact that Cummian endorsed
Patrick’s cycle in full knowledge of the existence of the Dionysiac cycle,
and while repeatedly proclaiming his support for church unity and uni-
versality, is quite remarkable. It surely requires that in an Irish context
Patrick’s cycle had maintained a formidable status as representing the Al-
exandrian paschal tradition from the fifth to the seventh century. Indeed,
the relative antiquity of Patrick’s Alexandrian tradition in Ireland may
well have been the crucial factor that persuaded the synod at Mag Léne
to adopt it, for it was in use in Ireland for more than a century before the

64
Ó Cróinín (1995), 154 (identity of monasteries); cf.  Walsh and Ó Cróinín
(1988), 90–91 nn. 261–62.


 DANIEL MC CARTHY

Roman church adopted the Alexandrian paschal tradition, and for over
half a century before Dionysius compiled his cycle. Regarding the ques-
tion of in which churches Patrick’s cycle had flourished in earlier times,
undoubtedly Armagh appears to be the most likely candidate. While it is
disputed whether Armagh was founded by Patrick, there can be no doubt
regarding its Patrician affiliation from at least the earlier sixth century.65
The Clonmacnoise group of annals, the Annals of Tigernach, Chronicum
Scotorum, and the Annals of Roscrea, record six of the seven annalistic
obits of Armagh principals between AD 469 and 534, two identified as
‘successor of Patrick’. For Clonmacnoise, a competing ecclesiastical insti-
tution, to transmit this substantial account of Armagh’s early ecclesiasti-
cal dynasty and Patrician affinity surely implies its substantial veracity.66
In these circumstances, whether it was founded by Patrick or by one of his
followers, at least by the mid-sixth century Armagh was an important ec-
clesiastical institution that was considered to maintain Patrick’s tradition.
While Bede’s brief summary in his Historia ecclesiastica of the letter
of Pope Honorius to the Irish gives no information regarding the pas-
chal disposition of any Irish foundation, his attribution to Honorius of
the statement, ‘nor should they [the Irish] celebrate a different Easter
contrary to the paschal tables and the decrees of the bishops of all the
world met in synod’, shows that Honorius was endorsing Alexandrian
paschal tradition, for no such universal episcopal decrees are known to
have supported the earlier Roman paschal tradition.67
Bede’s ensuing account of the letter of Honorius’ successor, Pope-
elect John, in AD 640, is more precise, first giving a summary of the
substance of John’s letter followed by a transcript of its beginning. In
the summary Bede wrote: ‘he [ John] showed clearly that Easter Sun-
day ought to be looked for between the fifteenth and twenty-first day
of the moon, as was approved in the Synod of Nicaea’ (evidenter astru­
ens quia dominicum paschae diem a quinta decima luna usque ad XXI,
quod in Nicena synodo probatum est, opporteret inquiri).68 The reference

65
Ó Cróinín (1995), 154–55 (foundation of Armagh).
66
Mc  Carthy (1999–2005), s.a. 469 (Benignus, successor of Patrick – CS/AR/
MB/AU/AI/AB), 479 (Brendann – AR/MB), 481 (Iarlath – MB/AU/AI), 497 (Cor-
mac, successor of Patrick – AT/CS/MB/AU/AI), 515 (Dubthach – AT/CS/MB/AU/
AI), 520 (Ailill – AT/CS/MB/AU), 534 (Ailill – AT/CS/MB/AU).
67
Bede, Historia ecclesiastica 2.19 (ed. and trans. by Colgrave and Mynors (1969),
198–99): neue contra paschales conputos et decreta synodalium totius orbis pontificum aliud
pascha celebrarent.
68
Bede, Historia ecclesiastica 2.19 (Colgrave and Mynors (1969), 200–01).


 THE PASCHAL CYCLE OF ST PATRICK

to luna 15–21 here shows unquestionably that John was endorsing the
Alexandrian paschal tradition and promoting paschal unity based
upon it. In the beginning of this letter, John addressed eleven Irish ec-
clesiastical principals by name, commencing with Tómíne, bishop of
Armagh, and placing Ségéne, abbot of Iona, in the penultimate posi-
tion. He then noted that letters from Ireland to his predecessor, Pope
Severinus, had been left unanswered by his death, and when these were
examined:69

repperimus quosdam prouinciae uestrae contra orthodoxam fidem nouam

ex ueteri heresim renouare conantes pascha nostrum, in quo immolatus est
Christus, nebulosa caligine refutantes ex XIIII luna cum Hebreis celebrare
nitentes.

‘we discovered that certain men of your kingdom were attempting to

revive a new heresy out of an old one and, befogged with mental blind-
ness, to reject our Easter in which Christ was sacrificed for us, contend-
ing with the Hebrews that it should be celebrated on the fourteenth day
of the moon.’

What could be the significance of John’s allegation that ‘certain men

of your kingdom were attempting to revive a new heresy out of an old
one […] that it [Easter] should be celebrated on the fourteenth day of
the moon’? This clearly identifies a seventh-century endeavour to main-
tain a paschal tradition that included luna 14 by means of an innova-
tion, and the only churches in these islands known to have celebrated
Pasch on luna 14 are those that followed the latercus tradition. In the
mid-seventh century those churches that maintained this paschal tradi-
tion faced serious difficulties on account of the inaccuracy of its 84-year
lunar cycle, resulting in the latercus moon advancing one day ahead of
the real moon about every 55 years. For example, by AD 640 the latercus
moon was on average about four days in advance of the real moon. In
consequence, when the latercus scheduled Pasch on luna 14, as it did
on 21 April AD 636 and 18 April AD 639, instead of rising fully illumi-
nated at sunset on the eve of Easter Sunday, the real moon rose around

69
Bede, Historia ecclesiastica 2.19 (Colgrave and Mynors (1969), 200–01); cf. also
Bede’s statement in the same chapter: Quo epistulae principio manifeste declaratur, et nu­
perrime temporibus illis hanc apud eos heresim exortam, et non totam eorum gentem sed
quosdam in eis hac fuisse inplicitos (‘At the beginning of this letter it is clearly asserted that
this heresy had sprung up among them very recently and that not all the race but only
certain of them were implicated in it’).


 DANIEL MC CARTHY

mid-day and was visibly not long after first quarter.70 Thus there was at
the time substantial pressure on those who wished to maintain the la­
tercus lunar limits of luna 14–20 to move to the much more accurate
19-year lunar cycle. Here Patrick’s lunar cycle offered them an attractive
possibility, for if all of Patrick’s luna 21 dates are advanced by one week
to luna 14, then the lunar limits will be the latercus luna 14–20, and the
resultant paschal termini will be the Alexandrian 22 March to 25 April,
as may be seen from Figure 2. Now, Cummian himself gives an explicit
statement that in the aftermath of the synod a strong counter-challenge
issued from one of the conservative faction, writing:71

Sed non post multum, surrexit quidam paries dealbatus, traditionem seni­
orum seruare se simulans, qui utraque non fecit unum, sed diuisit et ir­
ritum ex parte fecit quod promissum est.

‘But a short time afterwards a certain whited-wall arose, pretending to

preserve the tradition of our elders, who did not unite with either part
but divided them and partly made void what was promised.’

Cummian’s characterisation of the instigator of this challenge as ‘pre-

tending to preserve the tradition of elders’, but at the same time as not
uniting with ‘either part’, would accord well with an adaptation of Pat-
rick’s cycle that simultaneously incorporated the latercus luna 14–20 into
a 19-year lunar cycle with Alexandrian termini 22 March to 25 April.
I suggest, therefore, that this may have been the basis for John’s allega-
tion of ‘a new heresy out of an old one’.72 Whether this conjecture be
accepted or not, it emerges clearly from Cummian’s account that many
churches that faithfully maintained the latercus paschal tradition from

70
Blackburn and Holford-Strevens (1999), 873 (latercus paschal dates and moons
for AD 636 and 639). Simulation, using Voyager Interactive Desktop Planetarium, shows
that on the eve of Sunday 21 April 636, and Sunday 18 April 639, the moon arose at 11:06
am and 1:08 pm, and it was 63% and 77% illuminated respectively. Cf. http://eclipse.
gsfc.nasa.gov/phase/phasecat.html s.a. 636, 639 and s.v. ‘Full Moon’ after 25 March. See
also Mc Carthy (2011), 70–71 n. 99 (latercus lunar inaccuracy in AD 592, 595, 596, and
599).
71
Cummian, De controversia paschali ll. 270–72 (Walsh and Ó Cróinín (1988),
93–94).
72
Ó Cróinín (1985), 505–16, gives an ingenious interpretation of John’s allega-
tion of ‘new heresy’ based upon the one-day asynchronism between the Victorian and
Dionysiac paschal moon in 640, but it is implausible that anyone employing the Vic-
torian table could be accused of celebrating on luna 14; Warntjes (2010a), LXXXIII,
disputes the hypothesis on good grounds.


 THE PASCHAL CYCLE OF ST PATRICK

their foundation had, by c.AD  630, arrived at a profound theological

dilemma. Should they conserve their latercus luna 14–20 tradition by
modifying the lunar cycle, but nevertheless they would remain, as Cum-
mian remarks, ‘but pimples of the face of the earth’ (mentagrae orbis
terrarum).73 Or should they join the emerging western consensus led by
the church of Rome and accept the universality of the Alexandrian tra-
dition of luna 15–21? Cummian’s letter shows that at Mag Léne a con-
siderable number of these churches made the latter choice in the form
of Patrick’s cycle. Less than ten years later, John’s letter giving priority to
the bishop of Armagh suggests that in the papal view he was considered
pre-eminent in ecclesiastical matters in Ireland, and that John was pre-
pared to accept Patrick’s adaptation of Victorius’ cycle as an orthodox
celebration of the Pasch.
To consider what became of Patrick’s cycle after Cummian’s letter,
I examine the accounts of Patrick and his Pasch given in the seventh-
century Vitae, commencing with that given by Muirchú. In this account,
by advancing the chronology of Patrick’s mission to immediately follow
that of Palladius, and by representing so dramatically Patrick’s imposi-
tion of this ‘first Easter’ on the court of King Loíguire at Tara, Muirchú
expanded the paschal claim for Patrick far beyond that made by Cummi-
an. Simultaneously, Muirchú also consigned to oblivion two centuries
of Irish Christianity that had faithfully maintained the latercus paschal
tradition. For his part, Tírechán gave only a brief summary of this sup-
posed Easter triumph on the plain of Brega, and concentrated instead
on enumerating the distribution and number of Patrick’s churches and
clerics. These accounts both suggest that the cult of Patrick expanded
very greatly in the decades following the synod at Mag Léne, from which
it appears that the spectacular success of Patrick’s mission actually oc-
curred about 150 years after his death. The commencement of Patrick’s
exaltation to iconic status can be seen in Cummian’s choice of words,
sanctus Patricius papa noster, and this exaltation rose to a crescendo with
the Vitae of Tírechán and Muirchú.
What is also obvious, however, is that in this expansion Patrick’s pas-
chal cycle faced formidable competition. For example, Ó Cróinín point-
ed out that the exemplar of the Sirmond group of computistical texts
compiled in southern Ireland in c.AD 658 was ‘concerned almost exclu-

Cummian, De controversia paschali l. 110 (Walsh and Ó Cróinín (1988), 72–

73

75).


 DANIEL MC CARTHY

sively with the Alexandrian cycle’.74 This would then suggest that within
two decades of the synod at Mag Léne other southern Irish scholars were
seriously considering the cycle of Dionysius as a preferred alternative.
That this movement towards Dionysius prevailed by the end of the sev-
enth century can be seen from the fact that the three Irish computistical
textbooks, the Computus Einsidlensis, the Munich Computus, and De
ratione conputandi, dating from the later seventh to the earlier eighth
century, make no reference whatsoever to Patrick’s cycle, for this surely
signifies that their authors did not consider it appropriate to their pur-
pose.75 For these authors were all emphatically endorsing the Dionysiac
paschal cycle which observes both the Alexandrian limits luna 15–21
and the paschal termini 22 March to 25 April. At least by AD 703 this
southern Irish computus had reached Bede in Northumbria where he
first used it in the compilation of his De temporibus, and then again in
c.AD 725 for the compilation of his De temporum ratione.76 By the end
of the eighth century, with the support of these two works, the Diony-
siac cycle was extending its provenance into the proximate parts of the
Continent, to ultimately become the paschal cycle of the entire Western
church. One would judge from this sequence that, following the synod
of Mag Léne the actual use of Patrick’s cycle did not nearly match the
expansion in his cult status. Rather his cycle, with its incongruent pas-
chal termini of 23 March to 26 April, had eventually to be abandoned
by those Irish churches that had adopted it. The change from Patrick’s
to Dionysius’ cycle involved no controversial doctrinal or theological is-
sues, but as far as I am aware no documentary evidence of this change has
survived, so that we have no clear information as to how it proceeded in
Ireland. To effect a transition from Patrick’s cycle to that of Dionysius
required only the deferral of the saltus for thirteen years, for this will syn-
chronize the lunar epacts and result in the removal of epact luna 25 and
consequently the paschal termini will correspond to the Alexandrian
22 March to 25 April.

74
Ó Cróinín (1983), 242 (citation).
75
Warntjes (2010a), LIV–LV (dating of Einsiedeln, Munich, and De ratione con­
putandi), 1–317 (Munich Computus: edition, translation, and commentary); Warntjes
(2005), 61–64 (discovery of the Computus Einsidlensis); Walsh and Ó Cróinín (1988),
101–229 (De ratione conputandi: edition and commentary).
76
Bede’s De temporibus and De temporum ratione are edited by Charles W. Jones
in CCSL 123, 241–544, 579–611, with Theodor Mommsen in MGH Auct. ant., 223–
354.


 THE PASCHAL CYCLE OF ST PATRICK

However, the fact that the Collectio Canonum Hibernensis, a substan-

tial collection of ecclesiastical regulations compiled in the first half of
the eighth century, makes only one brief reference to the Pasch would
suggest that its compilers had not been able to reach agreement on the
cycle to be commended to its readers. The solitary reference, which oc-
curs only the B recension, reads as follows:77

Gildas ait: Britones toto mundo contrarii, moribus Romanis inimici, non
solum in missa, sed etiam in tonsura, cum Iudeis umbrae futurorum se­
ruientes, quam ueritati. Pascha cum Iudaeis XIIII luna celebrantes.

‘Gildas says: Britons contrary to the whole world, enemies of the Ro-
man customs, not only in the Mass but also in the tonsure, and with the
Jews serving the shadows of things to come, rather than truth. Celebrat-
ing Pasch with the Jews on luna 14.’

Thus, on a subject as germane to Christian rite and belief as the Pasch,

about which controversy had raged for the preceding century, the com-
pilers of the Hibernensis were only able to agree upon Gildas’ condem-
nation of Britons celebrating Pasch on luna 14. While this does indeed
suggest that the compilers favoured the Alexandrian luna 15–21, it
leaves the version of the cycle to be used as ambiguous. It is particularly
noteworthy that the compilers, despite a dozen citations of Patrick in
respect of such diverse matters as ecclesiastical judges, kings and their
works, repentance, wandering monks, excommunication, collection of
money, profanation, and tonsure, chose not to refer to either Patrick’s
paschal cycle or prologue.78 This would seem to suggest that in the first
half of the eighth century both Dionysius’ and Patrick’s paschal cycles
were in use amongst the influential Irish churches.
It emerges from the foregoing that the actual use of Patrick’s pas-
chal cycle did not flourish in the decades following the synod of Mag
Léne. However, it is equally clear that the decision of churches of the
latercus tradition at that synod to thenceforth identify Patrick as their
icon of ecclesiastical orthodoxy and unity was subsequently embraced
with ever increasing enthusiasm by many more Irish churches. Already

77
Collectio Canonum Hibernensis 52 (ed. by Wasserschleben (1885), 212 n. d; the
last sentence of this citation occurs only in the B recension).
78
Collectio Canonum Hibernensis 21.12, 25.3–4, 29.7–8, 37.27, 39.11, 40.9, 42.26,
45.9, 52.7, 56.18 (Wasserschleben (1885), 65, 77, 77, 101, 138, 151, 155, 169, 177, 213,
239): citations attributed to Patrick.


 DANIEL MC CARTHY

by the late seventh century, Muirchú had advanced Patrick’s mission to

immediately follow that of Palladius, and explicitly represented him to
be the founder saint of Irish Christianity: ‘when Patrick came the wor-
ship of idols was abolished and the catholic Christian faith spread over
our whole country’ (euersis enim in aduentu Patricii idulorum culturis
fides Christi catholica nostra repleuit omnia).79 At the same time, Muirchú
maintained absolute silence regarding the earlier arrival in Ireland of the
latercus, and the two centuries of its hegemony. By AD 730, these dis-
tortions had been imposed upon the Iona chronicle by the deletion of
the years AD 425–31, and the interpolation of entries asserting Patrick’s
alleged papal approval at AD 441 and the orthodoxy and success of his
mission at AD  443.80 Since then Muirchú’s seventh-century revision
of the chronology and nature of Patrick’s mission has prevailed as the
popular and scholarly consensus until the twentieth century. Cummian’s
letter, however, provides us with a unique and precious glimpse of the
initiation of this process of ecclesiastical revision, showing that it was
instigated by the churches founded in the latercus paschal tradition.

Conclusions
A critical appraisal of Cummian’s letter shows that, of the ten paschal
cycles that he enumerated, it was that of Patrick that he and his commu-
nity had chosen to adopt. This follows from the following details: Cum-
mian placed it first; he referred to Patrick as noster papa; he applied the
title sanctus only to Patrick; he supplied lunar limits and an equinoctial
date only for Patrick’s cycle, and these lunar limits agree with those that
he himself so passionately espoused. These details, together with Cum-
mian’s reference to the 532-year cycle adopted by his community, show
that Patrick had brought to Ireland a copy of the paschal cycle compiled
by Victorius in AD 457, and consequently that Patrick’s mission must
postdate that year. Cummian also indicated that Patrick had adapted
this cycle, and comparison of Patrick’s lunar limits with Victorius’ cycle
suggest that for epact luna 25 Patrick transferred celebration to the May
lunation, so that all instances of paschal luna 15–21 in his table fell after

79
Muirchú, Vita Sancti Patricii (Bieler (1979), 76–77).
80
Mc Carthy (2008), 140–48 (the revision and interpolation of Patrician entries
into the Iona chronicle AD 725–29), 143 (Patrick’s ‘approval’ at AD 441 and ‘success’ at
AD 443).


 THE PASCHAL CYCLE OF ST PATRICK

the equinox. Further, he removed the ambivalence of Victorius’ double-

dates by deleting all instances of luna 22 paschal dates. In consequence,
Patrick’s adaptation reproduced the essential features of the Alexandrian
paschal tradition, with the addition that in every instance the paschal
luna 14 fell after the equinox, so that his paschal termini were inevitably
23 March to 26 April. Cummian’s account of the synod of Mag Léne
shows that in c.AD 630 at least five influential churches, previously ad-
herents to the latercus paschal tradition, elected to adopt Patrick’s cycle.
Consideration of the paschal references in the Vitae Patricii by
Muirchú and Tírechán supports the conclusion that Patrick had been
the first to bring the Alexandrian paschal tradition to Ireland. These
Vitae also suggest that by the later seventh century the process of the
acceptance of Patrick as an icon of paschal orthodoxy by other Irish
churches was widespread, and that at that time Armagh was considered
the leading representative of the Patrician churches. This view is cor-
roborated by the letter of AD 640 written to the northern churches by
Pope-elect John in which he placed the bishop of Armagh as the first of
his eleven addressees. It is apparent that the seventh-century decision
of the Roman church to embrace the Alexandrian Pasch and to use this
to promote paschal unity had the immediate effect of placing the Pa-
trician churches in Ireland in a uniquely advantageous doctrinal posi-
tion.81 Since the later fifth century they had been an isolated outpost
celebrating according to the Alexandrian tradition, while about them in
Ireland and Britain the latercus tradition had flourished. But suddenly,
in the seventh century, they found their paschal tradition supported by
the Roman church, while at the same time the authority of the latercus
tradition languished as its paschal moon moved steadily and undeniably
further in advance of the real moon.
Patrick emerges from this examination of Cummian’s letter with his
intellectual status very significantly enhanced, for he is shown to have
engaged creatively and computistically with the most influential and
widely used Gaulish paschal compilation of the fifth century. He is also
seen to have recognized the pre-eminence of the Alexandrian paschal
tradition long before this was accepted by the Roman church, raising a
serious question-mark regarding the frequently supposed ‘Romanising’
character of his mission. This view of his mission appears rather to be
anachronistic, arising from the approval in the seventh century by the

81
Cf. Warntjes (2010a), XXXIX–XL n. 86 (‘It appears, then, very probable that
Rome officially switched from the Victorian to the Dionysiac reckoning in the 640s or
650s’).


 DANIEL MC CARTHY

Roman church of Alexandrian paschal principles. In view of Patrick’s

intellectual achievements, the prevailing custom of either ignoring or
dismissing Cummian’s account of Patrick’s cycle by historians of early in-
sular Christianity, and of Patrick in particular, is quite remarkable.82 For
Cummian not only provides essential details of Patrick’s cycle but also a
unique glimpse of the process of that cycle being adopted by influential
Irish churches that had hitherto supported the latercus tradition. This
was a crucial development in Irish ecclesiastical doctrine and politics.
Finally, when one considers that by the latter part of the fifth century
Ireland’s Christians had been confronted with the two major Gaulish
paschal compilations of that century, Sulpicius’ 84-year latercus and Vic-
torius’ 532-year cycle, it is clear that they were introduced to paschal
conflict at an early stage of their conversion to Christianity. Further,
Cummian’s letter shows that by c.AD 630 Ireland’s ecclesiastical scholars
had undertaken comprehensive study of the use of the 19-year cycle in
scheduling paschal celebration. Then, within two decades, other Irish
scholars had compiled the exemplar for the Sirmond computus. In these
circumstances it is not to be wondered at that by the end of the seventh

82
The following works treating early insular Christianity make no reference to
Patrick’s cycle: Williams (1912); Meissner (1929); Mac Neill (1934); O’Rahilly (1942);
O’Rahilly (1946); Bieler (1949); Ryan (1958); Carney (1958); Carney (1961); Binchy
(1962): de Paor (1964); de Paor (1998); Hughes (1963); Hughes (1966); Hughes
(1972); Barley and Hanson (1968); Thomas (1981); Bitel (1990); Etchingham (1999).
On the other hand the following works have effectively dismissed Cummian’s
account of Patrick’s cycle as irrelevant to seventh-century history: Plummer (1896), ii
26: ‘Cummian’s letter on the Easter question […] is still two hundred years later than his
[Patrick’s] supposed mission’; Mac Carthy (1901), cxl: ‘the description of the Patrician
cycle is demonstrably false’; Jones (1943), 95 n. 1: ‘it is clear that no one used it [Patrick’s
cycle] by the beginning of the seventh century’; Hanson (1968), 104 n. 4:‘Cummian in
his letter to Segene would, of course, ascribe to Patrick as traditional author any Paschal
cycle of which he approved and whose immediate origin he could not trace’; Thompson
(1985), 159: ‘No one else at any date mentions that Patrick introduced a new Paschal sys-
tem into Ireland, and Cummian says enough about this alleged Paschal system to suggest
to experts on the subject that he was in all probability misinformed’; Ó Cróinín (1986),
278: ‘Cummian refers to a cycle that passed under Patrick’s name’; Hughes (1963), 62:
‘Cummean who urged his fellow clergy to adopt the Roman Easter c 632’; Walsh and Ó
Cróinín (1988), 32: ‘Cummian was not prepared to follow the Patrician cycle’; Dumville
(1993), 85, 87: ‘Bishop Palladius’s computus?’, ‘‘Cummian’s papa noster was Palladius,
not Patrick’’, citing Ó Cróinín (1986), 282; Ó Cróinín (1995), 161: ‘The mention of
Patrick’s name comes almost as an aside: his Easter table heads the list of ten such tables
that Cummian consulted, but he does not champion its usage’; Charles-Edwards (2000),
406: ‘It has been strongly argued that this was in fact the cycle of Palladius rather than
Patrick’; O’Loughlin (2005), 102: ‘for whatever computus they [seventh-century Irish
churches] had in their possession which they considered to be their most ancient prac-
tice they linked it with Patrick’s name’. See also n. 49.


 THE PASCHAL CYCLE OF ST PATRICK

century Irish scholars were in a position to write a textbook such as the

Munich Computus, which deals so authoritatively with the details of
paschal celebration.

Acknowledgements
I wish to express my gratitude to the following: Immo Warntjes, then of
the Historisches Institut, University of Greifswald, and now of Trinity
College, Dublin, for drawing to my attention the reference to Patrick’s
book in Arno Borst’s edition of Das Kölner Lehrbuch von 805, and for
his assistance in obtaining a facsimile of MS London, British Library,
Cotton Vitellius A XII, 81v; Roy Flechner, then Trinity College, Cam-
bridge, now University College, Dublin, for drawing to my attention
the single reference to Pasch in the A and B recensions of the Collectio
Canonum Hibernensis; Leofranc Holford-Strevens of Christ Church
College, Oxford, for his helpful discussion of Cummian’s account of
Patrick’s lunar limits and the equinox; David Howlett, editor of the Dic­
tionary of Medieval Latin from British Sources, Oxford, for his assistance
with translations from the Sirmond MS, the Hibernensis, and the Kölner
Lehrbuch von 805; Dáibhí Ó Cróínín, whose own computistical research
and publications have been inspirational, for providing a wonderfully
stimulating environment for us in Galway in 2010 in which these ideas
were first discussed.


 LUCIANA CUPPO

FELIX OF SQUILLACE AND THE

DIONYSIAC COMPUTUS II: ROME,
GAUL, AND THE INSULAR WORLD

Abstract
The dissemination of the Dionysiac computus in Rome and in Gaul did not
happen in a vacuum, but found a strong competitor in the well-entrenched
Victorian computus. The first part of the paper considers such opposition on
the basis of three manuscript witnesses: Reg. lat. 2077 and Vat. lat. 1548 of the
Vatican Library, and MS 645 of the Burgerbibliothek at Bern.
The second part of the paper considers the dissemination of the Dionysiac
computus in the insular world. The main witness is MS Digby 63 of the Bodle-
ian Library, Oxford. Written in its present from in AD  867, the manuscript
includes various blocks of computistical material derived from earlier sources.
They include a dossier with the letters of Dionysius Exiguus and others on the
computation of Easter. The dossier ends with the prologue and preface to the
cycles of Felix of Squillace (AD 616).
Certain palaeographical traits betray the Roman provenance of the Dionysiac
dossier. While it is not possible to establish a definite date for the arrival of the
Dionysiac collection in England, there is a strong possibility that the dossier was
sent from Rome in the time of Pope Vitalian (AD 672–76) in the context of his
support for the Dionysiac computus and its adoption in Rome.
Keywords
Dionysius Exiguus, Felix of Squillace, Pope Vitalian; Oxford, Bodleian Library,
Digby 63, Vatican, Biblioteca Apostolica, Reg. lat. 2077, Vatican, Biblioteca
Apostolica, Vat. lat.1548, Bern, Burgerbibliothek, 645.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 138-181
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114736
 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Introduction
At the Second International Conference on the Science of Computus
(Galway 2008) we met Felix of Squillace and identified two discrete lines
of transmission of his computistical work, which includes a prologue and a
preface to the continuation of Dionysius’ cycles for an additional 95 years,
from AD 627 to 721.1 The first line of transmission is located at Bobbio
(Milan, Biblioteca Ambrosiana, H 150 inf., 108r–v), while the second
one, which had also reached Bobbio, as it is attested in the same codex
on fols 50r–v, has some traits in common with the manuscript Oxford,
Bodleian Library, Digby 63, 71r. These traits point to an exemplar whose
text was known in ancient Burgundy and in the insular world. Its prov-
enance cannot as yet be determined with certainty, but can be tentatively
located in ancient Burgundy. Therefore, Milan, Biblioteca Ambrosiana,
H 150 inf. presents two distinct recensions of Felix’s text, a (presumably)
local one and another one known in the insular world and Burgundy.
The reaffirmation of Dionysius’ principles and computistical meth-
ods by Felix in AD 616 occurred in an environment largely unfavourable
to the Alexandrian Easter reckoning. In Italy, Victor of Capua and Cas-
siodorus had supported Dionysius.2 Both scholars lived in Italy south of
Rome. In Rome itself, the climate was different and the Roman system
(the computus of Victorius of Aquitaine) was entrenched.3 Felix was
heir to the Dionysiac tradition promoted by Cassiodorus and Victor of
Capua half a century earlier and was still meeting with opposition on the
Roman part. Witnesses to the Romanist resistance against the Dionysi-
ac reckoning are the manuscripts Vatican, Biblioteca Apostolica, Vat. lat.
1548 and Reg. lat. 2077. The first part of this paper will examine opposi-
tion to the Dionysiac computus as evidenced by these two manuscripts.
The second part of the paper will consider what evidence we have for
the transmission of Felix’s computus in Gaul and the insular world. The
cycles of Felix, often transmitted anonymously, remained in the back-

1
My talk at the conference was entitled ‘Felix of Squillace and the Dionysiac
Computus I’. A revised version of this paper is now published as Cuppo (2011).
2
For Cassiodorus, see his eulogy of Dionysius Exiguus in Institutiones 1.23 (ed. by
Mynors (1937), 61–64), and his update of the Dionysiac computus published with Insti-
tutiones 2, known as the Computus of AD 562. It is discussed in Cuppo (2011), 126–29.
For Victor of Capua, see Bede, De temporum ratione 51 (ed.  by Charles  W. Jones in
CCSL 123B, 272–73).
3
The computus of Victorius saw its birth in Rome, at the behest of Archdeacon
Hilary. Vatican, Biblioteca Apostolica, Reg. lat. 2077, dated AD 585, uses the Victorian
dates in its continuation of the Chronicle of Prosper.


 LUCIANA CUPPO

ground, for the churches in Gaul upheld the system of Victorius of Aqui-
taine and the Irish had their own methods of computation. The popes,
the only persons whose intervention would had been considered authori-
tative, did not take an official stance on the matter for over a century after
Dionysius submitted to the papal chancery, without definite results, his
proposal for the computation of Easter.4 While keeping mum (and being
properly rebuked by Columbanus for such behaviour) on the Easter con-
troversy, the pope supported in effect the status quo, i.e. the primacy of the
Victorian system in Rome and in Gaul.5 The codices Vatican, Biblioteca
Apostolica, Vat. lat. 1548 and Reg. lat. 2077 give a realistic representation
of the mindset of the supporters of the Victorian computus in Rome or
in Gaul. If this was also the mindset of Laurentius, Mellitus, and Justus,
Pope Gregory’s envoys to England and successors to Augustine of Can-
terbury, it is not surprising that, far from accepting the Romans’ view on
Easter reckoning, the Irish Bishop Dagan even refused to eat with them.6

Vatican, Biblioteca Apostolica, Vat. lat. 1548

The manuscript is described in detail in the catalogue of the Vatican
Library written by Bartolomeo Nogara.7 Lynn Thorndike cited the
manuscript and transcribed its rubrics, but his description must be ap-
proached with caution, as it is not immune from transcription errors.8

4
The tract entitled ‘Ein Bericht der päpstlichen Kanzlei an Papst Johannes I. von
526’ by its editor, Bruno Krusch (1926), 56–57, mirrors the progress of the Alexandrian
computus as proposed by Dionysius Exiguus through the papal chancery. It is defined
as res(criptum) (‘written reply’) in London, British Library, Cotton Caligula A XV,
77r, and as scriptio (written memorandum) in Oxford, Bodleian Library, Digby 63, 60r.
Both manuscripts attribute to the primicerius the memorandum (or written reply) to the
opinion (suggestio) addressed by Dionysius to the primicerius and the secundicerius of
the papal chancery, and to Bonus, the secundicerius, the official transcript of Dionysius’
opinion with the reply rendered (exemplum). The memorandum had been prepared for
Pope John I, and according to Roman protocol it was subject to his approval. We have no
records of any action taken by Pope John I, who was reigning at the time, but the reason
may be that in those same months he was exiled at Ravenna and died there.
5
In his Epistola 1.5 (ed. and trans. by Walker (1957), 8–9), Columbanus addresses
Pope Gregory in no uncertain terms: Tuum itaque aut excusa aut damna Victorium (‘Do
you then either exonerate or condemn your Victorius’).
6
Bede, Historia ecclesiastica 2.4 (ed. by Plummer (1896), i 88).
7
Nogara (1912), 59–60.
8
Specifically, Ratio spere dionisii de circulo magno paschae becomes Ratio spere Pet-
rosulii (i.e. Petosiris) de celo magno pulche (sic) in Thorndike (1954), 228. On the same


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

The section of the codex that concerns us here is the treatise on com-
putus attributed to Mainfredus (fols  51r–75v). Written in verses, it is
partially published in volume 94 of the Patrologia Latina among the
spurious works of Bede, without any indication of the manuscript(s)
used.9 C. W. Jones, who commented on this work, knew it only through
the edition of Hervagius.10 Even so, Jones had doubts on the attribu-
tion to Manfred of Magdeburg. He noted that in Hervagius’ edition the
poem is preceded by a computistical tract, also attributed to Manfred,
extant in manuscripts of the early ninth century. This attribution would
rule out the authorship of Manfred of Magdeburg, who lived in the elev-
enth century.11
The codex gives scant but valuable, albeit mildly contradictory,
biographical data in two distinct introductions to Mainfred’s work
(fols 65r–68r): according to the first one, Mainfred flourished at Padua,
according to the other, at Mantua. Since in the case of Mantua there
is a reference to lares, the meaning may be that Mainfred returned to
Mantua, his hometown (lares); in this case, Padua ortus would mean ‘ar-
rived at Padua’. Mainfred’s origin is in northern Italy. Both biographies
agree in reporting that he travelled to Francia for reasons of studying
and gathering substantial computistical material there. The biographer
further adds that his work was disputed, but proved irrefutable; having
thus asserted himself, Mainfred proceeded to re-write his treatise in vers-
es (which was originally composed in prose). This is the version extant
in Vatican, Biblioteca Apostolica, Vat. lat. 1548 (I believe still unpub-
lished), generously annotated with glosses.
No dates are given for Mainfred. His treatise is certainly post-Bedan,
as evidenced by a considerable part of the accessory material (various
works on computus, mostly excerpted) attached to the main work. The
excerpts are closely inspired by Bede, as evidenced by the titles in the ru-
brics listed by Thorndike. This fact might indicate a date of composition
in the early Carolingian age for the accessory material (no mention of
Alcuin or later authors); on the other hand, the combination treatise on
computus + accessory material and sources is typical of treatises on com-

page, in the list of rubrics, Ad invenienda cetera (?) should be ad invenienda concurrentia;
Cur renovetur vel anqui’ dicatur prima should be cur renovetur luna an quare dicatur
prima.
9
PL 94, 641–55.
10
Jones (1975), 79, 93: ‘There is no extant manuscript and our only source for the
verses is Hervagius’ text […] There is no known manuscript [of Manfredi carmina].’
11
Jones (1975), 79–80.


 LUCIANA CUPPO

putus drafted in Carolingian times. Mainfred’s stated mobility between

northern Italy and Francia is also understandable in an early Carolingian
context, after the defeat of the Lombard kings of Italy by Charlemagne
and with Charlemagne promoting computistical studies. Since the bi-
ography attests to the existence of two recensions of the work, one in
verse and the other in prose, it is entirely possible that the tract in prose
attributed by Hervagius to Mainfred is an excerpt from the larger work.
Nonetheless, until Mainfred’s identity is better established, his chronol-
ogy must remain hypothetical.
For our purposes, the interest of Mainfred’s treatise lies in the first
of the computistical tracts attached to his work, Ratio spere Dionisii de
circulo magno paschae, a work still unpublished, it seems, and an exam-
ple of attempted adjustment of the Dionysiac argumenta to the Victo-
rian computus.12 It originally accompanied a Victorian Easter calendar,
which, however, is not extant in Vatican, Bibliotheca Apostolica, Vat.
lat. 1548; text and English translation are in Appendix 1. The tracts ap-
pended to Mainfred’s computus are arranged in chronological order,
with Ratio spere Dionisii coming first and Bede taking the lion’s share
of excerpts.13 The suspicion that the treatise is pre-Bedan becomes cer-
tainty when we read that the authors of the tract had known Diony-
sius Exiguus and had translated some of his works from Greek to Latin.
Dionysius’ activity in Rome is well attested until AD 526; in later years
he seems to have joined Cassiodorus’ community at Vivarium, where
he was buried.14 Highly esteemed as a translator from Greek to Latin,
there is no record of Dionysius’ composing texts in Greek and translat-
ing them to Latin—unless the original nine Argumenta are meant; but
Dionysius never claimed authorship for these, which he credited to the

12
According to Cordoliani (1961), 175, there existed computations (argumenta)
based on the year of the Passion: ‘On trouve parfois aussi un quatrième procédé: le calcul
d’après les années de la Passion. Le cas est rare et les seuls éléments aussi calculés sont: les
années elles-mêmes et l’indiction.’ Unfortunately, Cordoliani did not give the specific
location of the manuscripts mentioned, and in any case the dates indicated by him do
not coincide with those in Vatican, Biblioteca Apostolica, Vat. lat. 1548.
13
A detailed index in Thorndike (1954), 228–29.
14
This may be inferred from Cassiodorus, Institutiones 1.23.2 (Mynors (1937), 62;
my translation): interveniat pro nobis qui nobiscum orare consueverat (‘May he intercede
for us, who used to pray with us’); and Institutiones 1.23.3 (Mynors (1937), 64; my trans-
lation): Sed ille [= Dionysius], saeculi perversitate derelicta, praestante Deo in ecclesiae
pace susceptus inter Dei famulos credendus est habere consortium (‘But we must believe
that, having left behind the perversity of the world and having been received, by God’s
grace, in the peace of the church, he is in the fellowship of the servants of God’).


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Egyptians.15 Moreover, Dionysius would never have endorsed a system

of Easter reckoning based on the year of the Passion.
Thus, we can confidently say that the tract was never authored by
Dionysius; this false claim, however, provides an approximate date of
composition, which cannot be too far from the mid-sixth century, since
the authors claim to have known Dionysius. The environment is Roman:
Dionysius taught there and the tract mentions the paschal candle with
the year of the Passion—a Roman usage. On the other hand, the tract
was known in Francia. Thus, while there is a definite Roman compo-
nent, its transmission is French and originated perhaps in Gallo-Roman
circles.
In his Institutiones, Cassiodorus lashed out at those ‘evil ones’ who
‘strive to introduce slanderously some opinions under his glorious name,
and thus seem to justify somewhat their own errors’.16 It is quite possi-
ble that he alluded to these people or to someone of similar sentiments.
The argumenta in Vatican, Biblioteca Apostolica, Vat. lat. 1548, though
patterned after the Dionysiac ones and indeed imitating them, are in de-
clared opposition to those of Dionysius Exiguus; thus, the passage in
Institutiones with the eulogy of Dionysius may well be, beyond a generic
attestation of praise, a vindication of the scholar against those who, like
the authors of the tract under discussion here, used his name in support
of their own misguided notions.
The Victorian persuasion of the author(s) is the most remarkable fea-
ture of this tract. It is evidenced in the strong support for an Easter calen-
dar with the related argumenta patterned after the original ones of Dio-
nysius for the year AD 525, but based on the year of the Passion. Such
a year was chosen out of theological rather than scientific criteria, but
the theology is of a highly questionable kind: the idea that the passion
of Christ should be more profitable to mankind than his incarnation is
plain silly and certainly does not reflect orthodox Catholic doctrine.17

15
The Argumenta are specifically attributed to the Egyptians in the reply of the
primicerius to Dionysius’ Suggestio (ed. by Krusch (1926), 57; my translation): De qua re
iuxta Egiptiorum argumenta veracia talis supputacionis probantur adhibuisse compendium
(‘And at this point [= the celebration of Easter] they [= the Nicene fathers] provided a
summary of such computation, according to the truthful argumenta of the Egyptians’).
16
Cassiodorus, Institutiones 1.23.3 (Mynors (1937), 63–64): Cuius nomine glorioso
aliqua pravi homines calumniose nituntur ingerere unde sua videantur errata aliquatenus
excusare. I follow the reading nomine glorioso rather than nomini glorioso proposed by
Mynors.
17
I am assured of this by such a distinguished Roman Catholic theologian as
Msgr. Brunero Gherardini, who, when consulted on this point, answered wrily: ‘Si vede


 LUCIANA CUPPO

Yet these ideas, which we might call crypto-Pelagians, were for a long
time a concern of orthodox Catholics. As Dáibhí Ó Cróinín pointed
out a few years back, a celebration of Easter that failed to recognize the
divinity of Christ manifested in his incarnation and resurrection, was
‘pre-empting the pasch’.18 This could happen by observing luna XIIII, but
also by attributing the redemption to the human sufferings of Christ on
Good Friday. An exclusive devotion to the suffering humanity of Christ
could be tainted with Pelagianism, because it neglected Christ’s divin-
ity, the true (formal) cause of redemption; and orthodox doctrine had
always maintained that it was the divinity of Christ that made redemp-
tion effective. In AD 640 Rome expressed concern for the celebration
of Easter in connexion with the Pelagian heresy, but Columbanus and
others had not waited that long to voice their opinion on the matter.19

Bede’s De temporum ratione and the Ratio spere Dionisii

The Ratio spere Dionisii is interesting in yet another respect, for it was
one of the sources exploited by Bede for his De temporum ratione. The
evidence is found in chapter 47 (De annis dominicae incarnationis) and
again in chapter  65 (De circulo magno Paschae).20 In chapter  47 the
sources noted by the editor in the apparatus include Dionysius Exiguus
and—with some variants—Felix of Squillace.21 The variants to Felix’s
text are not due to Bede’s initiative, but reflect verbatim the text of the

che le sciocchezze non son una prerogativa dei teologi contemporanei’ (‘Apparently non-
sense is not a prerogative of contemporary theologians’).
18
Ó Cróinín (1985), 516.
19
By saying that the French bishops, ‘this soporific sting of Dagon has drunk in
this erroneous tumor’ (haec soporans spina Dagonis hoc imbibit bubum erroris), Colum-
banus, in Epistola 1.4 (ed. and trans. Walker (1957), 8–9), implies that the bishops are
heretics. The allusion is to the bubonic plague that had hit the worshipers of Dagon in
their groins, and in early Christian literature the plague is a frequent image for heresy. An
annotation in the codex St Petersburg, Nacionál’naja Bibliotéka, Q.v.I.6, 1r, attributed
to Vivarian scribes and dated late sixth-early seventh century, warns the reader from the
Pelagian venom: Hic liber qui attitulatur Rufini non te seducat o pie lector, quia pelagi-
anus est et blasphemiis pelagianorum plenus: simulans enim contra Arrianos disputationes
venena suae haereseos inseruit (‘Pious reader, this book, which is ascribed to Rufinus,
must not lead you astray, for it is Pelagian and full of blasphemies of the Pelagians: feign-
ing disputations against the Arians, it has in fact injected the venom of his heresy’; my
translation).
20
Bede, De temporum ratione 47, 65 (CCSL 123, 265, 290).
21
The text of Felix of Squillace was published in Cuppo (2011), 130–33.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Ratio. Specifically, where Felix had stated that Dionysius was endowed
with precise knowledge (eliganti praeditus scientia), Bede writes non
ignobili praeditus scientia; where Felix has sanctissimus abbas, Bede has
venerabilis abba, and the text of Felix does not include the word scribens.
Thus, chapter 47 of De temporum ratione tells us a few things: first,
that the Ratio spere was not limited to local dissemination but reached
Northumbria, where it was used by Bede. Second, that the variants be-
tween the text of Ratio spere Dionisii and that of Felix are not coinci-
dental, but carefully thought out: Felix (who wrote in AD 616) had the
Ratio present to his mind when writing his preface, and used a rhetorical
device cherished by classical authors: He imitated the text before him,
but with significant variants that could substantially alter its meaning.22
Felix states that Dionysius was endowed with eliganti scientia. Eligans,
Professor Howlett pointed out to me, was a word used by Columbanus
in reference to Gildas and indicates accuracy as well as precision. Well
before Dionysius and Felix, precision had been the hallmark of the Al-
exandrian Easter reckoning: Eusebius had outlined the opposing stances
of the bishops at Nicaea, distinguishing those ‘who wanted to celebrate
with the Jews’ (i.e. without calculating a different date for Easter every
year) and those who advocated a precise calculation (akribe logon).23 Eli-
gans scientia is the daughter of the akribes logos supported by the bishops
of one of the conflicting parties.
Eligans scientia is the counterpart of allegory, where—by defini-
tion—one says one thing and means another. Allegory, not science, was
highly prized by Pope Gregory and his followers. By replacing the non
ignobilis of Vatican, Biblioteca Apostolica, Vat. lat. 1548 with eligans,
Felix was postulating a return to science and a departure from Gregory’s
notion of allegorical interpretation of the Bible, an interpretation that
dispensed with the obligation to find the exact day of the full moon. By
replacing venerabilis of Vat. lat. 1548 with sanctissimus, Felix was giving
Dionysius a higher status than the authors of the Ratio had done. It was
a posthumous rehabilitation of a sort, akin to that done by Cassiodorus
half a century earlier.

22
One example from Ovid, a master in this art: in Metamorphoses 1.179–81, Ju-
piter is made to shake his head three or four times, signifying assent. The model is, of
course, Homer; but while in the Iliad a nod of assent from Zeus is sufficient to make
Olympus tremble, the repetition of this gesture in Ovid’s version removes all dignity
from the father of the gods.
23
Eusebius, Vita Constantini 3.5.


 LUCIANA CUPPO

Chapter  65 of De temporum ratione (De circulo magno Paschae)

shows its derivation from the Ratio spere Dionisii in the lines pertaining
to the great Easter cycle:24

Circulus paschae magnus est qui, multiplicato per invicem solari ac lunari
cyclo, dxxxii conficitur annis. Sive enim decies novies viceni et octoni seu
vicies octies deni ac noveni multiplicentur, dxxxii numerum complent.

This definition of the 532-year cycle is taken verbatim from Ratio spere
Dionisii, which reads:

Scripsit ergo circulum magnum paschae qui multiplicato invicem solari ac

lunari circulo DXXXII perficit annos. Sive enim decies novies viceni et oc-
toni seu vicies octies deni ac noveni multiplicentur, quingentorum triginta
duorum annorum complent numerum.

The Ratio attributes these lines to Dionysius. Although it cannot be con-

firmed from other sources, this assertion gives reasonable assurance that
Bede was not the inventor of the 532-year cycle.

‘Following from afar the Roman Computus’

(Vatican, Biblioteca Apostolica, Reg. lat. 2077)
Like Dionysius and the bishops at Nicaea, Felix insisted on the necessity
of accurate and exact study (eligans scientia). But there were those who
did not see such a necessity, indeed took exception to Dionysius for hav-
ing suggested that scientific precision was needed; and, while it may ap-
pear odd to discuss at a conference centered on the science of computus
texts that disclaim the need for such science, the discussion is necessary
to understand the amount and type of opposition met in certain circles
by Felix and others who, like Columbanus from Luxeuil, were serious
about the scientific determination of Easter.
The basic disregard for scientific inquiry typical of the Ratio spere
Dionisii is also typical of another paschal tract: the Ratio pascae (‘The
Reckoning of Easter’) of Vatican, Biblioteca Apostolica, Reg. lat. 2077,
79r–81r.

24
Bede, De temporum ratione 65 (CCSL 123, 290; trans. by Wallis (1999), 155):
‘When the lunar and solar cycles are multiplied together, a Great Paschal Cycle is com-
pleted in 532 years. For whether you multiply 19 by 28, or 28 by 19, it makes 532.’


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Characterized in Codices Latini Antiquiores as ‘probably? Italian’ and

dated ‘early saec. VII’, the manuscript was convincingly attributed to
Vivarium by Fabio Troncarelli.25 Two details not noted by that scholar
contribute to strengthen the attribution. One is the date of completion
of the codex, which can be inferred from an EXPL(icit) at the end of
the Easter calendar that originally ended with the year DLVIIII of the
Passion, which is the equivalent of AD 585 (fol. 98v);26 the second one
is given by the letters +IOEPSVC on folio 1r. Transcribed erroneously
by Bethmann as +OPFIOVC in the MGH edition of this text, these let-
ters, penned in a hand different from those of the two copyists, should
be read +IO(annes) EP(iscopu)S V(ir) C(larissimus). Ioannes episcopus is
bishop John of Squillace, known to us from the correspondence of Pope
Gregory I.27
The codex is structured in three units. The first one (fols 1r–78r) com-
prises De viris illustribus of Jerome / Gennadius. The second (fols 78r–
100v) is a historical unit that joins to the chronicle of Prosper two Easter
calendars; the second calendar, with numerous historical annotations, is
known as Paschale Campanum. The third unit (fol. 101r) consisted of
De haeresibus of Saint Augustine, but only the index page is extant.
For our purposes the pertinent unit is the second one (on history
and Easter reckoning). The computation of Easter is considered part of
human history, which began at creation and is a facet of the divine plan
for mankind—a plan centered upon redemption. Thus, human history
became sacred history: within such a framework the anonymous author
of this codex showed little concern for the calculation of Easter based
on the lunar phases, but aimed at relating the celebration of Easter to

25
CLA 1, 114–15; Troncarelli (1988). The attribution to Vivarium is based on
palaeographical grounds (same hands as in other known Vivarian manuscripts) and
contents and organization of the codex, both strongly, albeit posthumously, influenced
by Cassiodorus. An inventory of the contents was published by Theodor Mommsen in
MGH Auct. ant. 9, 372.
26
The EXPL is visible in the illustration of fol. 98v in Troncarelli (1998), plate 31,
next to the year DLVIIII (AD 585) that was reported on two consecutive lines. This was
not an error: DLVIIII on the first line referred to the Easter date, and on the second line
indicated the end of the computus, marked by an EXPL(icit). When someone decided
to prolong the Easter calendar tracing additional years in the bottom margin, DLVIIII
on the second line was corrected to make it a DLX, but the EXPL remained visible.
27
Gregory I, Registrum epistolarum 2.37 and 8.34 (ed. by Paul Ewald and Ludwig
Hartmann in MGH Epp. 1, 132–33; 2, 36–37). Troncarelli (2014) confirmed the iden-
tification of Ioannes Episcopus with Bishop John of Squillace, but read VC as VICARIVS.
The reading by Bethmann is reported by Mommsen in MGH Auct. ant. 9, 371.


 LUCIANA CUPPO

the divine plan that began unfolding at creation and would come to full
fruition with the second coming of Christ.28
Accordingly, the Ratio pascae opens with a pointed allusion to Dio-
nysius Exiguus, who had affirmed the necessity of scientific inquiry to
determine the first day of the first month, the basis for the calculation of
Easter. Dionysius had written:29

Tanta hac auctoritate divina claruit, primo mense, decimo quarto die ad
vesperum usque ad vigesimum primum festivitatem paschalem debere
celebrari. Sed quia mensis hic unde sumat exordium vel ubi terminetur
evidenter ibi non legitur, praefati trecenti et octodecim pontifices, antiqui
moris observantiam et exinde a sancto Moyse traditam, sicut in septimo
libro Ecclesiasticae fertur Historiae, solertius investigantes, ab octavo idus
Martii usque in diem nonarum Aprilium natam lunam facere dixerunt
primi mensis exordium.

‘It was clear by such divine authority that the celebration of the Pasch
should take place in the first month, in the evening of the fourteenth day
until the twenty-first day. But, since it is not clearly stated there [in Exo-
dus 12:2] where this month begins and when it ends, the aforesaid 318
bishops, in observance of the ancient custom (an observance handed
down by the venerable Moses from himself on, as stated in the seventh
book of Ecclesiastical History), inquiring most carefully, said that the
new moon between the eighth day before the ides of March and the day
of the nones of April marked the beginning of the first month.’

For the anonymous author of Vatican, Biblioteca Apostolica, Reg. lat.

2077 such inquiry was not necessary, because the first day of the first
month is—obviously!—the day when God first created the world in the
first month of creation:30

Dicente Domino Moysen: ‘mensis hic initium mensus primus erit vobis
in mensibus anni’, qui cum mundum statuerit per significantiam rerum
perfectarum, apert nobis declaravit, quando dixit: ‘producat terra herbam

28
The tract De duobus testibus (fol. 78r–v), on Enoch and Elias returning before
the second coming of Christ, is placed immediately before the introduction to the cal-
culation of Easter (fol. 79r). Both are followed by the chronicle of Prosper. I am prepar-
ing a new edition and English translation of De duobus testibus for the Library of Early
Christianity of the Catholic University of America Press.
29
Dionysius Exiguus, Epistola ad Proterium (Prologus; PL 67, 21; my translation).
30
Ratio paschae (Vatican, Biblioteca Apostolica, Vat. lat. 2077, 79r; the text is
ed. by Theodor Mommsen in MGH Auct. ant. 9, 740; my translation).


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

pabulum habentem semen secundum suum genus et lignum fructiferum

faciens fructum, cuius semen in ipso in similitudinem secundum genus, et
factum est sic’. Et in die, in qua agnus occisus est inmaculatus, Christus
Iesus Dominus noster, produxit terra herbam pabuli ovibus Christi ad re-
fectionem animarum in ipsum credentium. Quae ergo nisi corpus Christi
quod gestavit quod edent qui ovium innocentiam habuerint? Et lignum
fructiferum haben fructum quod lignum nisi crux Christi? Quae portavit
hominem fructum cuius semen, hoc est caro mortua vel sepulta incorrupta,
surrexit in eo semine vel in ea similitudine in qua ante fuit.

‘By saying to Moses, ‘This month shall be for you the beginning of the
months, the first among the months of the year’, God made this begin-
ning quite clear to us when he created the world, through the allegori-
cal meaning of the things created, by saying: ‘Let the earth bear pasture
with seed according to their kind, and a tree bearing fruit, whose seed
will be in it according to its kind, and so it was.’ And the day when the
immaculate Lamb, Jesus Christ our Lord, was killed, the earth produced
pasture for Christ’s sheep as nourishment for the souls who believe in
him. And what is this, if not the body of Christ that the earth bore, that
shall be eaten by those who have the innocence of the sheep? And the
fruitful tree bearing fruit, what tree is it, if not the cross of Christ that
bore as its fruit a man whose seed, that is his flesh, dead and buried in-
corrupted, rose again in that seed and appearance it had before?’

These opening words of the Ratio are an arenga, the statement of gen-
eral principles that prefaces the specific instructions to be imparted. The
pointed allusion to Dionysius Exiguus is in fact a refutation of his words
and an assertion of the principle of allegorical interpretation of the Holy
Writ. Allegory taken as a guiding light was highly prized in certain Ro-
man circles (suffice to think of the success of the Historia Apostolica of
Arator), but hardly a contribution to the science of computus.
Indeed, in this explanation of the reckoning of Easter the science of
computus is conspicuous for its absence. To be sure, the days of Easter
are duly noted, but the computist of Vatican, Biblioteca Apostolica, Reg.
lat. 2077 simply reproduces the old Roman Easter tables and the rules
for determining the time of Easter excerpted from the Supputatio Ro-
mana.31 The Ratio declares allegiance to the ancient Roman system, the
84-year cycle used in Rome prior to that of Victorius of Aquitaine. The
partial transcription of the Supputatio and the full transcription of the
84-year cycle from AD  354 to 437 (entirely past, and therefore of no

Published by Krusch (1880), 227–35, as Cölner Prolog.

31


 LUCIANA CUPPO

practical value to readers of this codex) witness to the attachment of this

author to a computus in the Roman tradition, an attachment reaffirmed
through the statement that the ordering of the computus in this manu-
script follows ‘at a distance’ (eminus) the old Roman one.32 This stated
attachment to the Roman tradition is remarkable in an environment
that in Cassiodorus’ time had been a gateway to the Byzantine East, but
was in keeping with the sentiments of Pope Gregory, a strong supporter
of what he considered Roman culture, as opposed to that of the Greeks.
From a scientific perspective the principles underlying the Ratio of
Vatican, Biblioteca Apostolica, Reg. lat. 2077 can be dismissed as irrel-
evant, and this was in fact the attitude of Columbanus when he wrote
that the Victorian computus was a matter of laughter with the Irish.33
From a historical perspective, however, those views cannot be ignored,
for they had high-placed supporters. This was true in Rome, witness
Arator and Eugippius, and was true in Francia, if I understand correctly
the sense of Columbanus’ remarks to the effect that all the Gallic bish-
ops could say in support of their computus was: ‘We must not celebrate
Easter with the Jews.’34 The prohibition to celebrate with the Jews is em-
phasized in the Acts of the Council of Caesarea, a piece close in inspi-
ration to Vatican, Biblioteca Apostolica, Reg. lat. 2077;35 and the same
letter of Columbanus includes, if I read it correctly, a sarcastic allusion
to the Gallic bishops as proponents of the Victorian system. Perhaps,
Columbanus wrote, Gregory did not take a stance on the question of

32
Ratio paschae (Vatican, Biblioteca Apostolica, Reg. lat. 2077, 79r; MGH Auct.
ant. 9, 740; my translation): Ideo in laterculum sive in volumen qua eminus iuxta con-
putem romanorum eas ordinatum (‘thus, in the calendar or roll to be followed, ordered
remotely according to the Roman computus’). In the manuscript the words are not sepa-
rated, and so it happened that Mommsen transcribed QUAEMINUS as quae minus,
which makes no sense.
33
Columbanus, Epistola 1.4 (Walker (1957), 6–7): Scias namque nostris magistris
et hibernicis antiquis philosophis et sapientissimis componendi calculi computariis Victori-
um non fuisse receptum, sed magis risu vel venia dignum quam auctoritate. (‘For you must
know that Victorius has not been accepted by our teachers, by the ancient Irish scholars,
and by the mathematicians most skilled in reckoning chronology, but has earned ridicule
or indulgence rather than authority.’)
34
Columbanus, Epistola 1.4 (Walker (1957), 6–7): Quia non mihi satisfacit post
tantos quos legi auctores una istorum sententia episcoporum dicentium tantum: Cum iu-
daeis facere Pascha non debemus. (‘For I am not satisfied, after reading such weighty au-
thorities, with the single judgement of those bishops who can only say: We must not
celebrate Easter with the Jews.’)
35
Acta synodi Caesarea 1 (Krusch (1880), 307; my translation): Quid nobis est a
xiiii. luna compotum cum iudaeis facere Pascha? (‘Why should we, reckoning from moon
14, celebrate Easter with the Jews?’)


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Easter because it was more appropriate for someone high-born (hones-

tius) not to incur the displeasure of the ‘Armorican novelty’.36 In eccle-
siastical language ‘novelty’ means ‘heresy’ and Armorica—if this emen-
dation from ‘Hermagorica’ is accepted—is another name for Brittany,
the home of Pelagius. In other words, what kept Gregory from speaking
out against the Victorian system was his allegiance to those of his own
senatorial kind: the Gallic bishops, heretics and Pelagians in the eyes of
Columbanus.
Fragments or traces of the historical-computistical section of Vatican,
Biblioteca Apostolica, Reg. lat. 2077 not known from any other sources
were handed down in later manuscripts. There is as yet no study on the
textual transmission of this codex, but I have run across a few manu-
scripts or texts partly derived from that codex and this is my modest list:

(1) Paris, Bibliothèque nationale de France, Lat. 4871: a tract on the

life of Christ added to the chronicle of Prosper in Vatican, Bibli-
oteca Apostolica, Reg. lat. 2077 was copied verbatim in the ab-
breviated chronicle of Prosper here, fol. 110r–v.37
(2) Wolfenbüttel, Herzog August Bibliothek, Cod.  Guelf. 532
Helmst. (579), 85r–v, has the preface to the Easter calendar from
AD 464 to 560 (Paschale Campanum).
(3) Bede, Explanatio Apocalypseos 17.132–49. In his commentary
Roger Gryson gives no sources for lines 132–34 in Bede’s text,38
but the similarities with the tract De duobus testibus of Vatican,
Biblioteca Apostolica, Reg. lat. 2077 are remarkable. The perti-
nent points are in bold letters.
Bede: Quidam duos prophetas Enoch et Heliam interpretantur,
qui tribus semis annis praedicantes contra mox secuturam anti-
christi perfidiam fidelium corda confirment.39

36
Columbanus, Epistola 1.4 (Walker (1957), 4–5; translation altered): Aliter ta-
men et honestius tua excusari potest peritia: dum forte notam subire times Hermagoricae
novitatis, antecessorum et maxime papae Leonis auctoritate contentus es. (‘Yet your states-
manship can be excused in a different way more appropriate for someone high-born:
while you perhaps fear to incur the displeasure of the Armorican novelty, you are content
with your predecessors’ authority, and especially that of Pope Leo.’)
37
Paris, Bibliothèque nationale de France, Lat. 4871 is available online at http://
gallica.bnf.fr/ark:/12148/btv1b8552412w.
38
CCSL 121A, 378–81.
39
Bede, Explanatio Apocalypseos 17 (ed. by Roger Gryson in CCSL 121A, 379).


 LUCIANA CUPPO

Reg. 2077: Dicitur venire Enoch et Helia praedicaturi adventum

Domini et diem iudicii mensibus XLII […] ut tribus et semis annis
per eundem Heliam Dei notitia confirmetur.40
(4) In his commentary to the Apocalypse Bede cited a vetus disticon
from an unknown source:41 Respice distinctis quadratum partibus
orbem, ut signum fidei cuncta tenere probes (‘Behold the world
divided in four distinct parts, that you may understand that the
cross holds all things together’). The allusion to the cross (signum
fidei) that holds together all things in a world divided in four
parts fits perfectly the cross at the centre of the wind-diagram in
this codex (see the following item).
(5) The wind-diagram in Vatican, Biblioteca Apostolica, Reg. lat.
2077, 99r, influenced the wind-diagram in Cologne, Dombiblio-
thek, 83-II, 141r.42
(6) The life of Pope Vitalian in the Roman Liber pontificalis seems
influenced by the imagery of the Paschale Campanum of Vatican,
Biblioteca Apostolica, Reg. lat. 2077, a point I discussed a few
years back.43

It seems, then, reasonable to assume as a working hypothesis that a twin

of Vatican, Biblioteca Apostolica, Reg. lat. 2077 or its copies found their
way to France and Germany, probably along the Via Francigena, but also
to England with Northumbria. And it makes sense to suggest that Pope
Gregory’s men—his missionaries to England, who, as we know from his
letters and from Bede, also spent time in Francia—were instrumental in
the dissemination of this text, eminently suited as a textbook.44 If this is
true, the dissemination of the Roman-Victorian computus came from a
40
De Enoc et Helia (Vatican, Biblioteca Apostolica, Reg. lat. 2077, 78r–v; ed. by
Theodor Mommsen in MGH Auct. ant. 9, 493).
41
Bede, Explanatio Apocalypseos 9 (CCSL 121A, 311).
42
More on this at the 2012 Conference. The Cologne codex is readily available
online at www.ceec.uni-koeln.de.
43
Cuppo (1998), 132–34.
44
In AD 601, Laurentius and Mellitus were sent by Pope Gregory to join Augustine
in England. Gregory wrote several letters to recommend them to various Gallic bishops
during their travel, making it very clear that material support (solacium) was expected
from the bishops. Epistola 11.41 (MGH Epp. 2, 315), written in June AD 601, lists among
the recipients of Gregory’s letter Menna Telonensis, Serenus Massiliensis, Lupus Caballon-
ensis, Agilfus Mettensis, Simplicius Parisiensis, Melantius Rotumensis, Licinius Andegaven-
sis, mapping out an itinerary that stretches from Marseille to Metz to Paris. In addition, in
Epistola 11.48 (MGH Epp. 2, 321), dated 22 June AD 601, Gregory recommends Lauren-
tius and Mellitus to Brunichilde, and in Epistola 11.51 (MGH Epp. 2, 323) to Chlotarius.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

milieu close to Pope Gregory I and to the Gallic bishops. If the mindset
that transpires from the pages of Vatican, Biblioteca Apostolica, Reg. lat.
2077 was also that of Gregory and his men, and if the ideas on Easter
reckoning expressed in that codex reflect those of the Roman experts on
Easter, it is not difficult to see why the Irish monks would dissent and
prefer to take their meals alone.

The Council of Caesarea

The paschal tract known as Acts of the Council of Caesarea expresses a
perspective on Easter that is akin to that of Vatican, Biblioteca Apostol-
ica, Reg. lat. 2077 inasmuch as it dispenses with any scientific investiga-
tion of the first day of the first month, defined (as in the Vatican codex)
as the first day of creation and thought to have occurred on March 25,
the day of the spring equinox.45 This simplistic view disposes of all dis-
putes about the exact occurrence of the first day of the first month in the
Latin calendar. In its most ancient A-version the Council of Caesarea is
akin to the Ratio spere Dionisii of Vatican, Biblioteca Apostolica, Vat. lat.
1548 in its emphasis on the Passion of Christ.
Too hastily considered a pastiche of medieval superstition, this
tract—in its uncompromising refusal of scientific evidence and exclusive
use of the Bible and of Roman tradition to set the proper time for Easter
celebrations—is in reality a deliberate and sophisticated rejection of any
striving for the scientific determination of the time of Easter.46 The A-re-
cension of Bern, Burgerbibliothek, 645, 72r–74v, and Vatican, Bibliote-
ca Apostolica, Reg. lat. 2077, 79r–v both proclaim: ‘We have the Bible,
and that’s enough’. In about those same years the future Pope Gregory I
was proclaiming to the world that the Bible cannot be constrained by
the rules of Donatus. Gregory aimed specifically at Cassiodorus, Reg.
lat. 2077 aimed at Dionysius: for questions of grammar and rhetoric,
as for questions of science, the Bible was enough.47 Felix of Squillace,

45
The Acta synodi Caesareae are ed. by Krusch (1880), 303–10. He distinguishes
three recensions (A, B, and C). Of these, only the first one (recension A) as found in
Bern, Burgerbibliothek, 645 uses the Victorian terms for Easter.
46
This was not always understood. André Wilmart (1933), 19, who published a
recension D from Vatican, Biblioteca Apostolica, Reg. lat. 39, referred to this piece as
‘littérature bizarre’ whose ‘barbarie n’en est pas moins éclatante’.
47
Pope Gregory, Epistola ad Leandrum 5, prefatory to the Moralia (ed. by Mar-
cus Adriaen in CCSL 143, 7). Gregory proclaimed that he disdained to observe situs


 LUCIANA CUPPO

who followed in the footsteps of both Dionysius and Cassiodorus, vin-

dicated—as they had done—the role of human reason.
Considered authentic by Bede, the Acts were deemed a fabrication
by their editor Bruno Krusch. Based on the fact that they were known
to Bede and dealt with Easter reckoning, Krusch proposed that the Acts
originated in the British Isles.48 But Bede had access to continental as
well as insular material, and the Easter controversy was by no means lim-
ited to the British Isles. The Acts were known to Cummian and Colum-
banus. In their A-recension they are close in inspiration and in some
details—specifically the language used in the translation of Genesis 1:11
and the references to the Passion of Christ, all eliminated or modified
in later recensions—to the texts in Vatican, Biblioteca Apostolica, Reg.
lat. 2077 and Vat. lat. 1548; this suggests that they were close in time
to those works. The manuscript tradition of recension A is distinctly
French—Bern, Burgerbibliothek, 645, the oldest extant manuscript, is
from the area of Lyon—and the tract reveals a very good knowledge of
church history in the light of Roman tradition.
The A-recension as handed down in Bern, Burgerbibliothek, 645
shows implicit opposition to the Dionysiac system. Its Victorian lunar
limits for Easter Sunday (moon 16–22) and its strong emphasis on the
Passion of Christ are both traits incompatible with the Dionysiac com-
putus, and are not present in other recensions.49 The Bern, Burgerbiblio-
thek, 645 version of the Acts uses a Vetus Latina translation of Genesis
1:11 close to that of Vatican, Biblioteca Apostolica, Reg. lat. 2077, of
the Supputatio Romana, and of Augustine (De Genesi ad litteram). The
reading lignum fructiferum faciens fructum is common to all these works,
while later versions of the Acts read lignum pomiferum ferentem fructum.
The various readings are:

(a) Bern, Burgerbibliothek, 645 (Wilmart (1933), 23 ll.  99–101):

Germinet terra herba foeni secundum genus et lignum fructiferum
faciens in se fructus.

motusque (‘location and motion’), the very same words used by Cassiodorus while cau-
tioning his monks against the confusion of ablative and accusative endings (Institutiones
1.15.9; Mynors (1937), 46).
48
Krusch (1880), 303–04.
49
The lunar limits of 16–22 are in Wilmart’s transcription of the Acts from Bern,
Burgerbibliothek, 645 (Wilmart (1933), 27).


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

(b) Augustine, De Genesi ad litteram 2.12 (ed.  by Joseph Zycha in

CSEL 28.1, 50): Germinet terra herbam pabuli ferentem semen
secundum genus […] et lignum fructiferum faciens fructum.
(c) Vatican, Biblioteca Apostolica, Reg. lat. 2077, 79r: Producat terra
herbam pabulum habentem semen secundum suum genus et lig-
num fructiferum faciens fructum cuius semen in ipso.
(d) Supputatio Romana:50 Germinet terra herbam pabuli ferentem se-
men saecundum genus et secundum similitudinem, et lignum fruc-
tiferum faciens fructum cuius semen ipsius in eo.
(e) Vulgate: Germinet terra herbam virentem et facientem semen et lig-
num pomiferum faciens fructum iuxta genus suum, cuius semen in
semetipso sit super terram.

In their A-recension, the Acts share with Vatican, Biblioteca Apostolica,

Reg. lat. 2077 and Vat. lat. 1548 the emphasis on the Passion of Christ
as the cornerstone of Easter celebrations. Once more, the remarks of
Dáibhi Ó Cróinín are pertinent: an over-emphasis on the Passion pre-
empts both Incarnation and Resurrection, hence it is heretical and a pos-
sible symptom of crypto-Pelagianism (exclusive stress on the suffering
humanity of Christ to the detriment of his divinity). Nor is it coinci-
dental that Columbanus had thundered against heresy and deprecated
Pope Gregory’s lack of action in the same geographical area where Bern,
Burgerbibliothek, 645 was disseminated. The following passages, indica-
tive of such emphasis on the Passion, were eliminated or modified in the
B and C recensions:

Quia CXVII psalmus totus de passione cantatur (‘because Psalm 117 is

all sung about the passion’),51 modified to totus de passione et de resur-
rectione cantatur (B-recension).
Haec de passione confirmat (‘this confirms the passion’)52 is omitted in
the subsequent recensions.
Ne passio omittatur (‘in order not to omit the passion’; from the paschal
terms, that is)53 becomes in the B-recension ne tantum sacramentum fo-
ras limitem excludatur (‘lest such a great mystery be excluded outside
the term’).

50
Prologus paschae 4 (Krusch (1880), 231).
51
Wilmart (1933), 25.
52
Wilmart (1933), 25.
53
Wilmart (1933), 26.


 LUCIANA CUPPO

Thus, on the one hand Bern, Burgerbibliothek, 645 shows a bent of the
theological mind akin to that branded as heresy by Columbanus and,
as the same Columbanus had not failed to note, totally ludicrous from
a scientific point of view. On the other hand, while the treatise seems
of French-Burgundian origin, its parallels with Vatican, Biblioteca
Apostolica, Reg. lat. 2077 and Vat. lat. 1548 suggest an environment
familiar with Roman tradition. We might call the recension of Bern,
Burgerbibliothek, 645 a Gallo-Roman piece, which was being trans-
mitted in the same environment where Columbanus of Luxeuil had
appealed to the pope against French heresy and the Victorian Easter
reckoning.
That piece was a far cry from the Dionysiac system; thus, chances
are that where the Acts in their A-recension as transmitted in Bern,
Burgerbibliothek, 645 found favour, Felix did not; and in late sixth- and
seventh-century Gaul the scale leaned heavily in favour of the Victorian
system and against Felix.
We know of two continental sources that show knowledge of Felix’s
cycles in Gaul. One is the prefatory letter, written in AD 721, to a trea-
tise on computus preserved in Vatican, Biblioteca Apostolica, Reg. lat.
123.54 The letter was written at the time of expiration of Felix’s cycles,
which apparently were known to the author when they were still used,
between AD 626 and 721.55 The second continental source is the pre-
sumably French exemplar of Milan, Biblioteca Ambrosiana, H 150 inf.,
50r–v: the author of the Liber de computo noted that Felix’s cycles were
omitted as being obsolete, thus showing that they were present in his ex-
emplar. Apart from these instances, Felix does not surface in Francia or
nearby areas until the Carolingian age, with the manuscripts Cologne,
Dombibliothek, 83-II and Reims, Bibliothèque municipale, 298 (now
destroyed).
The Acts in their A-recension testify to the doctrinal and pseudo-
scientific tendencies that had elicited the wrath of Columbanus and
eventually found an answer in the work of Felix of Squillace.

54
This prefatory letter is ed. under the abbreviated title Prol. Aquit. in Borst
(2006), 329–47.
55
Prol. Aquit. 2 (ed.  by Borst (2006), 339–40). For an extensive discussion see
Borst (2006), 300–02.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Felix in the insular world

Though meager, evidence for the knowledge of Felix in the insular world
in the seventh and eighth centuries does exist. It can be inferred from
the well-known words of Ceolfrid in his letter to Nechtan, king of the
Picts, to the effect that many in the insular world were able to continue
the Dionysiac cycles that were soon to expire.56 This assertion can be
literally true if the computists had at their disposal at least two consecu-
tive 95-year cycles to be taken as guidelines for the construction of sub-
sequent ones. In Ceolfrid’s time these two cycles could only be those of
Dionysius (AD 532–626) and Felix (AD 627–721), whose name does
not appear to have been known in the insular world.57 But two texts
that are unequivocally of Felix, written respectively as prologue and
preface to his cycles, appear in Oxford, Bodleian Library, Dibgy 63, 67r
and 70v. Thus, the question is: when did his texts come to the insular
world? The possibilities range from the ninth century, or more precisely
from AD 867, the annus praesens of the compilation of the codex, to the
early seventh century, close to the date of composition of the cycles in
AD 616.58 I hasten to say that I have no final answer on the matter, and
not even, I fear, irrefutable evidence for a firm date. But there are clues
that Felix may have come to the insular world rather early, and they shall
be discussed here as a starting point for reflection.
From the time of its donation to the Bodleian Library in 1634, co-
dex Digby 63 evoked contrasting reactions from different scholars and
has seen the hurling of beams and motes, sticks and stones by some very
distinguished scholars.59 On my part, I confess to love at first sight when,
after studying the manuscript on microfilm, I was allowed to view it in
the Duke Humfrey Room, and I can only hope that some of that glow
radiates on scholars of computus wherever they may be.
While they may appear heterogeneous to a casual observer, the con-
tents of Oxford, Bodleian Library, Digby 63 are clearly—if not regu-
larly—articulated and follow the customary practice of Carolingian

56
Bede, Historia ecclesiastica 5.21 (Plummer (1896), i 341).
57
It cannot be excluded that Cummian’s tenth computus refers to the cycles of
Felix, but without further study this must remain a conjecture; see Cummian, De contro-
versia paschali ll. 215–23 (ed. and trans. by Walsh and Ó Cróinín (1988), 86–87).
58
The year of composition of Oxford, Bodleian Library, Digby 63 was established
conclusively by Ó Cróinín (1983a), 258: ‘The only valid dating for Digby 63, on the
other hand, is the annus praesens of 867 in a formula on fol. 20v.’
59
Faithfully reported in Dumville (1983); Ó Cróinín (1983a).


 LUCIANA CUPPO

and post-Carolingian treatises: an initial miscellaneous composition on

computistical matters is followed by pertinent documents on computus.
The initial treatise (fols 9r–48v) is preceded by Dionysiac tables from
AD 513 to 892 (fols 1r–8v). While the final year of the Easter calen-
dar, drawn in advance, shows the expected life-span of the codex, a date
inserted as annus praesens in one of the Dionysiac formulae (Argumen-
ta) on folio 20v gives the year of composition of the text in its present
form: AD 867. The pieces that make up the treatise vary in age, contents,
origin, and provenance, and an edition that identifies the sources and
indicates the presence of similar pieces in other codices is still a deside­
ratum. The best known among the various components of this compu-
tistical anthology is the French calendar published by Francis Wormald
(fols 40r–45v),60 with a collectio temporum ending in AD 814, the year
of death of Charlemagne. This calendar and three other tables immedi-
ately following (fols 40r–48v), close to those found in manuscripts from
Fleury, mark the end of the computistical anthology compiled in its pre-
sent form in AD 867.
The ensuing section (fols 49r–70v) includes the Acts of the Council
of Caesarea in their B-recension and the Dionysiac dossier with letters of
Proterius, Paschasinus, Cyril, and Dionysius; this section ends with the
texts of Felix. Before being incorporated in Oxford, Bodleian Library,
Digby 63, the pages with the prologue and preface of Felix were the last
ones of a book that was the exemplar and may have reproduced the arche-
type of the Dionysiac dossier in the manuscript. This is shown by an ex-
plicit that marks the end of the Dionysiac dossier, followed by a colophon
with the name of the scribes who had copied this manuscript’s exemplar.
This proves that before being copied in Oxford, Bodleian Library, Digby
63 and in its exemplar(s), the works of Dionysius and Felix formed a
separate unit and were eventually attached to the treatise on computus.
The third and last section of the codex (fols 72v–87r, fol. 88r–v is
blank) has the Dionysiac Argumenta, variously expanded; it also has
another annus praesens, AD  675. The codex ends with the Disputatio
Morini and the tract known as the Paschal Prologue to Vitalis of Roman
origin and also found in Cologne, Dombibliothek, 83-II, 193v–197r;
but the recension in Oxford, Bodleian Library, Digby 63 is notably dif-
ferent. Additional smaller pieces are inserted among the main works of
the Dionysiac dossier.

60
Wormald (1934), 1–13. It has the siglum c12 in Arno Borst’s edition of the
Reichs­kalender (Borst (1998), 161–62).


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

The sheer presence of the Dionysiac dossier in this manuscript would

not say much about the chronology and attribution of the pieces there-
in. This is so because some of these pieces—particularly the letter of
Dionysius to Petronius—occur frequently and in various arrangements
in computistical works; and without a critical edition of the whole dos-
sier (yet another desideratum of scholarship) it is not possible to see the
connections among various families of manuscripts of the same work.61
It is, however, possible to discern some distinctive traits of the Digby
recension of the dossier that show it to be closer to the archetype than
other copies of the same texts spread in manuscripts throughout Europe.
They are: (a) the arrangement of the texts in logical-chronological order,
(b) the excerpt from the letter of Pope Vitalian, (c) archaic (late antique)
traits, both textual and palaeographical, presumably derived from the
archetype.

Arrangement in logical-chronological order

The Dionysiac dossier in Oxford, Bodleian Library, Digby 63 is out-
standing among other codices with the same works both because of its
completeness and because its pieces are arranged according to logical
chronological criteria. The dossier begins with the letter of Pascasinus
to Pope Leo I and ends with the Preface of Felix, written in AD 616; the
Acts of Caesarea, believed to have been composed in AD 196/98, are
prefixed to the Dionysiac material. In their A-recension the Acts, Victo-
rian in inspiration, were transmitted with the letter of Victorius of Aqui-
taine, but in Oxford, Bodleian Library, Digby 63 there are no writings
of Victorius and the Acts appear in their B-recension, immune from the
Victorian and possibly crypto-Pelagian emphasis on the Passion. Thus
removed from its Victorian habitat, this text, deemed authentic, could
be prefixed to those authored or cited by Dionysius and his successor.
The letter of Proterius to Pope Leo I (AD 455) follows, both logically
and chronologically, that of Pascasinus (AD 444). The next unit docu-
ments the steps taken by the Roman Church in favour of the Alexandri-
an computus: the recommendation (suggestio) made by the primicerius
of the Roman chancery and transcribed by the secundicerius Bonus, both
dated AD 525, and the excerpt from the letter of Pope Vitalian (reigned
AD  657–72). Thus, the dispositions of the Roman curia are gathered
together and placed before the instructional pieces on the Alexandrian

61
Krusch (1938), 60, used six manuscripts for his edition from French or Ger-
manic areas. He adduced no witnesses from Hiberic or Italian areas.


 LUCIANA CUPPO

computus (letter of Cyril of Alexandria, both letters of Dionysius and

both texts of Felix) that complete the dossier, originally a separate book
that ended at folio 71 with the explicit FINIT PRAESTANTE.62
Among the six manuscripts examined by Mommsen and listed by
Krusch,63 Oxford, Bodleian Library, Digby 63 and London, British Li-
brary, Cotton Caligula A XV reflect accurately the procedures of the
Roman chancery and the promotion (a meager one, all things consid-
ered) of the Dionysiac computus by the Roman church:

Oxford, Bodleian Library, Digby 63, 60r, reads: exemplum boni. scrip-
tio primiceri notariorum (‘official transcript [exemplum is the exemplary
copy to which all other copies should conform] of Bonus [Boni]. Writ-
ten memorial (scriptio) of the primicerius of the notaries’).

London, British Library, Cotton Caligula A XV, 77r, reads: Exem-

plum sugestionis boni. Res primiceri notariorum ad Johannem (‘official
transcript of the petition by Bonus. Reply [Res is the abbreviation of
res(criptum)] of the primicerius of the notaries’).

Both codices record the normal procedures of the Roman chancery:

from the initial opinion rendered by Dionysius called in as an expert
(Letter of Dionysius to Bonifatius and Bonus) to the written memorial
(scriptio) of the primicerius (not named) for submission of the file to the
pope (apostolatui vestro). Instead of ‘written memorial’, London, British
Library, Cotton Caligula A XV has ‘reply’ of the primicerius of the no-
taries to Pope John. There is no record of any action taken on it by Pope
John  I, but at about the time when the primicerius gave his response
prior to Easter AD 526, John was jailed at Ravenna and died there on
18 May. Thus, there was no official papal endorsement and there would
be none until the reign of Pope Vitalian, almost 150 years later.
The familiarity with Roman chancery procedures shown in Oxford,
Bodleian Library, Digby 63 is proof of close derivation from a Roman
archetype. But, while the Letter of Dionysius to Bonifatius and Bonus
was disseminated far and wide, the endorsement by the primicerius with
its summary of the main points of Dionysius’ letter never made it into
the edition of Wilhelm Jan or into the Patrologia Latina. We owe its

62
That this section was originally the last one in a separate book is confirmed by
the colophon of the scribe Regenbold, who added his own explicit to that already present
in the Dionysiac dossier.
63
Krusch (1926), 55–56.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

publication to Bruno Krusch, yet his meritorious edition is not without

some misunderstanding: Krusch assumed that the communication by
the primicerius to the Apostolic See was an independent document and
that it broke off at the end, and printed it as such.64 Neither assumption
is accurate.
The title given by Krusch to his 1926 edition is not a title at all in
Oxford, Bodleian Library, Digby 63. The words Exemplum boni scriptio
primiceri are written in the same insular minuscule and by the same hand
as the remainder of the text that follows, without any break of continu-
ity. The fact that there is no title suggests that this tract was appended
to a preceding text, one that is in fact included in the Dionysiac dossier:
the opinion rendered by Dionysius to Bonus and Boniface, a document
known to scholars as the second letter of Dionysius. The endorsement
(or reply) of the primicerius was in the format of a subscription to the
text of Dionysius, as customary in reply to petitions in Late Antiquity.
If the reading exemplum boni scriptio primiceri is quite clear in Ox-
ford, Bodleian Library, Digby 63, the same cannot be said for the other
manuscripts examined by Krusch, which present an impressive array of
variants: bonefacii; bonisancti; bonus sanctus; boni res.65 To this dazzling
variety Krusch added his own conjecture (ita conieci) and the result was
Exemplum Bonifati primiceri, while Digby’s scriptio, considered corrupt
by Krusch, was confined to the apparatus.66 But the reading in Oxford,
Bodleian Library, Digby 63, especially if taken together with the Cotton
manuscript that calls the document responsum, far from needing correc-
tions or conjectures is proof that the text reproduced faithfully a copy of
the Dionysiac dossier preserved in the papal archives. What we might
call the cover letter to such a copy varied according to the copyist, who

64
Krusch (1926), 57.
65
The manuscripts listed by Krusch (1926), 55–56, and their respective readings
are: Oxford, Bodleian Library, Digby 63, 59r (not 60, as noted by Krusch): exemplum
boni scriptio primicerii; London, British Library, Cotton Caligula A XV, 77r (not 75, as
given by Krusch): exemplum boni res(criptum) primicerii; then, as reported by Krusch:
Paris, Bibliothèque nationale de France, Nouv. acq. lat. 1613, 12v: bonefacii; Munich,
Bayerische Staatsbibliothek, Clm 14725, 23v: boni res; Paris, Bibliothèque nationale
de France, Nouv. acq. lat. 1615, 155r: boni sancti; and Paris, Bibliothèque nationale de
France, Nouv. acq. lat. 16361, 275r: bonus sanctus. Bonifati is a conjecture by Krusch
(1926), 56, note b: ita conieci.
66
Krusch (1926), 56; he justified this choice on p. 54: ‘So muss ihm [ Jan] zur Ent­
schuldigung dienen, dass die Oxforder Hs. den Namen in boni scriptio verdirbt; und die
Leseart boni res in der Münchener Hs. liess den Schreiber ebensowenig erkennen. Den
richtigen Namen hat erst meine Konjectur zum Vorschein gebracht, die jetzt auch ihre
handschriftliche Bestätigung gefunden hat.’


 LUCIANA CUPPO

might have defined the document scriptio or responsum, but the docu-
ment itself did not change.
The second misunderstanding by Krusch is the last line of the tract:
Haec veneracionis. Here, Oxford, Bodlein Library, Digby 63 stands
alone, for the other manuscripts omit altogether these two words that
are indeed puzzling. Krusch took them to be the first words of an unfin-
ished sentence and consequently printed them with ‘…’, but in the manu-
script (fol. 59r) there is no sign of interruption; on the contrary, the end
of the piece is underscored by the same ornamental motifs that appear
after almost every single work. Haec veneracionis was, then, a separate
sentence, albeit brief and with no verb, but nevertheless part of the text.
This is no mistake. Any student of Latin knows that, if there is no verb
in a Latin phrase, the verb ‘to be’ is implied. The sentence, then, reads:
‘These are [or: this is] an object of veneration’, where veneracio means
both ‘veneration’ and ‘observance or obedience’. And precisely the usage
of haec [these things] may be the key for a correct understanding of the
sentence.
The turn of phrase hoc (or haec, or hic) followed by a very short
sentence (typically without verb) is present as a marginal gloss in late
antique manuscripts. Examples are the glosses on Donatus in Monte-
cassino, Archivio dell’Abbazia, 150, and those in Verona, Biblioteca
Capitolare, XXII (20), 93r. Two glosses on folio 93r, still unpublished,
say respectively: hic dicit de peccato qui est orandum prout posse orare (‘this
concerns sin, that one must pray inasmuch as he can pray’) and hic dicit
quomodo fit remissio de spiritu sancto (‘this concerns how remission by
the Holy Spirit comes about’).67 Both have a signe de renvoie to the per-
tinent lines in the main text. The haec veneracionis in Oxford, Bodleian
Library, Digby 63 follows a similar pattern: the signe de renvoie indicates
the reference to the Council of Nicaea as the pertinent point in the main
text, and also shows that haec veneracionis, where haec refers to the de-
crees of the council, was present as a gloss in the margin of the exemplar.
The copyist of Oxford, Bodleian Library, Digby 63 transcribed faith-
fully the signe de renvoie on folio 59r, but moved the words of the gloss
from the margin to the end of the tract. Thus, haec veneracionis reflects a
stylistic convention used in marginal glosses in late antique manuscripts,
another trace of a possibly late antique archetype.

67
Publication of these and other glosses is now forthcoming in Cuppo (2017).


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

The excerpt from the Letter of Pope Vitalian

The excerpt from the letter of Pope Vitalian (Oxford, Bodleian Library,
Digby 63, 59v) gives the first known official endorsement of the Dio-
nysiac Easter reckoning by the see of Rome. The text, which is logically
linked to the recommendation by the primicerius that Easter reckoning
should follow the directives of Nicaea, is:

VITA.LINI PAPE URBIS ROMAE

Numquam enim celebrare sanctum pascae nisi secundum apostolicam et
chatholicam fidem ut in toto orbe caelebratur a christiane blebe; id est se-
cundum apostolicam regulam CCCXVIII sanctorum patrum ac compo-
totum sancti cirilli et dionisi. Nam in toto terrarum orbe sic Christi una
columba, hoc est ecclesia inmaculata, sanctum pascae resurrectionis diem
celebrat. Nam victoris sedis apostolica non adprobavit regulam pascae. Ideo
nec sequitur dispositionem eius pro pascae.

‘Holy Easter must never be celebrated except according to the Catholic

and apostolic faith, as it is celebrated in the whole world by the Chris-
tian people; to wit according to the apostolic rule of the 318 Holy Fa-
thers and the computus of the venerable Cyril and Dionysius. For the
one single dove of Christ, i.e. the immaculate Church, celebrates in this
way throughout the whole world the day of Holy Easter of Resurrec-
tion. Indeed, the Apostolic See did not approve Victor’s Easter rule;
thus, neither does it follow his dispositions for Easter.’

Apart from Bede,68 Oxford, Bodleian Library, Digby 63 is the only

extant witness for this letter of Pope Vitalian.69 The excerpt shows signs
of having been added to the Dionysiac dossier in times removed from
its transcription in Oxford, Bodleian Library, Digby 63: the editor who
penned this manuscript apparently had no idea that the author of the
fragment was Pope Vitalian, for he titled the piece VITA.LINI (fol. 59v).
Vita Lini is the ‘Life of Linus’: the scribe did not realize that Vitalinus
meant Vitalian; and the explanation papa urbis Romae (‘bishop of the
city of Rome’) caused the confusion of identity with Pope Linus, the

68
Bede, Historia ecclesiastica 3.29 (Plummer (1896), i 196–97).
69
Krusch (1926), 52–53, mentioned a manuscript from Whitby described by
Ussher as having the same contents as Oxford, Bodleian Library, Digby 63. Scholars
have assumed that Ussher’s manuscript and this manuscript are one and the same codex.
But the existence of two copies cannot be excluded, for there are some variants between
the two, such as caelebrare licet, cited by Krusch (1926), 53 n. 1. For the table of contents
drawn up by Ussher, see the article by Dáibhí Ó Cróinín in the present volume.


 LUCIANA CUPPO

second pope after Peter on the Roman see. Vitalinus, though, is not a
scribal error but an ancient form of the name Vitalianus attested for
Pope Vitalian in a recension of the Liber Pontificalis prior to AD 742.70
It is also attested in the entry of the Anglo-Saxon Chronicle on the same
pope in the edition of the work by Wilhelm Leibniz.71 It appears, then,
to be a Saxon (from Central Europe to Anglo-Saxon England) variant
of the Latin name, probably an indication of an early drafting of the text,
no longer understood in AD 867, when Oxford, Bodleian Library, Dig-
by 63 found its final form (another sign of antiquity is the word papa,
which in the sixth and early seventh centuries still meant ‘bishop’, while
by the ninth century it had come to mean ‘pope’).
In an oblique but very real way this fragment finds confirmation in
the words of Bede. Bede transcribes the letter of Pope Vitalian to Oswiu
with two omissions, both quite indicative: the first one is the full titula-
ture at the beginning of the letter, the second the fragment in Oxford,
Bodleian Library, Digby 63, replaced by an allusion: et post nonnulla qui-
bus de celebrando per orbem totum uno vero Pascha loquitur (‘and after a
few words in which he speaks of the necessity of celebrating only one
true Easter throughout the whole world’).
The letter to Oswiu is the only instance in the Historia ecclesiastica
where Bede does not give the full titulare of sender and addressee of the
letter. In his masterly work on papal diplomatic writing, Carolus Silva-
Tarouca discussed this fact and drew the conclusion that, if Bede, con-
trary to his habit, did not give any titulature, the reason was that he had
none; and if he did not have it, this was because he did not have access
to either the original or an authenticated copy of the original, which
would have included the titulature.72 What Bede had at his disposal was
the unofficial copy of a private letter that, precisely because it was not
official, did not record the titulature of the original. This was the nor-
mal procedure. The fact that the letter was addressed to a British king
and not to the bishops responsible for a given territory, much less to the
whole church, explains why its transmission was limited to the insular
world, and also tells us that the addition of Vitalian’s fragment to the
Dionysiac dossier in Oxford, Bodleian Library, Digby 63 was carried out

70
An example can be found in Bern, Burgerbibliothek, 408, cited extensively in
Duchesne (1886–92) as C3.
71
Leibniz (1698), 100: Annus Domini 652. Cunstans 11. Vitalinus papa 79 sedit
ann. 14. Martinus papa in exilio moritur.
72
Silva-Tarouca (1931), 37–44, for a discussion of Bede’s collection of papal let-
ters.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

there. We cannot tell when this happened, but—judging from the Vita.
Lini—it must have been at a time when the connection between Pope
Vitalian’s letter and the Dionysiac reckoning was perfectly understood.
At the time of final copying of Oxford, Bodleian Library, Digby 63, this
was no longer the case.
The second, and quite significant omission in Bede’s account, is the
words of Vitalian about the see of Rome not approving the Victorian
Easter reckoning: Bede paraphrases the sentence concerning the neces-
sity of celebrating only one Easter in the whole world but omits the sen-
tence referring to Victorius of Aquitaine. And if we venture to inquire
why, the answer might be that to transcribe such a sentence would have
exposed the contradictions of the Roman see on the whole issue of Easter.
The letter of Pope Vitalian is the first definite pronouncement of the
Roman Church that has come down to us in favour of the Alexandrian
system. Two verbal expressions make this quite plain: the gerund cel-
ebranda, which expresses duty or necessity (must, or is to be celebrated)
and the infinitive celebrare, which depends syntactically on the verb
oportet, not present in Oxford, Bodleian Library, Digby 63, but read-
ily retrievable in Bede. Now, a perusal of the Liber diurnus, the collec-
tion of formulae of the Roman church, shows that the verb oportet (‘it
is necessary’) is seldom used in papal diplomatic language; and when
used, it is typical of a firm command, while simple exhortations were
expressed with the subjunctive. Thus, these two grammatical construc-
tions in one brief sentence show that Pope Vitalian was serious about
the Easter issue. He finally provided the clarification that Columbanus
had demanded some sixty years earlier. Better late than never. Yes, but
Vitalian’s endorsement of the Alexandrian system was embarrassing for
the Romans who had hitherto supported the Victorian reckoning. This
was, it seems to me, the reason for its omission on Bede’s part; and the
omission is an eloquent one.
Strictly speaking, Vitalian was correct: Rome had never officially en-
dorsed Victorius’ computus. But therein lay the problem: Rome had not
provided guidance when it would have been not only necessary, but a
precise duty: only a few years after Vitalian, one pope—Honorius I—was
to be declared heretic by the Sixth Constantinopolitan Council because,
although he had not himself taught heresy, he did nothing to prevent its
spread, as was his duty. Meanwhile, the Victorian system was followed
not only in Gaul, where it had been officially adopted in AD 541, but in
Rome as well, as attested by the Victorian tables in Vatican, Biblioteca
Apostolica, Reg. lat. 2077, 96v–98r, and by the paschal candle in Rome


 LUCIANA CUPPO

with the year of the passion engraved on it. And, while Wilfrid assured
his listeners that Saint Peter had celebrated Easter in Rome according
to the Alexandrian reckoning of moon 15–21,73 Cummian and other
scholars were perfectly aware that the Acts of Caesarea stated that in ap-
ostolic times celebration differed from place to place and that Rome had
sought instruction on the matter from the Church of Jerusalem.74 And
while, strictly speaking, Wilfrid may have been correct in stating that
Saint Peter celebrated according to the Alexandrian system (for after all,
who can tell how Saint Peter celebrated in Rome?), the fact remained
that the system advocated by Wilfrid in AD 664 was not what Rome
had been using not too long before that year.

Peculiar traits, both textual and palaeographical

To look for traces of a seventh-century archetype in Oxford, Bodleian
Library, Digby 63 may be a thankless task. On the textual side—the
transmission of the text, that is—the identification of the archetype
is hampered by the lack of a critical edition and even a census of the
manuscripts of the Dionysiac dossier. Yet, even in the limited range of
observation afforded by the seven manuscripts of the letter of Dionysius
to Petronius collated by Mommsen and re-used by Krusch, some traits
stand out, indicating that the text of the letter in some manuscripts was
drafted in a recension more ancient than the others, regardless of the
time of transcription in the codices examined.
The seven manuscripts examined by Mommsen are, in the order in
which they are listed by Krusch:75

(1) Vatican, Biblioteca Apostolica, Reg. lat. 755, 1v–3v.

(2) Vatican, Biblioteca Apostolica, Reg. lat. 586, 108r–111v.
(3) Berlin, Staatsbibliothek, Phillipps 1830, 1v–2r.
(4) Oxford, Bodleian Library, Digby 63, 63r–67r.

73
Bede, Historia ecclesiastica 3.25 (Plummer (1896), i 188–89).
74
This is how Cummian, De controversia paschali ll. 86–88, alludes to this situation
(Walsh and Ó Cróinín (1988), 68–71): Hac de re sinodis in unum congregatis ob diuersi-
tatem successorum apostolorum (dum apostoli undique pressuris, ut legimus, acti et diuersis
limitibus sparsi, ordinare cyclum regulariter nequiuerunt) (‘In the canons of synods called
together because of disagreement on this matter amongst the successors of the apostles
(since the apostles, as we read, acting under strain from all sides and scattered over vari-
ous territories, were themselves unable to regularly compose a cycle)’). The Acta of the
Council of Caesarea, recension B, state (Krusch (1880), 306): Post transitum ergo de hoc
mundo omnium apostolorum per singulas prouincias diversa tenebant ieiunia.
75
Krusch (1938), 60.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

(5) Paris, Bibliothèque nationale de France, Lat. 5543, 7r–9r.

(6) Paris, Bibliothèque nationale de France, Lat. 5239, 4r–6r.
(7) Paris, Bibliothèque nationale de France, Lat. 4860, 148v–149v.

Within this group, two recensions may be discerned, one comprising the
manuscripts numbered 1–3–5–6 and the other that includes the manu-
scripts 2–4–7. In group 2–4–7 there are some distinctive traits not pre-
sent in the other. They are:

(a) The word for ‘week’ is septimana in group 2–4–7, hebdomada in

the other. Both would have been used from the fourth century
on, according to Souter;76 but after the Gregorian reform of the
liturgy hebdomada came to be used in the sense of ‘week’ in litur-
gical texts and may therefore have been preferred to septimana.
Also, Isidore of Seville used the word hebdomada in his De natura
rerum,77 and this may have influenced later editors. When Dio-
nysius wrote his letter the Gregorian reform and Isidore had yet
to come, while septimana had already been used in De paschate of
Pseudo-Cyprian, a treatise well-known in Rome. It is therefore
likely that septimana is the word used by Dionysius.
(b) 2–4–7 have adieciunt, 1–3–5–6 have adiciunt.
(c) 2–4–7 have dni nri ihu xti, the others only dni. This punctilious-
ness (also found in Cassiodorus and Felix) is understandable in
the light of the theological controversies on the Person of Christ,
quite lively in the sixth and seventh centuries. Hence, it may sig-
nal an older recension of the text.
(d) 2–4–7 preserve the Dionysiac cycles continued by Felix, the oth-
er manuscripts do not; possibly the cycles were omitted because
they were obsolete. If so, this would date the 1–3–5–6 group after
AD 721.78
(e) 2–4 have putius for potius. Given the frequent confusion between
o and u, this may seem a trivial detail, but concern for spelling
with u instead of o, even when it was not necessary, was high
at Vivarium: the codex Verona, Biblioteca Capitulare, 39 (the
Complexiones in Apocalypsi of Cassiodorus, copied at Vivarium

Souter (1949).
76

Isidore, De natura rerum 3 (ed. and trans. by Fontaine (1960), 182–83).

77

78
The omission was specifically noted in the Liber de computo of Milan, Biblioteca
Ambrosiana, H 150 inf., 50v: Hunc ergo cyclum [i.e. the years AD 627–721] praetermisi-
mus eoquod totus praeteriit (‘We omitted this cycle, because it is entirely past’).


 LUCIANA CUPPO

c.AD  575) has putantes where the text requires potantes (chap-
ter 13).79
(f ) finally, 4 is the only codex that refers to bishops as pontifices in-
stead of antistites. Pontifex (meaning ‘bishop’) is the word used by
Dionysius, and before him by Jerome and Augustine.80

On the palaeographical side, the first obvious fact is that the script of the
Dionysiac dossier in Oxford, Bodleian Library, Digby 63 is beautifully
(and expertly) insular; therefore, any traces of the exemplar / archetype
should not be sought in the script as such, but in layout, initials, margi-
nalia, or—possibly—indicative errors. This said, some remarks may be
ventured.
Besides haec veneracionis (which, as I proposed, appears to be a rem-
nant of a gloss in the exemplar), another remnant of the layout of the ar-
chetype is the FINIT PRAESTANTE on folio 71. The copyist working
in AD 867 reproduced or added on his own initiative ornamental motifs
to signal the end of each piece. He did this also on folio 71 to complete
the line only half filled by FINIT PRAESTANTE, which marked the
end of a whole book. The words FINIT and PRAESTANTE are spaced,
but while the book is finished, its last sentence is not: praestante is what
might be called a dangling ablative, which should find its logical comple-
tion in praestante Domino, a turn of phrase often used by Cassiodorus as
early as the Variae, also found in Expositio Psalmorum, Institutiones, and
Historia Ecclesiastica.81 The explicit should, then, read, Finit praestante
dno [= Domino], but the letters dno (nomen sacrum for Domino) are
omitted. Why?
I believe that this is due to the layout of the explicit in the archetype.
In Vivarian manuscripts the last letters of an explicit (or a citation, in
the case of marginal glosses) were inscribed in a frame in the shape of
a grape-cluster, or chalice, or vase—all of which are broader at the top

79
Cassiodorus, Complexiones in Apocalypsi 13 (ed. by Roger Gryson in CCSL 107,
120).
80
Krusch (1926), 63.
81
To cite only a few examples: superior liber Domino praestante completus (Institu-
tiones 2.praefatio.1; Mynors (1937), 89), ubi velut anachoritae praestante Domino feliciter
esse possitis (Institutiones 1.29.3; Mynors (1937), 74), quas uobis in annotato nuper codice
Domino praestante dereliqui (Institutiones 1.3.1; Mynors (1937), 18). The expression oc-
curs twice in the colophons of Verona, Biblioteca Capitolare, 39; Vatican, Biblioteca
Apostolica, Pal. lat. 824 has on fol. 1v: Incipiunt tituli ecclesiasticae historiae cum opere suo
ab epiphanio scolastico Domino praestante translati.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

and thin toward the base.82 Grape-cluster, cup, and vase are typically Vi-
varian motifs, but the stylistic convention of writing the last letters of
a gloss one on each line was also observed in manuscripts written in or
near Rome in the time of Pope Gregory I, for example in Troyes, Biblio-
thèque municipale, 604, where the marginal glosses are arranged in the
shape of a triangle turned upside down. If this was the style of the arche-
type of Oxford, Bodleian Library, Digby 63, dno [= domino] would have
been written:
d
 n
  o
A scribe not familiar with this late antique stylistic convention could
have easily omitted those last letters, not realizing that they were an in-
tegral part of the text.
In the introduction of Felix to the Dionysiac cycles there are two
errors. Both are common to Oxford, Bodleian Library, Digby 63 and
Milan, Biblioteca Apostolica, H 150 inf., signalling a common exemplar
for both manuscripts. The first error is a common one: ad instead of ut.
The fact that u was mistaken for a points to an exemplar written in a
script with an open a, easily (and often) mistaken for u. The second error,
however (Oxford, Bodleian Library, Digby 63, 67r; Milan, Biblioteca
Apostolica, H 150 inf., 108r), is not so easily explainable and an explana-
tion may be sought in a faulty reading of the archetype.
The error consists in the omission of the number ‘200’ (CC in Ro-
man numerals) from Olympiad 245, the time of the birth of Christ, so
that the year of the Nativity becomes Olympiad 45. A possible reason
for this error may have been the peculiar writing of the hundreds (CC,
CCC, CCCC = 200, 300, 400) in southern Italian uncial script, a stylis-
tic trait attested in Vatican, Biblioteca Apostolica, Reg. lat. 2077: in the
section of the codex that includes the continuation of the chronicle of
Prosper known as Paschale Campanum (fols 97r–98v) the Years of the
Passion are written in the margin of every decade. On folio 97r–v the
Cs in the dates AP 450, 460, and 470 are written one inside the other,
resembling a snail’s shell rather than a normal CCCC. Thus, it is possible
that a similar writing convention made the CC of the archetype unintel-

82
In Institutiones 1, Cassiodorus explained the characteristics and the symbolic
meaning of such figures, which he introduced in the books produced at Vivarium (Insti-
tutiones 1.3.1; Mynors (1937), 18). One original (Verona, Biblioteca Capitolare, 39) has
survived and later copies are fairly numerous.


 LUCIANA CUPPO

ligible to a scribe not used to that writing system: not realizing that the
snail-shell was the numeral ‘200’, he may have simply bypassed it.
Finally, we must consider the possibility that the siglum XB (Ch(riste)
b(enedic) = Christ bless), which appears in the Dionysiac dossier and in
another piece of Roman derivation in the top right margin on the verso
of nearly every page, and almost nowhere else in the codex, derives from
a late Roman archetype.83 While any conclusions would be premature
at this stage, particularly since there has not been to my knowledge a
systematic study of such sigla, the time is ripe for some considerations
on the possible late antique origin of XB.
Such origin is definitely attested for XF (Ch(riste) f(ave) = Christ be
propitious), the twin siglum of XB: the two are similar in concept but
different in their verbal expression. XF is attested in some late antique
manuscripts discussed extensively by E.  A. Lowe;84 in addition, Dom
Michel Huglo and D. A. Bullough provided additional information on
XF (or its full form, XRE FAVE) in inscriptions in Lombard Italy.85 In
the case of XB a late antique origin, though not directly attested, is sug-
gested by the fact that the siglum XB appears in medieval manuscripts
unrelated to one another in transmission or location and cannot be asso-
ciated with a single scriptorium or a given local tradition. The one thing
that such works have in common is that they are all copies of late antique
texts. Some examples will make this clear:

(a) Oxford, Bodleian Library, Junius 11 (Genesis A);86 suggested

provenance from Canterbury, Winchester, or Malmesbury. The
siglum xb (in minuscule) is found at the beginning of the preface
(page 1), then again after the preface, at the beginning of the text
proper (page 2), and lastly in the top left margin of page 52.

83
‘Almost’ because XB is also found in a piece of ancient Roman derivation, the
Cölner Prolog (Oxford, Bodleian Library, Digby 63, 80v–87v), so named by its editor
Bruno Krusch (1880), 227–35, who published the recension of the work extant in Co-
logne, Dombibliothek, 83-II, 193v–197r: he was not aware that the same text is also
extant in Oxford, Bodleian Library, Digby 63, but in a remarkably different recension.
In Oxford, Bodleian Library, Digby 63 XB is found consistently in the right top margin
of fols 49v, 53v, 54v, 55v, 56v, 57v, 58v, 59v, 60v, 61v, 62v, 63v, 64v, 65v, 66v, 67v, 68v, all
containing texts from the Dionysiac dossier; and on fols 80v, 86v, and 87v, pertaining to
the Roman paschal prologue.
84
Lowe (1935), xv.
85
Huglo (1954); Bullough (1960).
86
Available online at http://image.ox.ac.uk/show?collection=bodleian&manuscr
ipt=msjunius11.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

(b) Karlsruhe, Badische Landesbibliothek, Aug. perg. 195, 42r.87 The

text is based on Cassian (Collatio V), but derived from Evagrius,
a Spanish bishop contemporary of Pope Gregory I and Leander
of Seville.88 Given the known exchange of manuscripts between
Gregory and Leander, the Epistola of Evagrius may well have been
known in Rome and transmitted by Gregory or his followers. The
provenance is Reichenau and the script is Caroline, presumably
from a Germanic area.
(c) St Gall, Stiftsbibliothek, 904 (Priscian);89 from St Gall, but writ-
ten in Ireland in the late eighth century. This manuscript presents
a pot-pourri of invocations: XB is by far the most frequent (thirty
occurrences), followed by FAVE (ten occurrences), XF (six occur-
rences), and adiuba addressed to Christ or saints such as Brigit or
Patrick (six occurrences). These were cited by Lindsay with a few
spicy comments in his description of the codex.90

These manuscripts differ in origin and / or provenance: Oxford, Bodle-

ian Library, Digby 63 and Junius 11 are Anglo-Saxon, St Gall, Stiftsbib-
liothek, 904 is Irish, Karlsruhe, Badische Landesbibliothek, Aug. perg.
195 is Irish in the section pertaining to Augustine, but the XB appears at
the beginning of the Germanic section. It is, then, impossible to consider
XB a local mark identifiable with a given scriptorium or institution, as
suggested by earlier scholars, and it does not seem unreasonable to apply
to this siglum the tried-and-true methodological principle that, if the
same palaeographical traits appear in manuscripts that are not mutually
related, chances are that such traits go back to a common exemplar or
archetype.
Further, the character of XB should be considered. This can be done
by an examination of its contents, which—though it seems the obvi-
ous thing to do—has been done neither for XB nor for XF. Lowe was
more concerned with establishing the geographical coordinates of XF—
which he did admirably—than the content of this ‘pious invocation’;91
others, primarily Michel Huglo, were concerned with the expansion of

87
Available online at http://digital.blb-karlsruhe.de/blbhs/content/pageview/​
3301580.
88
It can be found in PL 80, 9–14.
89
Available online at http://www.e-codices.unifr.ch/en/description/csg/0904.
90
Lindsay (1910), 57.
91
Lowe (1935), xv.


 LUCIANA CUPPO

XF beyond the boundaries of southern Italy and in monuments and

inscriptions. Earlier scholars, such as James, studied the provenance of
manuscripts with XB; to James ‘provenance’ meant a single scriptorium,
not a Schriftprovinz. Accordingly, he identified XB with a Canterbury
mark without inquiring what the meaning of the mark might be. Study-
ing St  Gall, Stiftsbibliothek, 904, Lindsay was quick to point out the
frequency of XB, characterized as an invocation by the fact that it was
mixed not only with XF, but with the much more explicit adiuva and
occasionally with the name of the scribe and that of the saints expect-
ed to help him. A staunch Presbyterian such as Lindsay could only be
mildly amused at these medieval superstitions, but not take them seri-
ously. His stern disapproval is understandable and makes for delightful
reading, but his dumping all detectable sigla and inscriptions into one
big cauldron of popish superstition is not conducive to a proper under-
standing of the palaeographical material involved.
Such an understanding begins, I believe, with the Lapalissian realiza-
tion that XB and XF are prayers. This must be said, because little atten-
tion has been devoted to this aspect of the sigla; and when it has, as in
the meritorious case of W. M. Lindsay, no difference among the three
formulae has been perceived. Yet, XF and XB differ in concept from adi-
uva inasmuch as they invoke divine help not for the writer or scribe (as
adiuva does) but for the work at hand. There is, then, a conceptual dif-
ference between XF and XB on one side and adiuva on the other. To the
contrary, in XF and XB the concept is the same, but there is a difference
in expression, for XF belongs to the literary language and is not specifi-
cally Christian, while XB is colloquial and definitely Christian in charac-
ter. In a later stage of development, reflected in St Gall, Stiftsbibliothek,
904, the three expressions are mixed and the scribes used the invocations
as they pleased, adding to the traditional ones some of their own with
their favourite saints. But this uninhibited use of the invocations, as well
as the mix-and-match approach, belong to a later stage of development.
In the early stage (up to the sixth or early seventh century for XF used in a
Christian context), the siglum is used to invoke divine protection for the
work, while adiuva invokes divine help for the writer.92 We know of no
manuscripts of Late Antiquity that add names of saints to the customary
formulae or pile up two and even three invocations on the same page, as
St Gall, Stiftsbibliothek, 904 does in a uniquely uninhibited way.

92
Vatican, Biblioteca Apostolica, Reg. lat. 2077, 101r, provides one example: Xre
adiuba desiderantes te nosse, penned on the initial page of an epitome of Augustine’s De
haeresibus (only the initial page is extant).


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

In older manuscripts that feature XF, the invocation is situated in

initial position at the beginning of the work. It is the same position as
that reserved in ancient poetry for the invocation to God (or, in pagan
poetry, to the Muse) and is a sign that help is requested for the work, not
the worker (another such sign is that the worker is never mentioned).93
The presence of FAVE or FAVE VOTIS in pagan literature as an invo-
cation for divine protection on the work or the worker is well attested,
from Vergil (Eclogae 4.10: nascenti puero casta fave Lucina) to Ovid
(Tristia 4.2.55: delubra faventia votis; Metamorphoseon libri 13.334: fa-
veat fortuna) to Tibullus (Carmina 2.51: Phoebe fave) to Ennius (An-
nales 291: Romanis Iuno coepit placata favere). It lived on in Christian
times, no longer as an invocation to pagan gods, but to a symbolic Muse
or to a powerful patron to whom the literary work was being offered, or
as a prayer to God. Thus, the fifth-century Merobaudes (Carmina 4.5)
and Martianus Capella (De nuptiis Philologiae et Mercurii 2.220) wrote
favete Musae, as Statius (Silvae 2.7.20) had done centuries before. Their
near-contemporary Servius Honoratus used laboris mei […] faveas voto
in the dedicatory preface of his work to his patron Albinus, and Vegetius
used fave in the dedication of his Epitoma de re militari to the emperor.
The appropriation of faveo for expressions of Christian worship is at-
tested as early as Paulinus of Nola (Carmina 27.499: Favet […] Christus).
XF has, then, a rich classical and literary background, and perhaps it is
not coincidental that its known witnesses originated in southern Italy, a
geographical area open to Greek culture and classical learning, and later
flourished in Lombard Italy, where Irish influence (and with it the love
for classical learning) was strong.
No such background can be claimed for XB, which appears only in
a Christian context. And, while XF is attested in sixth-century manu-
scripts, XB survives—as far as we know—only in later copies, so that

93
Some examples gleaned from CLA 2: XRE F(ave) occurs at the beginning of
each Gospel in Oxford, Bodleian Library, Auct. D. 2. 14; XF in the hand of the annota-
tor (none other than Victor of Capua) is found in the Codex Fuldensis of the Bible at the
beginning of the Acts and Letters of the Apostles; Donatus (the annotator of Ambrosi-
aster; Montecassino, Archivio dell’Abbazia, 150) wrote XF before the initial line of the
commentary to the Letter of Saint Paul to the Romans; in the Codex Amiatinus (Flor-
ence, Biblioteca Laurenziana, Amiatino 1) XRE FAVE appears at the opening of Saint
Matthew’s Gospel (fol. 805r). Apart from the Bible, XRE D(omi)NE FAVE VOTIS is
found at the beginning of the Acts of the Fifth Ecumenical Council of Constantinople
in 553. The Martyrologium of Willibrord (Paris, Bibliothèque nationale de France, Lat.
10837), copied in the eighth century from an earlier exemplar, also has XRE FAVE VO-
TIS at the beginning of the text (fol. 2r, top margin).


 LUCIANA CUPPO

direct evidence is not available. Once more, we have to be content with

indications preliminary to further inquiry.
The affinity between XB and XF is unmistakable and was promptly
noted by Bernhard Bischoff, who noted that both expressions have the
same meaning.94 Precisely the affinity of concept between the two sug-
gests that XB developed as a variant of XF when the latter had become
a recognized Christian invocation. The shift from one to the other oc-
curred at various levels: at the religious level (FAVE, though no longer
used in a pagan sense, could call to mind its former pagan connota-
tions); at the level of appreciation of classical culture (FAVE was a liter-
ary expression, certain to be appreciated by literate persons, but not a
household word among the simpler folk); at the biblical level (FAVE
did not belong to biblical language). In contrast, benedic (‘bless’) was a
biblical word par excellence, it belonged to the everyday language heard
in church at religious services, and was not tainted with reminiscences
from pagan literature.
And here one name comes to mind: that of Pope Gregory I. His ap-
proach to classical culture was cautionary to say the least; to him, the Bi-
ble was paramount. He confirmed as the official rite of the mass for the
Roman church the prayers of the canon that had existed for centuries.95
And in the canon of the mass—the holiest part of the rite, which in-
cludes the words of the consecration—Gregory retained the invocation
benedic: to this day, the priest pronounces it five times before the conse-
cration, never in the sense of invoking blessings for himself or the faith-
ful, always in the sense of invoking God’s intervention on the bread and
the wine, that they may become the body and blood of Christ.96 The
blessing is, then, sought for the work—the consecration—that the priest

94
Bischoff (1990), 44: ‘For the invocation the form ‘Christe fave’ (XF) occurs,
and the group xb- (‘Christe benedic’) with the same meaning is frequent in Irish manu-
scripts.’
95
The words of the Canon can be found in Ambrose’s De sacramentis 5.21–23 (PL
16, 451–52).
96
The use of benedicere (‘to bless’) in the sense of ‘to consecrate’ is well attested.
Some examples drawn from the Thesaurus linguae latinae: (a) Vrsinum Paulus Tiburtinus
episcopus benedicit (‘Paul, bishop of Tivoli, consecrated Ursinus’; Avellana collection);
(b) vidua, quae benedicta non fuit, quare non debet maritum accipere? (‘why should not
a widow, who has not taken vows, take a husband?’; Concilium Turonensis AD 567); (c)
benedixit Deus diem septimum (‘God consecrated the seventh day’; Gen. 2:3 in the vetus
translation reported by Augustine, De civitate Dei 22.30); (d) benedixerit, id est sanctifi-
caverit (‘consecrated, i.e. sanctified’; Augustine, Contra Faustum 16.6); (e) sacri mysterii
benedictionem (‘the consecration of the sacred mystery’; Gregory the Great, Epistola
14.2).


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

is performing, that it be made perfect. And it is fitting that, as bread and

wine are blessed to become acceptable to God, so literary works written
for the glory of God should be blessed to become pleasing to God.
Thus, to think that Pope Gregory was responsible for the substitu-
tion of the semi-pagan XF with the thoroughly Christian XB, a thor-
oughly Roman formula patterned on the ritual for the Roman canon
confirmed by Gregory, seems a reasonable hypothesis. In support of
it—which, because of the scarcity of witnesses, cannot as yet be called
evidence—it should be noted that XB is found in manuscripts compat-
ible with a Roman archetype no earlier than the time of Pope Gregory I.
A connection with Rome is assured for the Dionysiac dossier of Oxford,
Bodleian Library, Digby 63 and is plausible for Junius 11 and Karlsruhe,
Badische Landesbibliothek, Aug. perg. 195 (Evagrius). The versification
of Genesis in Oxford, Bodleian Library, Junius 11 is, of course, Anglo-
Saxon, but the English text presupposes a Latin exemplar of the Bible. If,
as scholars have suggested, this codex comes from Canterbury or from
the Winchester area, its author worked in the area of operations of the
Roman bishops sent by Gregory and their successors, who used books of
Roman provenance.
In Karlsruhe, Badische Landesbibliothek, Aug. perg. 195, XB pref-
aces a text on the capital vices sent by Evagrius, the Spanish bishop con-
temporary of Pope Gregory, to another bishop. The capital vices were a
theme dear to Gregory’s heart, and the exchange of literary works be-
tween the Roman curia and Spanish bishops—specifically Gregory and
Leander of Seville—is well documented. Thus, a Roman origin for the
tract of Evagrius is a strong possibility.
While there are very good reasons to think that some material in Ox-
ford, Bodleian Library, Digby 63 is of Frankish provenance, there is no
reason to assume that it came all from the same location in Francia, or
at the same time. On the other hand, the fragment of Vitalian’s letter
suggests that at least part of the material for the collection was assem-
bled in Anglo-Saxon England, while the massive presence material of
Frankish provenance suggests a centre open to exchange with the Conti-
nent, perhaps along the pilgrim route (Dover-Canterbury-Winchester).
Modest traces of the archetype were preserved in the Dionysiac dossier
of Oxford, Bodleian Library, Digby 63 through we do not know how
many intermediate copies. This may be a sign that such an archetype was
considered authoritative; the traces point to stylistic conventions alive
in Rome and at Vivarium. Computistical material bearing Rome’s seal
of approval was needed in England particularly, but not exclusively, at


 LUCIANA CUPPO

the time of the Easter controversy at Whitby in AD 664. This is a likely

date for the copies of the texts of Felix to reach England, but it cannot
be excluded that they came earlier—or later. All in all, Felix met with a
happier lot in England than on the Continent: his work was preserved
as he had wanted it to be, in a reasoned sequence after that of Dionysius,
and he remained known for what he had wanted to be: successor Dionisii.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Appendix
RATIO SPERE DIONISII DE CIRCULO MAGNO PASCHE
(Vatican, Biblioteca Apostolica, Vat. lat. 1548, 68r–69r)
Dionisius venerabilis abbas cuiusdam monasterii urbis Romae. utri-
usque linguae grecae videlicet et latine non ignobili praeditus scientia
paschales scribens ciclos voluit in eis annorum memoriam incarnationis
Christi tradere christianis, ut vero passio Christi et redemptio humani
generis et tempus in quo quanta pro omnibus salvandis filius Dei sus-
tinere voluit. ad memoriam reduceret. Scripsit ergo circulum magnum
paschae qui multiplicato invicem solari ac lunari circulo dxxxii perficit
annos. Sive enim decies novies viceni et octoni seu vicies octies deni ac
noveni multiplicentur, quingentorum triginta duorum annorum com-
plent numerum. In quo circulo notavi, quem tu si vis cognoscere potes,

1. quot sint semper anni ab incarnatione Domini vel a passione,

2. et qualis sit indicio in unoquoque anno,
3. et quales concurrentes
4. et qualis sit epacta
5. et quot claves terminorum
6. et in qua feria termini eveniant
7. et dies festorum quos termini indicant.

Anno igitur septimo cicli decemnovenalis natus est Christus. Annus

apud Romanos habet inicium in kalendis ianuariis. Inito itaque supra-
dicto anno quo Dominus Iesus natus est, quarta fuit concurrens in kalen-
dis marci et epacta sexta in kalendis septembris et indicio xv in kalendis
octobris. Deinde in kalendis ianuariis natus est de virgine filius Dei. Post
cuius nativitatem in kalendis marciis sexta fuit concurrens et fuit bis-
sextus anno illo, epacta vii x in kalendis septembris et indicio prima in
kalendis octobris. Completis sic post Domini nativitatem xxxiii annis
et tribus mensibus, passus est ipse Dominus anno secundo cicli decem-
novennalis in mense marcio octavo kalendis aprilis luna xiv concurrente
v epacta undecima indicione tertia. Ideo hii numeri temporis passionis
Christi notati scribuntur in cereo paschali, ut semper recordemus nos
illius passione liberatos esse de servitute diaboli. et a poenis inferni . sci-
entes quia si socii passionis eius fuerimus, simul et resurrectionis erimus.
Atque ideo non a nativitate, ut quidam aberrantes estimant, set a pas-
sione annos Christi numerare debemus et ad indictionem inveniendam
vel concurrentem vel epactam et cetera ut in sequentibus ostendimus.


 LUCIANA CUPPO

Venerabilis itaque Dionisius ipse nobis in paschalibus argumentis

ostendit. quatenus
(1) ad inveniendum annum cicli decemnovenalis sumamus annos a
passione Domini et unum superaddamus (fol. 68v) et per xix partiamus.
Ideo unum addimus. quia unus annus iam transierat de circulo decem-
novennali quando passus est Dominus, non quando natus; quia tunc
septimus annus fuit.
(2) Item ad inveniendum solarem ciclum monet ab annis a passione
subtrahere octo et ceteros per viginti octo partiri. Ideo octo subtrahimus
quia octo supererant de ciclo solari quando passus est Dominus, non
quando natus; quia tunc xv annus fuit.
(3) Item ad sciendum quotus sit annus cicli lunaris sumere monet
annos a passione Domini. et subtrahere duos ex ipsis. et ceteros per xix
dividere. Ideo duos subtrahimus quia adhuc duo anni supererant de ciclo
lunari quando passus est Dominus, non quando natus; quia tunc quartus
annus fuerat.
(4) Item ad inveniendas epactas monet annos a passione Domini per
xix partiri et quod superfuerit per undecim multiplicare. et multiplica-
tos rursus per xxx dividere. Per xix dividimus quia per xix annos currunt
epactae.
(5) Item ad indicionem inveniendam regulares monet annis a pas-
sione Christi addere. et ipsos per xv partire. Ideo tres regulares iunguntur
quia tercia erat indictio quando passus est Dominus, non quando natus;
quia tunc xv fuit.
(6) Item ad concurrentes inveniendum monet annos Domini a pas-
sione per quattuor dividere. Quartam partem ideo apponimus, quia in
quarto anno semper bissextus. Quattuor regulares ideo iu(n)gimus, quia
quarta concurrens praecesserat annum in quo passus est Dominus.
(7) Item ad bissextilem annum inveniendum. annos Christi per quat-
tuor dividere monet. et si unus remanserit, primus erit annus post bissex-
tum. Si duo remanserint, secundus erit. Si tres, tertius erit. Si quattuor,
bisextus erit. Notate ergo, si annis supradictos regulares iuncxerimus,
nulla omnino extra conveniencia iunctus et temporibus; si vero annis a
passione illos addidimus, singula convenient singulis temporibus.
Sic enim Deus voluit homines passionis suae sepius quam nativitatis
recordari, quia plus passio nobis quam nativitas contulit. Nichil enim
eum nobis nasci profuisset nisi per mortem suam nos redemisset. Per na-
tivitatem suam nos revocavit. per passionem liberavit. per resurrectionem
ad fidem confirmavit. per ascensionem spem nobis in celestibus dedit.
per sancti Spiritus igneam missionem credentes ipse omnes illuminavit.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

Haec ad instruendum nos reliquit nobis scripta merito honorandus

doctor noster Dionisius. docuitque nos de annis Domini a passione ar-
gumenta facere ad invenienda omnia illa quae supra sunt dicta. et omnia
fere illa notavit in spera quam graecis litteris de circulo magno paschae
composuit. Quam et nos de grecis litteris in latinis transtulimus. et per
easdem notas signavimus. In capite notantur anni a passione Domini. et
ibidem in capite semper ciclus decemnovennalis incipit. Ubi est color
croceus, ibi est concurrens prima. Ubi est color de azurio, ibidem est
prima (fol.  69r) indicio. Epactae habentur desuper, et deorsum claves
terminorum.

‘EXPLANATION OF THE SPHERE OF DIONYSIUS ON

THE GREAT EASTER CYCLE
Dionysius, venerable monk of a certain monastery in the city of Rome,
endowed with no mean knowledge of both languages—Greek, that is,
and Latin—, writing the Easter cycles wanted to hand down with them to
Christians the memory of the years of Christ’s incarnation, yet so that the
passion of Christ, and the redemption of mankind, and the time when
the son of God willed to suffer so much for the salvation of all, would
come back to memory. Thus he wrote the Great Easter Cycle, which by
the reciprocal multiplication of the solar and lunar years totals 532 years;
for, whether one multiplies 19 by 28 or 28 by 19, they total the number of
532 years. And in this cycle I noted, and you may find it, if you so desire:

1. How many years there are always are from the incarnation and
the passion of the Lord,
2. Which is the indiction in every year,
3. Which are the concurrents,
4. Which are the epacts,
5. Which are the keys to the Easter limits,
6. And on which day the Easter limits fall,
7. And the feastdays signalled by the limits.

Christ was, then, born in the seventh year of the 19-year cycle. For the
Romans the year begins on the kalends of January. Thus, after the begin-
ning of the aforementioned year in which our Lord was born, the fourth
concurrent was on the kalends of March and the sixth epact on the kal-
ends of September and the 15th indiction on the kalends of October.
Then, on the kalends of January, the Son of God was born of the Virgin.


 LUCIANA CUPPO

The sixth concurrent was on the kalends of March after his nativity, and
that was a bissextile year with the 17th epact on the kalends of September
and the first indiction on the kalends of October. Thus, upon comple-
tion of 33 years and 3 months after the nativity of the Lord, the Lord
suffered in the second year of the 19-year cycle, in the month of March,
on the kalends of April on the 14th moon, on the fifth concurrent, in the
11th epact, in the third indiction. Therefore these numbers of the time of
Christ’s passion have been noted and are being inscribed in the paschal
candle, so that we may always remember that through his passion we
have been freed from subservience to the devil and the torments of hell;
knowing that, if we have been sharing in his passion, we will also share
in the resurrection. Consequently we must not count the years of Christ
from the nativity, as some deem in error, but from the passion, in order
to find the indiction and the concurrents and the epact, etc., as shown
below.
Thus the venerable Dionysius shows us in the Easter argumenta that,

1. In order to find the year of the 19-year cycle we must take the
years from our Lord’s passion and add one,
2. Next, in order to find the solar cycle he instructs us to subtract
8 years from the passion and divide the remaining ones by 28. We
subtract 8 for the reason that there were 8 years left in the solar
cycle when the Lord suffered, not when he was born, because it
was then the 15th year.
3. Next, in order to find out which year of the lunar cycle it is, he
instructs us to take the years from the Lord’s passion and subtract
two from them, and divide the others by 19. We subtract two be-
cause there were still two years of the lunar cycle left when our
Lord suffered, not when he was born, for then it was the fourth
year.
4. Next, to find the epacts he instructs us to divide by 19 the years
from the Lord’s passion and to multiply by 11 what is left, and to
divide again by 30 the years multiplied. We divide by 19 because
the epacts run for 19 years.
5. Next, in order to find the indiction he instructs us to add the reg-
ulars to the years from Christ’s passion and to divide them by 15.
Three regulars are added because it was the third indiction when
the Lord suffered, not when he was born; for then it was the 15th
indiction.


 FELIX OF SQUILLACE AND THE DIONYSIAC COMPUTUS II

6. Next, in order to find the concurrents he instructs us to divide

by 4 the years from the Lord’s passion. We use the fourth part
because in the fourth year there is always a bissextile year. We add
the four regulars because the fourth concurrent had preceded the
year when our Lord suffered.
7. Next, in order to find the bissextile he instructs us to divide the
years of Christ by four, and if the remainder is one, it will be the
first year after the bissextile. If the remainder is two, it will be the
second. If three, it will be the third. If four, it will be a bissex-
tile. Then take notice: if we add to the years the aforementioned
regulars, in no addition whatsoever there is lack of concordance
with the time reckoned; for if we add those years to the years of
the passion, each individual result will be concordant with each
individual time reckoned.

For in this way God wanted men to remember his passion more often
than his nativity, because the passion brought us more profit than the
nativity. Indeed, it would have profited nothing to us that he was born,
if he had not redeemed us with his blood. He called us back through his
nativity, he freed us through his passion, he confirmed us in the faith
through his resurrection, he gave us hope in heavenly things through his
ascension, he enlightened all believers by sending the fire of the Holy
Spirit.
Our truly honourable doctor Dionysius left us these things in writing
for our instruction and taught us to compose the argumenta about the
years from the Lord’s passion, in order to find all the aforesaid items, and
noted them almost all in the sphere that he composed in Greek about
the Great Easter Cycle, and we translated it from Greek to Latin and
marked it with the same notes. In the beginning are noted the years from
the Lord’s passion, and the 19-year cycle always begins from that same
point. Where the colour is yellow, there is the first concurrent. Where
the colour is blue, there is the first indiction. All epacts are above, and
below are the Easter limits.’


 BRIGITTE ENGLISCH

OSTERFEST UND WELTCHRONISTIK

IN DEN WESTGOTISCHEN REICHEN

Abstract
In Visigothic Spain and its post-AD 711 successor kingdoms, a system for the
calculation of Easter different from the early medieval standard reckonings by
Victorius of Aquitaine and Dionysius Exiguus  / Bede appears to have been
used. The crucial parameters for determining Easter, the sun and the moon,
held a special place within Visigothic religious belief, especially in respect of
its resurrection theology and contemporary apocalyptic views as expressed by
Beatus of Liébana. Because of contacts with the rival calendars of the Jews and
Muslims, it seems that the neo-Gothic kingdoms relied on actual observations
of these two heavenly bodies. Key for our understanding of this approach are
the intellectual productions of the reign of Alfonso  III, especially the codex
León, Archivo de la Catedral, 8 of the early tenth century. Thus, a previously
unknown Easter reckoning is to be assumed for the Visigothic kingdoms of the
seventh to eleventh centuries.
Keywords
Alfonso III, Wamba, Reccared, Isidore of Sevilla, Dionysius Exiguus, Victorius
of Aquitaine, Bede, Martin of Braga, Julian of Toledo, Migetius, Elipandus
of Toledo, Wamba, Beatus of Liébana; Santa Maria de Quintanilla las Viñas;
council, synod, liturgy, incarnation era, Hispanic era, sun, moon, colophon;
4th council of Toledo, 2nd synod of Braga; Commentary on the Apocalypse;
Computus Cottonianus, Mozarabic Chronicle, Antiphonary of León, Tapís de
la creació of Girona.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 182-211
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114737
 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

Einleitung
Richtet man aus der Perspektive der insularen Reiche oder des Fran-
kenreichs den Blick auf das westgotische Spanien, so könnte man der
Auffassung sein, nach der für sein Volk Weisung gebenden Konversion
König Rekkareds im Jahr 589 vom häretischen Arianismus zum etablier-
ten Katholizismus sei das religiöse Leben und mithin auch die Bestim-
mung des Osterfestes durch ordnende Regulative strukturiert gewesen.1
Spätestens mit dem Bekenntnis des die Bildung des gesamten Mittel-
alters prägenden Isidors von Sevilla zu der auf Cyrill von Alexandrien
projizierten Osterrechnung des Dionysius Exiguus und die Fortführung
seiner Ostertafel2 müsste die christliche Festberechnung in ruhige Bah-
nen geraten sein. Doch weit gefehlt: Die Bestimmung des Osterfestes
bietet, zumindest dem ersten Anschein nach, ein chaotisch anmutendes
Szenario. Das 4. Konzil von Toledo (633) spricht explizit von miteinan-
der konkurrierenden Modi, das Auferstehungsfest festzulegen.3 Davon,
dass man bestrebt war, dennoch eine einheitliche Regelung zu erreichen,
zeugt die 2. Synode von Braga (572), auch wenn der modus operandi aus­
sergewöhnlich anmutet: Damit das Osterfest nicht überraschend ein-
trete (sic!), solle vom Bischof verkündet werden, mit welchem Datum
und welchem Mondalter man umgehen müsse. Der übrige Klerus hätte
diese Daten zu notieren, um sie dann in der jeweils eigenen Kirche am
Weihnachtstage zusammen mit dem adäquaten Beginn der Fastenzeit in
jedem Jahr an prominenter Stelle, nämlich nach der Lesung des Evan-
geliums, dem Volke kundzutun.4 Diese immer wiederkehrenden Forde-
rungen lassen darauf schließen, dass die Zustände im Hinblick auf ein
1
Zur Herrschaft Rekkareds siehe Claude (1971), 77–91; Schäferdiek (1967), 192–
233; Scheibelreiter s.v. ‘Reccared’ in Reallexikon der Germanischen Altertumskunde 24,
200–03; sowie mit besonderem Blick auf die Bedeutung des die gotische Identität prägen-
den arianischen Bekenntnisses und der Konversion unter Rekkared zuletzt Faber (2014).
2
Im wesentlichen bewegt sich der Stand der diesbezüglichen Forschung immer
noch auf den Ergebnissen von Krusch (1884), 115–22 (II. Spanien) und Schmid (1907),
58–108. Zu diesem Fragenkomplex zuletzt Englisch (2011), 104–05.
3
Concilium Toletanum 4.5 (hg. von Vives (1963), 191): Solet in Spaniis de solemni­
tate paschali varietas exsistere praedicationis, diversa enim observantia laterculorum pascha­
lis festivitatis interdum errorem parturit. Proinde placuit, ut ante tres menses Epiphaniorum
metropolitani sacerdotes litteris se invicem inquirant, ut communi scientia edocti diem resur­
rectionis Christi et conprovincialibus suis insinuent et uno tempore celebrandum adnuntient.
4
Concilium Bracarense 2.9 (hg. von Vives (1963), 84): Placuit […] ut superventu­
rum ipsius anni Pascha, coto Kalendarum die vel cota luna debet suscipi a metropolitano
episcopo nuntietur; quod ceteri episcopi vel reliquus clerus breviculo subnotantes unusquis­
que in sua ecclesia, adveniente Natalis Domini die, adstanti populo post lectionem evangeli­
cam nuntiet, ut introitum Quadragesimae nullus ignoret.


 BRIGITTE ENGLISCH

eindeutiges Osterfest im westgotischen Spanien keineswegs als abge-

schlossen und homogen betrachtet werden müssen.
Wiewohl dieser Feststellung im Kern nicht zu widersprechen ist, soll
untersucht werden, ob dieser erste Eindruck bezüglich der Festlegung
des Osterfestes den historischen Verhältnissen entspricht. Davon, dass
man der Frage nach einem eindeutigen Osterfest keineswegs mit Gleich-
gültigkeit begegnete, sondern nachhaltig versuchte, die oben genannte
Methode umzusetzen, zeugt das Faktum, dass auf gleich 6 Konzilien von
572–691 solche oder ähnliche Anweisungen ergehen.5 Geht man den
Dingen weiter auf den Grund, stellt man rasch fest, dass das Westgo-
tenreich, neben der relativen Isolation im Land ‘hinter den Pyrenäen’, in
vielerlei Hinsicht in einer besonderen Position war, die sich mithin auch
im Osterfest widerspiegelte. Bedingt wurde diese anders geartete Ent-
wicklung zu einem erheblichen Teil durch die relativ späte Konversion

5
Hierbei handelt es sich neben dem schon erwähnten 4. Konzil von Toledo um die
2. Synode von Braga (s.u.), um das 9. Konzil von Toledo; Concilium Toletanum 9.17 (hg.
von Vives (1963), 306): Antiquitatis dein hinc ordinem saluberrime retinentes, postquam
rationem festi paschalis fraternitas vestra cognovit, noverit se anno venturo die Kalendarum
Novembrium causa peragendi concilii in hac urbe, favente Domino, congregari; um die Pro-
vinzialsynode (= 3. Konzil) von Zaragoza 1.11.691, Kanon 2 (hg. von Vives (1963), 476):
Siquidem sancta instituto patrum de sollemnitate paschali praecipiat, ut omnes confinitimi
episcopi annua vicissitudine primatum suum inquirant quo tempore Paschae festum pera­
gant; tamen quia didicimus nonnulus episcopos sententiam patrum transgredientes propter
quod etiam debuerunt usquequaque vigilantes manere, placuit omnibus nobis ut abinceps
cuncti confinitimi episcopi nullam sibi occasionem obicientes aut longinquitate itineris pra­
caventes annua recursione de festivitate paschali tempore congruo primatem suum, sub cuius
potestatis manserint regimine, inquirant; ut quo die et tempore illis Paschae festum pronunti­
averit sollicita veneratione peragant; um die Synode von Merida von 666, Kanon 6 (hg. von
Vives (1963), 329–30): Comuni deliberatione censemus et sententiae huius ordinem serva­
ndum instituimus, ut dum quisque conprovincialis episcopus metropolitani sui ammonitio­
nem acceperit pro diebus festis Nativitatis Doimini et Paschae cum eo peragendis, veniendi
ad eum ullam faciat ex cusationem. Quod si contigerit eum ab aegritudine esse detentum vel
per nimiam intenperantiam non habere qualiter ad praesentiam eius possit venire, epistolam
manu sua subscriptam dirigere debebit, in qua huius rei causam verissime notescat; und um
das 13. Konzil von Toledo von 683; Concilium Toletanum 13.8 (hg. von Vives (1963),
424–25): Hac de re nascitur et difficultas ordinibus et contemtus maioribus: et ideo si quis
episcoporum a principe vel metropolitano suo admonitus desingato sibi dierum rationabili ad
veniendum spatio sive pro festivitatibus summis, Pascha scilicet, Petecoste et Nativitate Domi­
ni celebrandis, sive pro causarum negotiis seu pro pontificibus conscrandis vel pro quibuslibet
ordinationibus principis, excepta inevitabili necessitate infirmitatis quae testibus possit con­
probari idoneis, ad constitutum diem venire distulerit, contemtorum se noverit excommuni­
catione multari; si tamen fluminum aut aerum procellosa inmensitas ad praefinitum diem
aditum ei properandi subduxerit, quod tamen et hoc ipsut convinci idoneis testibus oportebit.
Siehe hierzu auch Orlandis und Ramos-Lissón (1981), 262–71, besonders 263, die dort
die Bedeutung dieses Konzils hervorheben, dessen Akten erstmals seit 653 wieder die An-
wesenheit fast des gesamten Episkopats der hispanischen Kirche belegten.


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

zur etablierten Variante des Christentums und der sich erst im frühen
7. Jh. dezidiert stellenden Frage, wie nun das Osterest im Einklang mit
den neu errungenen katholischen Bräuchen zu feiern sei. Man muss zu
diesem Zeitpunkt bereits aufgrund der demographische heterogenen
Struktur zumindest drei in Spanien bereits vorhandene Festtagsgebräu-
che annehmen: die Variante der aus spätantiken Tradition erwachsenen
altkatholischen Christen, die Osterechnung der Arianer und den Kom-
putus des Victorius von Aquitanien.6 Hinzu treten dann noch Regio-
nen,7 über die generell wenig ausgesagt werden kann, da sie erst spät in
das Westgotenreich inkorporiert wurden, wie die von den Byzantinern
besetzten Landstriche in Südspanien oder das 585 letztendlich eroberte
Königreich der Sueben,8 zu deren Mission und Konversion zum Katho-
lizismus wesentlich Martin von Braga9 beitrug. Ein Versuch der Eini-
gung wurde zu Beginn des 7. Jh., offenkundig auf Initiative Isidors von
Sevilla, angestrebt, indem diese unterschiedlichen Varianten durch die
neue Lösung des Dionysius Exiguus ersetzt werden sollten.10 Genau dies
geschieht aber, wie die Konzilientexte belegen, allem Anschein nach
nicht,11 so dass für das westgotische Spanien eine andere Lösung, Ostern
zu bestimmen, angenommen werden muss. Dennoch muss von einer
regen Auseinandersetzung mit der Osterfestfrage ausgegangen werden,

6
Zur Bevölkerungsstruktur im westgotischen Spaniens, geprägt von einer roma-
nischen angestammten Bevölkerung, der im Kontext der Westgoten hinzukommenden
Vandalen, Alanen, Sueben und Westgoten, für die allesamt ein arianisches Bekenntnis
anzunehmen ist, aber auch immer wieder auftauchende Bezüge zum Frankenreich, wie
sie beispielsweise die Mission der suebischen Gebiete prägt oder auch das ambivalente
Gebiet Septimaniens; siehe z.B. Collins (2004), 38–81 sowie Orlandis (2006), 21–46.
Die Existenz divergenter Osterrechnung ist für das Reich der Westgoten belegt durch
Gregor von Tours, Historia Francorum 5.17 (hg. von Bruno Krusch in MGH SS rer. Me-
rov. 1,1, 215); zur arianischen Osterrechnung im Vergleich zu den anderen hispanischen
komputistischen System zuletzt Englisch (2011), 76–109.
7
Claude (1970), 75–83.
8
Claude (1970), 127–28.
9
Martin von Braga werden verschiedentlich Schriften zum Thema der Osterfestre-
chung zugeschrieben (De paschate), deren Nachweis jedoch nicht gesichert ist. Eines die-
ser Werke findet sich in der Handschrift Paris, Bibliothèque nationale de France, Lat. 609
(zu dieser siehe unten, Anm. 11), 54r–55v. Siehe hierzu Cordoliani (1956). Diese dürften
aber, selbst bei angenommener Autorenschaft nur bedingt etwas über die spanische Oster-
rechnung aussagen, da der aus Pannonien stammende Martin nicht im Auftrag der West-
goten handelte; zu seiner Lebensbeschreibung siehe Rambaud-Buhot (1968), 16–19, so-
wie Collins (1995), 80–83; Alonso-Núñez s.v. ‘Martin von Braga’ in LM 6, 343–44.
10
Englisch (2011), 106–07.
11
Dies bemerkte bereits Krusch (1884), 119–20; siehe hierzu zuletzt García Avilés
(2001), 40.


 BRIGITTE ENGLISCH

auch wenn sich die frühen Belege dieser Diskussion nur teils ungeklär-
ter Deszendenz wie der Codex Paris, Bibliothèque nationale de France,
Nouv. acq. lat. 2169,12 teils außerhalb Spaniens, im Frankenreich, erhal-
ten haben, wie in der Handschrift der Kölner Dombibliothek 83-II13
oder Paris, Bibliothèque nationale de France, Lat. 609,14 auch wenn diese
aufgrund ihrer externen Überlieferung nur noch als Zeugen dafür gelten
können, dass es diesbezügliche Bemühungen gab, diese aber nicht hin-
sichtlichen Einsatz und Rezeption geltend gemacht werden können. Es
wird mithin zu klären sein, wie diese spezielle Lösung der Osterfestfrage
aussah. Dies beinhaltet zunächst die Analyse des religiösen Deutungsge-
haltes des Auferstehungsfests in der westgotischen Gesellschaft, gerade
auch im Hinblick auf die kulturelle Situation Spaniens sowohl vor als
auch nach der arabischen Invasion von 711. Hinzu tritt die Untersu-
chung charakteristischer Datierungsgepflogenheiten bei den Westgoten,
speziell bezüglich der Einführung der Inkarnationsära, die als zentrale
Innovation des Dionysius Exiguus betrachtet und als Indikator für die
Übernahme seiner Osterrechnung15 gewertet werden kann. Schließlich

12
Wie Cordoliani (1942), 67, betont, kann aufgrund eines Vermerks fol.  2v le-
diglich ausgesagt werden, dass die Handschrift einmal der Abtei Silos gehörte. Sie er-
hält fols 5v–9v Teile des sogenannten Computus Cottonianus, ohne dass der genaue Weg
der Informationswanderung bisher regklärt worden wäre; siehe hierzu Millares Carlo
(1961), 403–04 (Nr. 138); Gómez Pallarès (1986), der dort die diesbezüglichen Hand-
schriften London, British Library, Cotton Caligula A XV; Paris, Bibliothèque nationale
de France, Nouv. acq. lat. 2169; und León, Archivo de la Catedral, 8, von der nachfol-
gend noch zu handeln sein wird, vergleichend betrachtet. Eine Transkription des Textes
findet sich bei Gómez Pallarès (1999), 67–76. Siehe hierzu ferner Cordoliani (1958),
125–26 sowie zum Computus Cottonianus und der kritischen Bewertung von Cordo-
lianis diesbezüglicher Auffassung durch die Forschungsergebnisse von Gómez Pallarès
auch García Avilés (2001), 42, sowie in deren Folge Warntjes (2011), 179–81, der daher
diese Handschrift auch nicht in seine Analyse inhaltlich miteinbezieht.
13
Zu dieser Handschrift van Euw (1998); Englisch (2002), besonders 11–17; so-
wie Warntjes (2012), 74–80.
14
Zu dieser Handschrift siehe Millares Carlo (1961), 393–94 (Nr. 121). Zur Da-
tierung siehe Alturo (1994a), 41–43, mit umfassendem Forschungsüberblick. Dass die
Handschrift dem späteren karolingischen Einflussbereich nahesteht, ist dadurch evident,
dass sie von einer westgotischen und einer fränkischen Hand, letztere in karolinigscher
Minuskel, erstellt worden ist. Über die Datierung gibt der Eintrag der Tafel (fol. 97r) zur
Festlegung der Quadragesima Aufschluss, wiewohl sich beim Jahr 815 ein Kreuz findet.
Lindsay (1915), 417, nahm an, dass sich dies nicht auf die Inkarnationsära sondern die
spansiche Ära bezog und datierten die Handschrift daher in das Jahr 777. Alturo (1994a),
43, betont, dass die Handschrift einstmals dem Kloster San Martial in Limoges gehörte.
Díaz y Díaz (1990), 308, fixiert ihren Ursprung Anfang des 9. Jh. im Norden von Cata-
lonien. Hierfür spricht auch, dass im angefügten Glossar sich in einer Glosse in ‘carolina
Narbonense’ ein Verweis zum Jahr 825 findet; siehe hierzu Alturo (1994b), 185–86.
15
So auch bereits Krusch (1884), 121.


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

sollte nach Hinweisen auf den Inhalt der speziellen Form der Osterrech-
nung in den westgotischen Reichen bis in das 11. Jh. in der schriftlichen
Überlieferung der Epoche geforscht werden, um eine Antwort auf die
Frage zu finden: Wie haben sie’s gemacht?

Zwischen Schöpfung, Ostergeschehen und

Endzeiterwartung—die Darstellung der komputistischen
Bestimmungsgrößen Sonne und Mond im Kontext der
hispanischen Ikonographie
Betrachtet man die theologischen Werke Spaniens vom 7. bis zum 11. Jh.,
so beeindruckt der intensive Bedeutungsgehalt, der den Zentralgestirnen
Sonne und Mond im Kontext der westgotischen Religiosität zugemessen
wird. Diese sind, wie der Schöpfungsteppich in der Kathedrale von Ge-
rona16 aus dem 10. Jh. zeigt, Zeugen der Schöpfung und sie sind—parallel
zum Evangelienbericht nach Luk. 23, 44–45 in Gestalt der Verdunkelung
der Sonne während der Passion und der Einbindung des Geschehens
in den Kontext des an Vollmond gebunden Passahfestes—Zeugen von
Kreuzigung und Auferstehung, wie es schon das berühmte Encolpium
von San Pere de Rodes zum Ausdruck bringt.17 Diese Art der Darstel-
lung gehört zum ikonographischen Standardrepertoire des westgotischen
Spaniens, welches nicht nur den religiösen Bereich dominiert, wie eine
Elfenbeinplatte aus dem 9. Jh. aus Narbonne18 oder ein Kapitell aus der
hochmittelalterlichen Kathedrale von Tarragona (Abbildung 1) oder in
Reduktion auf die den Zeitenlauf bewachenden Himmelskörper ein Ka-
pitell aus dem katalanischen Kloster Santa Maria d’Estany,19 beide aus
dem 11. oder 12. Jh., zeigen. Insofern nimmt es auch kaum Wunder, dass
schon in der ältesten Tradition, für die hier die asturische Kirche Santa
Maria de Quintanilla las Viñas20 aus dem 7. Jh. stehen mag, Sonne und

16
de Palol (1986), 26–27, 91–107; Abbildung unter https://es.wikipedia.org/
wiki/Tapiz_románico_de_la_Creación.
17
Vidal Álvarez (2000), 255–59.
18
Tresor de la cathédral de Saint Just Saint-Pasteur, Narbonne; siehe hierzu Camps
i Sòria (2000), 377.
19
Massons i Rabassa (2002–03), 11–37; Abbildung unter http://kw1.uni-pader-
born.de/uploads/RTEmagicC_Santa_Maria_de_l_Estany3_01.JPG.jpg.
20
Schlunk und Hauschild (1978), 94–97; Fontaine (1973), 199–238; Abbildun-
gen unter http://www.jdiezarnal.com/quintanillasol01.jpg (Sonne) und http://www.
jdiezarnal.com/quintanillaluna01.jpg (Mond).


 BRIGITTE ENGLISCH

Abbildung 1  Sonne und Mond als Zeugen der Kreuzigung, Kapitell, 11.
oder 12. Jh., Kathedrale von Tarragona (Quelle: B.E.).

Mond zu den Garanten jeglichen religiösen Geschehens werden. Ist dies

in gewisser Weise noch mit den kontinentalen Entwicklungen in Ein-
klang zu bringen, existiert jedoch noch eine weitere Assoziationsebene,
die mit dem Osterfest verknüpft ist: die des Jüngsten Gerichts.
Dieses Thema der Offenbarung des Johannes beschäftigt die west-
gotischen Theologen in nachhaltigem Maße, für die der Apokalypsen-
kommentar des Beatus von Liébana von 776 ein bedeutsames Beispiel
darstellt. Diese Affinität wird gemeinhin als Reaktion auf die arabische


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

Invasion von 711 gewertet, die einen großen Teil der westgotischen In-
telligenz auf der Flucht vor diesen apokalyptischen Reitern in die hohen
Berge A­ sturiens verdrängte.21 Diese auf den ersten Blick sehr einleuch­
tende Verknüpfung verstellt jedoch den Blick dafür, dass die Ausein-
andersetzung mit dem prophezeiten Weltende schon vorher im west-
gotischen Reich von großer Bedeutung war. Dass dies gerade auch in
Verbindung mit dem österlichen Geschehen von Relevanz war, macht-
eine Weisung des schon erwähnten 4. Toletanums von 633 deutlich,
welches die Lektüre der Apokalypse in der Osterzeit bei Strafe der Ex-
kommunikation anordnet.22 Diese besondere Betonung der Apokalypse
könnte dahingehend erläutert werden, dass die Weltzeit bei Übernahme
der konventionellen auf Eusebius zurückgehen Berechnungen im begin-
nenden 7. Jh. sich allmählich ihrem Ende näherte, und es ratsam schien,
auf diesbezügliche Zeichen zu achten.23 Diese Indikatoren des Weltendes
zeigen sich (für die hier das älteste Manuskript, der sog. Maius-Beatus
aus dem frühen 10. Jh. angeführt sei)24 insbesondere an den Zentralge-
stirnen Sonne und Mond, die so zu Garanten des mit Schöpfung und
Kreuzigung errichteten Handlungsstranges bis zu seinem Endpunkt wer-
den. Sonne und Mond avancieren in den Illustrationen zur Apokalypse
zu permanenten Zeugen des apokalyptischen Weltendes werden, sei es
bei der Öffnung des sechsten Siegels, bei der die Sonne schwarz und der
Mond wie Blut werden25 (Abbildung 2), im Maius-Beatus auf Folio 112r
als ringförmige Sonnenfinsternis und als Mondfinsternis realitätsnah ins
Bild gesetzt, oder sei es bei der Erschienung der mit der Sonne bekleideten
Frau, den Mond zu ihren Füßen Folio 152v, die den Drachenkampf des

21
Herbers (2006), besonders 102–30; Collins (1995), 225–46; Vones (1993), 35–
40; Williams (1994–2003), i 10–15; Dubler (1962); Sénac (2000), 11–12, 19–32, 85–89.
22
Concilium Toletanum 4.17 (Vives (1963), 198): Apocalypsum librum multorum
conciliorum auctoritas et synodica sanctorum praesulum Romanorum decreta Ioannis evan­
gelistae esse praescribunt, et inter divinos libros recipiendum constituerunt: et quia plurimi
sunt qui eius auctoritatem non recipiunt atque in ecclesiam Die praedicare contemnunt, si
quis cum deinceps aut non receperit aut a Pascha usque ad Pentecosten missarum tempore in
ecclesia non praedicaverit, excommunicationis sententiam habebit.
23
Zur Endzeitvorstellung im frühen Mittelalter, mit besonders mit Blick auf den
Apokalypsenkommentar des Beatus und seine Quellen siehe Frederiksen (1992), 20–37;
Matter (1992), 38–50; McGinn (1998), besonders 2–28 und 77–79; sowie zuletzt das
Überblickswerk von Palmer (2014).
24
Zur Geschichte der Handschrift siehe García Lobo (1979), 213–15 (mit aus-
führlicher Bibliographie der älteren Literatur); Williams (1991), 17–30; Williams
(1994–2003), ii 21–27. Da keine Handschrift aus der Zeit des Beatus überliefert ist,
wird hier das älteste Textzeugnis angeführt.
25
Apok. 6:12–17.


 BRIGITTE ENGLISCH

Abbildung 2  New York, Pierpont Morgan Library, 644, 112r: Beatus von Lié-
bana, Commentarius in Apocalypsin, ältestes illuminiertes Manuskript, herge-
stellt in der Provinz León, vermutlich in San Salvador de Tábara von einem Au-
tor namens Maius (daher auch oft Maius-Beatus) zwischen 940-45. Im Kontext
der Öffnung des sechsten Siegels wird die Verdunkelung der Zentralgestirne,
die das Weltende im apokalyptischen Geschehen anzeigen, realistisch als ring-
förmige Sonnenfinsternis und Mondfinsternis dargestellt.


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

Erzengels Michael ankündigt,26 um nur einige Beispiele zu präsentieren.

Die Zyklen von Sonne und Mond sind es aber, die auch das Osterfest
prägen; durch die Verzahnung mit der Apokalypse wird ihre Berechnung
doppelt bedeutsam für das Seelenheil des Einzelnen. Durch die Nachbil-
dung der kosmischen Situation von Kreuzigung und Auferstehung wird
nicht allein das neutestamentarische Geschehen in die Gegenwart über-
führt, sondern das Ganze geschieht auch mit Blick auf die Anzeichen des
geoffenbarten Endes der Welt im Jüngsten Gericht. Die adäquate Bestim-
mung des Festtermins wird angesichts dessen zur conditio sine qua non,
was die Annahme, der westgotische Klerus hätte aufgrund mangelnden
Interesses chaotische Verhältnisse zugelassen, verbietet.

Zeitermittlung und -problematik in den westgotischen

Nachfolgereichen
Daneben dürfte es aber auch noch einen höchst realen Grund gege-
ben haben, sich mit der zutreffenden Osterterminansetzung akribisch
zu beschäftigen: Einer besonderen kulturellen Situation, gemäß der die
katholischen Westgoten vor und nach dem katastrophalen Ereignis der
arabischen Invasion sich stets weiteren religiösen Konkurrenten gegen-
übersahen, vor 711 den Juden (die, und hierfür spricht die immer wieder
neu aufgelegte und intensivierte Judengesetzgebung, einen nicht uner-
heblichen Bevölkerungsanteil ausgemacht haben dürfte) und nach 711
Juden und Muslimen. Dabei handelt es sich nicht nur um im Grunde
zwei mit dem Christentum konkurrierende monotheistische Religio-
nen, sondern auch um solche, deren Kalender auf einer an der sichtbaren
Realität orientierten Mondkalender fußt.27
Besonders bedeutsam ist der jüdische Mondkalender, der das Jahr
mit der Beobachtungen des Mondneulichtes im Frühjahr, mit dem der
erste Tag des Frühlingsmonats begann, insofern als durch die Knüp-
fung des Ostergeschehens an das jüdische Passahfest—begangen am 14.
Mond des ersten Kalendermonats Nisan—28 dieser Modus Projektions-
folie und Korrektiv der eigenen Osterfestberechnung gewesen sein dürf-

26
Apok. 12:1–8.
27
Zu den unterschiedliche Kalendertypen siehe die einführenden Handbücher
Schlag (2008), 152–65, 182–83; Görke (2011), 43–82; sowie ergänzend die weiter-
führenden Studien zur jüdischen Chronologie von Mahler (1916) und zur arabischen
Chronologie von Grohmann (1966).
28
Huber (1969), 2–8.


 BRIGITTE ENGLISCH

te. Berechnungsfehler oder Ungenauigkeiten bei der Terminansetzung,

wie sie allen zyklischen Ostertabellen der Zeit aufgrund ihrer langfristi-
gen Struktur systemimmanent innewohnen, sind daher für die Christen
im Westgotenreich aus zweierlei Gründen inakzeptabel. Zum einen ist
die einzige direkte Weisung des Konzils von Nicäa das Verbot, mit den
Juden zu feiern, wie es insbesondere Isidor von Sevilla hervorhob.29
In die gleiche Richtung zielt auch der Brief des Mönches Leo an den
Archidiakon Sesuldus.30 Dieser wendet sich nicht nur einerseits auf-
grund ihrer mangelnden Eindeutigkeit gegen die Tafel des Victorius,
sondern andererseits gegen die, die Ostern an luna 14 zulassen und somit
mit den Juden feiern.31 Doch nicht nur die irreguläre Annäherung an jü-
dische Festgebräuche ist es, die wie ein Damoklesschwert über der west-
gotischen Festberechnung schwebt, sondern auch, dass der Anspruch
auf Vorherrschaft des Christentums exakt dann unglaubwürdig wird,
wenn man an dem, was die anderen Religionsgemeinschaften offenkun-
dig ohne Probleme können, nämlich das Mondalter korrekt bestimmen,
im Zusammenhang des höchsten Festes der Christenheit scheitert.
Diese Konfrontation ist für die Westgoten insofern zwingend, als die
Gesellschaft, gerade was jüdische Bevölkerungsteile, aber auch nach 711

29
Isidor, Etymologiae 6.17.10 (hg. von Oroz Reta, Marcos Casquero und Díaz y
Díaz (1993–94), i 600): Antiquitus Ecclesia pascha quarta decima luna cum Iudaeis cele­
brabat, quocumque die occurreret. Quem ritum sancti Patres in Nicaena synodo prohibu­
erunt, constituentes non solum lunam paschalem et mensem inquirere, sed etiam et diem
resurrectionis Dominicae observare. Zu den Beschlüssen des Konzils von Nicäa bezüglich
des Osterfestes siehe Englisch (2011), 87–89; tatsächlich war diese Forderung bereits
auf dem Konzil von Arles, bezeichnenderweise im Kanon 1, erhoben worden (hg. von
Mansi (1901–23), ii 471): Primo loco de observatione paschae dominici, ut uno die & uno
tempore per omnem orbem a nobis observetur, & juxta consuetudinem literas ad omnes tu
diregas; siehe hierzu auch Huber (1969), 69–75.
30
Bisher galt der Text aus der Handschrift 83-II der Kölner Dombibliothek,
fols 184r–185v komplett belegt; siehe hierzu Krusch (1880), 298–302, mit der hier zi-
tierten Edition des Textes sowie Schmid (1907), 94–95. Unlängst wurde jedoch eine
weitere Abschrift des Briefes in der Handschrift Bremen, Universitätsbibliothek, msc
0041 ausfindig gemacht; siehe hierzu Warntjes (2010c).
31
Leo monachus, Epistola ad Sesuldum 4, 6 (hg. von Krusch (1880), 300–02): Sed
vulgus omne in commune, qui in usu est conpotum, diversis inter se discrepant modis, et
alii lune cursum utpota XX adiecerunt, alii XXI, alii XXII. […]  Ex quo pervenit illud,
quod dictum quoque nefas est, ut, dum XXI. luna in pascha esse contingerit, illis fallentibus,
XXII. aut XX. […] Iam si Ebreorum pascha dominico die provenerit, celebrare nos in uno
eodemque die non convenit, cum presumentes ista damnabile sunt notati et XIIII. canoni­
cum institutione vocati. Möglicherweise wollte man Anfang des 7. Jh. nicht mehr offiziell
zu noch bestehenden arianischen Gruppen respektive Traditionen Stellung beziehen,
musste sich ihrer aber im Kontext der Osterrechnung annehmen; zur Sache siehe, wenn
auch ohne Bezugnahme auf eine arianische Osterrechnung, Krusch (1884), 120.


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

arabische Einflüsse angeht, trotz aller Restriktionen offenkundig un-

geheur durchlässig war. So kann trotz der strengen Judengesetzgebung
Julian von Toledo, ein konvertierter Jude, unter König Wamba zum Pri-
mas Spaniens aufsteigen.32 Von der regen jüdischen Präsenz auch in den
westgotischen Nachfolgereichen, gerade auch in Verbindung mit für
alle Augen sichtbaren Datierungen, künden viele Grabsteine, wie das
Beispiel einer auf 1026 datierte hebräische Grabinschrift aus León bele-
gen, die das Grab eines Mar Yaáqob bezeichnet, der am 6. Wochentag,
am 20. Tag des Monats Tammuz, im Jahr 786 entweder auf der Straße
nach Sahagún oder Santiago einen gewaltsamen Tod gefunden habe.33
Zudem kommt es immer wieder zu Kontakten zum arabischen Spa-
nien, vornehmlich da dort die Christen unbehindert ihre Religion aus-
üben konnten, und folglich in das öffentliche Leben der arabischen Welt
und damit auch ihrer Art der Datierung eingebunden waren. Hiervon
zeugt nicht nur der Sachverhalt, dass im frühen 8. Jh. vor dem Bau der
Mezquita in Cordoba Muslime und Christen sich die dortige Haupt-
kirche des hl. Vinzenz teilten und wechselseitig benutzten.34 Auch die
Nachfolgeauseinandersetzung hinsichtlich des Primats über Spanien
zwischen dem uns schon durch seinen Apokalypsenkommentar bekann-
ten Beatus von Liébana für Asturien, Migetius für Cordoba und Bischof
Elipandus von Toledo für das arabisch besetzte Spanien, besser als sog.
Adoptianismusstreit bekannt, zeugen von einem lebendigen Christen-
tum im islamisch dominierten Teil Spaniens,35 wobei sicherlich, ange-
sichts seiner integrativen und identitätsstiftenden Bedeutung leicht er-
klärbar, auch über das Osterfest gehandelt wird.36 Dass es sich dabei um
des arabischen Kalenders und der Bestimmung des zutreffenden Mon-
dalters kundigen Christen resp. Mönche handelt, mag das Beispiel des

32
Bronisch (2005a), zu Julian von Toledo 102–05.
33
Arbeiter und Noack-Haley (1999), 41, mit deutscher Übersetzung der hebräi-
schen Inschrift: ‘Dies ist das Grab von Mar Yaáqob […] der auf der Straße nach S. (Saha-
gun oder Santiago) ermordet wurde. Jahwe möge sein Blut rächen! […] Und er verschied
am sechsten Wochentag, am 20. Tag des Monats Tammuz des Jahres 786 und starb im
Alter von 45 Jahren.’
34
Sanchez Velasco (2000), 9–26.
35
Immer noch lesenswert ist die Studie von de Abadal y de Vinyals (1949). Siehe
hierzu ferner Schäferdiek (1969–70); Heil (1965). Zur Geschichte des Adoptianismus
siehe zuletzt Nagel (1998), 21–138. Das beste Zeugnis für die Position des Erzbischofs
von Toledo bildet die Epistola ad Migetium des Elipandus (hg. von Gil (1973), i 67–78);
für die Position der Gegenpartei, siehe Beatus von Liébana und Eterius von Osma, Ad­
versus Elipandum libri duo (hg. von Bengt Löfstedt in CCCM 59).
36
Krusch (1884), 121–22; Schäferdiek (1969–70), 382–84.


 BRIGITTE ENGLISCH

Bischofs von Elvira, Reccesmunth alias Rabi ibn Zaid belegen, der den
sogenannten Mozarabischen Kalender von 961 verfasste.37 Diese ‘Moza-
raber’ prägten die nördlichen Reiche verstärkt seit dem 9. Jh., als nach
der die Lebensbedingungen zeitweise verschlechternden Episode der
freiwilligen Märtyrer von Cordoba 858 größere monastische Einheiten
nach Norden flohen und von Alfons III. und seinen Nachfolgern in den
im Zuge der Repoblación restaurierten Klöstern im Gebiet des Duero
und rund um die neue Hauptstadt León angesiedelt wurden.38 Diese
wurden dann zu den prägenden intellektuellen Zentren, vermutlich den
Denkzentralen auch hinsichtlich des Osterfestes, für die hier stellvertre-
tend die Klöster Sahagún, Tábara und San Miguel de Escalda zu nennen
sind, darunter also das Kloster, in dem der eingangs vorgestellte Mai-
us-Beatus hergestellt wurde. Und trotz der Ablehnung der islamischen
Invasoren von offizieller Seite, im Kontext der intellektuellen Reflexi-
on, wie in der Chronik von 754,39 zeigen gerade die Emigranten, sogar
Geistliche einen relativ unbefangenen Umgang mit Traditionen der
muslimischen Welt. So enthält die Unterfertigung einer im Jahre 931
ausgestellten Schenkungsurkunde für Peñalba de Santiago eine Zeugen-
liste, die davon spricht, dass die meisten der Signierenden der arabischen
Kulturwelt entstammten und dies auch gar nicht zu verbergen suchten:
Didacus Ibenfroila, Zuar ibn Mohaiscar, Zuloeiman Ibenapelia, Abomar
Handinit, Alvaro Ibnzalem, Mahacer Zibalur und Pelagius Presbyter
Ibanzaute.40
Angesichts dieser direkten Konfrontation nimmt es nicht Wunder,
dass die Entscheidung zwischen den einzelnen Osterregelungen, die im
7. Jh. zur Verfügung standen, keineswegs einfach war. Auf den Konzilien
der Westgotenzeit ist, auch in den Fällen, in denen keine dezidierte An-
weisung erging, die Festlegung des Ostertermins und die Bestimmung
des Mondalters immer wieder ein Thema gewesen. So erscheint es als
keineswegs zufällig, dass seit der die Innenpolitik nachhaltig ordnenden
Herrschaft Chindaswinths, also seit 642 in jedem der Fälle, in dem Vic-
37
Arbeiter und Noack-Haley (1999), 46. Text bei Férotin (1904), XXXIII–
XXXIV, 305–52.
38
Zu den Märtyrern von Cordóba siehe Franke (1958); Pochoshajew (2007);
Monferrer Sala (2004); Cutler (1965).
39
Die Mozarabische Chronik von 754 ist hg. von Gil (1973), i 15–54; eine Ein-
führung und englische Übersetzung findet sich in Wolf (1999), 25–42, 111–60; siehe
auch Arbeiter und Noack-Haley (1999), 46.
40
Arbeiter und Noack-Haley (1999), 46, 296. Die Verfasser kommen dort zu dem
Schluss: ‘Es kann keinerlei Zweifel daran bestehen, daß in diesen kulturell maßgeblichen
Kreisen die mozarabische Präsenz besonders spürbar wurde.’


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

torius ein Doppeldatum präsentiert, ein Konzil in zeitlicher Nähe abge-

halten wird (Tabelle 1).41

Jahre mit Doppeldatum Zeitnahes westgotisches Konzil

bei Victorius
645 7. Konzil von Toledo, 18.11.646,
655 10. Konzil von Toledo, 1.2. 656,
675 11. Konzil von Toledo, 7.11.675,
685 14. Konzil von Toledo, 14.-20.11.684,
689 15. Konzil von Toledo, 11.5. 688,
692 Synode von Zaragoza 1.11.691
693 16. Konzil von Toledo, 25.4.-2.5. 693

Tabelle 1  Korrespondenz von Doppeldaten bei Victorius und westgotischen

Konzilien.

Ebensowenig als Produkt der Willkür erweist sich das 12. Konzil von
Toledo, welches die unrechtmäßige Thronfolge Erwigs legitimieren soll-
te. Dieses dauerte vom 9.1.–25.1.68142 und umfasste damit exakt eine
halbe Lunation vom Vollmond, der Donnerstag, den 10.1.681, eintrat,
bis zum Neumond, am Freitag, den 25.1.681, so als wolle man durch die
akribische Dokumentation der Beobachtung des Mondlaufes die Ein-
heit mit dem kosmischen Weltenplan in einer politisch heiklen Situa-
tion dokumentieren oder auch der Gefahr vorbeugen, parallel zu einem
illegitimen Herrscher auch noch einen sich in einer fehlerhaften Mond­
alterbestimmung zeigenden illegitimen Ostertermin zu haben. Ein
ähnlicher Fall liegt beim 16. Konzil von Toledo vor, welches am 25.4.,
zugleich letztmöglicher (alexandrinisch-dionysischer) Ostertermin und
Vollmond, beginnt, und an dem für die Osterrechnung der Westgoten
41
Zu den Daten siehe Vives (1963), 249–59 (7. Konzil von Toledo), 308–24 (10.
Konzil von Toledo), 344–69 (11. Konzil von Toledo), 441–48 (14. Konzil von Toledo),
449–74 (15. Konzil von Toledo), 475–81 (Synode von Zaragoza), 482–522 (16. Konzil
von Toledo).
42
Die Angaben zur Dauer des Konzils lassen sich ableiten aus der einleitenden Pas-
sage zum Concilium Toletanum 13 (hg. von Vives (1963), 411; dann die weiteren An-
gaben 414, 440): In nomine Domini incipiunt gesta synodalia habita in urbe toletana sub
era DCCXXIa anno regni excellentissimi Ervigii principis quarto. […] in ecclesiam videlicet
sanctorum apostolorum petri et Pauli anno regni quarto serenissimo Ervigii principis sub die
nonas novembres era 721. […] Datum sub die pridie nonas novembres anno feliciter quarto
regni gloriae nostrae in Dei nomine Toleto. […] edita lex in confirmatione concilii sub die idus
Novembris era DCCXXI anno quoque feliciter quarto regni gloriae nostrae in Dei nomine.


 BRIGITTE ENGLISCH

wichtigen 2.5. endet,43 ein Faktum, was im Kontext der westgotischen

Komputistik noch bedeutsam sein wird. Und auch die Feststellung von
Lunation (luna 11) und Zyklus (lunae cursum IV) bezüglich der pro-
grammatischen Weihe der dem Apostel Jakobus geweihten Kirche, über-
liefert im Kontext der Aufzeichnungen Konzils von Oviedo 873 durch
Alfons III., dürfte in den Osterkontext einzureihen sein.44
Für die tatsächliche Bestimmung des Auferstehungsfestes ist zudem
von Bedeutung, dass die Majorität der Konzilien im Mai oder Oktober
stattfand, ein Tatbestand auf den ebenfalls noch zurückzukommen sein
wird.
Dass sich zumindest die Osterrechnung nach Dionysius Exiguus trotz
mehrfacher Versuche, sie zu etablieren nicht durchsetzte, kann anhand

43
Auch hier finden sich die zeitlichen Angaben bezüglich der Dauer in der Vorre-
de zum Text des Concilium Toletanum 13 (hg. von Vives (1963), 482; dann die weiteren
Angaben 488, 518): In Nomine Domini Iesu Christi incipit Synodus Toletana XVIa Dum
sexto anno inclyti et orthodoxi domni et princeps nostri Egicani sub die VI nonas Maias
era DCCXXX unanimitatis nostrae conventus in praetoriensis baselica sanctorum videlicet
Petri et Pauli adgregatus consisteret atque unusquisque nostorum ex more secundum ordi­
nationis suae tempus in locis debitis resideret. […] Datum sub die VII Kalendas maias anno
feliciter VIo regni mansuetudinis nostrae Toleto. […] Datum sub die Kalendas maias anno
feliciter VIo regni nostri, in Dei nomine sedis nostrae Toleto.
44
Das Problem bei der Bewertung dieser Quelle ist, wie für alle auf die frühe Santia-
go-Geschichte rekurrierenden Texte, die gerade bezüglich der Datierung komplexe Über-
lieferungslage. So liefert die Edition des Concilium Ovetense von Mansi (1901–23), xviiA
265–66, audiendum verbum Domini in prima die, quot erat nonas Maii anno incarnationis
Domini era DCCCCLXXIX secunda feria deducebat animum ad lunae cursum IV. Luna
XI consecratum est. Das Faktum, dass dieses deutlich später als das Konzil von Oviedo 873
hier Berücksichtigung findet, mag darin liegen, dass auf besagtem Konzil der Entschluss
zum Neubau der Jakobuskirche gefasst wurde. Jedoch verändert bereits die Parallelüberlie-
ferung der von Mansi herangezogenen Chronik des Sampirus aus dem frühen 11. Jh. (die
wichtigste historiographische Überlieferung für die Region nach dem Ende der Crónica
albeldense 883) die Lesart in: In prima die quod erat Nonis Mai, anno Incarnationis Domini,
Era DCCCCXXXVII, lässt dabei aber auch Wochentag und Mondalter nicht außer Acht,
welche nun als Secunda feria deducebat, annum ad Lunae cursum III., Lunae XI. Erwäh-
nung finden; siehe hierzu Sampirus von Astorga, Chronicon regum Legionis (hg. von de
Ferreras (1700–27), xvi 31; auch de la Fuente (1855–59), i 510–11). Wiewohl man folg-
lich über das tatsächliche Weihedatum diskutieren kann, bedeutsam ist, dass stets auch das
Mondalter und der Mondzyklus nicht nur Erwähnung fand, sondern bei einem veränder-
ten Datum auch dieser ‘lunare Datierungsparameter’ akribisch berücksichtigt wurde, was
auf dessen besonderen Stellenwert, zumal im Kontext der forcierten Jakobus-Verehrung
hindeutet. Interessant ist, dass der 7.5.899 tatsächlich ein Montag war, jedoch das sichtbare
Mondalter mit dem Neumond am 13.5. und mithin luna 25 am 8.5. markant von dem
von Sampiro vorgefunden Wert abweicht. Jedoch findet im Vorjahr 898 der (sichtbare)
Vollmond am 9.5. statt, was angesichts einer gewissen Beobachtungsvarianz durchaus zu
luna 11 am 7.5. geführt haben könnte. Zudem sei darauf hingewiesen, dass diese Parame-
ter nicht aus der dionysisch-bedanischen Osterrechnung stammen können, die für 899 die
Epakte 6 aufweisen und mithin den Ostervollmond am Freitag, den 30. März.


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

der Datierungsgepflogeheiten der Westgoten dahingehend belegt wer-

den, als die Inkarnationsära in praxi nahezu unbeachtet bleibt. Diese
hatte sich im weltlichen Bereich zwar auch bei den Franken erst unter
Ludwig dem Frommen in den Urkunden durchgesetzt,45 im Rahmen der
komputistischen Schriften im Gefolge von Dionysius Exiguus und Beda
war die Inkarnation aber schon viel früher von zentraler Bedeutung, als
sie dort in der Weltzeitalterberechnung die ältere Variante, die die Tren-
nungslinie zwischen den letzten Weltzeitaltern an der Kreuzigung / Auf-
erstehung ausrichtete, durch die Inkarnation ablöste.46 Im Westgoten-
reich bleibt dies fast gänzlich aus, was sich schon bei der für die folgenden
Jahrhunderte prägende klerikale Bildungsschicht, repräsentiert durch
Isidor von Sevilla und Julian von Toledo, zeigt. So definiert Isidor47 in
seiner Chronik zwar, im Einklang mit der Überlieferung durch Eusebi-
us-Hieronymus, die Geburt Christi für das 42. Jahr des Octavian.48 Und
seine Kreuzigung im 18. Jahr des Tiberius, jedoch ohne Bezugnahme auf
ein Weltzeitalter, lediglich unter Angabe der Jahre seit der Schöpfung.
Wie schwierig es war, die Auffassung von der Zeitenwende ausgerichtet
an der Fleischwerdung des Herrn zu übernehmen, demonstriert Julian
von Toledo. Dieser sucht in seinem De comprobatione aetatis sextae die
Inkarnation durch die Assoziation mit dem 41. Jahr des Tiberius zeitlich
zu definieren und die Relation zur spanischen Ära herzustellen; jedoch
bildet bei ihm der Tod Cäsars und die Herrschaft von Octavian Augustus
die Trennlinie zwischen dem 5. und dem 6. Weltzeitalter.49

45
Declercq (2000a), 186–90.
46
Bestes Beispiel für eine solche Form der Weltärenberechnung ist Gregor von
Tours, Historia Francorum 10.31 (MGH SS rer. Merov. 1,1, 537), der dort nicht die Ge-
burt, sondern die Auferstehung als Merkmal für die Epochentrennung wählt: Ab e hoc
maris transitu usque ad resurrectionem dominicam anni MDXXXVIII. A  resurrectione
dominica usque ad transitum sancti Martini anni CCCCXII.
47
In den Etymologiae enthielt er sich jedoch einer eindeutigen zeitlichen Zuord-
nung und begnügte sich mit der Feststellung, dass Christus in der Herrschaft Octavians
geboren worden sei und zu Zeiten des Tiberius gekreuzigte wurde: Isidor von Sevilla,
Etymologiae 5.39 (Oroz Reta, Marcos Casquero, and Díaz y Díaz (1993–94), i 560):
Sexta aetas. 26 Octavianus ann. LVI. Christus nascitur. VMCCX. Tiberius ann. XXIII.
Christus cruci figitur. VMCCXXXIII.
48
Isidor von Sevilla, Chronica 1.235–39 (hg. von José Carlos Martin in CCSL
112, 112, 114): 235 Octauius Augustus regnavit annos LVI. […] 237 et cessante regno a
sacerdotio Iudaeorum Dominus Xristus ex Virgine nascitur anno regni eius quadragesimo
secundo. […] 238 Tiberius filius augusti regnavit annos XXIII. 239 Huius octauo decimo
anno Dominus crucifixus est peractis a principio mundi annis V ̅CCXXVIII.
49
Julian von Toledo, De comprobatione aetatis sextae 3.34 (PL 96, 568): Jam vero
residuus annorum numerus a tempore nativiatis Christi usque in praesens in promptu est uni­
cuique, et scire si volet, et supputare si placet, assumptis videlicet annnis secundum aeram ad


 BRIGITTE ENGLISCH

In Spanien datiert man indes durchgängig bis weit in das 12. Jh. hin­
ein nach der spanischen Ära, die, ohne den Ausgangspunkt in Gestalt
des Epochentags konkret zu bestimmen, 38 Jahre vor der Inkarnation
einsetzt.50 Dennoch war es diese spanische Ära, die das langfristige Da-
tierungsgerüst für die nachfolgenden Jahrhunderte bereit stellte.
Im Allgemeinen wurde, wie Urkunden aber auch Inschriften belegen,
noch im 9. Jh. ausschließlich das Tagesdatum und die Ärenzählung her-
angezogen. So lautet die Unterfertigung einer Urkunde von Alfons III.
vom 20. August 883: Facta scriptura concessionis et confirmationis: sub
die XIII kalendas septembris luna X. Era DCCCCXXI. Adefonsus rex.51
In gleicher Weise datiert derselbe Herrscher in der von ihm errichteten
Kirche San Salvador der Valdedios: Sub era DCCCCXXX prima die
XVIo klds octobris.52 Dass dies keineswegs eine allein offiziöse Datierung
war belegt die Vielzahl der Grabinschriften, für die hier der Epitaph der
Nonne Mumadonna, der das Todesdatum mit den 12. Kl. Nov. in der
Ära 988 (= 950 n. Chr.) ohne Bezugnahme auf die alternative Inkarnati-
onsdatierung angibt, als Beispiel angeführt sei.53
Der Umgang ausschließlich mit der spanischen Ära war ein so so-
lides Datierungsfundament, dass man im Laufe der Zeit vielfach dazu
überging, das jeweilige Jahr in Gestalt einer Rechenaufgabe vorzufüh-
ren, so im schon erwähnten Maius-Beatus, in dem das Entstehungsjahr
Ära 964 als duo gemina ter terna centiese ter dena bina angegeben wird,
wobei nicht unterschlagen werden sollte, dass in der Beatusforschung
diese Datierung in höchst vielfältiger Weise aufgelöst und so ein Zeit-
raum von 894–952 abgeleitet wird.54

ipsa Domini incarnatione. Aera enim inventa est triginta et octo annos, quam Christus nasce­
retur. Nunc autem acclamatur aeram esse 724. Detractis igitur triginta et octo annis, ex quo
aera inventa est, usque ad nativitatem Christi, residui sunt 686 anni. Krusch (1884), 122,
sieht dies als frühen Beleg für die Auseiandersetzung mit der Inkarnationsära in Spanien,
der dann aber als zu singulär zu betrachten wäre, um als wirklich folgenreich zu gelten.
50
Bereits für Isidor stellte sich dieses Problem, der in den Etymologiae eine Volks-
zählung des Kaisers Augustus als diesbezüglich epochemachendes Ereignis anführt; aus-
gehend von dieser Volkszählung sei in jedem Jahr eine Steuer in Form von Erz zu entrich-
ten gewesen, wobei Isidor Ära von aeris = ‘Erz’ herleitet. Isidor von Sevilla, Etymologiae
5.36.4 (Oroz Reta, Marcos Casquero und Díaz y Díaz (1993–94), i 548): Aera singulorum
annorum est constituta a Caesare Augusto, quando primum censu exagitato Romanum or­
bem descripsit. Dicta autem aera ex eo, quod omnis orbis aes reddere professus est reipublicae.
51
Floriano (1949–51), ii 144.
52
Arbeiter und Noack-Haley (1999), 194.
53
Arbeiter und Noack-Haley (1999), 40.
54
Diese löste bereits Lowe (1910), 65 (Nr. 43), als 894  n.  Chr. (3×3×100 und
3×10+2, d.h. 932 der spanischen Ära) auf. Unter Einbeziehung von duo gemina ge-


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

In diesem System fungiert allem Anschein nach das Jahr 1000 als
Wendemarke der Ärenzählung, worauf die Memorialinschrift im Berg-
kloster Penalba de Santiago hinweist. Diese lautet:55

Annum centenum duc, septies adito s(err. D) enum mille

quibus sociis que fuit era scies tertio decimo kalendas ju-
lii obiit stephanus abbas, era CLXX

‘Nimm das Jahr hundert, füge sieben mal zehn hinzu, füge 1000 hinzu,
und du wirst wissen, in welcher Ära wir uns befinden [1170=1132]. An
den 13. Kl. d. Juli, in der Ära 170 ist Abt Stephanus gestorben. [1000
fehlt].’

Dass parallel dazu das Bewusstsein wuchs, sich im letzten Weltzeitalter

zu befinden, dass also das Ende der Welt nahe sei, macht die Weihein-
schrift der Kirche Santa Cruz in Cangas de Onis deutlich, die auf das
Jahr 737 zurückreicht, wobei das für das westgotische Selbstverständnis
traumatische Erleben der arabischen Eroberung sicherlich eine nicht un-
erhebliche Rolle gespielt haben dürfte.56 Diese Inschrift ist insofern be-
deutsam, als das hier genannte Tagesdatum des 300. Tages des Jahres 775
der spanischen Ära (= 737 n. Chr.) auch einen möglichen Hinweis auf
den, ansonsten bislang unbekannten Epochentag der spanischen Ära ge-
ben könnte. Zählt man nämlich vom antik-römischen bürgerlichen Jahr-
beginn, dem 1. März 300 Tage vorwärts landet man beim 25. Dezember;
die Weihe einer Kirche, die der göttlichen Intervention im Kampf gegen
die Mauren unter dem Symbol des Kreuzes gewidmet war, am Tag der
Geburt des Herrn, scheint einer gewissen Strahlkraft nicht zu entbehren.
Fragt man nach Belegen für die allmähliche Übernahme der Inkar-
nationsära, so scheint es immer wieder Phasen gegeben zu haben, die
die Inkarnationsära durchzusetzen; dass dies in praxi bis in das 12. Jh.

langte Gómez-Moreno (1919), 131 Anm. 2, auf Ära 964 (2+2=4; 3×3×100=900;
3×10×2=60), also das Inkarnationsjahr 926. Neuss (1931), 13–14, schließt das duo ge­
mina aus und gelangt so zu dem Wert 960 resp. 922 n. Chr. Camón Aznar (1960), 32,
ermittelt aus dem Kolophon indes das Jahr 952 n. Chr.; in Camón Aznar (1975), 93,
korrigierte er dieses Jahr indes in 958 n. Chr. (332×3=996). Siehe hierzu auch Williams
(1994–2003), ii 24.
55
Martínez Fuertes (1936), 109–13; die Transkription dieser Passage der Inschrift
findet sich 111.
56
Arbeiter und Noack-Haley (1999), 98; die Datierung am Ende des Textes lau-
tet: Hic vate Asterio sacrata sunt altaria Christo diei revolutioni temporis
anni ccc seculi etate porrecta oer hordinem sexta currente era septingen­
tesima septagesima quintaque.


 BRIGITTE ENGLISCH

nie wirklich gelang, machen die vielen Datierungsfehler deutlich, die

immer dann auftreten, wenn die Zählung in Jahren nach Christi Geburt
ins Spiel kommt. Hierfür finden sich Inschriftenbeispiele wie der Grab-
stein für den vorletzten König der Westgoten, Witiza (702–10), der
posthum in Narbonne im 9. Jh. angefertigt wurde.57 Dieser enthält be-
zeichnenderweise nicht nur eine dreifache Datierung, nämlich Ära 938,
das Inkarnationsjahr 890 und das 2. Regierungsjahr Karls III., sondern
auch just das Inkarnationsjahr falsch mit einem um 10 Jahre zu geringen
Wert.58 Ein reiner Flüchtigkeitsfehler scheint aber bei einem doch eher
arbeitsaufwendigen Grabstein eher wenig wahrscheinlich.
Trotz dieser nur zögerlichen Annäherung an die im restlichen Europa
im 9. Jh. etablierten Osterrechnungen respektive deren Parameter, dürfte
das Bedürfnis nach einer zutreffenden Bestimmung des Osterfestes auch
in Spanien von zentraler Bedeutung gewesen sein, insbesondere für Herr-
scher, die die Machtkonsolidierung Asturiens vorantreiben wollten. Für
diese Strömung steht wie kein zweiter Alfons III.59 Dieser Herrscher kann
in mehrfacher Hinsicht als prägend angesehen werden, besonders hinsicht-
lich einer religiöse Neuausrichtung der neogotischen Reiche.60 Insofern
nimmt es auch kaum Wunder, dass unter seiner Herrschaft einige für die
westgotischen Reiche programmatischen Manuskripte entstehen, nämlich
zur Historiographie (die berühmte Chronik von Albelda: El Escorial, Real
Biblioteca de San Lorenzo, d.I.1)61, zur Rechtstradition (El Escorial, Real
Biblioteca de San Lorenzo, d.I.2)62 und zur christlichen Zeiterfassung und

57
Museu d’Historia de Barcelona 10051; siehe hierzu Arbeiter und Noack-Haley
(1999), 69, Abb. 31; die Inschrift lautet: Hic Requiescat VVitiza, Filius Teodere­
di Dimittat Ei Deus Amen. Era DCCCXXXVIII, ab Incarnatione Domini Anni
DCCCXC, Anno II Regnante Karulo Rege, Die XIII Kalendas Aprelis sic obiit.
58
Arbeiter und Noack-Haley (1999), 69.
59
Vones (1993), 35–40; García Toraño (1986), 283−341; Sánchez-Albornoz
(1972–75), iii 493−962.
60
Bronisch (2005b), 182–87.
61
Ediert wurde die Handschrift mehrfach, so als Die Chronik Alfons’ III. (Prelog
(1980)), Chroniques asturiennes (Bonnaz (1987)), Crónicas asturianas (Gil Fernández
(1985)). Hierbei handelt es sich um den Codex Aemilianensis, aus dem Kloster San
Millán de la Cogolla aus dem 10. Jh, der die sog. Chronik Alfons III., eine Beschreibung
der westgotischen Geschichte von ihren Ursprüngen bis zu Alfons III. in ihrer vollstän-
digste Abschrift des ursprünglichen Werkes, enthält.
62
Die Chronik erfährt im sogenannten Codex Albeldensis seu Vigilanus aus dem 10.
Jh., entstanden im Kloster San Martín de Albelda, eine weitere Redaktion und wird bis
zum Jahr 976 fortgeführt. Insgesamt handelt es sich bei dieser durch den Mönch Vigila und
seine Assistenten Sarracino und García erstellten Sammelschrift aber um ein beeindruk­
kendes Kompendium kanonischen und zivilen Rechts, welches u.a. den berühmten Liber


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

zur Liturgie (das berühmte Antiphonar von León, heute León, Archivo de
la Catedral, 8),63 eine auf die Zeit König Wambas zurückgehende Gottes-
dienstordnung. Diese Schriften eint nicht nur das Ziel, die sie charakteris-
tischen Elemente mit einer Norm zu versehen, sondern auch das Faktum,
dass alle drei Handschriften daneben einen Kalender und diverse kompu-
tistische Schriften enthalten.64 Diese sind bislang, aufgrund ihrer gänzlich
von der abendländischen Komputistik abweichenden Gestaltung, nicht
als kohärente, eigenständige Form der Festberechnung erkannt worden.65
Dass sie aber, mit geringen Variationen, just in allen drei Manuskripten
enthalten sind, mag verdeutlichen, als wie bedeutsam die Verbreitung ei-
ner einheitlichen Osterrechnung im Reiche Alfons III. angesehen wurde.
Es war nicht möglich, eine neogotische Ordnung zur Geschichte, Recht
und Liturgie durchzusetzen, wenn nicht die zentrale Grundlage geschaf-
fen war: das Osterfest eindeutig und im Einklang mit den nicäischen Wei-
sungen zu definieren, welches allein die Erlösungsgewissheit im Angesicht
des drohenden Endes des 6. Weltzeitalters beinhaltete.

Das Antiphonar von León—Liturgie und Komputistik

unter Alfons III
Eine dieser Berechnungsformen soll als Beispiel hierfür herangezogen
werden. Diese ist dem sogenannten Antiphonar von León vorangestellt,

Iudiciorum (hg. von Karl Zeumer in MGH LL nat. Germ. 1, 33–456), die westgotische
Gesetzessammlung aus der Zeit König Rekkeswinths, enthält. Zu beiden Handschriften
siehe ferner Antolin (1910–23), i 320–68 (d.I.1) und 368–404 (d.I.2 / Codex Vigilanus).
63
Millares Carlo (1961), 362–63; Zapke (2007), besonders 252; beide mit umfas-
sender Bibliographie. Einen Überblick über die Handschrift vermittelt Alvarez Miran-
da (1928). Die Handschrift ist inzwischen in hervorragender Qualität komplett im
Internet verfügbar unter der von staatlicher spanischer Seite eingerichteten Biblioteca
Virtual del Patrimonio Bibliográfico http://bvpb.mcu.es/es/catalogo_imagenes/grupo.
cmd?path=26408, die damit die älteren Faksimile-Editionen ersetzt; Instituto de Histo-
ria Eclesiástica Padre Enrique Flórez Madrid (1953); Brou und Vives (1959).
64
Zu den wenigen Studien, die sich bislang diesem Thema partiell widmeten,
gehören die vergleichende Analyse der Handschriften des Escorial d.I.1 und d.I.2 von
Cordoliani (1951) und die von Joan Gómez Pallarès in einem Sammelband (1999)
zusammengefassten Einzeluntersuchungen u.a zu León, Archivo de la Catedral, 8. Die
Studien konzentrieren sich jedoch vornehmlich auf die Bestimmung bereits bekannter
komputistischer Texte innerhalb der Handschrift und weniger auf die Analyse der west-
gotisch-neogotischen Elemente der Osterfestberechnung, wie sie nachfolgend im Blick-
punkt stehen sollen; siehe hierzu z.B. Gómez Pallarès (1989–89), besonders 77–91.
65
Die diesbezüglichen Texte werden derzeit von mir bearbeit und in einer separa-
ten umfassenden Studie in nächster Zeit veröffentlicht werden.


 BRIGITTE ENGLISCH

welches Anfang des 10. Jh. für Abt Ikila (León, c. 917 oder 970) her-
gestellt wurde und damit in seinen älteren Teilen vermutlich genau in
die Zeit Alfons III. fällt;66 hierauf nimmt auch die Eingangsillustration
Bezug (León, Archivo de la Catedral, 8, 1v), die die Widmung des An-
tiphonars an Abt Ikila zeigt (Abbildung 3). Diese Handschrift ist nicht
nur insofern besonders interessant für die frühmittelalterlich-westgo-
tische Zeitauffassung, als sie bislang noch nicht in einem umfassenden
Zusammenhang erkannt und analysiert worden ist. Je nach Intention ist
bislang allein ihre Bedeutung als Musikhandschrift,67 als bedeutender
Repräsentant frühmittelalterlicher Kalenderwerke68 oder auch als kom-
putistische Handschrift, als Überlieferung des Computus Cottonianus,69
behandelt worden. Was jedoch völlig offen ist, ist die Zusammenfügung
all dieser Bestandteile in einem umfassenden Kontext wie auch die Un-
tersuchung aller Elemente der Handschrift. So erscheint es als keines-
wegs zufällig, dass alle Teile des Manuskriptes sich unter dem Aspekt
des Gottesdienstes respektive der Liturgie subsumieren lassen,70 eine
Notwendigkeit für Alfons III. und seine Nachfolger71 zur Durchsetzung
seines Konzeptes einer westgotischen Renaissance, besonders wenn
man die ehemals besondere Position der westgotischen Herrscher auch
im kirchlichen Bereich72 mit in die Überlegungen einbezieht. Insofern
nimmt es auch kaum Wunder, dass das Manuskript bei genauerer Be-
trachtung nicht nur einen Komputus erhält,73 sondern gleich drei Anlei-
tungen das Osterfest zu bestimmen, die keineswegs deckungsgleich, son-

66
Die exakte Datierung der Handschrift ist nach wie vor ungeklärt, da die Hand-
schrift zum einen in mehreren Stufen erstellt wurde und auch für die älteren Teile anzu-
nehmen ist, dass sie Abschriften von Vorgängerhandschriften sind. Üblicherweise wird
das Manuskript in das frühe / mittlere 10. Jh. datiert, mit Ergänzungen aus dem späteren
10. und 11. Jh.; siehe hierzu Zapke (2007), 252 mit umfassender Literatur; ferner Des-
warte (2013), 67–84; Díaz y Díaz (2007), 93–112; Díaz y Díaz (1983), 308–09, 390–91.
67
Zapke (2007), 252.
68
Férotin (1904), XXXII, 305–52 (D); Borst (2001), 170–72; Vives Gatell und
Fábrega Grau (1949), 344–47 (Nr. 7).
69
S.o., Anm. 12.
70
Zu den Aspekten der Liturgie bei den Westgoten und auch in den nordspani-
schen Reichen siehe Gómez-Ruiz (2007), besonders 45–67.
71
Siehe hierzu Bronisch (1998), 124–27; Valdeon Baruque (1994).
72
Zur Position des Herrschers, insbesondere in Bezug auf die Kirche, siehe Claude
(1971); Herwig (2005); Orlandis (2003); García Volta (1979), 175–79; de Palol und
Ripoll (1999), 112.
73
Zu den komputistischen Teilen dieser Handschrift siehe ferner Millares Carlo
(1954); Alvarez Miranda (1928), XII–XV; sowie zum Computus Cottonianus Gómez
Pallarès (1986), 43–60; Gómez Pallarès (1989).


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

Abbildung 3  León, Archivo de la Catedral, 8, 1v: Widmung des Antiphonars

an Abt Ikila.

dern eher zusammengefügt sind, um möglichst alle Aspekte der im 10.

Jh. im Machtbereich Alfons III. geltenden Festtraditionen zu umfassen:
Folios 5v–9v, 9v–19v und 20r–27v.74 Hierbei handelt es sich bei den bei-

Transkription von fols 20r–25r bei Gómez Pallarès (1988–89), 77–91.

74


 BRIGITTE ENGLISCH

den ersten Abschnitten um zwei bislang noch nicht klassifizierte Varian-

ten der Osterfestberechnung, letztere, zuletzt als Computus Cottonianus
bezeichnet, stellt eine in einen separaten Überlieferungskontext einge-
bettete Hinzufügung aus dem 11. Jh. dar. Diese ist zwar durch die Werke
des Dionysius Exiguus  / Beda beeinflusst; die vielen dort enthaltenen
Fehler lassen aber eher auf eine bearbeitete Adaption jenseits der Pyrenä-
en schließen. Dennoch finden sich auch in diesem Teil nicht nur Bezüge
zu Spanien, wie die Nennung eines Schreibers namens Arias für 1069,75
sondern auch Hinweise darauf, dass sich mindestens eine Schicht des
Textes auf die Westgotenzeit vor 711 zurückführen lässt. Dies macht die
dezidierte Bezugnahme auf König Wamba (672–80)76 in dem der Welt-
zeitalterberechnung gewidmeten Abschnitt deutlich, die vom Anbeginn
der Welt bis zum gegenwärtigen und ersten Jahr der Herrschaft Wambas,
entsprechend der Ära 710 (= Inkarnationsjahr 672), 5872 Jahre ermit-
telt.77 Nachfolgend soll aber, vornehmlich aufgrund des Faktums, dass
diese Teile der Handschrift bislang unberücksichtig blieben, der zwei-
te komputistische Abschnitt behandelt werden, wobei, angesichts des
beschränkten Raumes, exemplarisch ein charakteristischen Beispiel der
rota auf Folio 19v betrachtet werden soll, um hieraus Erkenntnisse über
die westgotische Osterrechnung abzuleiten.
Auf Folio 19v findet sich ein Kreisdiagramm (siehe Tabelle 2 und
Abbildung 4), welches einen direkten Einblick in die Praxis der westgoti-
schen Osterfestansetzung bietet. Das Diagramm bietet eine Zuweisung
der 35 möglichen Ostertermin zu den einzelnen Tagen einer Woche,
wobei gleichmäßig jeweils fünf Termine auf je einen Wochentag fallen.
Ergänzt werden diese Angaben um die parallelen Werte für den Beginn
der vorösterlichen Fastenzeit sowie für jeden der 35 Osterdaten diejeni-
gen Jahre des 19jährigen Mondzyklus, an denen dieser Termin eintreten
kann, wobei sich für jeden Paschaltermin 1–5 mögliche Mondalter er-
geben können.

75
Gómez Pallarès (1989), 58.
76
Zu diesem Herrscher siehe Claude (1971), 154–68; Collins (2004), 100–07.
77
León, Archivo de la Catedral, 8, 25v (hg. von Alvarez Miranda (1928), LVII): Ab
exordium autem mundi ad presentem et primum UUanbanis annum, qui est era DCCX,
colliditur annos V ̅MDCCCLXXII.


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

Abbildung 4  León, Archivo de la Catedral, 8, 19v: Die westgotische Berech-

nung des Osterfestes im Überblick als komputistisches Diagramm.


 BRIGITTE ENGLISCH

Mondjahr Beginn der Fastenzeit Ostern

DOMINICA (Sonntag)
1, 4, 9, 12 4. Kl. Mar. 26.2. 5. Id. Apr. 9.4.
2, 5, 13, 16 2 Id Febr 12.2. 5. Kl. Apr. 26.3.78
3, 6, 11,79 14, (17)80 3. Non. Mar. 5.3. 16. Kl. Mai 16.4.81
7, 10, 15, 18 11. Kl. Mar. 19.2. 4. Non. Apr. 2.4.
8, 19 4. Id. Mar. 12.3. 9. Kl. Mai 23.4.82
Secunda feria (Montag)
1, 4, 9, 12, 15 5. Kl. Mar 25.2. 6. Id. Apr. 8.4.
2, 7, 10, 18 12. Kl. Mar 18.2. K. Apr. 1.4.
3, 6, 14, (11), 17
83 84
4. Non. Mar. 4.3. 17. Kl. Mai 15.4.85
5, 13, 16 3. Id. Febr. 11.2. 8. Kl. Apr. 25.3.
8, 11, 19 5. Id. Mar 11.3. 10 Kl. Mai 22.4.

78
Da der in der Hs. vorgeführte Wert der 6. Kl. Apr. identisch mit dem Oster-
termin von Sabbato 2 ist, jedoch die Angabe der Fastenzeit um einen Tag differierte
(Dominica 2: II. Id. Febr., Sabbato 2: Id. Febr.) und der 26.3. als Termin nicht in der
rota erfasst war, ist es angebracht, einen Kopistenfehler (V. Kl April verschrieben zu VI.
Kl. April) anzunehmen; die Angabe der Mondzyklen ist für beide Tage identisch.
79
In der Hs. X statt XI, vermutlich Kopistenfehler.
80
Die Hs. zeigt hier eine falsche XVII, vermutlich die in der nächsten Spalte feh-
lende Angabe; s.u. Anm. 85.
81
In der Hs. steht die nicht homogene Angabe der II. Non., 15. Kl. April und damit
ein deckungsgleicher Wert zu Sabbato 3. Die Angabe des Mondzyklus ist für beide Termi-
ne gleich. Da jedoch für Sabbato 3 dort die Angabe systemimmanent zutreffend ist, scheint
hier ein Fehler vorzuliegen; richtig wäre für den Beginn der Fastenzeit 3. Non. März (5.3.)
und für den Ostertermin 16. Kl. Mai (16.4.). Dieser Wert des 16.4. findet sich stattdessen,
ebenfalls in diesem System nicht zutreffend, bei Secunda 3, so dass hier möglicherweise der
Kopist sich nicht in der Spalte, sondern beim Tag (Secunda 3 statt Dominica 3) geirrt hat.
82
Die Hs. zeigt an dieser Stelle die Id.  Mar. (15.3.) und 8. Kl.  Mai (24.4.) und
mithin den gleichen Ostertermin wie auch an Sabbato 5, dort aber zutreffend mit den 4.
Id. März (12.3.). Es scheint sich folglich in diesem Fall, aufgrund der systemischen Un-
vereinbarkeit beider Werte und des Faktums, dass der 23.4. zudem in der Tabelle fehlt,
diesen hier, insbesondere da er sich der Struktur der Tabelle entsprechend homogen ein-
fügen lässt, als ursprüngliche Eintragung anzusetzen. Der Beginn der Fastenzeit muss
dann anlog zu den 4. Id. März verschoben werden.
83
Fehlt in der Hs.
84
In der Hs. XVIII statt XVII, vermutlich Kopistenfehler aus vorhergehender Spalte.
85
Durch Verschreibung des Termins aus Dominica 3 (s.o., Anm. 81) ‘verschwand’
hier und somit aus der ganzen Liste der 15.4., der aus diesem Grund hier für den unzu-
treffenden Wert eingesetzt wurde.


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

Mondjahr Beginn der Fastenzeit Ostern

Tertia Feria (Dienstag)
1, 4, 12, 15 6. Kl. Mar 24.2. 7. Id. Apr86 7.4.
2, 7, 10, 13, (18) 87
13. Kl. Mar 88
17.2. 2. Kl. Apr. 31.3.
3,89 6, 9, 14, 1790 5. Non. Mar 3.3. 18. Kl. Mai 14.4.
5, 1691 4. Id. Febr 10.2. 9. Kl. Apr. 24.3.
11, 19, 8 6. Id. Mr. 10.3. 11. Kl. Mai 21.4.
Quarta feria (Mittwoch)
1, 4, 7, 12, 15 6. Kl. Mar 24.2. 8. Id. Apr. 6.4.
2, 10, 13, 18 14. Kl. Mar 16.2. 3. Kl. Ap. 30.3.
3, 8, 11, 19 92
7. Id. Mar 9.3. 12. Kl. Mai 20.4.
5, 16 5. Id. Febr. 9.2. 10. Kl. Apr. 23.3.
6, 9, 14, 17
93
5. Non. Mar. 3.3. Id. Apr. 13.4.
Quinta feria (Donnerstag)
1, 6,94 9, 17 Kl. Mar. 1.3. 2. Id. Apr. 12.4.
2, 5,95 10, 13 15. Kl. Mar. 15.2. 4. Kl. Apr. 29.3.
3, 8, 11,96 14, 19 8. Id. Mar. 8.3. 13. Kl. Mai 19.4.
4, 7, 12, 15, 18 8. Kl. Mar. 22.2. Non. Apr. 5.4.
16 6. Id. Febr. 8.2. 10. Kl. Apr. 22.3.97

86
In der Hs. stehen die 6. Id. April, was gemäß der Parallelangabe des Fastenzeitbe-
ginns als Kopistenfehler erkennbar ist.
87
Die XVIII fehlt in der Hs., möglicherweise ist diese in die nächste Spalte ‘verrut-
scht’ und ersetzt dort fälschlich die korrekte XVII.
88
Verschreibung zu 14. Kl. März.
89
In der Hs.  II, schwer leserlich; möglicherweise eine durch Abnutzung ver-
schwundene I oder Kopistenfehler (II entspricht der vorherigen Spalte).
90
In der Hs. XVIII, vermutlich Kopistenfehler; s.o., Anm. 84.
91
In der Hs. XVII, vermutlich Kopistenfehler.
92
In der Hs. XVIII, vermutlich Kopistenfehler, gleicher Wert wie in der vorherge-
henden Spalte.
93
In der Hs. VIII, vermutlich Kopistenfehler.
94
In der Hs. V, wohl vertauscht mit der folgenden Spalte; richtig VI.
95
In der Hs. VI, wohl vertauscht mit der vorherigen Spalte; richtig V.
96
X von XI stark verblasst.
97
In der Hs. steht hier 5. Id. Febr. (9.2.), 10. Kl. Apr. (23.3.) und mithin ein de-
ckungsgleicher Wert für den Ostertermin wie bei Quarta 4, so dass hier erneut eine Ver-
schreibung anzunehmen ist. Als hilfreiche erweist sich in diesem Fall die Angabe des


 BRIGITTE ENGLISCH

Mondjahr Beginn der Fastenzeit Ostern

Sexta Feria (Freitag)
1, 6, 9, 12, 17 2.Kl. Mar. 28.2. 3 Id Apr.  11.4.
2, 5, 10, 13, 16 16. Kl. Mar.98 14.2. 5. Kl. Apr.  28.3.
3,11, 14, 1999 Non. Mar. 7.3. 14. Kl. Mai  18.4.
4,7, 15, 18 8. Kl. Mar. 22.2. 2. Non. Apr.  4.4.
8 2. Id. mar. 14.3. 7. Kl. Mai 25.4.100
Sabbato (Samstag)
1, 9, 12, 17 3. Kl. Mar. 27.2 4 Id. Apr. 10.4.
2, 5, 13, 16 Id. Febr. 13.2. 6. Kl. Apr. 27.3.
3, 6, 11, 14 2. Non Mar. 6.3. 15. Kl. Mai 17.4.
4, 7, 10, 15, 18 10. Kl. Mar. 20.2. 3. Non Apr. 3.4.
8, 19 4. Id. Mar.101 12.3. 8. Kl. Mai 24.4.

Tabelle 2  Transkription der rota in León, Archivo de la Catedral, 8, 19v.

Der Sinn einer solchen Zuweisung erschließt sich nicht unmittelbar. So

findet sich Vergleichbares nicht in den komputistischen Quellen der Zeit,
wie z.B. bei Beda oder Isidor, erst aus der Zusammensicht des Inhaltes
dieses komputistischen Diagramms ergibt sich aufgrund des Fehlens von
den Wochentag definierenden Strukturen, wie Konkurrente oder Sonn-
tagsbuchstaben, dass hiermit eine Zuweisung des Wochentags zu einem
bestimmten fixen Termin gegeben sein muss. Gemäß logischer Prämissen

Mondzyklus, die bei Quinta 5 nur 16 ist; tatsächlich kann der Ostertag des 22.3. nur im
16. Jahr des Mondzyklus eintreten, so dass oben der Wert entsprechend korrigiert wurde.
98
Hier hatte der Schreiber ursprünglich eine I vergessen (XVI verschrieben zu
XV); diese wurde bei einer Korrektur des Textes zeitnah hinzugefügt.
99
In der Hs. stand hier ursprünglich XVIII, wurde dann durch ein darüber einge-
fügtes I zu XVIIII korrigiert.
100
In der Handschrift findet sich die in keiner Hinsicht stimmige Angabe der 2.
Id. März (14.3.) für den Beginn der Fastenzeit und für den Ostertermin der 6. Kl. Mai
(sic!), also der 26.4., der per definitionem außerhalb der bedanischen Ostergrenzen liegt
und eher auf das System des Victorius von Aquitanien verweist. Systemimmanent be-
trachtet scheint es sich aber um einen (wiewohl auch dies an sich schon aufschlussreich
ist) mehrfach verschriebenen Termin zu handeln, insbesondere auch, da der 25.4. in der
Tabelle keine Berücksichtigung fand. Da aber der 25.4. innerhalb der hier vorgeführten
Struktur zur Quinta feria gehört (der 7.5. ist ein Freitag, wenn Ostersonntag auf den
25.4. fällt) und auch die Angabe des Mondzyklus mit 8 für den 25.4. zutrifft, scheint es
statthaft, besagten Wert an dieser Stelle einzufügen.
101
Systemimmanent verschreiben; in der Handschrift findet sich der Wert der IIII.
Id. März (14.3.) statt der III. Id. März (13.3.).


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

kommt nur ein Stichtag in der Nähe der möglichen Ostertermine—jedoch

außerhalb der Ostergrenzen—in Frage; hieraus ergibt sich eine Erklärung
zwanglos. Unter der Annahme dass es sich bei dem Basistermin um den
7. Mai handelt,102 für den eine Reihe von Gründen sprechen, resultiert
hieraus—das vermag ein Blick in die Grotefendschen Tabellen ohne wei-
teres zu belegen—dass, wenn der 7.5. z.B. ein Sonntag ist, sich als mögliche
Ostersonntage innerhalb der konventionellen bedanischen Ostergrenzen
22.3.–25.4. der 26.3., der 2.4., der 9.4., der 16.4. und der 23.4. ergeben. Die
Zuordnungen im Diagramm der genannten Handschrift treffen, wie die
mitgeteilte Tabelle 2 zeigt, insgesamt gesehen zu; Abweichungen um +1/–
1 dürften hier auf im römischen Zahlensystem oft zu beobachtende Kopis-
tenfehler zurückzuführen sein, was dem pluralen Charakter der gesamten
Handschrift entspricht. Die entsprechenden Osterdaten ergeben sich ohne
weiteres aus der allgemeine Kalenderarithmetik, wenn man 35 fortlaufende
Sonntagstermine in Bezug auf den Wochentag eines bestimmten Stichtages
verteilt. Somit fallen jeweils—wie im Diagramm richtig aufgelistet—fünf
Termine auf einen bestimmten Wochentag des Stichtages. Diese unge-
wöhnliche, der Kalenderarithmetik jedoch entsprechende Zuordnung hat
gegenüber den übrigen abendländischen Zuordnungen wie Sonntagsbuch-
stabe und Konkurrente nicht unerhebliche praktische Vorteile:

1. Durch die Kenntnis des Wochentags des Stichdatums reduzieren

sich von vorn herein die möglichen Ostertermine von 35 auf fünf.
2. Aus der Kenntnis des Stichtag-Wochentags des laufenden Jahres
lässt sich problemlos der entsprechende Wochentag des folgen-
den Jahres ermitteln; folgt ein Gemeinjahr, so ist um eins zu er-
höhen, folgt ein Schaltjahr, jedoch um zwei.

Dies mag an einem einfachen Beispiel erläutert sein. Im Jahr 2010 fiel
der 7. Mai auf einen Freitag, folglich fällt der 7.5. des Jahres 2011 auf ei-
nen Samstag. Damit können gemäß Diagramm—und wie die Kontrolle
in Grotefends Tafeln zeigt103—als Ostertermine nur der 27.3., der 3.4.,
der 10.4., der 17.4. und der 24.4. in Frage kommen. Diese Zuordnungen
lassen sich auch an beliebigen anderen Daten verifizieren.

102
Die Nutzung des 1. Januar, der ansonsten in der Komputistik als ‘Ankerpunkt’
der Wochentagsbestimmung zur Anwendung kommt, scheint insofern wenig sinnvoll,
als dann auf das Faktum des Schaltjahres separat Rücksicht genommen werden muss; of-
fenkundig wurde hier aber ein von diesem Erwägungen unabhängiges System generiert.
103
Grotefend (1991), 154–55 (Tafel 6), 168–69 (Tafel 13), 182–83 (Tafel 20),
196–97 (Tafel 27), 210–11 (Tafel 34).


 BRIGITTE ENGLISCH

Unter der Voraussetzung visueller Beobachtung der Mondphasen,

lässt sich nun leicht aus diesen fünf Terminen der richtige für die Fei-
er des Osterfestes herausfinden. Wir haben zahlreiche Hinweise darauf,
dass sich der westgotische Klerus im November traf, um den Osterter-
min für das darauf folgende Jahr zu verkünden. Diese Terminierung ist
in vielerlei Hinsicht günstig:

1. Es blieb genug Zeit den Termin des Osterfestes rechtzeitig bis in

den letzten Winkel des Reiches zu verbreiten.
2. Es ergab sich mit dem Fest des bei den Westgoten hochverehr-
ten hl. Martin ein herausragender Stichtag zur Verifizierung der
Mondphase.
3. Dieser Termin lag zeitlich hinreichend nahe an der folgenden
Frühlingslunation.

Man kann davon ausgehen, dass dem Klerus—im wesentlichen auf-

grund des erwähnten Zusammenlebens einerseits mit den Muslimen
und andererseits mit den Juden und deren Mondkalendern—die prak-
tische Beobachtung der Mondphase geläufig war. Aus dem Turnus der
Mondphasen ergibt sich als Referenztag für den 11. November, so ein
Gemeinjahr folgt, der 9. März, d.h. die am 11. November zu beobacht-
ende Mondphase entspricht der des 9. März des darauf folgenden Jahres.
Zwischen beiden Terminen liegen exakt vier Mondphasen, so dass vom
9. März sehr leicht weitergezählt werden kann, bis zur Erreichung des
Frühlingsvollmonds. Der darauf folgende Sonntag aus den genannten
fünf Möglichkeiten ist dann der Ostersonntag. Auch das sei an einem
aktuellen Beispiel verifiziert. In diesem Jahr—wie sich leicht an jedem
Wandkalender feststellen lässt—tritt der Vollmond am Sonntag, dem
21.11.2010 ein. Am 11. November wird man ihn also unter dem Mond­
alter 4 beobachten können, also vier Tage nach Neumond. Folglich gilt,
dass am 9.3.2011 der Mond sich am 4. Tag der zunehmenden Phase be-
findet und wir den nächsten Vollmond am 19. März erwarten dürfen.
Daraus ergibt sich, da dieser Vollmond vor der Ostergrenze des 21. März
liegt, durch einfaches Abzählen der nächste Vollmond am 18.4.2011. Da
der 7. Mai ein Samstag ist, sind, wie bereits gezeigt, folgende Ostertermi-
ne möglich: Der 27.3., 3.4., 10.4., 17.4., 24.4. Als nächster Sonntag nach
dem 18. April kommt also nur der 24.4.2011 als Ostertermin in Frage,
was in der Tat so ist, wie ein Blick in den Grotefend zeigt.
Somit eröffnet dieses Diagramm eine einfache Festlegung des folgen-
den Ostertermins unter der Voraussetzung einer zeitnahen annähernden


 OSTERFEST UND WELTCHRONISTIK IN DEN WESTGOTISCHEN REICHEN

Beobachtung der Berechnung104 der Mondphase, wobei nicht nur der

schlecht zu beobachtende Vollmond, sondern auch die genauer zu fixie-
renden Mondphasen ins Kalkül gezogen werden können. Es ist also zu-
sammenfassend zu konstatieren, dass der westgotische Klerus eine leicht
zu handhabende Methode entwickelt hat, den Ostertermin von Jahr zu
Jahr festzustellen ohne die komplexen arithmetischen Strukturen der
Osterfestrechnung eines Beda Venerabilis oder eines Dionysius Exigu-
us heranzuziehen, was erklären könnte, warum sich diese über mehrere
Jahrhunderte nicht durchsetzen können. Zugleich spiegelt sich dieses
Verfahren im eingangs zitierten Quellentext: Im November wird durch
die Versammlung der Bischöfe der Ostertermin festgestellt, dieser dann
an die einzelnen Priester übermittelt, welche diesen zusammen mit dem
Beginn der davon abhängigen Fastenzeit und Pfingsten am Weihnachts-
tage dem Volk kundtun.

Zusammenfassung
Man muss also zusammenfassend festhalten, dass die eingangs vorge-
führte Schilderung der Osterfestansetzung keineswegs einem chaoti-
schen Szenario entspringt, sondern die Osterefestrechung in den west-
gotischen Reichen von einem eigenen komputistischen System geprägt
wurde. Hierbei handelt es sich, wie an einem Beispiel exemplarisch ge-
zeigt werden konnte, um eine einfach zu handhabenden, aber ebenso
klug erdachten Methode, das Auferstehungsfest im Einklang mit dem
Nicäum und den religiösen Weisungen, aber auch unter Berücksich-
tigung der Realität zu bestimmen. Gleichzeitig werden die Informati-
onswege bestimmt, um den aktuell astronomisch verifizierten Termin
rechtzeitig allen Bevölkerungsteilen mitzuteilen, genau so, wie es das
zweite Konzil von Braga beschreibt. Diese Methode war, angesichts des
engen Gegenübers von Juden, Muslimen und Christen, und der daraus
resultierenden Notwendigkeit, die Osterlunation realiter zu bestimmen,
so überzeugend, dass sie gegenüber der dionysisch-bedanischen Berech-
nung bis in das 11. Jh. nachhaltig bevorzugt wurde. Damit ist in den
Reichen der Westgoten vom 7.–11. Jh. eine eigenständige, bislang unbe-
rücksichtigte Variante der Komputistik anzunehmen.
104
Hierauf deuten, zumindest als Korrektiv von beobachteten Daten, die Integra-
tion von den Mond betreffenden Parametern in besagtem Kreisdiagramm hin; eine um-
fassende Analyse dieser vielschichtigen, keineswegs homogenen Bestandteile der Hand-
schrift, die den hier gesteckten Rahmen sprengen würde, wird derzeit von mir vorbereitet.


 DAVID HOWLETT

AN ADDITION TO
THE HIBERNO-LATIN CANON:
DE RATIONE TEMPORUM

Abstract
The essay provides a text, translation, and analysis of the poem, fixing its origin
in a milieu of seventh-century Hiberno-Latin compositions.
Keywords
Alliteration, alphanumeric value, assimilation, calendrical and arithmetic com-
position, diaresis, elision, fractions, gematria, rhyme schemes, synizesis; De ra-
tione conputandi, De ratione temporum, Versus de annis a principio.

Introduction
The text of De ratione temporum that follows is based upon Karl Streck-
er’s edition.1 It includes only the unbracketed verses from eight sources,
six manuscripts written between the tenth century and the twelfth, one
manuscript written during the fourteenth century, and one book print-
ed in 1612, the consensus of which Strecker labelled w, representing the
original form of the poem. The only change to Strecker’s text here is col-
lapse of the Classical diphthong ae to e, based upon internal evidence.
Roman numerals designate numbers of couplets. Within the text italic
marks rhyme and boldface marks alliteration within lines and couplets.
To the right of the text columns note numbers of lines in Arabic numer-
als, and for each line the rhyme scheme, the number of words, and the
number of letters.

The poem De ratione temporum is ed. by Karl Strecker in MGH Poetae 4,2.3,
1

682–86. I owe thanks to Dan Mc Carthy for illuminating discussion of the poem.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 212-228
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114738
 AN ADDITION TO THE HIBERNO-LATIN CANON

Text
DE RATIONE TEMPORUM 3 17
I
Annus solis continetur quatuor temporibus a 5 37
Ac deinde adimpletur duodecim mensibus a 5 34
II
Quinquaginta et duabus currit ebdomadibus a 5 37
Trecentenis sexaginta atque quinque diebus a 5 38
III
Sed excepta quarta parte noctis atque diei 5 b 7 36
Que dierum superesse cernitur seriei b 5 32
IIII
De quadrante post annorum bis binorum terminum c 7 40
Calculantes colligendum decreuerunt bis sextum c 5 42
V
Hinc annorum diuersantur lune longitudines d 5 38
Quorum quidam embolismi quidam fiunt communes 10 d 6 40
VI
Breuis quippe qui uocatur communis lunaribus a 6 39
Solis semper duodenis terminatur mensibus a 5 37
VII
Longus autem qui omnino embolismus dicitur e 6 37
Lune tribus atque decem cursibus colligitur e 6 38
VIII
Breuioris anni totus terminatur circulus 15 a 5 36
Trecentenis quinquaginta atque quater diebus a 5 40
VIIII
Longus uero lune annus in dierum termino f 7 34
Continetur trecenteno octogeno quaterno f 4 36
X
Uno nempe atque decem diebus in ordine g 7 32
Breuis annus anni solis superatur agmine 20 g 6 35
XI
Nouem uero embolismus atque decem diebus a 6 35
Peruidetur anni solis eminere cursibus a 5 34
XII
Ac per istam sui semper incrementi copiam h 7 35
Breuiorum longiores compensant inopiam h 4 35


 DAVID HOWLETT

XIII
Dehinc decem atque nouem annorum statuitur 25 e 6 37
Tempus certum quo lunaris terminus porrigitur e 6 40
XIIII
In quo lune replicantur cursu quodam subtili i 7 38
Bis centeni terque deni semel quini circuli i 7 37
XV
Hüic quippe ut exigit perscrutanda ratio j 6 35
Se per partes dies saltus interserit spatio 30 j 7 37
XVI
Intra quoque supra dictum habent semper spatium c 7 41
Dies solis atque lune equum pondus partium c 7 36
XVII
Huius cycli pars uocatur ogdoas anterior k 6 35
Ac deinde appellatur endecas ulterior k 5 33
XVIII
Ogdoadi deputantur octo anni priores 35 d 5 32
Endecadi reliquorum destinatur series d 4 34
XVIIII
In hoc cursu fiunt anni communes duodecies d 7 36
Anni uero embolismi supputantur septies d 5 35
XX
Talis quippe cum annorum calculus extenditur e 6 39
In eundem lune cursum circulus reuertitur 40 e 6 36
XXI
Solis uero in id ipsum non recurrit series d 8 35
Donec anni reuoluantur octies et uigies d 6 34
XXII
Adimpleto sane solis supra scripto tempore l 6 37
Reclinatur absque ullo resistenti rancore l 5 37
XXIII
His itaque reciprocis alternata cursibus 45 a 5 36
Solis lune diuersantur tempora temporibus a 5 37
XXIIII
Sed ad prima post annorum recurrent initia m 7 36
Quingentorum ac triginta atque duum spatia m 6 37
XXV
In hoc cyclo per concursum celi luminarium c 7 36
Manifesta fiunt festa dierum paschalium 50 c 5 35


 AN ADDITION TO THE HIBERNO-LATIN CANON

XXVI
Qui porrectus per extensam annorum uertiginem n 6 40
In eandem ut predixi recurrit originem n 6 33
XXVII
Pasche uero longos breues intellectu uarios o 6 38
Hoc excepto multi plures conscripserunt circulos o 6 43
XXVIII
Quos dispono preterire breuitatis gratia 55 m 5 36
Quorum nobis necnon nota numeri peritia m 6 34
XXVIIII
Altum celum qui creauit terras atque equora p 7 37
Doxa Regi per eterna Deo soli secula 58 p 7 30

AMEN 1 4

Translation
ON THE RATIO OF TIMES
I
A year of the sun is contained in four seasons (lit. ‘times’),
and then it is fulfilled in twelve months.
II
In fifty and two weeks it runs,
in three hundred sixty and five days,
III
but with the fourth part of a night and a day excepted,
which is beheld to be in excess of the series of days.
IIII
About the fourth part after the term of twice two years
those calculating have decreed the bissextus (lit. ‘twice-sixth’) to be tabu-
lated.
V
Hence are distinguished the lengths of the years of the moon,
of which some are made embolismic, some common.
VI
The short one, indeed, which is called common,
in twelve lunar months alone is always terminated.


 DAVID HOWLETT

VII
The long one, however, which is spoken of entirely as embolismic,
is computed in three and ten courses of the moon.
VIII
The complete circle of the shorter year is terminated
in three hundred fifty and four days.
VIIII
The long year of the moon in truth in a term of days
is contained in the three hundredth, eightieth, fourth.
X
In one doubtless and ten days in order
a short year is exceeded by the sequence of a year of the sun.
XI
In truth an embolismic (year) by nine (and) ten days
is seen clearly to stand out from the courses of a year of the sun.
XII
And through this supply of its own increment always
the longer (years) compensate the defect of the shorter.
XIII
From this ten and nine years is established
the certain time in which the term of the lunar (year) is set forth,
XIIII
in which by a certain subtle course of the moon
two hundred and three times ten together with five cycles (lit. ‘circles’)
are turned back on themselves.
XV
To this space, indeed, as a reckoning that must be thoroughly scrutinized
demands,
the day of the saltus (lit. ‘leap’) inserts itself among the parts.
XVI
Also within the above-said space always
the days of the sun and the moon have an equal weight of parts.
XVII
Of this cycle the anterior part is called the ogdoad,
and then the ulterior is named the hendecad.
XVIII
The eight prior years are deputed ogdoadic,
and the series of those remaining is designated hendecadic.


 AN ADDITION TO THE HIBERNO-LATIN CANON

XVIIII
In this course twelve years are made common,
seven years, in truth, are reckoned embolismic.
XX
When, indeed, such a calculation of years is extended
into the same course the circle of the moon is returned to.
XXI
The series of the sun in truth does not run back to itself
until eight and twenty years are revolved.
XXII
With the above-written time of the sun soundly fulfilled
it is laid back without any pausing (or ‘resisting’) grudge.
XXIII
And so with these reciprocal courses
alternated times are distinguished in the times of the sun, the moon.
XXIIII
But they run back after to the first beginnings of the years
of five hundred and thirty and two spaces.
XXV
In this cycle through the concourse of the luminaries of heaven
the festivals of Easter days are made manifest,
XXVI
which (cycle) set forth through the extended gyration of the years
runs back to the same origin, as I said before.
XXVII
Many men wrote very many cycles (lit. ‘circles’), with this excepted,
of Easter, in truth, long, short, varied in the understanding,
XXVIII
which I dispose to pass over for the sake of brevity,
and of which knowledge of the number (is) known to us.
XXVIIII
Glory to God the King alone through eternal ages,
Who created lofty heaven, lands, and seas.

AMEN.


 DAVID HOWLETT

Analysis
The poet uses in twenty-nine couplets sixteen end-rhymes, that extend
from one syllable and two letters to three syllables and six letters. But he
extends rhyme further backwards from the ends of lines to make the cae-
suras rhyme in couplets I, IIII, XI, XVI, XVII, XVIIII, XX, and XXVII.
He makes the caesuras of the first hemistichs rhyme in couplets IIII, V,
VIII, XVIII, and XXVII. He introduces internal rhymes within the first
hemistichs in couplets VIII, X, XIII, XIIII, XV, and XXV, and chiastic
rhymes ABBA -i-a -orum -orum -i-a within the first hemistichs in cou-
plet XXIIII. There is rhyme in about 40% of the syllables and alliteration
in about 30% of the syllables. The pentadecasylabic metre is identical
with that of Versus de annis a principio, Deus a quo facta fuit huius mundi
machina, a Hiberno-Latin poem about the Six Ages that dates itself in
three ways to AD 645.2
Pure rhyme, dense alliteration, metre, and consistently faultless gram-
mar and syntax help to secure the text. Noting the parallelism of sed 5
with que, not quae 6, extensam 51 with predixi, not praedixi 52, pasche,
not paschae 53 with excepto 54, preterire, not praeterire 55 with necnon
56, celum, not caelum 57 with regi 58, and creauit 57 with eterna, not
aeterna 58, we may infer that in the poet’s Latinity the Classical diph-
thong ae had collapsed to e. Alliteration of calculantes with colligendum
8, not conligendum, suggests that our poet assimilated his consonants.
He practised diaeresis in hüic 29 and synizesis in duodecies 37. He did
not practise elision in lines 2, 3, 4, 10, 13, 16, 17, 18, 19, 21, 25, 29, 31,
32, 34, 35, 38, 39, 44, 48, 51, 52, 57. All these phenomena are not only
consistent with composition in a seventh-century Hiberno-Latin mi-
lieu; together they are difficult to parallel outwith these islands during
the seventh century. An extensive corpus of comparable Hiberno-Latin
verse survives, but the present writer knows of no centre elsewhere in
which such densely rhyming and alliterating verse can be proved to have
been composed during the seventh century.
The argument of the poem follows the order of instruction in com-
putus set forth in the seventh-century Hiberno-Latin treatise De ratione
conputandi, written by someone in the circle of Cummian, if not by
Cummian himself, before his death, recorded in the Annals of Ulster s.a.
661:3 4 couplets about the solar year (I–IIII), 8 couplets about the lunar

2
Howlett (1996b), 1–6.
3
De ratione conputandi is ed. by Ó Cróinín in Walsh and Ó Cróinín (1988), 115–
213; for the dating, see Ó Cróinín (1982), especially 121.


 AN ADDITION TO THE HIBERNO-LATIN CANON

year (V–XII), 8 couplets about the 19-year lunisolar cycle (XIII–XX),

2 couplets about the 28-year solar-ferial cycle (XXI–XXII), 4 couplets
about the 532-year lunisolar-ferial cycle and its relation to the 532-year
paschal cycle (XXIII–XXVI), 2 couplets about the poet’s knowledge of
all proposed paschal cycles (XXVII–XXVIII), and 1 couplet of Doxol-
ogy to the God of the cosmos (XXVIIII). With the poet’s remarks in
couplets XXVII–XXVIII about other systems that he knows but de-
clines to discuss one may compare Cummian’s summary of his knowl-
edge of ten paschal cycles in the Epistola.4 The account of the solar cycle
occupies 4 couplets, coincident with the period of the bissextus, and the
computistical section occupies 28 couplets, coincident with the period
of the solar-ferial cycle. From  | recurrent 47 to  | ut predixi recurrit 52
there are 28 words.
The poem contains 29 couplets, including the title 30 units, includ-
ing the Amen 31 units, 28–29–30–31 representing, perhaps, the num-
bers of days in a month.
Let us pursue analysis of the text by considering the placement of
words for numbers and their values in discrete lines and couplets.5
In line 1 the 4th word is quatuor, after which follow 4 syllables.
In line 2 the 12th syllable is the last of duodecim.
In line 3, the first of couplet II, there are 2 words before duabus and
2 words after it.
In line 4 the 5th syllable from the end is the first of quinque.
In line 5 there are 4 syllables before quarta and 4 words after it.
In couplet IIII the 4th syllable is the last of quadrante, and after 4
words one reads bis binorum, 4, the last of which is 4th syllable from the
end. The couplet ends with the words bis sextum, the last the 12th word.
The number of syllables in the couplet divides by the ratio 2:1 at 20 and
10 in the 20 syllables from | bis 7 to bis | 8. The number of letters in the
couplet divides by the same ratio at 54.66 and 27.33 in the 54 letters
from | bis 7 to bis | 8.
In line 12 the 12th letter is the first of duodenis.
In line 14 the 3rd syllable is the first of tribus, and the 13th syllable from
the end is the 1st of tribus atque decem, which follows 13 lines of verse.

4
Cummian, De controversia paschali ll. 208–20 (ed. and trans. by Walsh and Ó
Cróinín (1988), 84–86). I owe thanks to Dan Mc Carthy for reminding me of these
parallels.
5
A single late antique source for this phenomenon, widely practised in Insular
Latin compositions, is the Late Latin translation of Anatolius’ De ratione paschali, for
which see Mc Carthy and Breen (2003); Howlett (2007).


 DAVID HOWLETT

In line 16 the 4th word is quater, the last of which is the 4th syllable
from the end.
In line 18 there are 8 syllables before octogeno, after which there are 8
letters and spaces to the end.
In couplet X the first word is uno.
In line 21 there are 19 letters between nouem | and | decem.
In line 25 the 19th letter from the beginning and from the end is the
u of decem atque nouem.
The 2nd line of couplet XIIII begins bis; the 3rd word is terque; and the
5th syllable from the end is the 1st of quini.
In line 32 the number of letters divides equally at | equum.
In line 33 there are 8 syllables before | ogdoas and after ogdoas | 8 let-
ters to the end.
In line 34 the 11th syllable is the last of endecas.
In line 35 there are 8 words after ogdoadi | and 8 syllables before | octo.
In line 36 there are 11 syllables after endecadi |.
In line 37 the 12th syllable is the 1st of duodecies.
In line 38 there are 7 letters in septies.
In line 42 there are 8 syllables before | octies and 28 letters before oc-
ties et | uigies. In couplet XXI the 28th syllable is octies et |ui|gies.
Now let us consider the placement of words for numbers and their
values in units larger than the line and the couplet.
From | duodenis 12 to tribus atque decem | 14 there are 13 words.
Before couplet XI the poet writes uno atque decem 19. In couplet XI
he writes nouem atque decem 21, then after it 19 couplets to the end of
the poem.
From | trecenteno octogeno quater 18 to bis centeni terque deni semel
quini circuli | 28 there are 384 letters, and from | quingentorum ac trigin-
ta atque duum spatia 48 to quorum nobis necnon nota numeri peritia | 56
there are 384 letters and spaces between words.
After nouem … atque decem | 21 the 19th word is the first of decem
atque nouem 25.
We noted above that in line 32 the number of letters divides equally
at | equum. More comprehensive order appears in that in the 8 couplets
XIII–XX from the space before | Dehinc 25 to the space after equum | 32
there are 339 letters and spaces between words, and thence to the space
after reuertitur | 40 there are also 339 letters and spaces between words.6
From | ogdoas 33 to ogdoadi | 35 there are 8 words.

6
For similar play in a later Hiberno-Latin poem see Howlett (1996a), 87.


 AN ADDITION TO THE HIBERNO-LATIN CANON

From | endecas 34 to | endecadi 36 there are 22 (11×2) syllables.

In couplets XVII and XVIII, in which the poet writes about a cycle
that divides into ogdoadic and hendecadic parts, the 20 words divide by
sesquioctave ratio or epogdous, 1 1/8:1 or 9:8, at 11 and 9, in the 11 words
between ogdoas | 33 and | endecadi 36. Those 11 words divide in turn by
the same ratio at 6 and 5, in the 5 words from ogdoas | 33 to endecas | 34
and in the 5 words from | ogdoadi 35 to | endecadi 36.7
Between couplet VI, which ends Solis semper duodenis terminatur
mensibus and couplet XVIIII, which begins In hoc cursu fiunt anni com-
munes duodecies there are 12 couplets.
From the end of couplet XI line 21 Nouem uero embolismus atque
decem diebus to the end of the account of the 19-year cycle in couplet
XX line 40 In eundem lune cursum circulus reuertitur there are 19 lines.
After octies et uigies | 42 the 30th word is triginta | 48, and the 2nd word
after triginta | is duum.
From | Huius cycli pars 33 to | quingentorum ac triginta atque duum
48 there are 532 letters.8
Note recurrence of the word bis in lines 8 and 28. The 337 words of
the entire poem divide by duple ratio 2:1 at 225 and 112. From | bis 8
to bis | 28 there are 112 words. The 2445 letters and spaces of the entire
poem divide by the same ratio at 1630 and 815. From | bis 8 to | bis 28
there are 815 letters and spaces.
Between line  15, which ends with the word circulus, and line  27,
which ends with the word circuli, there are 12 lines, one for every month
of a year. Between couplet XIIII, which ends with the word circuli, and
couplet XXVII, which ends with the word circulos, there are 12 couplets,
one for every month of a year. Between line  15, which ends with the
word circulus, and line 40, which ends with the words circulus reuertitur,
there are 24 lines, one for every hour of a day. The 24 lines between 15
and 40 contain 360 syllables, and the 12 couplets between XIIII and
XXVII contain 360 syllables, one for every degree of a circle.
In the 8 couplets V–XII about the lunar cycle note the alteration
lune longitudines 9, breuis 11, longus 13, breuioris 15, longus 17, breuis
20, breuiorum longiores 24.

7
For other examples of double or multiple diminution by the same ratio in Hiber-
no-Latin texts, see Howlett (1995a), 213–16; Howlett (2002), 85–88; Howlett (2008),
84–86.
8
For other examples of this in seventh-century Hiberno-Latin texts, see Howlett
(2010b).


 DAVID HOWLETT

The poet relates in line 28 the completion of a lunar cycle in 235 years,

and he relates in line 42 the completion of a solar cycle in 28 years. The
latter is confirmed in the line number of the former, 28.
Let us consider next fractions. The 340 words of the poem divide by
twelfths at 28.33. From | duodecim 2 to duodenis | 12 there are 56 words,
that is from the first of these words for 12 to the second 2/12 of the words
of the poem.
In stanzas III–IIII from | quarta parte 5 to quadrante 7 there are 12
words, that is as one fourth part and one fourth part make one half, ex-
actly half of the 24 words of stanzas III–IIII.
Another locution for 12 is | bis sextum 8. From | duodecim 2 to bis
sextum | 8 inclusive there are 36 words (12×3). From | bis sextum 8 to
decem | 14 inclusive there are 34 words, exactly 1/10 of the poem.
The fourth locution for 12 is duodecies | 37, from which to the end of
the computus and the beginning of the doxology at line 56 there are 112
words (56×2), 4/12 of the words of the poem.
From | tribus atque decem 14 to | uno nempe atque decem 19 there are
26 words, that is 1/13 of the poem.
From uno nempe atque decem | 19 to | dehinc decem 25 there are 31
words, 1/11 of the poem.
From uno nempe atque | decem 19 to decem | 25 there are 34 words,
exactly 1/10 of the poem.
The poet arranged much of his diction in chiastic and parallel
patterns:9
A1  1  temporibus
A2   5  sed
A3   7  post
A4   7  annorum
B   9  diuersantur
C   9  lune
D 12  solis (‘alone’)
E 14  cursibus
F1 15  circulus
F2 17  uero
F3 17  annus
F4 18  octogeno
F5 20  solis

This also is a recurrent basic feature of Insular Latin composition; cf. Howlett

9

(1995a), 1–10; Howlett (1997), 1–11.


 AN ADDITION TO THE HIBERNO-LATIN CANON

F6 20  superatur
F7 26  tempus
G1 26  lunaris
G2 27  cursu
H 29  huic
I 32  partium
J 33  huius cycli
I’ 33  pars
H’ 37  in hoc
G’1 40  lune
G’2 40  cursum
F’1 40  circulus
F’2 41  uero
F’3 42  anni
F’4 42  octies
F’5 43  solis
F’6 43  suprascripto
F’7 43  tempore
E’ 45  cursibus
D’ 46  solis (‘of the sun’)
C’ 46  lune
B’ 46  diuersantur
A’1 46  tempora temporibus
A’2 47  sed
A’3 47  post
A’4 47  annorum

A widespread tradition of gematria, the reckoning of numerical values

of letters of the alphabet, in Hebrew, Greek, and Latin, was understood
and practised in Ireland from the beginnings of the Hiberno-Latin tra-
dition.10

10
For a tabulation of the Hebrew system see Kautzsch and Cowley (1985), 26
(§ 5.2). For evidence of knowledge of the shapes and names of letters of the Hebrew al-
phabet, see the seventh-century Old Irish text Auraicept na n-éces (ed. by Calder (1917),
86–87, 229–30). For early insular explications of the Greek and Latin systems, see
Bede, De temporum ratione 1 (ed. by Charles W. Jones in CCSL 123, 272–73), and the
seventh-century Auraicept na n-éces (Calder (1917), 230–31). For actual calculation of
Greek values, see the eighth-century Hiberno-Latin manuscript Milan, Biblioteca Am-
brosiana, F 60 sup., 61rb; Howlett (1995a), 29–32; Howlett (1997), 34–37; Howlett
(2006). For the phenomena in fifth-century Hiberno-Latin texts see Howlett (1998b);
Howlett (1994).


 DAVID HOWLETT

In the 23-letter Latin alphabet, the alphanumeric value of ANNUS is

1+13+13+20+18 or 65, that of ANNORUM 1+13+13+14+17+20+12
or 90.
Between annus | 1 and | annus 17 there are 90 words. From the space
before | annorum 7 to the space after annorum | 9 there are 90 letters
and spaces between words. Between annorum  | 9 and  | anni 15 there
are 90 syllables. After annus | 17 the 90th letter is the a of annus 20. Be-
tween anni | 22 and | anni 37 there are 90 words. From the space before |
anni 37 to the space after annorum | 39 there are 90 letters and spaces
between words. From | anni 38 to the space after anni | 42 there are 180
(90×2) letters and spaces between words.
Between anni | 20 and | anni 22 there are 65 letters. From | anni 38
to the space after annorum | 39 there are 65 letters and spaces between
words. From | annorum 47 the 65th syllable is in annorum 51. From | an-
norum 25 to annorum | 47 there are 130 (65×2) words.
The alphanumeric value of DIES is 4+9+5+18 or 36, that of DIEI
4+9+5+9 or 27, that of DIERUM 4+9+5+17+20+12 or 67, that of
DIEBUS 4+9+5+2+20+18 or 58.
The 27th word of the poem is diei 5. From | diebus 16 to | dierum 17
there are 27 letters.
Between dierum | 6 and | diebus 16 there are 58 words.
Between diebus | 4 and | diebus 16 there are 67 words. After diebus |
4 the 67th word is dierum 17. From | dierum 17 to | diebus 19 or from
dierum | 17 to diebus | 19 there are 67 letters. From | diebus 19 to dies |
30 there are 67 words. From | dies 30 to | dies 32 or from dies | 30 to dies |
32 there are 67 letters.
From | dies 32 to dierum | 50 there are 108 (36×3) words.
The alphanumeric value of EBDOMA is 5+2+4+14+12+1 or 38,
that of EBDOMADIBUS 5+2+4+14+12+1+4+9+2+20+18 or 91.
From the beginning of the poem to quinquaginta et duabus | currit ebdo-
madibus there are 38 syllables and 91 letters and spaces between words.
The alphanumeric value of EMBOLISMUS is
5+12+2+14+11+9+18+12+20+18 or 121. From embolismi | 38 to the
end of the poem there are 121 words. From | embolismus 13 to embolis-
mus | 21 there are 120 syllables.
The alphanumeric value of LUNA is 11+20+13+1 or 45, that of
LUNE 11+20+13+5 or 49, that of LUNARIS 11+20+13+1+17+9+18
or 89, that of LUNARIBUS 11+20+13+1+17+9+2+20+18 or 111.


 AN ADDITION TO THE HIBERNO-LATIN CANON

The first 49 words of the poem bring one to lune longitudines  | 9.

From | lune 14 to | lune 17 or from lune | 14 to lune | 17 there are 49
syllables.
From | lune 32 to lune | 40 there are 45 words.
From | lunaribus 11 to lune | 27 there are 89 words. From | lunaris 26
to lune | 32, scanning huic as a monosyllable, there are 89 syllables.
From | lune 27 to | lune 46 there are 111 words.
The alphanumeric value of MENSIS is 12+5+13+18+9+18 or 75.
From mensibus | 2 to mensibus | 12 there are 150 (75×2) syllables.
The alphanumeric value of PASCHE is 15+1+18+3+8+5 or 50,
that of PASCHALIS 15+1+18+3+8+1+11+9+18 or 84.
Line 50 ends with the word paschalium. From | paschalium 50 to the
end of the poem there are 50 words.
Between paschalium | and | pasche there are 84 letters and spaces be-
tween words.
The alphanumeric value of SOLIS is 18+14+11+9+18 or 70.
From | solis 20 the 70th letter is the first of solis 22. Between solis | 20
and | solis 32 there are 70 words.
The alphanumeric value of TEMPUS is 19+5+12+15+20+18 or 89,
that of TEMPORIBUS 19+5+12+15+14+17+9+2+20+18 or 131.
From tempore | 43 to the end there are 89 words. From | tempore 43 to
tempora temporibus | 46 there are 132 letters and spaces between words.
This accounts for every recurrence of the words annus, circulus, dies,
ebdomadibus, luna, lunaris, mensibus, pasche, paschalium, and every re-
currence but one of the word embolismus in the poem.
The text of the poem from the title to Amen inclusive contains 2130
letters in 60 lines, an average of 35.5 letters per line, with which one may
compare the carmen figuratum addressed by Joseph Scottus to Charle-
magne that contains 35 letters in each of its 35 lines.
The Book of Cerne (Cambridge, University Library, L1.1.10) con-
tains on folio  21r a prayer with an acrostic that reads AEDELUALD
EPISCOPUS.11 On folio  87v it contains an introduction to excerpts
from the Psalter: hoc argumentum forsorii oeðelwald episcopus decerp-
sit (‘this matter of a book of (biblical) verses Aedeluald the bishop ex-
cerpted’). The manuscript is a ninth-century copy of a book of Aedel­
uald, bishop of Lindisfarne AD 721–40, successor to Eadfrith, bishop
of Lindisfarne probably AD 698–721, the scribe and illuminator of the

11
The Book of Cerne is ed.  by Kuypers (1902); Brown (1996). See Howlett
(1997), 128–35; Howlett (2005), 157–59.


 DAVID HOWLETT

Lindisfarne Gospels.12 Aedeluald was a pupil and correspondent of Ald-

helm, bishop of Sherborne (†AD 709), the first English man of letters,
both Anglo-Latin and Old-English, the most illustrious pupil of the
school at Canterbury of Theodore and Hadrian, who from AD 670 en-
couraged their pupils to read Cambro- and Hiberno-Latin texts.13 The
thirty-second prayer in Aedeluald’s book is a poem:14

Christum peto Christum preco

Christo reddo corde leto
Gratis homo imo fono
Uti latro tetro metro
Pendens ligno petit regno 5
Fore uiso paradiso
In clalisso in abysso
Hoste truso ac deluso
Sic et ego quantum queo
Manus Deo leuo meo 10
Quo cum reo fruar eo
Gignans Huio Patri pio
Flatus riuo quando uiuo
Sarcem trunco Pauli culmo
Stirpem limeo uti uideo 15 lines
Altum celum qui creauit terras atque equora. 15 syllables

‘I ask Christ, I pray to Christ,

I give back to Christ with a happy heart
thanks, [I] a man with a very low voice
in four-part metre, like the thief
hanging on the wood [i.e. the penitent thief on the cross];
he asks
to be in the Kingdom,

12
Aedeluald’s extant letter and verses are ed. by Rudolf Ehwald in MGH Auct. ant.
15, 495–97, 528–37. For an account of his contribution to the making of the Lindis-
farne Gospels see the colophon of Aldred, glossator of the Gospels, in Howlett (2005),
216–20.
13
Examples include Gildas, from whose work De excidio Brittanniae batches of
glosses survive in the Leiden Glossary (Leiden, Universiteitsbibliotheek, VLQ 69,
7r–47v), originally from the school of Theodore and Hadrian, for which see Howlett
(1998a), 49–52; Laidcenn mac Baíth Bannaig, Lorica, cited in the archetype of the
Épinal and Erfurt Glossaries (Épinal, Bibliothèque municipale, 72; Erfurt, Universitäts-
bibliothek, Amplonianus fol. 42) and in chapter 11 of Aldhelm’s prose De virginitate
(MGH Auct. ant. 15, 240), for which see Jenkinson (1908), xxii; Howlett (1995b), 9
n. 18; and the oldest manuscript of De ordine creaturarum, Paris, Bibliothèque nationale
de France, Lat. 9561, written in England during the eighth century.
14
Howlett (1996b), 28.


 AN ADDITION TO THE HIBERNO-LATIN CANON

as a perfected man in beheld paradise,

with the enemy thrust
in the abyss and deluded [or ‘crushed down’];
so also I, as much as I can,
lift hands to my God,
so that with the guilty [thief ] I shall [or ‘may’] enjoy Him;
bringing forth for the Son, the Father, the holy
flowing of the Spirit, as long as I live,
[my] flesh, I prune Paul’s
stem grafted in the muddy stubble [sc. of the gentile poet’s
body; cf. Romans 9:24 and 1 Cor 9:27],
so that I see
the Lofty One who created heaven, lands, and seas.’
I have presented elsewhere varied evidence that Christum peto was com-
posed in a seventh-century Hiberno-Latin milieu.15 An additional argu-
ment is that for an Irish poet voiced D and unvoiced T were compara-
ble sounds, illustrated by the adjacent position of the graphs for their
sounds in ogam, dair being || two strokes to the left of or above the base
line, and tinne being ||| three strokes to the left of or above the base line.16
On such a criterion the rhyme of reddo with leto in line 2 is pure.17 Pro-
nunciation of G as palatal, not velar, makes a pure rhyme of ego and queo
in line 9.18 A function of the last line is to conclude the poem, summing
up in its 15 syllables the preceding 15 lines.19 As the preceding lines are
all octosyllabic, it is likely that the last line was not composed for Chris-
tum peto, but borrowed into it from an already existing poem of penta-
decasyllabic verses. An indication of the direction of borrowing is that in
our poem the line is a self-contained unit, in which altum is an adjective
modifying celum, while in Christum peto it is used as a substantive, the

15
Howlett (1996b), 28.
16
Consider by contrast the differences in shapes and the distance apart of the
runes dæg and tir in the English futhorc; see Howlett (2005), 197–200.
17
This phenomenon recurs in other seventh-century Hiberno-Latin poems:
Nonae Aprilis, published in AD 640; In primo certe canone, a poem on the Eusebian Can-
ons; and Benchuir bona regula, composed certainly in Bangor, perhaps in AD 686; also
in the Old Irish poem Brigit bé bith maith by Bishop Ultán of Ardbraccan (†AD 657),
for which see Howlett (2011–12), 183.
18
Such a rhyme would be unlikely for an English poet, especially one who knew
that the expanded English futhorc included distinct runes for palatal G (X) and velar G
(<X>).
19
For other examples of compositions in which the number of one sort of element
prefigures the number of another sort of element that follows, or recapitulates the num-
ber of another sort of element that precedes, see Howlett (1998b); Howlett (2010a).


 DAVID HOWLETT

object of uideo in the preceding line of a different metre. The verse altum
celum qui creauit terras atque equora must have existed before Aedeluald
made his collection of prayers early in the eighth century or late in the
seventh, and before the author of Christum peto incorporated it into his
poem earlier in the seventh century. As all the prosodic and structural
phenomena of our poem have close analogues in other works that were
composed undoubtably in a Hiberno-Latin milieu (and could have been
composed in no other place known to the present writer), this poem also
is likely to have issued from Ireland in the seventh century, earlier than
composition of Christum peto.

Conclusion
The canon of Hiberno-Latin mathematical poetry now includes two
didactic computistic poems: Nonae Aprilis, published by Mo-Cuoróc
maccu Net Sémon at Crannach of Downpatrick in AD 640,20 and De
ratione temporum (Annus solis), composed also during the seventh cen-
tury; two poems about the Eusebian Canons: Ailerani sapientis canon
Euangeliorum, composed by Ailerán the Wise, lector of Clonard, some-
time before his death in AD 661, and In primo certo canone, composed
also during the seventh century;21 two poems about Problemata arith-
metica, quadam nocte niger Dub nomine Candidus alter by Clemens
Scottus, and Bis duo nam niuei praesunt et quinque nigelli by Thomas
Scottus, both about the middle of the ninth century.22
To a post-romantic sensibility the verses of Annus Solis may seem to
transport us not far up the slopes of Helicon. But considering their origi-
nal purpose, to inculcate basic computistic learning among schoolboys
in snappy, easily memorable form, they function more than competently.
If, like Nonae Aprilis, they were set to music, they would once have pro-
vided teachers with a formidable pedagogic tool and students with a use-
ful mnemonic device.

20
Rankin (forthcoming); Howlett (2013).
21
Howlett (2010c).
22
Howlett (2010d).


 MARINA SMYTH

ONCE IN FOUR: THE LEAP YEAR IN

EARLY MEDIEVAL THOUGHT

Abstract
A survey of early medieval computistical works into the early Carolingian pe-
riod reveals a number of interesting and unexpected themes on the subject of
the leap year. Representative examples are presented in this paper. It was com-
mon knowledge that Julius Caesar, in order to keep the calendar aligned with
the seasons, had introduced the practice of inserting an additional day in Febru-
ary every four years, so that the date we would call 24 February occurred twice
in that fourth ‘bissextile’ year. There was uncertainty as to why this additional
day was called bissextus. More curiously, there were two schools of thought on
the duration of the bissextus: was it a 24-hour day or a 12-hour day? The more
scientifically inclined understood that the leap-year day was necessary because
the solar year is a quarter of a day, a quadrans, longer than the 365 days of a nor-
mal calendar year. From that perspective, it followed that each year contributed
a 6-hour quadrans toward the bissextus. There was, however, a long tradition,
which was imported into Britain and Ireland, that the annual contribution to
the bissextus was a 3-hour quadrans. Some of the justifications, implications and
consequences of this erroneous belief are examined, and it is noted that these
are mostly found in texts with insular connections.
Keywords
Leap year, Julian calendar, equinoctial hours, bissextus, quadrans; Anatolius,
Macrobius, Victorius of Aquitaine, Bede, De temporibus, De ratione temporum,
Augustinus Hibernicus, De mirabilibus sacrae scripturae, Joshua, Joca monaco­
rum, Disputatio Chori et Praetextati, Argumentum XVI, Computus Einsidlensis,
Munich Computus, De ratione conputandi, Burgundian Dialogue (Dial. Burg.),
Neustrian Dialogue (Dial. Neustr.), Cologne Computus (Comp. Col.), Bobbio
Computus.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 229-264
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114739
 MARINA SMYTH

Introduction
The pre-Julian Roman calendar was determined by political considera-
tions rather than scientific principles and bore no relationship to the cy-
cle of the seasons.1 Julius Caesar determined to put an end to this chaotic
situation and called in the experts, who solved the problem for many
centuries2 by creating a completely new calendar, the Julian calendar we
all know, which approximates the solar year by normal years of 365 days,
with an extra day of 24 equinoctial hours added every fourth year (what
we call the leap year). It was decided at the time of this Julian reform
that the additional day would be inserted ante quinque ultimos Februarii
mensis dies (‘before the last five days of February’), in other words, after
the seventh kalends of March. Since February was a month of 28 days,
the intercalary day was thus added the day after what we call 23 Febru-
ary. During leap years there were two ‘sixth kalends of March’ in the new
Julian calendar, and the intercalary day was given the date ante diem bis
sextum Kalendas Martias,3 and therefore came to be called the bissextus,
as recounted by Macrobius.4 The epitome of Saturnalia 1.12–15 known
as the Disputatio Chori et Praetextati5 circulated in Britain and Ireland
by the early eighth century: one of the ‘Sirmond’ group of texts, it was
certainly known to Bede, who cites it briefly in his early work De tem­
poribus (AD 703), then mentions it by name and cites it extensively in
De temporum ratione (AD 725).6 The author of De ratione conputandi,

1
For more detail on Roman pre-Julian calendars, see Stern (2012), 205–11.
2
Barring the much later refinement of the Gregorian calendar, which assumes
a solar year of 365.2425 days, much closer to the now generally accepted value of the
mean tropical year (that is, ‘the average time from one year to the next between each
of the four equinoctial and solsticial points’), namely 365.24219  days; see Mossham-
mer (2008), 38–39. The Julian calendar reform was not consistently implemented until
AD 8, when regular intercalation of the leap-year day every four years was resumed and
continued thereafter until the Gregorian reform, with the intercalary day added to years
corresponding to multiples of 4 in our current system; see Blackburn and Holford-Stre-
vens (1999), 671; Mosshammer (2008), 36, 376.
3
Mosshammer (2008), 34–36; Blackburn and Holford-Strevens (1999), 669–73.
During the Middle Ages it was unclear which one of the two days was the bissextile
insertion.
4
Macrobius, Saturnalia 1.14.6 (ed. by Willis (1994), i 66–7).
5
Now edited separately by Leofranc Holford-Strevens; it will appear as a subsidi-
ary volume in the series STT.
6
Bede, De temporibus 6 (ed. by Jones (1943), 297); De temporum ratione 12–13
(ed. by Jones (1943), 206–10­).


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

­ riting between AD 719 and 727, also knew the Disputatio and includ-
w
ed the following passage:7

MACROBIUS: Ne quadras deeset, statuit ut quarto quoque anno sacer­

dotes, qui curabant mensibus ac diebus, unum diem interkalarent, eo scilicet
mense ac loco quo, apud ueteres, mensis interkalabatur, id est ante quinque
ultimos dies Februarii mensis, idque bissextum censuit nominandum.

‘MACROBIUS: So that the quarter [day] should not be missing, he

decreed that every fourth year the priests, who were responsible for the
months and days, should intercalate one day, in that same month and
place where the ancients intercalated a month, that is, before the last five
days of the month of February, and he declared that it should be called
the bissextus.’

Immo Warntjes has recently shown8 from the evolution of the content in
the three known early Irish computi—the Computus Einsidlensis (com-
posed between AD 699 and 7199), the Munich Computus (dated inter-
nally to AD 71910), and De ratione conputandi (between AD 719 and
72711)—that there was confusion in Ireland at the end of the seventh
century as to when the bissextile day should be inserted: should it be at
the sixth kalends of March (24 February) as the Romans did according
to Macrobius, or at the sixth nones of March (2  March) as Isidore of
Seville seemed to say,12 or on the day that the world was believed to have

7
De ratione conputandi 52 (ed. by Walsh and Ó Cróinín (1988), 164; the transla-
tion is mine); Ó Cróinín (2010), 332–33; Warntjes (2010a), LXVII n. 171. As noted
by Ó Cróinín, Macrobius is named at the beginning of the Disputatio (Walsh and Ó
Cróinín (1988), 137 n. 16). The authors of the Irish computi of the end of the seventh
and early eighth century usually mention Macrobius himself as the originator of infor-
mation they drew from the Disputatio.
8
Warntjes (2010a), CXLIII–CXLIV, CXCVIII.
9
Warntjes (2010c), 278. Einsiedeln, Stiftsblibliothek, 321 (647), 82–125; the
content of this manuscript can be consulted at www.e-codices.unifr.ch/en/list/one/
sbe/0321.
10
Munich, Bayerische Staatsbibliothek, Clm 14456, 8r–46r (ed. and trans.  by
Warntjes (2010a), 1–317; the manuscript can be viewed online at: http://daten.digitale-
sammlungen.de/~db/0004/bsb00046449/images/). For the date and its incorporation
of a Victorian computus of AD 689, see Warntjes (2010a), LVII–LXI.
11
Edited by Ó Cróinín in Walsh and Ó Cróinín (1988), 99–213; Warntjes
(2010a), CC.
12
Isidore, De natura rerum 6.7 (ed. by Fontaine (2002), 197): Bissextus autem a
sexto nonas martias usque ad diem pridie kalendarum ianuariarum in lunae cursu ad­
ponitur; Etymologiae 6.17.27 (ed. by Lindsay (1911), i): A vi autem Non. Mart. usque


 MARINA SMYTH

been created, that is, at the spring equinox, then assumed to be on the
twelfth kalends of April (21 March). As Warntjes noted, this last option
was entirely logical since the bissextile increment would have begun to
build on the very day of Creation, adding up to a full day at the end of
the fourth year thereafter.13 The Computus Einsidlensis even considers
this a valid option for calculations.14 The Munich Computus eliminated
21 March ne primi mensis initium turbet (‘lest [the bissextile day] disturb
the beginning of the first month’), that is, lest it confuse establishing the
date of the beginning of the first month, calculating backwards from the
Easter full moon.15 De ratione conputandi, which favours the Dionysian
reckoning for the date of Easter, gives serious consideration only to the
date given by Macrobius.16 Already in AD  703, Bede appears to have
taken it for granted that this was the appropriate placement for the bis-
sextile day in the Julian calendar.17
In the early Middle Ages, scholars knew that the regular leap-year
adjustment was essential for keeping the calendar in agreement with

in diem prid. Kal. Ian., in lunae cursu bissextus adponitur atque inde detrahitur. In these
passages, Isidore appears to be talking about the lunar rather than the solar bissextile
day (Warntjes (2010a), 130). This sentence is preceded in a number of the manuscripts
of the Etymologiae (and in the edition by Lindsay) by the following statement refer-
ring unambiguously to the solar bissextus: sive quod nequeat anno suo introduci, nisi bis
sextum nonas Martias conputaveris, hoc est et primo die sexto nonas Martias, et addito bis
sexto, alio die sexto nonas Martias iteraveris—further evidence for the belief that Isidore
assigned the leap-year insertion to 2 March, and looking very much like the resulting
adaptation of the usual formula assigned to the sixth kalends of March.
13
Warntjes (2010a), 130–31 n. 43–49. The Cologne Computus of AD 805 (ed. by
Borst (2006), 885–950) refers to this concept as the natural bissextus, byssextus secun­
dum naturam (Comp. Col. 3.3B (Borst (2006), 902)).
14
Warntjes (2010a), CXLIII–CXLIV.
15
Munich Computus 41 (Warntjes (2010a), 130–31).
16
De ratione conputandi 57 (Walsh and Ó Cróinín (1988), 168–69).
17
Bede, De temporibus 10 ( Jones (1943), 299); see also De temporum ratione 11,
38, 40, 41 ( Jones (1943), 205, 251, 254–55). He notes that the Egyptians set the in-
tercalated day at the end of their year, at the fourth kalends of September (29 August),
whereas the Romans set it at the sixth kalends of March (24 February), and, in De tem­
porum ratione 11, he specifies that the latter is the date at which ‘we make the inter-
calation’ (nostrae tempus intercalationis, quae fit sexto kl. martiarum die; Jones (1943),
205). Nonetheless, in De temporum ratione 6 ( Jones (1943), 193), Bede agrees with the
idea that, as far as nature is concerned, the year and the leap-year increments start at the
spring equinox. This is consistent with his belief that the sun and the moon were created
on 21  March, marking the beginning of time and of its measurement (De temporum
ratione 6 ( Jones (1943), 190)).


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

the seasons. They liked to quote St Augustine on this point18 and they
learnt from Macrobius that other cultures dealt with this issue in differ-
ent ways. Thus, like Bede, the author of De ratione conputandi was aware
that Rome and Alexandria set their four-yearly intercalary day at differ-
ent times, the ‘Egyptians’ inserting it before the beginning of the fifth
year, which then started on the third kalends of September (30 August)
according to the Julian calendar.19
The practice of presenting different opinions was also popular in
computistical works composed on the Continent, where it was only by
the end of the eighth century that the Dionysian Easter reckoning gen-
erally prevailed. The Merovingian Neustrian Dialogue of AD 737,20 for
instance, gathered information from several earlier computistical trea-
tises and stated that, whereas the Romans opted for duplicating the sixth
kalends of March (24 February), Isidore associated the bissextus with the
sixth nones of March (2 March), and others preferred 21 March, since
they believed the world was created on that date. It also notes that others

18
Augustine, De trinitate 4.4 (ed.  by William  J. Mountain in CCSL 50, 171):
Quattuor enim quadrantes faciunt unum diem quem necesse est intercalari excurso quadri­
ennio quod bissextum uocant ne temporum ordo turbetur. The Einsiedeln Computus (Ein-
siedeln, Stiftsbibliothek, 321 (647), 107–08) speculates on the practical consequences of
neglecting the bissextile insertion, concluding that in 360 years, because 90 days would
be missing in the calendar, the seasons would shift by one place with respect to the cal-
endar, with the first of January falling at the beginning of autumn, for example: Secundo
quaeritur si non conputaremus bissextum, quid esset? Dum est, hoc est tempus in aliut re­
diendo retrorsum mutaretur; similarly, the Munich Computus 41 (Warntjes (2010a),
132–33); De ratione conputandi 56 (Walsh and Ó Cróinín (1988), 168): AGUSTINUS
dicit: Quattuor quadrantes unum diem efficiunt, quem necesse est excurso quadriennio
interkalari, ne ordo temporum turbetur; Bede, De temporum ratione 39 ( Jones (1943),
252–53), cites the entire passage from De trinitate, and the reference was transmitted
into the Carolingian period.
19
De ratione conputandi 39 (Walsh and Ó Cróinín (1988), 149). The Alexandrian
fourth year, like that of the Romans, had 366 days, but since the year started at the end of
summer (according to the Julian calendar on 29 August in three years out of four), and
the months all had 30 days, with the intercalary days necessary to keep attuned to the
seasons inserted at the end of the year, and moreover since the days of the month were
simply numbered consecutively, the dating system was quite different; see Macrobius,
Saturnalia 1.15.1 (Willis (1994), i 69). Bede, De temporum ratione 11 and 38 ( Jones
(1943), 205, 251), knew that the Egyptian intercalary day occurred on the fourth kal-
ends of September (29 August); see Blackburn and Holford-Strevens (1999), 710.
20
Dial. Neustr. 27a (Borst (2006), 411–12): De locis bissexti. Secundum Romanos,
ut diximus, in VI. Kalendas Martii. Secundum Isidorum in VI. Nonas Martii [2 March].
Alii in XII. Kalendas Aprilis, quia initium mundi est. Alii in XI. Kalendas Ianuarii
[22 December] vel in VIII. Kalendas Ianuarii [25 December] vel in II. Kalendas Ianua­
rii [31 December], quia in illo tempore sol incipit ascendere. Item alii arbitrantur: Promp­
tum est tibi ponere in quocumque loco vis in anno quarto.


 MARINA SMYTH

again chose 22 December, or the 25th, or the 31st, ‘because that is when
daylight begins to increase’ (quia in illo tempore sol incipit ascendere).
And indeed others again, we are told, said that it does not matter when
you choose to insert the extra day, which is of course correct from an
astronomical point of view.
It certainly mattered, however, when it came to calculating the date
of Easter. In the Julian calendar, if the year immediately before a leap year
begins, say, on a Tuesday, it will end on a Tuesday, but the leap year fol-
lowing will start on Wednesday and end on Thursday, and the year there-
after will begin and end on a Friday.21 A practical consequence is that if
Christmas, for example, is celebrated on Tuesday in the year preceding a
leap year, it will be celebrated on Thursday during the leap year—that is,
one day of the week has been skipped, a fact which has been offered as a
possible explanation for the English name ‘leap year’.22 More generally, the
insertion of the bissextile day every four years means that it takes 28 years
(7×4) to get the same sequence of weekdays in the Julian calendar.
In order to show how all the days of the week are shifted by one after
the leap-year duplication of the sixth kalends of March, the Neustrian
Dialogue offers the strange ‘real-life’ analogy of military leaders seated
side by side in order of rank and all shifting to the adjacent vacated seat
when the leader with highest rank invites the leader second in rank to
join him in the first chair. These lateral moves leave an empty seat at the
end, which is then occupied by a new leader, representing the first day
of the following year! Such a bizarre example suggests that some consid-
ered this very arcane matter indeed.23

21
Early computi elaborated on how this affected the day of the week on particular
dates during the year, especially at the beginning of the year, when this was important for
determining the date of Easter. See, e.g., Computus Einsidlensis (Einsiedeln, Stiftsbiblio-
thek, 321 (647), 97, 106); Munich Computus 29 (Warntjes (2010a), 82–83); De ratione
conputandi 35 (Walsh and Ó Cróinín (1988), 146–47). See also, using more sophisti-
cated methods, Bede, De temporum ratione 21 ( Jones (1943), 222).
22
For this common explanation, see, for instance, the entry ‘leap year’ in D. Harp-
er’s Online etymology dictionary, consulted in January 2016; for a closer investigation of
the origins of the English name, see Blackburn and Holford-Strevens (1999), 677–78.
23
Dial. Neustr. 28a (Borst (2006), 413–14): Quomodo praeparatur sedis bissexti?
Id: Augetur dies diei in una die mensis. Quae dies augetur? Id est: V. Kalendas Martii super
VI. Kalendas Martii, in similitudine augmenti lunae. Haec res habet similitudinem in re­
bus humanis. Id est, ceu esset exercitus extensus in campo. Sedentes omnes principes, primus
sedens deseruissetque secundam sedem suam et venisset in sedem principis primi, ut essent
in una sede, et unusquisque deseruisset secundum ordinem sedem suam et venisset in sedem
alterius usque ad novissimum. Novissimus autem deserens sedem suam, alius princeps in
locum illius venisset, id <est> dies principalis alicuius anni secondi. Cf. Krusch (1910), 235.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

The additional day in the leap year also affects the age of the moon,24
which could radically affect the date of Easter Sunday, depending on the
theological rules for the limits of the age of the moon. Consequently,
the leap year might be identified with a large B in the margin of Easter
tables, as a reminder of the reason for an otherwise surprising jump in
the pattern of data.25 And then there was the extreme case of Anatolius,
in the second half of the third century, attempting to reconcile with the
Julian calendar his lunar 19-year cycle based on a solar year of 364 days
(i.e. 52 weeks) as defined in 1 Enoch 72:8–32, while allowing only two
leap years during his 19-year cycle.26
The leap year certainly complicated matters for computists27 and
Victorius even reckoned with the bissextus retroactively when discussing
the dating of events in the Old Testament: ‘with the same necessary pat-
tern of the bissexti’ (bissextorum pariter necessitate decursa).28 Cummian,
writing in Ireland in AD 632 or 633, knew that Victorius’ 532-year cycle
contained exactly 133 bissextile years29 and even credited Victorius with
being aware that the reason his Easter data repeated after 532 years was
that it combined the 19-year lunar cycle with the 28-year cycle of week-
days in the Julian calendar.30

24
Munich Computus 62 (Warntjes (2010a), 288–89): Sciendum si bissextus ali­
quid mutat in cursum lunae. Id est, mutat lunam, commutat Kalendas. See also Bede, De
temporum ratione 41 ( Jones (1943), 254–55).
25
See, e.g., the 532-year table of Victorius of Aquitaine as edited in Krusch (1938),
27–52, or the conveniently printed representation of the first 19-year cycle constructed
by Dionysius to follow the Cyrillan table in Declercq (2000), 102–03. See also the in-
complete set of tables for the Dionysian system on pp. 137–45 in Einsiedeln, Stiftsbib-
liothek, 321 (647), or the paschal table associated with Anatolius in fols  74v–75r of
Padua, Biblioteca Antoniana, I 27, reproduced in Mc Carthy and Breen (2003), 60–61
(with only two bissextile years in 19 years; see the following).
26
Anatolius, De ratione paschali 11 (ed. and trans. by Mc Carthy and Breen (2003),
51, 68, with commentary 99–100, 119–20, 164, 171). As the editors noted on p. 171:
‘the cycle was quite useless to all those countless people across the Roman Empire who
observed the Julian solar year with a bissextile every four years’.
27
For instance, to quote the Munich Computus 58 (Warntjes (2010a), 238–39):
Si quis exordio XVIIII annorum sollicite usque in finem disputet, cautus in bissextili anno.
28
Victorius, Prologus 9 (ed. by Krusch (1938), 24; repr. in PL Supplementum 3, 385).
29
This number could readily be determined from Victorius’ 532-year table starting
with the year of the Passion set in the bissextile year we call AD 28.
30
Cummian, De controversia paschali ll. 225–226 (Walsh and Ó Cróinín (1988),
88): quem <Uictorius per uicesimas> et oc<tau>as uic<es> cum kalendis .dxxxii. et bissext<is
.cxxxiii. i>n id i<psum>, unde ortus est, redire fecit, where the reconstruction is based in part
on the Prologue of Victorius. It is likely that Victorius himself did not understand why the
Easter dates repeat after 532 years; see, e.g., Blackburn and Holford-Strevens (1999), 808.


 MARINA SMYTH

While actual calculation of the date of Easter assumed that the bis­
sextus is a full calendar day, computus manuals of the early Middle Ages
reveal that this was by no means a universal assumption, and that some
questions could be given surprising answers.

What is the duration of the bissextus?

In late-antique and medieval texts, discussions of the bissextus often in-
cluded a statement such as: ‘the annual course of the sun lasts 365 days
and a quadrans’31 and it was this quadrans that required adding a calen-
dar day every four years.32
For those more knowledgeable of the prevailing cosmology, the an-
nual course of the sun was defined as the time it takes the sun to return
to the very same place against the pattern of the fixed stars, specifically
in the band of the zodiac.33 In such a scientific context, a ‘day’ was under-

31
Macrobius, Saturnalia 1.14.3 (Willis (1994), i 66), speaking of Caesar’s reform
of the calendar: ad numerum solis, qui diebus tricentis sexaginta quinque et quadrante
cursum conficit, annum dirigere contendit. This sentence is copied in the Disputatio Chori
et Praetextati (see, e.g., Oxford, Bodleian Library, Bodley 309, 103v). In Dionysius Ex-
iguus’ Epistola ad Bonifatium we read (ed.  by Krusch (1938), 83): Similiter octo anni
solares, si in summam redigantur, id est octies trecenteni sexageni quini et quadrantes, faci­
unt simul II̅ DCCCCXXII. Similarly, in the tract after Prologus Cyrilli 8 (ed. by Krusch
(1880), 343): Nam et sol in anno lucet dies CCCLXV et quadrans. Bede, De temporibus 9
( Jones (1943), 298): Annus solaris vel civilis est dum sol ccclxv diebus et quadrante zodia­
cum peragit. The Einsiedeln Computus seems to be strongly influenced by Anatolius and
I have not found a corresponding statement. On the other hand, the Munich Computus
21 (Warntjes (2010a), 90–91) reads: Annus certus solis diebus CCCLXV et quadrante, id
est quarta pars diei; similarly, De ratione conputandi 46 (Walsh and Ó Cróinín (1988),
154): Sciendum nobis quid sit annus solis. Id est cum solis cursus finitur, peractis .ccc.lxv.
diebus et quadrante.
32
De ratione conputandi 28 (Walsh and Ó Cróinín (1988), 139): Unus uero mensis
.xxviii. dierum, id est Februarius. Et in isto anno .ccc.lxv. dies sunt et quadras, de quo bis­
sextus quarto anno efficitur. Et haec ordinatio manebit apud Latinos usque ad diem iudicii,
et est qua nos utimur.
33
Macrobius, Saturnalia 1.13.9 and 1.14.4,6 (Willis (1994), i 62, 66–67), makes
it clear that the solar year corresponds to the time it takes the sun to traverse the signs
of the zodiac, and this information is taken over in the Disputatio Chori et Praetextati
(see, e.g., Oxford, Bodleian Library, Bodley 309, 103v–104r). Isidore, De natura rerum
6.4 (Fontaine (2002), 195), Etymologiae 6.36 (Lindsay (1911), i); Bede, De temporum
ratione 16, 36 ( Jones (1943), 215, 249): tantum nunc dicamus quia sol ccclxv diebus et
vi horis, luna xxvii diebus et viii horis, zodiaci ambitum lustrant. […] Item solis est annus
cum ad eadem loca siderum redit, peractis ccclxv diebus et sex horis, id est quadrante totius
diei. See De ratione conputandi 53 (Walsh and Ó Cróinín (1988), 165): Sciendum nobis
quid efficit bissextum. Id est solis naturalis cursus, quis duodecim sidera sunt in celo contra


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

stood to consist of 24 equinoctial hours,34 so that a quadrans—a quarter

of a day—would last 6 hours.
Early medieval discussions of units of time make it clear, however,
that whereas quadrans simply means ‘a quarter’,35 it was used in early
medieval computistical texts discussing the leap year to refer both to a
quarter of a 24 hour-day and to a quarter of a daylight day of 12 hours,
that is, to a period of 3 hours, usually assumed to be equinoctial hours.
While it is true that the concept of a 3-hour quadrans building up to the
leap-year intercalary day occurs mostly in medieval texts with insular,
and more specifically with Irish connections, it did not originate in Brit-
ain or Ireland.
We do not know where the third-century computist variously known
as Pseudo-Cyprian or as ‘the Computist of AD  243’ wrote a preface
to his proposed improvements on methods for calculating the date of
Easter,36 but there can be no doubt that he understood that the ‘quarter
day’ associated with the bissextus meant 3 hours. He explicitly stated:37

sol annuus per CCCLXV dies et quartam partem diei consummat. quae
ipsa quidem pars quarta diei habet tres horas. tres autem horae imaginem
portant illorum trium dierum qui in principio saeculi sine sole et luna

quae currit sol et quae ita uocantur, id est: Aquarius, Pisces, Aries, Taurus, Gemeni, Can­
cer, Leo, Uirgo, Libra, Scorpius, Sagittarius, Capricornus. Unsurprisingly, such statements
continue to appear in later computistical works, e.g., in the Cologne Computus of AD
805; Comp. Col. 2.2 (Borst (2006), 894): quoniam appareret de solis cursu, qui trecentis
sexaginta quinque diebus et quadrante zodiacum conficit.
34
Both the Irish Augustine, De mirabilibus sacrae scripturae 1.7 (PL 35, 2159),
and the anonymous author of Liber de ordine creaturarum 9.4–6 (ed. by Díaz y Díaz
(1972), 148–50) describe the duration of the ebb and flow of the tide in what must be
equinoctial hours, suggesting that this was normal in their environment (Smyth (1996),
251–58). Bede, in De temporum ratione 3, 24 ( Jones (1943), 183, 227), assumed that his
readers were familiar with the concept of equal equinoctial hours.
Bede, De temporum ratione 4, 38 ( Jones (1943), 184–86, 251). De ratione con­
35 5

putandi 21 (Walsh and Ó Cróinín (1988), 129) ascribes to St Augustine the statement
that quadrans refers to the quarter of anything: AGUSTINUS uero ait quod quarta pars
uniuscuiusque rei, quae in quattuor aequales diuiditur partes, quadras dicitur. See Pseu-
do-Balbus (saec. 3med.–4in.), Liber de asse minutisque eius portiunculis 7 (ed. by Hultsch
(1864–66), ii 73): item cuiuslibet rei integrae in quatuor partes aequales divisae quarta
pars quadrans vocabitur, reliquae dodrans.
36
Mosshammer (2008), 125–27; see also Mosshammer’s article in the present vol-
ume, pp. 43–70.
37
Pseudo-Cyprian or Computist of AD 243, De pascha computus 19 (ed. by Wil-
helm von Hartel in CSEL 3.3, 266; trans. by Ogg (1955), 17; for a discussion of the
weaknesses in von Hartel’s edition, see Mosshammer’s article in the present volume on
p. 45).


 MARINA SMYTH

fuerunt. et ideo quomodo tres horae in quadriennio quater computatae

unum diem duodecim horarum effecerunt, sic et ipsi tres dies, per quatuor
tempora ternos menses sibi defendentes annum post XII menses suppletum
demonstrauerunt. et sic per hanc multiformem trinitatem et ipsae duo­
decim horae euangelium unum in quatuor partes diuisum ostenderunt, et
tres menses per quatuor tempora id est per quatuor euangelistas electos a
Christo XII apostolos nobis manifestauerunt.

‘the sun year by year completes its course in 365 days and the fourth part
of a day. The same fourth part of a day comprises three hours. Now the
three hours bear the image of the three days in which the beginning of
the age is without sun and moon. And just as the three hours reckoned
four times in a space of four years amount to one day of twelve hours,
so these three days, standing for the three months in each of the four
seasons, indicate the completion of a year after twelve months. So also
by this multiform trinity the same twelve hours make known one Gos-
pel divided into four parts, whilst the three months of each of the four
seasons, which answer to the four Evangelists, make manifest to us the
twelve Apostles chosen by Christ.’

The two manuscripts transmitting this text date from the ninth century,38
and it is tempting to explain the error as resulting from a later ‘correc-
tion’, but a characteristic of this prologue is the consistent emphasis on
the biblical associations of all numbers under consideration and the con-
text of the ‘3-hour quadrans’ statement is a discussion of the number 3
which is seen, among other things, as ‘bearing the image’ of the three
days before the two great luminaries were created.39 The author might
have been thinking of John 11:9, where Jesus is reported to say: ‘Are
there not twelve hours in a day?’,40 so that a quarter of a day would indeed
be 3 hours, since the day mentioned in the Gospel was a daylight day
divided into 12 equal daylight hours (not to be confused with the equi-
noctial hours implied in the scientific explanation of the leap year). Or
must we understand that the author was assuming that to each 3 hours

38
One of these manuscripts, from Reims, is now lost, but it was transcribed twice
in the seventeenth century (CSEL 3.3, lxv; Ogg (1955), vii; Mosshammer, in this volume
p. 46).
39
It is curious that there is no mention of the well-known biblical 3-hours span of
darkness associated with the crucifixion, as mentioned in Luke 23:44.
40
Nonne duodecim horae sunt diei?, which later computi will give as an illustration
of the natural hours in the day, with duration changing with the seasons. See, for exam-
ple, Bede, De temporum ratione 3 ( Jones (1943), 182–3).


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

of daylight there corresponded 3 hours of night? Perhaps he did,41 but be

that as it may, the statement is confusing and reflects a common percep-
tion, based on the practical reality that daylight was the only time that
really counted in ordinary life. As the Irish Munich Computus phrased
it, claiming to quote Isidore: ‘darkness is nothingness’.42 Indeed the same
computus claims that Augustine said: ‘The bissextile day is a famous so-
lar feature, which has no effect on the moon’ (Bissextus famosus solaris
euentus, cui nulla uis est secundum lunam).43
In the common Roman duodecimal system of measurements, quad­
rans was the name for three twelfths. While the practice of partitioning
a unit, that is, an as or an assis, into twelve equal parts usually known
as unciae44 was ordinarily applied to weights, it was also applied to the
measurement of time, as the following examples will show.
During the second half of the third century, Anatolius of Laodicaea
wrote De ratione paschali,45 a justification for his attempt at creating a
19-year paschal cycle adapting the Alexandrian 19-year lunar cycle to
the Julian calendar. In the difficult section dealing with the annual as-
cent and descent of the sun,46 he refers to 25 March as the day ‘in which
12 hours are completed, that is, an assis’ (in qua consummantur xii horae
et assis), where the 25 March context makes it clear that the hours are
equinoctial hours. We shall see below that a number of early medieval
texts will associate Anatolius with the 3-hour quadrans.47

41
That is the understanding in the Carolingian Comp. Col. 3.3 (Borst (2006), 902)
in the description of the bissextus secundum naturam, assumed to be on 21 March after
growing per quattuor annos quattuor quadrantes noctis et quattuor quadrantes diei.
42
Munich Computus 41 (Warntjes (2010a), 120–21): Bissextus quasi bis sex­
ies ductus assem facit, quod unus dies [so far = Isidore, Etymologiae 6.17.26], id est XII
horarum, quia tenebrae nihil sunt. Warntjes notes that it is Augustine, not Isidore, who
wrote tenebrae nihil sunt—in De Genesi contra Manicheos 1.4.7 (ed. by Dorothea Weber
in CSEL 91, 74).
43
Munich Computus 41 (Warntjes (2010a), 120–21).
44
As shown by Bede in De temporum ratione 4 ( Jones (1943), 184–85).
45
Cited already by Columbanus and Cummian, the treatise of Anatolius is dis-
cussed in insular computistical works starting in the late seventh century, specifically in
the Einsiedeln and Munich computi, in De ratione conputandi, in the works of Bede, and
well into the Carolingian period; see Mc Carthy and Breen (2003), 39–42.
46
Mc Carthy and Breen (2003), 52; the relevant passage is quoted at n. 92, below.
See also Howlett (2008a), 146.
47
Even authors believing in the 6-hour quadrans could associate Anatolius with
the annual bissextile quadrans (e.g., Bede, De temporum ratione 6 ( Jones (1943), 193)).
On the Continent, see Quaest. Austr. 2.9 of AD 764 (Borst (2006), 492): Anatolius au­
tem noctem in praeparatione bissexti non procurat, et propterea secundum illum quadrans


 MARINA SMYTH

Isidore of Seville, writing his Etymologiae in the early seventh centu-

ry, is anything but clear on his understanding of the bissextile quadrans,
but the evidence suggests that he too is thinking in terms of a 12-hour
bissextus:48

Bissextus est per annos quattuor unus dies adiectus. Crescit enim per singu­
los annos quarta pars assis. At ubi quarto anno assem conpleverit, bissex­
tum unum facit. Dictus autem bissextus quia bis sexies ductus assem facit,
quod est unus dies; sicut et quadrantem propter quater ductum; quod est
bissextus quem super dierum cursum in anno sol facit.

‘The bissextus is the day added every four years, for in each year it grows
a quarter of a whole unit (quarta pars assis), but when it has completed
a unit in the fourth year, the bissextile day is made. It is called bissex­
tus because twice reckoned six times makes a unit (assem), which is one
day—which is also four times (quater) a quadrans. This is the bissextus
which the sun makes beyond the course of the days in the year.’

Indeed, in his discussion of the olympiad in the De natura rerum, Isi-

dore refers explicitly to an annual increment of 3  hours in the differ-
ence between the duration of the solar year and the normal Julian year of
365 days.49 Elsewhere, Isidore clearly distinguished between a day prop-

tres tantum horae sunt, quae in quarto anno duodecim horas, diem integrum, sed sine nocte
efficiunt. Similarly, among the numerous earlier comments on the bissextus collected in
the Bobbio Computus, chapters 33 and 35 (PL 129, 1292–94) speak of the leap year and
the quadrans, referring back to Anatolius though explicitly reckoning on a 6-hour quad­
rans. On the other hand, Bobbio Computus 39–43 (PL 129, 1295–96) which assume a
3-hour quadrans, make no mention of Anatolius.
48
Isidore, Etymologiae 6.17.25–26 (Lindsay (1911), i; the translation in mine).
Up to the first facit, this passage is quoted in De ratione conputandi 53 (Walsh and Ó
Cróinín (1988), 164–65), but the ending there reads bissextilem annum facit. Note that
in Munich Computus 41 (Warntjes (2010a), 122) this citation is attributed to Augus-
tine, though it occurs in essentially the same form as in De ratione conputandi where it
is correctly assigned to Isidore (Warntjes (2010a), CXXVI n. 367). The Computus Ein­
sidlensis (Einsiedeln, Stiftsbibliothek, 321 (647), 105), on the other hand, rejected false
etymology and the notion of a 12-hour bissextus (see the insistence on dies abusivus on
p. 108): Bis autem aduerbium numeri. Sextus non ordinis numeri. Bissextus autem quasi
bis sexies ductus, non quia dies .xii. horarum sit, quae bis sex dicuntur.
49
Isidore, De natura rerum 6.5 (Fontaine (2002), 195–97): Olympias enim est apud
Graecos annus quartus ab Olympio agone, qui uenit transactis annis quattuor, in cuius fi­
nem sortitur agonis tempus propter quadrienni cursum solis, et propter quod singulis annis
trium horarum consumptione in quadriennium dies unus conpleatur. Fontaine (2002),
343 n. 52, stressed that manuscript witnesses are unanimous in reading trium. This pas-
sage is influenced by Censorinus, De die natali liber 18.3–4, 12 (ed. by Ivan Cholod-
niak (1889), consulted on the LacusCurtius website http://penelope.uchicago.edu/


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

erly so called—from sunrise to sunrise—and a day abusive, that is, from

sunrise to sunset.50 Was there perhaps some confusion in his sources
between Roman numerals iii and vi? It seems rather that Isidore him-
self was confused on the matter of the quadrans and, perhaps because of
some false etymologizing (bis sexies ductus assem facit), he assumed that
the solar year lasts three hours more than 365 days.
The Prologus Cyrilli—written at the end of sixth century according
to Jones, but more likely composed already in the fifth century accord-
ing to Mosshammer51—was known in Ireland by the end of the seventh
century.52 A tract was added at the end of this Prologus, and the attribu-
tion to Cyril comes after the combined text in both Cologne, Dombibli-
othek, 83-II (early saec. IX),53 215r, and Milan, Biblioteca Ambrosiana,
H 150 inf. (saec. IX), 4v. It is assumed the two texts were linked very
early in their transmission,54 though it should be noted that the tract is
missing in the Sirmond manuscript.55 The tract assumes that hours and
unciae are the same thing, and it is clearly talking about a bissextus of
12 hours only when it affirms:56

Quadrans uero, qui super CCCLXV dies est, bissexti facit post quadrien­
num crescere diem. Ipsa uero quadrans III habet uncias, quae in anno uno

Thayer/E/Roman/Texts/Censorinus/home.html; I have not seen the 2012 edition by

Kai Brodersen), where the connection is made between the 4-year cycles of the Olympi-
ads and of the leap-year.
50
Isidore, De natura rerum 1.1 (Fontaine (2002), 173): Dies est solis orientis
praesentia, quousque ad occasum perueniat. Dies gemine appellari solet: proprie a solus ex­
ortu, donec rursus oriatur, abusiue a solis ortu usquequo ueniat ad occasum. Spatia diei duo
sunt, interdianum et nocturnum; et est dies horarum XXIIII, spatium horarum XII.
51
Jones (1943), 49–54, 373; Mosshammer (2013) reviews the earlier scholarship
on this text and argues from previously overlooked manuscripts that the text should be
called Praefatio sancti Cyrilli.
52
The Prologus Cyrilli (ed. by Krusch (1880), 337–42) is included in the so-called
Sirmond group of texts which would have been circulating in Ireland at the end of the
seventh century. In Cummian’s De controversia paschali ll. 223–243 (Walsh and Ó
Cróinín (1986), 86–89), the citation attributed to Cyril is from pseudo-Cyril’s Epistola
de pascha (ed. by Krusch (1880), 344–49).
53
See the online facsimile at www.ceec.uni-koeln.de.
54
Jones (1943), 373: ‘The Prologus Cyrilli, to which this tract is attached so tightly
that the explicit with Cyril’s name comes after it’.
55
The tract is not copied in Oxford, Bodleian Library, Bodley 309, 101r, at the end
of the Prologus proper (fols. 99v–101r).
56
Ratio solis vel lunae cursus atque bissexti (ed. by Krusch (1880), 343; the transla-
tion is mine).


 MARINA SMYTH

crescunt. Item in alio anno tres, in tertio III, in quarto III, et fiunt in unum
horae XII, faciuntque diem unum, qui uocatur bissextus.

‘As for the quadrans which is above the 365 days, after a period of four
years it grows into a day. For the quadrans has 3 unciae, which grow dur-
ing the first year. There are 3 again in the second year, 3 in the third year,
3 in the fourth year, all of which adds up to 12 hours, which constitute a
day which is called bissextus.’

Jones presented this passage as evidence that the use of unciae as a unit
of time, not of weight, is associated with the 84-year Easter cycle,57 but
we have seen above that this was a common understanding in the duo-
decimal system and need not be associated with any particular paschal
system.
The manuscript Oxford, Bodleian Library, Bodley 309 contains, on
folios  64v–73v, a version of a treatise De divisionibus temporum com-
posed in Ireland during the second half of the seventh century and fa-
vouring the Dionysian reckoning.58 This text gives the following defini-
tion of quadrans, answering the needs of both Christian exegetes and
computists:59

Quadrans dicitur a quarta parte unciae, qui alio nomine dicitur silicus,
id est sex scrupuli, quem Hebraei quadrantem, Graeci vero dodrantem,
Latini quadrantem vocant. Dicitur etiam et quadrans major, qui habet
tres uncias, hoc est quarta pars assis; assis enim duodecim uncias habet. Sic
etiam et quadrans in tempore sex horarum, quarta pars est diei naturalis,
quia quatuor quadrantes unum diem perficiunt, qui dies habet horas vi­
ginti quatuor.

‘A quadrans is so called as the fourth part of an ounce, which is also

called a silicus, that is, six scrupuli, and the Hebrews call it quadrans, the
Greeks dodrans, and the Latins quadrans. There is also a quadrans major,

57
Jones (1943), 334, 372–74; see also Wallis (1999), 324 n. 148.
58
Graff (2010), 113. In view of the more recent work by Warntjes demonstrating
extensive debate in Ireland between proponents of Victorius and Dionysius at the end of
the seventh century (Warntjes (2010c), 267–68, and, in more detail, Warntjes (2015)),
I favour a dating close to AD 700. However, as Eric Graff has observed, the Dyonisian
elements are inserted in an already existing school treatise (Graff (2010), 121–23), and
the section on the quadrans cited above occurs in this earlier treatise.
59
Oxford, Bodleian Library, Bodley 309, 66r (= PL 90, 655; after PL 90, 657, the
Patrologia edition (PL 90, 653–64) no longer coincides with the text in Bodley 309; see
Graff (2010), 113 n. 2; Wallis (1999), lxxiii n. 172; the translation is mine).


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

which contains three ounces, that is the fourth part of an as, since the as
contains twelve ounces. Similarly, when referring to time, the quadrans
contains six hours, the fourth part of a natural day, since four quadrantes
make up a day, and this day has twenty four hours.’

Both the Munich Computus and De ratione conputandi will later state
that the word as is one of the alternative names for the bissextus, thereby
revealing that some thought of the bissextus as composed of 12 equal
units of time.60
While the Pseudo-Dionysian Argumentum XVI61 is repeated in
­several early medieval continental computi, it occurs in only one of the
manuscripts of the Dionysian argumenta, namely in the Computus Dig­
beanus of AD 675 in the ninth-century manuscript Oxford, Bodleian
Library, Digby 63, 72v–79r. As Immo Warntjes noted, it is the only ar­
gumentum with a title (De racione bissexti) in this section of the Digby
manuscript, suggesting that it originally circulated on its own and was
simply added to the list of Dionysian argumenta. It was most likely com-
posed in Ireland.62 Certainty that 3 hours is the annual advance of the
sun on the calendar is a feature of Argumentum XVI:63

Sed ab hoc dicitur bissextus, quod in unumquemque mensem punctus unus

adcrescit. Punctus vero unus quarta pars horae est. IIII vero puncti unam
horam faciunt; XII vero puncti III horas explicant. Ergo in IIII annis ter­
nae horae, quae sunt XII, diem faciunt I, qui addatur Februario, cum VI
Kl. Mar. habuerit, ut in crastino sic habeat. Verbi causa, si hodie VI Kl. Mar.
additur ille dies in IIII anno expleto, nihilominus et in crastino VI Kl. Mar.
habeatur. Et ideo bissextus dicitur, quia bis VI Kl. Mar. habet Februarius.

‘It is called bissextus because one punctus is added to it every month. And
a punctus is the fourth part of an hour, so that four puncti make an hour

60
Munich Computus 41 (Warntjes (2010a), 122): apud calcentores asis uocatur.
See also the reconstructed reading in De ratione conputandi 52, based on manuscripts
containing the same material (Walsh and Ó Cróinín (1988), 163): Quot nomina bissex­
tus habet? Quinque. Id est interkalares dies apud Aegyptios; <as> apud alios, superfluum
lucis apud Latinos, aeque nundinarum diem et bissextum.
61
Argumentum XVI is ed. in Krusch (1938), 80–81, from Oxford, Bodleian Li-
brary, Digby 63 (saec. IX); an earlier edition from the same codex can be found in PL
67, 506–07; the text is now readily available online at www.nabkal.de/dionys.html.
62
Warntjes (2010b), 45, 53, and 92–95, esp. 92–93; Walsh and Ó Cróinín (1988),
125 n. 15.
63
Argumentum XVI (Krusch (1938), 80; the translation is mine). See, e.g., Quaest.
Austr. 2.9 (Borst (2006), 491).


 MARINA SMYTH

and twelve puncti make up three hours. So that in four years, those three
hours add up to twelve hours, and make one day, which is added in Feb-
ruary, when the sixth kalends of March [24 February] is reached, so that
it will also occur on the following day. For example, if today is the sixth
kalends of March [24 February] and that day is added in the fourth year,
then tomorrow will also be the sixth kalends of March [24 February].
It is called bissextus because February contains twice the sixth kalends
of March.’

The section of Argumentum XVI in the Irish computus copied in the

Einsiedeln manuscript64 clearly states that an annual bissextile contribu-
tion of 3 hours builds up every four years only to a day abusive: per quad­
riennium diem bissexti licet abusiuum faciunt et non diem proprium (‘in
four years these [increments] add up to the day of the bissextus, though
that day is incorrectly called ‘day’, it is not truly a day’). However, a cita-
tion from Augustine in De ratione conputandi shows that there circu-
lated in Ireland in the early eighth century a corrupt text of De Genesi
ad litteram, which would have the bishop of Hippo assert that each year
contributes only 3 hours to the bissextus:65

Sciendum nobis quomodo quadrantem in natura temporis esse super .ccc.lxv.

dies, in quibus currit sol, scire possumus. […] Et ita inuestigatur illud quos
diximus, id est solem contra singula signa .xxx. diebus et .x. horis et puncto
currere et inde efficere .ccc.lxv. dies et quadrantem ita, id est .xxx. enim dies
duodecies .ccc.lx. dies efficiunt. Decim uero hore duodecies .v. dies sunt. Duo­
decim uero puncti tres horae fiunt, et de his tribus horis quadras efficitur.
Et ita in naturali solis cursu deprehenditur quadras, ut Agustinus dicit in
Exameron: ‘Annus solis est cum ad eadem loca siderum’ sol ‘redit, quod non
facit nisi peractis .ccc.lxv. diebus et tribus horis, id est quadras’, quarta pars
diei bissexti ‘totius diei, quae pars quater ducta cogit interponi unum diem,
quem Romani bissextum uocant, ut ad eundem circuitum rediatur’.

‘We wish to know by how much the quadrans, as a measure of time,

exceeds the 365 days in which the sun has its course. […] And we then

64
Computus Einsidlensis (Einsiedeln, Stiftsbibliothek, 321 (647), 108). See also
n. 48, above, for the rejection in that computus of the false etymology of bissextus. For a
transcription of this passage, see Warntjes (2010a), 138 n. 92–106.
65
De ratione conputandi 55 (Walsh and Ó Cróinín (1988), 166–67; bold face and
the translation are mine); cf. Augustine, De Genesi ad litteram 2.14 (ed. by Joseph Zycha
in CSEL 28.1, 55): peractis trecentis sexaginta quinque diebus et sex horis, id est quadrante
totius diei, quae pars quater ducta cogit interponi unum diem, quod Romani bissextum
uocant, ut an eundem circuitum redeatur.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

investigate what we have said, that is, that the sun travels through each
sign [of the zodiac] for 30 days and 10 hours and a punctus,66 and this
builds up to 365 days and a quadrans as follows: 12 times 30 days gives
360 days; 12 times 10 hours gives 5 days; 12 puncti make 3 hours, and
those 3  hours make up a quadrans. And thus the [duration of the]
quadrans is made clear from the natural course of the sun, as Augus-
tine says in the Hexameron: ‘The solar year is when the sun returns to
the same place among the stars, which happens only after the passing
of 365 days and 3 hours, that is, a quadrans,’ the fourth part of a day,
‘of the whole bissextile day, and this part multiplied by four makes it
necessary to intercalate one day, which the Romans call bissextus, so that
it [the sun] can return to the same course’.’

The Irish author of De ratione conputandi generally favoured the quad­

rans of 6 equinoctial hours, but in his effort to present as complete an
overview as possible of all existing theories associated with the measure-
ment of time, he nonetheless gave a careful explanation of the 3-hour
quadrans based on multiplication by 12 of the punctus, the monthly con-
tribution to the bissextus according to Argumentum XVI—even citing
Augustine in support of that theory.67 The extant manuscripts of Augus-
tine’s treatise De Genesi ad litteram normally read that the annual incre-
ment is 6 hours, so that the text available in Ireland would have been
‘corrected’ by someone who believed in the 3-hour theory. Such a faulty
text survives in the ninth-century manuscript St Gall, Stiftsbibliothek,
161,68 where the text on p. 44 reads: peractis trecentis sexaginta quinque
diebus et tribus horis, id est quadrante totius diei. This same totally unam-
biguous statement is also attributed to Augustine in the section headed
Augustinus de bissexto in the early fragment Vatican, Biblioteca Apos-
tolica, Reg. lat. 586, at folio 125r–v (as part of a miscellaneous set of
fragments dating from the ninth to the twelfth century and associated

66
A punctus is a quarter of an hour. See Argumentum XVI, at n. 63 above.
67
De ratione conputandi 56 (Walsh and Ó Cróinín (1988), 168) will later quote
Augustine, De Trinitate 4.4, as: Quattuor quadrantes unum diem efficiunt, quem necesse
est excurso quadriennio interkalari, ne ordo temporum turbetur; but that citation says
nothing about the duration of the quadrans and the Irish author does not mention that
in that passage the bishop of Hippo was extolling the perfection of the number 6, for
which the 6-hour quadrans was presented as a God-given illustration. Bede cites that
same De Trinitate passage at length when discussing the 6-hour quadrans in De tempo­
rum ratione 39 ( Jones (1943), 252–53); see below, at n. 71.
68
The content of this manuscript can be consulted on the Codices Electronici San­
gallenses website, at www.e-codices.unifr.ch/en/list/one/csg/0161.


 MARINA SMYTH

with Fleury69). These incorrect citations from De Genesi ad litteram 2.14,

referring to tribus horis instead of sex horis, are further evidence that Au-
gustine could be erroneously associated in the early Middle Ages with a
three-hour annual contribution towards the bissextus.
Bede provided a table of duodecimal fractions in his De temporum
ratione and clearly stated that the duodecimal system also applies to
the measurement of time, but he found the idea of a 12-hour bissextus
derisory:70

Nec omnino putamus eorum suscipiendam esse sententiam qui, quasi nocti
nihil tribuentes, tres tantum horas per annum bissexto accrescere confir­
mant. Quod si ita esset, non ante annorum viii circuitum dies qui creverat
totus compleretur; diem namque totum, id est cum sua nocte, xxiiii habere
horas, etiam vulgus ignobile novit. Qui dum totus per quadriennium non
negetur impleri, qua ratione quarta pars eius per singulos quattuor anno­
rum negatur impleri?

‘As for those who say that only three hours accrue to the bissextus each
year, as if ascribing nothing to the night, we do not think that their judg-
ment is to be accepted at all. If this were so, the whole day which accu-
mulated would not be complete before the passage of eight years. For
even the base herd [uulgus ignobile] know that a whole day, that is, [a
day] together with its night, has 24 hours. If they do not deny that the
whole [day] is filled up in four years, why do they deny that one-fourth
is filled up by each of the four years?’

Bede followed this with a lengthy quote from Augustine’s De Trinitate

4.4 which, in the course of extolling the wondrous perfection of the
number six, states that ‘the four quarter-days (quadrantes) make one
day that must be intercalated in the course of the fourth year’ (Quat­
tuor enim quadrantes faciunt unum diem quem necesse est intercalari ex
cursu quadriennio), clarifying further that ‘the quarter-day itself has six
hours, for the whole day with its night has 24 hours, whose fourth part,
which is a quarter-day, is found to be six hours’ (Ipse autem quadrans sex
horas habet; totus enim dies, id est cum sua nocte, xxiiii horae sunt quarum
69
Mostert (1989), 271. I thank Leofranc Holford-Strevens for drawing my atten-
tion to the incorrect citation from De Genesi ad litteram 2.14 in this manuscript. The
content of this manuscript (including these fragments, though they are not associated
with Lorsch) can be viewed on the Bibliotheca Laureshammensis website, at http://
bibliotheca-laureshamensis-digital.de/bav/bav_reg_lat_586.
70
Bede, De temporum ratione 39 ( Jones (1943), 252; the translation is from Wallis
(1999), 108 (with seven changed to eight)).


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

pars quarta, qui est quadrans diei, sex horae inveniuntur).71 Could this ex-
plicit extensive selection from a major theological treatise by Augustine
be intended to outweigh the brief statement claiming a 3-hour annual
contribution to the bissextus which Irish computists falsely attributed to
Augustine in De Genesi ad litteram? Moreover, Bede’s citation gives the
full context of the brief quote from De trinitate in De ratione conputandi
56. In any case, Bede’s emphatic opposition to the notion of a 3-hour
bissextile increment is further evidence for the presence of this belief in
insular circles.
While the notion of a 3-hour bissextile increment dropped out of
mainstream scholarship,72 it persisted, at least as an option, in a number
of early continental computi incorporating Irish material, such as the
eighth-century Burgundian Dialogue (also known as the Merovingian
Computus of AD 727)73 and the Neustrian Dialogue of AD 737,74 as
well as in the Bobbio Computus of the early ninth century.75

71
Bede, De temporum ratione 39 ( Jones (1943), 253; Wallis (1999), 108–09). See
Augustine, De trinitate 4.4 (CCSL 50, 171–72).
72
Thus, for example, the Cologne Computus of AD  805 emphasizes that the
quadrans building to the bissextile day has 6 hours, that is, 3 hours during the day, but
also 3 hours during the night; Comp. Col. 3.2–3 (Borst (2006), 901–02).
73
We find the 3-hour theory in Dial. Burg. 12 (Borst (2006), 363): Secundum solis
cursum ratione annus solaris per menses duodecim trecentos sexaginta quinque dies habet
et horas tres. Annus, ut diximus, solaris ad annos viginti octo revertitur propter bissextum,
quia per ternas horas superiores in quattuor annos coniunctas unum diem efficiunt, et exple­
tur annus vicesimus octavus. A little later, Dial. Burg. 14 (Borst (2006), 366) gives a much
abbreviated version of the computation in Argumentum XVI (see pp. 257–8, below; see
also Krusch (1880), 55–56). For a similar explanation of the bissextus as formed by an-
nual 3-hour increments, see Bern, Burgerbibliothek 645, 50v (Krusch (1910), 235). See
Jones (1943), 373; Walsh and Ó Cróinín (1988), 161 n.  21: ‘these continental tracts
clearly derive from Irish exemplars’; Wallis (1999), 324 n. 148.
74
Dial. Neustr. 27 (Borst (2006), 410–11): De bissexto: Dum de saltu pauca os­
tendimus, de bissexto alia ostendere debemus pauca. Bissextus <dicitur>, ut alii putant,
quia bis VI. Kalendas Martii habet Februarius. In unoquoque mense ad materiam bissexti
unus punctus adcrescit. Unus uero punctus quarta pars horae est. Quattuor autem puncti
unam horam faciunt. Ergo in quattuor annis horae duodecim fiunt et unum diem faciunt.
Qui adicitur Februario, et ita dictum est: VI. Kalendas Martii hodie, sic et cras VI. Kalen­
das Martii.
75
The Bobbio Computus preserved in Milan, Biblioteca Ambrosiana, H 150 inf.
(partly edited as Liber de computo in PL 129, 1274–372) was ‘written at Bobbio about
810 and comes largely from an exemplar written in the Columban Burgundian founda-
tions’ ( Jones (1943), 373). It follows the normal Carolingian practice of relaying several,
sometimes even conflicting, opinions on particular topics. Many of the opinions cited in
this computus assumed that 3 hours is the annual contribution to the bissextus (chapters
39, 40, 41, 42, 43; PL 129, 1295–96).


 MARINA SMYTH

How does the extra time build up,

whether 3 or 6 hours every year?
Variations on the scientific explanation for the leap year
The more expert computists understood the consequences of the solar
year being 6 hours longer than a calendar year of 365 days. Bede can be
paraphrased as follows:76 ‘Say the moment of the equinox occurs at sun-
rise (i.e., 6am) in year 1. It then occurs at noon in year 2, and at 6pm in
year 3. In year 4, it occurs at midnight, and in year 5, it will occur again
at 6am, with the course of a whole day completed (utpote completo diei
totius circuitu recurrat).’
Such scientific considerations on the length of the natural solar year,
led to another type of speculation on the nature of the quadrans: Was it
a block of time somehow added to each year, and if so when? Or did it
grow continuously over the duration of the year?
The Einsiedeln Computus—assuming that a new 24-hour day starts
at the beginning of the night—expressed Bede’s idea in different terms:77

Ipse autem Augustinus diem bissexti per .iiii. annos comminuit, uerbi gra­
tia Kalendas Ianuarii dominicus et finitur dominico die annus primus
diebus .ccclxv. Adde .vi. horas usque ad dimidium noctis secundae feriae,
finiturque .iii. feria conpletis diebus .ccclxv. post sex in nocte .iii. feriae.
Adde sex horas usque ad ortum solis tertiae feriae et finitur .iiii.ta feria
eadem hora diebus conputatis .ccclxv. Adde .vi. horas a solis ortu usque ad
dimidium diei. Annus .iiii. incipit a medio diei quartae feriae finiturque
eadem hora quinta feria. Adde .vi. horas usque ad uesperam. Hinc com­
minuisti diem, id est quintam feriam, uerbi gratia ita donec .v.tus annus
incipiat a .vi.ta feria et non a quinta.

‘Augustine himself broke down the bissextile day throughout four years.
If January 1 is on a Sunday for example, then the first year ends on a Sun-
day after 365 days. Add 6 hours until the middle of the night of Monday.
After 365 days (the year) finishes on Tuesday after 6 hours in the night
of Tuesday (i.e. at midnight). Add 6 hours until sunrise on Tuesday, and
after 365 days (the year) ends at that hour on Wednesday. Add 6 hours
from sunrise to midday. Year four begins from midday Wednesday and
ends at the same time on Thursday. Add 6 hours until evening. In that

76
Bede, De temporum ratione 38 ( Jones (1943), 251).
77
Einsiedeln, Stiftsbibliothek, 321 (647), 107–08 (transcribed in Warntjes
(2010a), 124 n. 25–28; the translation is mine).


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

way you have assembled a day, that is, Thursday, and in this example
therefore the fifth year begins on Friday, not on Thursday.’

The Munich Computus also addresses this shift in the end of the natural
solar year due to the quadrans:78

Ubi ponantur hii quadrantes donec dies de his fit? Id est initium anni
et annum initium mutat. Tamen non facile, quia ortus solis anni solaris
orbem in confusa uarietate admutat, ut Augustinus dicit.

‘Where are these (bissextile) quarter-days placed until the time when a
day is produced by them? At the beginning of a year, and (every) year
(this) beginning varies. (This is,) however, not easy, since sunrise changes
the cycle of the solar year by a confusing variation, as Augustine says.’

Similarly:79

Primus annus incipit a uespere in VIII Kalendas Iulii et in media nocte in

VIII Kalendas eiusdem Iulii finitur, et reliqua.

‘The first year begins in the evening of 25 June and it is terminated at

midnight of 25 June (of the following Julian calendar year), and so on.’

Early medieval scholars might also be concerned about the nature of

the first quadrans in the leap-year day: is it a day quadrans or a night
quadrans? This question is discussed, though rather confusedly, in the
Munich Computus:80

Quo quadrante quaeritur bissextus initiatur. Ut alii quadrante diurnali.

Et non est facile eum in uernali tempore poni, nisi a quadrante nocturno
appareat incipi. Hinc nocturnum precedere bissextum propterea dicunt.
Diurnum enim preesse mundum ferunt.

78
Munich Computus 41 (Warntjes (2010a), 124–25). Warntjes has identified the
reference ascribed to Augustine as probably originating in the Missale mixtum secundum
regulam B.  Isidori dictum Mozarabes (see PL 85, 225). The phrase also occurs in the
Cologne Computus (AD 805), in the context of the discussion of Anatolius’ supposed
views on the quadrans; Comp. Col. 3.2C (Borst (2006), 901).
79
Munich Computus 41 (Warntjes (2010a), 136–37).
80
Munich Computus 41 (Warntjes (2010a), 128; my tentative translation, bor-
rowing elements from Warntjes’s translation and notes on pp. 128–29, as well as Graff ’s
(2010), 117–21, discussion of the meaning of the word mundum as it occurs in De diui­
sionibus temporum).


 MARINA SMYTH

‘It is asked, by what kind of quarter-day the bissextile day is initiated. As

some (say,) by a diurnal quarter-day. But the bissextile day cannot easily
be placed in springtime, unless it appears to be initiated by a nocturnal
quarter-day. For this reason, they argue that the bissextile day begins
with a nocturnal (quarter-day). They say that all of time began with a
diurnal (quarter-day).’

See also an earlier passage in the same text:81

Unde quadrans in initio mundi primum est finis lucis angelicae antequam
oreretur sol in VIIII Kalendas Aprilis.

‘Therefore the quarter-day at the beginning of the world constituted the

end of the angelic light before the sun rose on 24 March.’

While this is less than clear, it appears to be consistent with the views
on Creation expressed in Munich Computus 43, where it is stated that
the days of Creation all started at noon and that the sun was set in the
sky at noon on 24  March. In this scheme, the first quadrans would
presumably be a daytime quadrans, from noon till evening.82 However,
for those who believed that the world was created at midnight at the
spring equinox—a common view in early medieval Ireland83—it was
obvious that the first 6-hour quadrans was a night quadrans, ranging
from midnight till sunrise. Ps-Alcuin’s De bissexto—most likely a Car-
olingian composition incorporating Irish material—then goes on to
point out that since the sun was created at sunrise, the first full solar
year ended at noon (since a quadrans had to be added to the 365 days),

81
Munich Computus 7 (Warntjes (2010a), 18–19).
82
It remains difficult to understand what was meant by the ‘hour and a half ’ men-
tioned in Munich Computus 44 (Warntjes (2010a), 146–49): Hora et dimidium horae
uenit de luce ubi materiam quadrantis apparari dicunt, et finis trium luminum pallide
lucis, quia angelorum erat. The ‘light of the angels’ presumably alludes to Augustine’s idea
that the ‘days’ in the seven-day account of Genesis referred to the angels and their clear
knowledge after they were ‘illuminated’ when God said: Let there be light (see, e.g., De
Genesi ad litteram 1.1.32, 4.24.41). This would have been the only light in existence until
the sun and the moon were created on the fourth day, marking he beginning of time and
of the materia quadrantis.
83
Numerous texts associated with Ireland—including De diuisionibus temporum
(PL 90, 657), the Munich Computus 8, and De ratione conputandi 26—attribute to Je-
rome the statement: Media nocte factus est mundus. See Walsh and Ó Cróinín (1988),
134 n. 26; Warntjes (2010a), CXXXII and 27 n. 38–43.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

the second at 6pm, the third at midnight, the fourth at 6am84—anoth-

er variation on the scientific explanation of the origin of the 24-hour
bissextile day.
Then there were those who felt that it was not good enough to think
in terms of an annual 6-hour increment towards the leap-year day, but
preferred to think rather that there must be a combination of annual
quadrantes noctis (‘quarters of the night’) increments and annual quad­
rantes diei (‘quarters of the day’) increments, building up to a full 24-
hour day after four years.85 A variation on this occurs in the Bobbio
Computus, relaying a garbled understanding by first asserting that the
quadrans is forming throughout the entire year, in each month and in
each day, but then going on to say that in year one the quadrans lasts
from sunset till midnight, in year two it lasts from midnight till sunrise,
in year three until midday, in year four from midday till sunset.86
Indeed simple acknowledgment of the annual 6-hour contribution
toward the bissextus did not suffice for many a scholar, and even Bede
saw the need to break the quadrans down into smaller regular incre-
ments: one hour every two months, or half an hour in every natural solar
month, or indeed the fourth part of an hour (that is, a punctus) in half a
month, all adding up to six hours every year. Bede also explained the an-
nual quadrans in terms of the time it took the sun to traverse one of the

84
See, e.g., Pseudo-Alcuin, De Bissexto (PL 91, 996), which Warntjes holds to be
a Carolingian continental compilation ‘drawing on Irish sources’ (Warntjes (2010a),
XXII n. 37, in his discussion of the contents of Oxford, Bodleian Library, Bodley 309).
For more information on this text, see Springsfeld (2002), 203–14.
85
Einsiedeln, Stiftsbibliothek, 321 (647), 105: Hinc si conputaueris .iiii. quadrantes
noctis, uidebis quartum quadrantem noctis deesse, ueluti quadrans diei et quadrans noctis
in primo anno, quadrans diei et quadrans noctis in .ii. anno, quadrans diei et quadrans noc­
tis in .iii. anno, quadrans diei in .iiii. anno. [there follows an explanation of why the quad­
rans noctis for the fourth year should be set in first position] Et si .iii. quadrantes noctis et
diei conputaueris, .iiii.to anno congrue uidebis .iiii. quadrantes noctis et diei conuenire, qui
diem bissexti conplent quadrante diei terminante diem ordinaliter. See also Comp. Col. 3.3
(Borst (2006), 902): Deinde principium anni XII. Kalendas Aprilis <est>. et inde incipi­
unt per quattuor annos quattuor quadrantes noctis et quattuor quadrantes diei, et faciunt
diem in XII. Kalendas Aprilis.
86
Whereas Bobbio Computus 44 (PL 129, 1296) clearly assumes that the quad­
rans contributing to the bissextus is 6 hours long, the explanation is rather simplistic:
Fit quadrans per totum annum, et mensem, et diem. Primo anno incipit quadrans a solis
ortu [sic; should read occasu] usque in medium noctis. Incipit quadrans anno secundo a
dimidio noctis usque ad solis ortum. Quadrans tertio anno usque ad sextam horam. Quad­
rans quarto anno a sexta hora usque ad solis occasum, where the text goes on to explain
why the intercalary day was added at 24 February.


 MARINA SMYTH

signs of the zodiac, which was assumed to be 30 days and 10 ½ hours.87

These notional monthly figures were no doubt the result of dividing
365 days and a 6-hour quadrans by 12, assuming twelve months of equal
duration and a continuously regular progression of the sun. While the
punctus seemed to be the preferred unit for this incremental approach,
there was nothing to stop computists from calculating the bissextile ac-
cretion for ever shorter periods. The Munich Computus for instance,
figured out that since half an hour is 20  momenta, it follows that the
daily bissextile contribution is ⅔ of a momentum.88 Bede had little pa-
tience with such calculations.89 It was nonetheless only natural that some
computists should introduce the analogy of the child gradually and con-
tinuously growing in the womb until it is born for all to see.90

The quadrans associated with increasing and decreasing daylight

This approach to the bissextus based on the changing duration of the
period of daylight during the year derived from that passage by Anato-
lius where he discussed the lengthening and shortening of the period of
daylight throughout the year, starting at the winter solstice.91 Anatolius
divided the year into four periods of some ninety days each, bounded
by the solstices and the equinoxes. He erroneously assumed a constant
rate of elongation of the daily period of daylight in the months from the
winter solstice to the spring equinox, and said that daylight increased
then by four momenta each day:92

87
De temporum ratione 39 ( Jones (1943), 252).
88
Munich Computus 41 (Warntjes (2010a), 136–37).
89
Bede, De temporum ratione 38 ( Jones (1943), 251), where he ‘derides those
who practised excessively minute calculation of the bissextile accretion’ (Walsh and Ó
Cróinín (1988), 161 n. 15).
90
The comparison with a pregnant woman apparently first occurred in a somewhat
different form in Munich Computus 41 (Warntjes (2010a), 126): Bissextus plenus est se­
metipse, ut mulier concipiens primo unum corpus, postea II corpora plena. It ‘became fairly
popular in eighth-century Frankish computistics’ (Warntjes (2010a), 126). For example,
in Dial. Neustr. 28a (Borst (2006), 414), we read: Cui simile est incrementum bissexti et
ortus eius? Id: Mulieri concipienti in utero et parturienti.
91
Anatolius, De ratione paschali 13 (Mc Carthy and Breen (2003), 52–53, 69–70).
This treatise was already quoted by Columbanus (Mc  Carthy and Breen (2003), 39).
Cummian later names Anatolius as one of his sources of information (Walsh and Ó
Cróinín (1988), 84).
92
Anatolius, De ratione paschali 13 (Mc Carthy and Breen (2003), 52; the transla-
tion is mine and differs from that offered in the edition at p. 69). A moment is a unit of
time, where 40 moments=1 hour.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

Per xv dies et horae dimidium sole ascendente per singula momenta, id

est per iiii in una die, at ab viii Kal. Ianuarii in viii Kal. Aprilis hora de­
minuitur, in qua consummantur xii horae et assis, id est prima pars, et ex
ea inchoatur particula prima ex xii partibus.

‘As the sun ascends for 15 days and half an hour, [the period of daylight
increases] gradually by momenta, that is, by four [momenta] each day.
From the eighth kalends of January [25 December] up to the eighth kal-
ends of April [25 March], the [duration of the night] hour[s] decreases.
At that time, 12 hours and one unit (xii horae et assis) are completed,
that is, the first sign [of the zodiac], and from that is begun the first little
part of the 12 signs.’

Now Anatolius assumed only two bissextile years in his attempt to rec-
oncile the 19-year lunar cycle with the Julian calendar, so that he does
not speak here about the bissextus, but rather of the travel of the sun
through the signs of the zodiac. However, a computist familiar with Isi-
dore’s terminology and confused by this difficult passage in the treatise
by Anatolius, might naturally assume that the mention of a 12-hour as­
sis must refer to the bissextile intercalary day—all the more so that this
passage refers to that part of the year in which the leap-year day was
normally inserted. Also, given Anatolius’ assumption that the duration
of daylight went from 6 to 12 hours in that period, this 6-hour increase
in the duration of daylight would imply a daily increase of 2 ⅔ momenta.
Later computists would therefore be puzzled by the 4-momenta daily
increment proposed by Anatolius in his complex system. These compu-
tists, probably in seventh-century Ireland, combined all this information
by positing a virtual daily increase of (4–2 ⅔) momenta=1 ⅓ momenta in
the duration of daylight, which adds up to 3 hours in the 90-day period
from the winter solstice to the vernal equinox. Since this was the only
place Anatolius spoke of the bissextus—as they thought—in his break-
down of the year, they apparently assumed that this is how he explained
a 12-hour bissextus:93

93
Einsiedeln, Stiftsbibliothek, 321 (647), 99 (transcription in Warntjes (2010a),
110; the translation is mine). The Einsiedeln and the Munich Computi speak of ‘pale
light’ and of a mysterious ‘cloud’. The Munich Computus 39 (Warntjes (2010a), 110–
12), for instance, says that these three hours are ‘in a cloud’ (in nube): VI horae praesen­
tiae solis adunantur, et III in nube noscuntur […] Momentum uero et tertia pars momenti
ad preparationem bissexti, ut alii dicunt. In De ratione conputandi 50–51 (Walsh and Ó
Cróinín (1988), 160, 163), we read: Ordinauit autem Anatolius annum solis, qui diebus
.ccclxv. et quadrante peragitur, in quatuor nonaginta aequalia […] Et in primo nonaginta
crescentis lucis frangit quadrantem preparationis bissexti ideo, quia in primo nonaginta, id


 MARINA SMYTH

Hic interrogandum est, quomodo crescit solis lumen ab XI Kalendas Ianu­

arii usque in XII Kalendas Aprilis. Hoc est per II momenta et tertiam par­
tem bis momenti. Cuius probatio haec est: II momenta in unoquoque die
crescentia per XV dies XXX momenta faciunt. Tertia autem pars bis mo­
menti coniunctis ternis singulis diebus per XV dies X facit momenta. Additis
his X ad XXX, XL faciunt momenta. Hinc VI horae ab XI Kalendas Ianu­
arii usque in XII Kalendas Aprilis crescunt, duae enim horae per menses
singulos augentur. Est quoddam augmentum, quod quadrantem efficit, id
est momentum uniuscuiusque diei et tertia pars momenti ante ortum solis
et post occasum. Cuius probatio haec est: Per XV dies, hoc est momentum
uniuscuiusque diei, XV momenta sunt. Tertia pars momenti per XV dies, V
facit momenta. Coniunctis ternis singulis diebus inde horae III per III men­
ses efficiuntur ab XI Kalendas Ianuarii usque in XII Kalendas Aprilis, quae
lucis pallide sunt. Inde haec dies horas XV horas habet, nox uero VIIII horas.

‘We ask how much the light of the sun increases from the eleventh
kalends of January [22  December94] till the twelfth kalends of April
[21  March]. This is by 2  momenta and the third part of 2  momenta.
This is shown as follows: 2 momenta each day add up to an increase of
30 momenta in 15 days. As for the third part of 2 momenta, adding the
thirds every single day for 15 days gives 10 momenta. Adding these 10
to 30 gives 40 momenta.95 Thus there is a 6-hour increase from the elev-
enth kalends of January [22 December] till the twelfth kalends of April
[21  March], since 2  hours are added each month.96 It is that increase
which makes the quadrans, that is, a momentum each day and a third of
a momentum before sunrise and after sunset. This is shown as follows:
Over 15 days, the momentum every single day adds up to 15 momenta
and the third parts of a momentum add up to 5 momenta. Adding the

est in sexto Kl Martii in Februario, Iulius Cesar diem qui de eo efficitur possuit. Et unus­
quisque dies in nonaginta momentum et terciam partem momenti de quadrantis fractione
habet, id est dimedium horae in .xv. diebus. Et quadras trium horarum his frangitur et hora
de tenebris ad praesentiam lucis a .xv. diebus et dimedio horae quadrantis crescit. […] Sci­
endum nobis quod ideo Anatolius quatuor nonaginta aequalia et fractionem quadrantis
<et> .v. dierum fecit, id est ad aequalitatem crescendi lucis et tenebrarum in quatuor no­
naginta diebus. Sciendum nobis quod Anatolius artificalem quadrantem et artificales dies
hic frangit, neque fractionem in natura temporis esse. See the excellent account by Warntjes
of later interpretations of Anatolius’ division of the year (Warntjes (2010a), 110–11).
94
Since 21 December was the winter solstice, daylight begins to increase the fol-
lowing day.
95
Bede knew that there are 40 moments in an equinoctial hour (De temporum
ratione 3 ( Jones (1943), 183)); this is also the understanding, for example, of the author
of Munich Computus 6 (Warntjes (2010a), 16).
96
More accurately, in a non-leap year, there are six 15-day periods from 22 Decem-
ber till 21 March.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

thirds every single day, makes up 3 hours in three months from the elev-
enth kalends of January [22 December] till the twelfth kalends of April
[21  March], and these are [hours] of pale light. So that that day has
15 hours and that night has 9 hours.’

While the logic is less than clear, one cannot fault the arithmetic. Some-
what later, with increased realization that it was more correct to think of
a 24-hour bissextile day, it was claimed that Anatolius asserted that for
the first 90-day period after the winter solstice, the materia bissexti was
increasing during the day, whereas it increased at night during the third
90-day period. That is what the Neustrian Dialogue says:97

De incremento solis et bissexti

Ab XII. Kalendas Ianuarii usque XII. Kalendas Iulii sol duobus momen­
tis et duabus partibus momenti ab ortu solis usque ad occasum adcrescit.
Id est: Per quindecim dies hora facitur. Hoc est: Sex horae crescunt ab XI.
Kalendas Ianuarii usque XII. Kalendas Aprilis.
Et ab XI. Kalendas Aprilis usque ad XII. Kalendas Iulii altere sex horae
crescunt. At ab illo loco sic sol decrescit.
Ad materiam autem bissexti.
De XI. Kalendas Ianuarii usque XII. Kalendas Aprilis momentum et ter­
tia pars momenti adcrescunt.
Ab XI. Kalendas Aprilis usque XII. Kalendas Iulii materia bissexti non
crescit.
Et ab XI. Kalendas Iulii usque XII. Kalendas Octobris materia bissexti in
nocte crescit.
Ab XI. autem Kalendas Octobris usque ad XII. Kalendas Ianuarii bis­
sextus non crescit. Sic praeparatur bissextus et crescit per quattuor annos,
mane et vespere per quattuor annos, ut alii dicunt.

‘On the increase of the sun and the bissextus

From the twelfth kalends of January [21 December] until the twelfth
kalends of July [20 June], the sun adds two momenta and two parts of a
momentum from sunrise to sunset. That is: an hour is made each 15 days.
That is: there is a 6-hour increase from the eleventh kalends of January
[22 December] till the twelfth kalends of April [21 March].
And from the eleventh kalends of April [21 March] till the twelfth kal-
ends of July [20 June], there is another 6-hour increase. From then on
the sun decreases in a similar way.
Regarding the materia bissexti

Dial. Neustr. 28–28a (Borst (2006), 413; the translation is mine).

97


 MARINA SMYTH

From the eleventh kalends of January [22 December] until the twelfth

kalends of April [21March], a momentum and a third of a momentum
are added.
From the eleventh kalends of April [22 March] till the twelfth kalends
of July [20 June], the materia bissexti does not increase.
And from the eleventh kalends of July [21 June] until the twelfth kal-
ends of October [20 September], the materia bissexti increases during
the night.
From the eleventh kalends of October [21 September] until the twelfth
kalends of January [21 December], the bissextus does not increase. This
is how the bissextus is prepared, and it grows for four years, morning and
evening for four years, as some say.’

There clearly was confusion on whether the bissextile increment was

time that should be added all at once or whether it should be partitioned
out somehow over the year:98

Alii dicunt <incrementum> bissexti adicere, alii non adicere, sed frangere
et dispergere in dies et menses et tempora, ne crescat numerus dierum.

‘Some say the increment of the bissextus should be added, others say that
it should not be added, but rather broken up and spread out over the days,
months and seasons, so that the number of days would not increase.’

Leofranc Holford-Strevens has drawn my attention to the last fragment

in Vienna, Österreichische Nationalbibliothek, Ser. Nov. 37, which he
assigns to Salzburg, saec. VIIIex., derived from an Irish exemplar.99 The
text ends with a section from Epitoma 11 of Virgilius Marro Gram-
maticus, a not infrequent source for computistical treatises. However
the content of this particular passage—stella derived from Stillone; dif-
ference between stella, astra, and sidus100—would be more appropriate
as an addition to De cursu stellarum of Gregory of Tours, fragments of
which also occur in this manuscript.101 The computistical section, poorly
copied and in poor Latin, speaks unambiguously of a 12-hour bissextile
day, which is built over four years by 3-hour annual increments. It goes

98
Dial. Neustr. 28a (Borst (2006), 414; the translation is mine).
99
Private communication 24 August 2010.
100
Virgilius Maro Grammaticus, Epitomae 11 (ed. by Polara (1979), 148–50; Löf-
stedt (2003), 229).
101
Gregory of Tours, De cursu stellarum is ed. by Bruno Krusch in MGH SS rer.
Merov. 1,2, 404–22; Obrist (2002) has identified additional manuscripts.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

on to discuss the increase in daylight from 25 December till 24 June, ap-

parently ending with ‘a day of 15 hours and a night of 8 hours’. Clearly
at least one of these numbers must be wrong, but the statement might
better approximate observed reality for the author (or the scribe) than
would Anatolius’ theoretical figures for the summer solstice: 18 hours
of daylight and 6  hours of night. These computistical fragments—for
which I have not been able to identify an immediate source—are further
evidence for the association of the 12-hour bissextus with the discussion
of the annual pattern of increasing and decreasing daylight as initiated
by Anatolius.

Arithmetical computations
Yet another type of reasoning occurs first in the pseudo-Dionysian Ar­
gumentum XVI and resurfaces, among others, in the Bobbio Compu-
tus.102 It is based on calculations involving the God-given number 7.
First there is the simple division by 7 of the number of hours in the day:
24/7=3 with remainder 3, and these remaining 3 hours are said to be the
source of the bissextus. A much more complex computation103 amounts
ultimately to dividing by 7 the total number of hours in the 365-day
year (8760 hours). This yields 1251 hours, again with remainder 3. This
remainder of 3 hours is then assumed to be the annual increment adding
up eventually to the bissextus:104

Sex diebus fecit Deus mundum, septimo requieuit. Ut ergo plenius intel­
legatur, conputa qantas horas habet unus dies, et divides illas in VII partes
et qantas remanent, exinde fit bissextus. Primo conputa dies CCC, quo­
modo horas habent, decies trecenti, sunt III ̅. Iterum facis; bis tricenteni,
sexcenteni, fiunt in tricentis diebus horae III D̅ C. Iterum facies; decies
se­xageni DC, et bis sexageni CXX. Fiunt ergo in sexagenis diebus hor­
ae DCCXX. Iterum facis: decies quini L, et bis quini X.  Ecce habes in
quinque diebus horas LX. Fiunt simul integro anno in diebus CCCLXV
horae IIII ̅CCCLXXX, et alias tantas in nocte, fiunt simul dierum et noc­

102
Pseudo-Dionysius, Argumentum XVI (Krusch (1938), 80–81); Bobbio Com-
putus 43 (PL 129, 1296). See Warntjes (2010b), 93 n. 151, for Irish or Irish-influenced
computistical works—the oldest being the Computus Einsidlensis (Einsiedeln, Stiftsbib-
liothek, 321 (647), 108)—containing these calculations characteristic of Argumentum
XVI. See also Bisagni and Warntjes (2008), 89–90; Springsfeld (2002), 204.
103
Assisted perhaps with the multiplication tables from ×2 to ×50 in the Calculus
of Victorius of Aquitaine (floruit c.AD 450), edited on pp. 1–37 of Peden (2003).
104
Pseudo-Dionysius, Argumentum XVI (Krusch (1938), 80–81; the translation is
mine).


 MARINA SMYTH

tium totius anni horae VIII̅DCCLX. Divide illas in VII partes. Primum
facis: septies milleni VII̅, remanent I D
̅ CCLX. Item facis: septies ducen­
teni, fiunt I C
̅ CCC, remanent CCCLX. Item facis: septies quinquageni,
fiunt CCCL, remanent X. Item facis: septies as, VII, remanent III. Iste
tres horae faciunt in III [sic; must read IIII] annis diem.

‘God created the world in six days, and rested on the seventh. To under-
stand this more fully, count the number of hours in a day, divide them
in 7 parts and the bissextus is made up of the hours that remain.105 First
compute the number of hours in 300 days: ten times three hundred,
that is, 3,000; then twice three hundred is six hundred, so that there are
3,600 hours in 300 days. Now take ten times sixty is 600 and twice sixty
is 120. Thus there are 720 hours in sixty days. Now take ten times five is
50 and twice five is 10. And so you have 60 hours in five days. Adding it
all up over the 365 days of a full year, there are 4,380 hours, and another
equal amount in the night. So that there are altogether 8,760 hours of
day and night in an entire year. Divide these in 7 parts. First you take:
seven times a thousand is 7,000, there remains 1,760. Then take: seven
times two hundred is 1,400, there remains 360. Then take: seven times
fifty is 350, there remains 10. Then take: seven times a single unit is 7,
there remains 3. In 4 years, these three hours make up a day.’

Just as those who started with the punctus were thinking in terms of ide-
alized months and divided the duration of the year by twelve, so this
method seems to be thinking in terms of idealized weeks spread evenly
throughout the year and the calculation is determining how many hours
would be assigned to such a week. But there are not exactly 52 weeks in
the year, and anyway the computation has no bearing on reality.

Is the leap year part of the created order of the universe?

During the early Middle Ages, not everyone believed that the discrep-
ancy between the duration of the solar year and the calendar year was
part of the created order. We know from the version of Argumentum
XVI in the Computus Digbeanus of AD 675, probably compiled in an
Anglo-Saxon centre in Ireland,106 that some must have believed that the

24/7=3 with remainder 3.

105

Warntjes (2010b), 95.

106


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

pattern of leap years started when the sun stopped to allow Joshua to
completely rout his enemies:107

Bissextum non ob illum diem fieri, ut quidam putant, iessue oravit solem
stare, credendum est.

‘We should not believe, as some do, that the bissextus occurs because of
that day when Joshua prayed that the sun would stop.’

The argumentum gives a very simple and very sound argument against
that theory: ‘because that day happened then, and is now gone’ (quia
dies ille et fuit et praeteriit).108 In other words, that was a one-time event,
and one should not assume that it was ever repeated. Evidence for the
association of the leap year with Joshua is found in a number of early
computistical works, often with similar disapproval.109
According to the Latin biblical account of the miracle, in response
to Joshua’s prayer, ‘the sun stood in the middle of the sky and hastened
not to go down the space of one day’.110 Ecclesiasticus also speaks of ‘one
day becoming as two’.111 This duplication of the day must have triggered
the association with the leap year, that year in which two days have the
same name in the Julian calendar. Given the ambiguity of the ‘day’ in the
biblical texts—is it a 24-hour day or ‘day’ as opposed to ‘night’?—this

107
Pseudo-Dionysius, Argumentum XVI (Krusch (1938), 80; the translation is
mine). Note that the entire first paragraph of the Digby 63 version of Argumentum XVI
is absent from the Computus Einsidlensis.
108
Pseudo-Dionysius, Argumentum XVI (Krusch (1938), 80; the translation is
mine). Copied in Bobbio Computus 42 (PL 129, 1296). See also Joshua 10:14.
109
Quaest. Austr. 2.9 (Borst (2006), 491); Comp. Col. 5.7 (Borst (2006), 936), pos-
sibly referring to the discussion by Augustinus Hibernicus (see n. 119 below), concludes:
Item alii male putant hanc stationem prestare byssextum; De argumentis lunae (PL 90,
701–24: 722; note that when citing Argumentum XVI, this 10th-century text speaks of
the 12-hour bissextus as diem artificialem); see also the notes concerning the bissextus in
Vatican, Biblioteca Apostolica, Reg. lat. 39 (saec. IXmed., probably French, according to
Wilmart), which are derived from Argumentum XVI (Wilmart (1933), 27).
110
Joshua 10:13. The Vulgate text reads: Stetit itaque sol in medio caeli, et non festi­
nauit occumbere spatio unius diei. Whichever version(s) of this text was available in Ireland
in the seventh century, the Irish Augustine, De mirabilius sacrae scripturae 2.4 (PL 35,
2175), writing in AD 654, understood the sun to remain in the sky twice as long as usual
during this miracle: Nam sol duos dies in uno conclusit, et luna diei spatio non occurrit.
111
Ecclesiasticus 46:5: Et una dies facta est quasi duo (according to the Vetus Latina
database, this is the translation from the Septuagint in the Codex Amiatinus). Jerome,
Commentarii in Isaiam 9.28.21 (ed.  by Marcus Adriaen in CCSL 73, 365): Stetit sol
spatio unius diei.


 MARINA SMYTH

connection could be made whether one believed in a 24-hour bissextus

or a 12-hour bissextus. Why this one-time event should have initiated
the pattern of leap years is not obvious—except that there was a ten-
dency in some circles to assume that when Scripture first mentioned a
commonly recurring natural phenomenon such as rain or the rainbow,
it was recording the first occurrence of that phenomenon.112
Association of the leap year with Joshua was both common and per-
sistent. Combined with belief in the 12-hour bissextus, it underlies one
of the items in several versions of the early medieval sets of questions and
answers on biblical trivia known as the Joca monachorum:113

Quis tres horae solem fecit stare in celo? Iosue, minister Moysi, in pugna.

‘Who made the sun stand still in the heavens for three hours? Joshua,
the assistant of Moses, in the battle.’

The Joca are assumed to have insular connections, though the particular
question concerning Joshua is not contained in the core that has been
definitely identified as insular.114 The question shows that it was well
known that Joshua stopped the sun for exactly 3 hours! It is as if belief
that Joshua’s miracle brought about the first bissextus was somehow car-
ried one step further: since the bissextus is built from four quadrantes,
it must be that the sun then stopped for the first time for the supposed
annual 3 hours of the quadrans, thereby initiating the process of building
the 12-hour bissextus for all time thereafter! While this idea has little to
recommend it, it still surfaces in a section of the Leofric Missal, written
around the year AD 1000:115

Inquirendum est quare dicitur bisexus propter his [sic: should probably
read bis] kalendas nominatas et ut dii quando opugnauit Iosue in terra

112
See, e.g., Augustinus Hibernicus, De mirabilibus sacrae scripturae 1.6 (PL 35,
2157); Smyth (1996), 204–07.
113
Joca monachorum (ed.  by Suchier (1955), 33 and 39, no.  55; 56, no.  43 (in
French); 109, no. 12; 121, no. 32; see also Wilmanns (1872), 167, no. 11; Omont (1883),
69, no. 79). This particular question does not occur in the selection of Joca monachorum
copied in the Bobbio Missal; see Wright and Wright (2004).
114
There seems to be an insular core to which this question does not belong—but it
is difficult to determine the origin of individual accretions; see Bayless (1998), especially
pp. 17–18.
115
Leofric Missal (ed. by Orchard (2002), ii 55; the translation is mine). I thank Pe-
ter Pesic for drawing my attention to the presence of this belief in Anglo-Saxon England.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

gabaon orauit ad dominum, ut staret sol tribus oris in caelo et per optineret
uictoriam. Et ita factum est quasi annis singulis ipsi hore adcrescunt. Et
in tres annos ad quartam faciunt unum diem, et ipse dies dicitur bisexus.

‘We wish to know why it is called bissextus: because [the date is named
twice?] and because of the day when Joshua was attacking the land of
Gabaon and prayed to the Lord that the sun would stop for three hours
in the sky to allow him to gain victory. And this happened, so that those
hours are added each year. And in three years these quarter days add up
to a day, and that day is called bissextus.’

Further evidence that this was a commonly accepted notion in elev-

enth-century England is provided by Aelfric (†c.AD 1025), who, in his
De temporibus anni, rejected the opinion of the ‘unlearned’ (ungelaere­
dan) who claimed that the bissextus originated with the Joshua epi-
sode.116 Similar disapproval, this time using the words of Argumentum
XVI, is registered in the Missal of Robert of Jumièges (†AD 1055).117
The notion of a 3-hour pause even contradicts the biblical account
of the miracle, which asserts that the duration of the day doubled in re-
sponse to Joshua’s prayer. Augustinus Hibernicus, writing in Ireland in
AD  654,118 was eager to point this out. He reminded his readers that
not only did the sun ‘bring together two days into one, but the moon
also did not move for the duration of a day’ and he went on to assert
that this long period of daylight was followed by a night of normal du-
ration since both luminaries had stopped in their course and the moon
did not change phase.119 As a result, the relative paths of the sun and of
the moon remained unchanged, they merely stopped during the miracle,
and thereafter the year continued as normal:120

116
Aelfric, De temporibus annis 62–63 (ed. by Henel (1942), 52, 54; Blake (2009),
88, 120–21).
117
Robert of Jumièges, Missal (ed. by Wilson (1896), 42).
118
Warntjes (2015), 50–52.
119
De mirabilibus sacrae scripturae 2.4 (PL 35, 2175) summarized as follows: Nam
sol duos dies in uno conclusit, et luna diei spatio non occurrit. See also Hilary of Poitiers,
Tractatus super psalmos, Ps. 135:7 (ed. by Anton Zingerle in CSEL 22, 718): Sol et luna in
diei unius moram ex cursus sui lege detenti sunt.
120
De mirabilibus sacrae scripturae 2.4 (PL 35, 2175; Smyth (1996), 165). This
miracle is discussed in Munich Computus 66, which refers to the Irish Augustine as
Augustinus (Warntjes (2010a), 306–09, with extensive notes on these pages).


 MARINA SMYTH

Haec luminarium mora nihil nouum in natura commisit, etsi in ministe­

rio aliquid uarium ostendit. Sed et illa uarietas nihil in anni cursu et reli­
quorum dierum commouit, dum pariter sol et luna unumquodque in suo
ordine requieuit. Si enim unum luminare curreret, dum alterum interim
requiesceret, dierum et mensium et annorum assuetum cursum contur­
baret. Dum autem utrumque moram hanc habuit, quasi post consuetum
diem in occasus sui limitem perrexit.

‘This standing still of the luminaries involved nothing new in their na-
ture, though it clearly involved a change in their ministration. But this
change did not disturb anything in the course of the year and of the
other days since the sun and the moon paused both together, each in its
own order. For if one of the luminaries had continued in its course while
the other remained stationary, this would have disturbed the normal
course of the days, of the months and of the years. But since both paused
in this way, they reached thereafter the limit of their setting as though
after a normal day.’

Augustinus Hibernicus did not leave it at that: this is where he inserted

the long sequence itemizing the 532-year (Easter!) cycles since Creation
in order to show that the cycles were not disturbed in any way at the time
of the Joshua miracle:121

Non enim quod ad belli illuminationem luna tunc in praesentia solis pro­
ficeret stare imperatur; sed ne quid in congruo luminarium meatu per unius
quietem et alterius cursum destrueretur. […] Ut enim hoc manifestis approba­
tionibus pateat, cyclorum etiam ab initio conditi orbis recursus in se breuiter
digeremus, quos semper post quingentos triginta duos annos, sole ut in prin­
cipio, et luna per omnia conuenientibus, nullis subuenientibus impedimen­
tis, in id unde coeperant, redire ostendemus. […] quorum unusquisque uni­
formi statu, peractis quingentis triginta duobus annis in semetipsum, id est,
in sequentis initium reuoluitur, completis uidelicet in unoquoque solaribus
octouicenis nonodecies, et in lunaribus decemnouenalibus uicies octies circulis.

‘In the presence of the sun, the moon would have been of no use for
­casting light on the battle. Therefore, that was not the reason why it was
commanded to stand still, but lest anything should be destroyed in the
congruent course of the lights by the pausing of one and the moving of
the other. […] In order to show this by clear proof, we shall briefly de-

121
De mirabilibus sacrae scripturae 2.4 (PL 35, 2175–76; Smyth (1996), 167–68).
This list has allowed scholars to date the treatise to AD 654; see Warntjes (2010a), LXX-
VIII–LXXIX n. 209; Warntjes (2015), 50–52.


 ONCE IN FOUR: THE LEAP YEAR IN EARLY MEDIEVAL THOUGHT

scribe the return of the cycles into themselves since the beginning of the
created world. We will show that always, after 532 years, the sun and the
moon coming together in all ways as in the beginning, return unimpeded
to where they began. [The lengthy list of cycles is given here.] Each of
these cycles, in the very same way, having completed 532 years, turns back
into itself—that is, into the beginning of the next cycle—after completing
both nineteen 28-year solar cycles and twenty-eight 19-year lunar cycles.’

Why the emphasis on the unperturbed course of the 532-year Easter cy-
cles at the time of Joshua’s miracle? I used to think that Augustinus Hi-
bernicus was simply sticking to his project of showing that nothing new
was ever created since the original act of creation.122 And it is of course
true that this agrees with the purpose of the treatise: since the sun, the
moon, and the cycles were stopped during the miracle, nothing at all
was changed in their created nature. It now appears to me, however, that
Augustinus Hibernicus was also aware of the association of the Joshua
miracle with the origin of the leap year and wanted to register his disa-
greement. He therefore stressed the unperturbed pattern in the 532-year
cycles because this demonstrated that the pattern of leap years remained
undisturbed since Creation.
The author of De ratione conputandi, among others, accepted this
continuity since Creation, and even extended it to the full duration
of the world, that is, for a total of 6000 years. Simple calculation then
yielded a total of 1500 leap years from Creation to Judgement Day.123 By
AD 737, there was less certainty on the expected duration of the world,
so the corresponding computation in the Neustrian Dialogue was more
modest, determining that there were 1307 leap years from Creation till
the Passion, which was assumed to have occurred in AM 5228.124
In the perspective of these early medieval scholars, Julius Cesar may
have introduced the leap year into the calendar, but in doing so he was
merely uncovering the divine patterns underlying the unfolding of time
from the very moment of Creation.

122
See De mirabilibus sacrae scripturae, Prologue (PL 35, 2151): Cuncti uero laboris
hoc magnopere intentio procurat, ut in omnibus rebus, in quibus extra quotidianum admin­
istrationem aliquid factum uidetur, non nouam ibi Deum facere naturam, sed ipsam quam
in principio condidit, gubernare ostendat.
123
De ratione conputandi 58 (Walsh and Ó Cróinín (1988), 170): Sciendum nobis
quot sunt bissexti ab exordio mundi usque in finem. Id est i̅.d., quia hic numerus quarta pars
est sex milium, id est bissextus inter quartosquosque annos.
124
Dial. Neustr. 28a (Borst (2006), 415): Item interrogandum est; quot sit numerus
bissextorum a principio mundi usque ad crucem Christi? Id est: Mille trecenti et septem.


 MARINA SMYTH

Conclusion
Together with a more correct understanding of the leap year and its ne-
cessity to keep the calendar in line with the solar year, early medieval
Ireland inherited from the late antique world the belief that the bissextus
(that is, the intercalated bissextile day) lasted only 12 hours. This view
was even given an exegetical / arithmetical justification in the pseudo-
Dionysian Argumentum XVI, probably composed before AD 675 in an
Anglo-Saxon centre in Ireland.125 By the end of the seventh century, the
notion of a 12-hour bissextus appears to have been generally dismissed
by Irish computists and was explained away as referring only to daytime
hours, so that the corresponding night hours should also be assumed to
belong to the bissextus (Computus Einsidlensis, Munich Computus, De
ratione conputandi). This correction appears to have been prompted by
careful consideration of the gradual shift of the end of the solar year last-
ing a quadrans longer than 365 days, so that when considering the time
of day, the end of a solar year occurred some hours later in the day than
the beginning of that same year. There followed an increased awareness
that there must be night quadrantes contributing to the bissextus. The
12-hour bissextus continued to be mentioned in Merovingian and early
Carolingian manuals presenting alternative theories. It must still have
had some currency, as it would explain the persistence of the popular
belief—documented into the eleventh century—that leap years started
when God stopped the sun for 3 hours [sic] at Joshua’s request ( Joshua
10:12–14).

Warntjes (2010b), 92–94.

125


 C. PHILIPP E. NOTHAFT

CHRONOLOGICALLY CONFUSED:
CLAUDIUS OF TURIN AND THE DATE
OF CHRIST’S PASSION

Abstract
The chronicle of the world, written in c.AD 814 by Claudius of Turin, is note-
worthy for its use of computistical operations to establish Julian calendar dates
for pivotal events in the Old Testament such as Noah’s Flood, the Exodus, and
the destruction of Solomon’s Temple. These attempts to fuse chronography and
computus into a coherent system were highlighted by the late Arno Borst, who
claimed that Claudius’ chronicle proposed an unorthodox and controversial dat-
ing of Christ’s Passion to 23 March AD 31. A closer look at Claudius’ methods
of calculation instead reveals that the actual Passion date intended by the author
was the chronologically impossible 21 March AD 34. I will explain the rationale
behind the calculations that led to this result and situate them within the broad-
er context of early medieval ‘computistical chronography’. As will be shown, the
latter field found itself in a state of ‘crisis’ during the Carolingian period, caused
by the difficulties of accommodating the accepted chronological data for the life
of Jesus Christ within the numerical framework of the Easter computus.
Keywords
Venerable Bede, Claudius of Turin, Charlemagne, Louis the Pious; world
chronicle, Hebrew calendar, chronography; chronology of Jesus, date of cruci-
fixion, Passover; Arno Borst.

Introduction
The reign of Charlemagne has long been recognized as an important pe-
riod for the development of Christian time-reckoning. Under Carolingi-
an sponsorship, the art of computus became part of the required curricu-
lum in monastic and cathedral schools, which in turn contributed to the

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 265-292
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114740
 C. PHILIPP E. NOTHAFT

standardization and dissemination of computistical knowledge in the

realm of the Franks. In the wake of this development, the scientific writ-
ings of the Northumbrian monk Bede (c.AD 673–735) became medie-
val Europe’s doctrinal foundation for the calculation of Easter and other
aspects of chronology.1 Bede’s importance as a teacher of the computus
lies chiefly in his advocacy of the Alexandrian style of Easter reckoning,
based on the 19-year lunar cycle inscribed into the Easter table of Dio-
nysius Exiguus (c.AD 525), which is the source for our present system of
counting the years BC / AD. Another area where Bede’s influence made
itself felt was Christian world chronography. In his works De tempori-
bus (AD 703) and De temporum ratione (AD 725), Bede championed a
count of years based on Jerome’s Vulgate translation of the Hebrew Old
Testament, which reckoned only 3952 years between the creation of the
world and the birth of Christ. Compared to the scheme accepted previ-
ously, which stemmed from the chronicle of Eusebius (c.AD 325) and
counted roughly 5200 years for the same period, Bede’s era secundum he-
braicam veritatem rejuvenated the world by more than twelve centuries.2
Among the earliest Carolingian works to endorse Bede’s chronology
of world history to its full extent is the little-known chronicle of Claudi-
us of Turin, written in or around the year AD 814. Claudius, a man of
Visigothic origin, made his career as a courtier to Louis the Pious, by
whom he was appointed bishop of Turin between AD  816 and 818.
Aside from his highly prolific output of scriptural commentaries, he is
today chiefly remembered as a maverick theologian, who railed against
the use of images in churches, the veneration of the cross, relics, pilgrim-
ages, the intercession of saints, and the supremacy of the papacy. His
iconoclastic leanings were denounced as heretical by Dúngal of St Denis
and Jonas of Orléans, but Claudius died in office in either AD 827 or
828, without incurring deposition or official condemnation.3 Modern

1
See, e.g., Jones (1963); Jones (1976); Borst (1993); Borst (1998); Springsfeld
(2002); McKitterick (2004); Contreni (2005); Englisch (2010); Warntjes (2012).
2
Bede, De temporibus 22 (ed. by Charles W. Jones in CCSL 123C, 607); Bede, De
temporum ratione praef., 66, 67 (ed. by Charles W. Jones in CCSL 123B, 263–65, 463,
495, 535–37); Bede, Epistola ad Pleguinam (ed. by Charles W. Jones in CCSL 123C,
617–26). According to Mc Carthy (2008), 118–52; (2010), Bede was probably not the
originator of this Vulgate chronology. On the background, see further von den Brincken
(1957); Siniscalco (1973); Landes (1988); Allen (2003); McKitterick (2006); Palmer
(2011a).
3
For biographic and bibliographic information, see Dümmler (1895); Manitius
(1911), 390–96; Wemple (1974); Matter (1985); Heil (1997); van Banning (1997);
Gorman (1997); Boulhol (2002); O’Brien (2011); Dorfbauer (2013).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

research on the exegete’s world chronicle, which is preserved in three

manuscripts from the ninth, tenth, and eleventh centuries, effectively
began in 1973, when Mirella Ferrari made a decisive case for Claudius’
authorship of the work, which had been denied in earlier scholarship.4
Further significant contributions have since been made by Michael Id-
omir Allen, who surveyed the chronicle’s contents, and by Elisabetta Bel-
lagente, who published a listing of Claudius’ sources.5 More recently, in
2006, Arno Borst included a heretofore unknown excerpt of this chroni-
cle in his great edition of Schriften zur Komputistik im Frankenreich.6
Besides an epistolary preface, dedicated to the priest Ado, the chronicle
of Claudius of Turin consists of two distinct parts. The first part, which
Michael Allen has dubbed the ‘display section’,

‘consists of a series of verso-recto openings used as frames for a left-to-

right progression of inscribed genealogical roundels accompanied by
discursive exposition. The result is a commentated genealogical tree
from Adam to Emanuel whose textual component includes patristic ex-
tracts and occasional borrowings from the reckoning section that now
figures spatially as the tailpiece, though it is textually prior.’7

The second part, referred to by Allen as the ‘reckoning section’, is a con-

cise chronological overview of the six ages of the world, marked by a
heavy focus on numbers, which establishes the intervals between impor-
tant biblical events. For some of these, Claudius supplied computistical
calculations with the aim of finding their Julian calendar dates and cor-
responding weekday data (see below). The final calculation refers to the
annus praesens AD 814, the year of Charlemagne’s death.

4
Ferrari (1973). The three MSS in question are: Paris, Bibliothèque nationale de
France, Lat. 5001, 1r–8v (saec. IX3/4); Monza, Biblioteca Capitolare, c-9/69, 66ra–83vb
(saec. X1/2); Madrid, Biblioteca Nacional, 9605, 104r–112r, 113r–117v (AD  1026).
Claudius’ authorship was influentially rejected by Ernst Dümmler (1895), 442, who
surmised that the ‘dürftige chronologische Schrift […] einem jüngeren Träger dieses Na-
mens anzugehören [scheint], weil sie in einer zweiten Ausgabe bis 854 reicht’. Dümmler
accordingly omitted the chronicle’s letter-preface from his 1895-edition of Claudius’
correspondence (MGH Epp. 4, 586–613). His judgment was based on a misunderstand-
ing of the interpolated chronological notes in Monza, Biblioteca Capitolare, c-9/69,
115v–116r.
5
Ferrari (1973); Allen (1998); Bellagente (1999). See also Heil (2000), 74–83.
A critical edition of the text by Michael I. Allen is in preparation.
6
Claudius Ser. (ed. by Borst (2006), 1335–49).
7
Allen (1998), 291.


 C. PHILIPP E. NOTHAFT

Borst’s theory
According to Arno Borst, Claudius wrote his work with the intention
of criticizing an earlier chronicle of the six ages of the world, which had
been incorporated into the famous Seven-Book-Computus (Lib. comp.)
of AD 809/12, but this contention is based on no solid evidence and
therefore seems debatable.8 In addition, Borst attributed to Claudius
an unorthodox dating of Christ’s Passion to 23  March AD  31, which
would have been liable to cause serious irritation among his contempo-
raries. Borst went as far as speaking of a ‘revolution of the time scale’,
which shattered conventional views of the time of Christ’s crucifixion
and resurrection, as well as an ‘explosive truth […] bound to make him
new enemies’.9 The sole basis for these assertions was a passage at the
beginning of the chronicle’s ‘reckoning section’, which establishes a
chronological parallelism between the six days of creation and the week
of Christ’s Passion. Borst’s analysis of this passage rested on a tenth-
century manuscript from Monza’s Biblioteca Capitolare, which can be
shown to have undergone interpolation:10

Et ita ratione deducta X Kal. Aprilis luna septima decima invenitur pro-

toplastus Adam ex terrae limo esse formatus pariter atque animatus. Se-
cundum hanc nostram supputationem, secundus Adam, id est Christus,
Dei filius […] eodem tempore, sed non eadem feria, id est Xo Kal. Apriles,
invenitur fuisse passum et VIII resurrexisse a mortuis.

‘And so it follows by reckoning that the first man Adam was formed
from the slime of the earth and fitted with a soul on 23 March, the sev-
enteenth day of the moon. According to this our calculation, the second
Adam, that is Christ, the Son of God […] is found to have suffered at

8
Borst (2006), 1336–38. The chronicle is labelled Ser. nov. in Borst’s edition
(2006), 951–1008; in its revised version it was included in Lib. comp. 1.15 (ed. by Borst
(2006), 1108).
9
Borst (2006), 61–62: ‘Doch rang sich Claudius bei seinen komputistischen
Rechnungen […] zu einer brisanten Wahrheit durch, die ihm neue Feinde einbringen
mußte. Sie […] erschütterte aber die Überzeugung des Gottesvolkes vom Zeitpunkt der
Kreuzigung und Auferstehung Christi, somit die derzeit herrschende Osterberechnung
und Osterfeier. Bewirkt wurde dieser Umsturz der Zeitskala auf einfachste Weise, durch
komputistische und typologische Verknüpfung der ersten weltzeitlichen Datenreihe mit
der zweiten.’ See also Borst (2006), 1337, 1341. Borst’s claim appears to have been in-
spired by a suggestion made by Allen (1998), 314 n. 128: ‘Claudius’ ‘calculation’ reflects
the date for ‘Easter’ in AD 31 according to Bede’.
10
Monza, Biblioteca Capitolare, c-9/69, 79rb–va. See Allen (1998), 314 n. 128;
Borst (2006), 62.


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

the same time, but not on the same weekday, that is on 23 March, and to
have risen from the dead on 25 March.’

Just like Claudius’ entire chronicle, the above passage betrays the influ-
ence of the Venerable Bede, who argued that the sun was at the point of
the vernal equinox and the moon was full when both luminaries were
created on the fourth day (Genesis 1:14). Guided by the Alexandrian
rules for dating the Easter full moon, Bede assumed that the vernal equi-
nox in the year of creation had fallen on 21 March. From this followed
that the hexaëmeron had begun on Sunday, 18 March, and ended with
the creation of man on Friday, 23 March.11 In the preface to his Chronica
maiora, which constitutes the 66th chapter of De temporum ratione, Bede
revisited this chronology of the creation and noted that it favoured a
date of the crucifixion on 23 March, as this led to a neat typological con-
nection between the fall of mankind and its salvation through Christ’s
suffering.12 While the Monza manuscript thus converges perfectly with
Bede’s words, the oldest surviving manuscript—Paris, Bibliothèque
nationale de France, Lat. 5001, which was transcribed and printed by
Philippe Labbé in 1657—shows that Claudius deviated from the North-
umbrian monk more strongly than Borst realized:13

Secundum hanc nostram supputationem, secundus Adam, id est Christus,

Dei Filius […] eodem tempore, sed non eadem feria, id est X Kal. Aprilis
invenitur resurrexisse a mortuis.

‘According to this our calculation, the second Adam, that is Christ, the
Son of God […] is found to have risen from the dead at the same time,
that is 23 March, but not on the same weekday.’

23  March is here regarded as the date of the resurrection rather than
that of the Passion. This can also be seen from Claudius’ remark ‘not
on the same weekday’, which makes it clear that, although Adam was
brought to life on the sixth day of creation and hence on a Friday, the
typological recurrence of the corresponding calendar date (23 March) in

11
Bede, De temporum ratione 6 (CCSL 123B, 290–95).
12
Bede, De temporum ratione 66 (CCSL 123B, 464–65). The passage is repeated
almost verbatim in the ninth century in Ado of Vienne’s Chronicon in aetates sex divisum
(PL 123, 24).
13
Paris, Bibliothèque nationale de France, Lat. 5001, 1r = Labbé (1657), i 309 =
PL 104, 917D–918D. I have slightly modified punctuation. See also Madrid, Biblioteca
Nacional, 9605, 113ra.


 C. PHILIPP E. NOTHAFT

the year of the Passion did not apply to the Friday of Christ’s crucifixion,
but to his resurrection two days later. Claudius thus adopted Bede’s date
for Adam’s creation, but decided to put the crucifixion two days earlier,
on 21  March, which must have irritated later readers and commenta-
tors.14 The scribe of the Monza manuscript accordingly tried to correct
the above passage by inserting the words fuisse passum et VIII between
invenitur and resurrexisse. This brought Claudius’ words in line with
Bede’s, but at the same time rendered them incoherent, as the resulting
new date presupposed what Claudius explicitly denied: that 23 March
had occupied the same weekday in the weeks of creation and the cruci-
fixion. There is, therefore, every reason to believe that the Monza codex
is interpolated and cannot serve to correct the Parisian version.

Calculating biblical calendar dates

An important hint as to how Claudius approached the date of Christ’s
Passion comes from the preface to his chronicle, in which he tells his oth-
erwise unknown friend Ado that the number of years between the crea-
tion of the world and the Lord’s Passion was ‘not greater than 3986 years’.15
This figure is once again based on Bede, who, in his Chronica maiora, dat-
ed the nativity of Jesus to AM (Annus Mundi) 3952 and the crucifixion to
AM 3984. It would appear that Claudius corrected this interval of years
on the basis of Bede’s claims in chapter 47 of De temporum ratione, where
he assigned to Christ a life-span of 33 ½ years, leading to a crucifixion
in AD 34. If Jesus was born toward the end of AM 3952, his crucifixion
was hence bound to have taken place no earlier than AM 3986.16 Moreo-
ver, if Jesus died in AD 34 and lived for 33 ½ years, it also followed that
his nativity took place in 1 BC and, accordingly, that AM 3952 was the
year immediately prior to AD 1. This interpretation of the data is indeed
reflected by Claudius’ chronological system, which identifies the annus
praesens as both AD 814 and AM 4766 (4766–814=3952).17 Based on

14
von den Brincken (1957), 117, still erroneously identified 23 March as Claudius’
date of the Passion despite working from Labbé’s edition.
15
Edited in Ferrari (1973), 307: Nos vero, secundum divinos quos in manu tenemus
codices ex hebrea auctoritate per beatum Iheronimum christianum interpretem, non plus
a conditione mundi usque ad passionem domini invenire possumus quam III. DCCCC
LXXXVI.
16
Bede, De temporum ratione 47, 66 (CCSL 123B, 430–31, 495–96).
17
Claudius of Turin, Brevis chronica (PL 104, 924D).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

these conclusions, it becomes highly unlikely that Claudius would have

chosen a Passion date in AD 31, as claimed by Borst, since this year in
his chronology would have corresponded to AM 3983, which is hardly
what the Carolingian exegete had in mind when he spoke of an inter-
val ‘not greater than 3986 years’. It is also worth noting that the correla-
tion AM 3952=1 BC provides an important justification for Claudius’
claim that Adam was created on 23 March. If the world was created in
3952 BC, then the events described in Genesis chapter 1 took place in
the 306th year of the 532-year cycle, in which 18 March fell on a Sunday,
thereby validating Bede’s chronology of the hexaëmeron.18 This would
not have been the case if AM 3952 had been made equal to AD 1.
Moreover, since 3952 is evenly divisible by 19, Claudius’ correlation
made sure that AM 1 and AD 1 occupied the same position in the 19-
year lunar cycle, namely the second year. The same holds true for the se-
quence of leap years, as 3952 is also evenly divisible by 4, with both AM 1
and AD 1 corresponding to the first year after a leap year. This meant
that many of the same argumenta computists applied when they worked
with years AD, could also be used in connection with years AM.19 Such
formulae were of no small importance to Claudius, who invested much
effort in determining Julian calendar dates and weekdays for important
biblical events, in particular those that mark the beginning of a new age
of world history: Noah’s Flood (2nd age), the Exodus (4th age) and the
destruction of Solomon’s Temple (5th age).20 These calculations reflect
an underlying assumption according to which the months and days
mentioned in the Old Testament could be directly identified with the
months in the ecclesiastical lunar calendar. Wrong as this assumption
may have been, it was quite in tune with Bede’s teachings in De tempo-
rum ratione, which sometimes created the impression that the ancient
Israelites used the same 19-year lunar cycle as the Alexandrian Church.21

18
This can be shown thus: 1 BC is equivalent to AM 3952. If this is the 1st year
of the 19-year cycle, so was AM  228, which fell 7×532=3724  years earlier (3952–
3724=228). Hence, AM  227 will have been the 532nd year and AM  1 the 306th year
(532–226=306).
19
Mid-ninth-century computists were certainly aware of this; cf. Lect. comp. 3.4–5,
MSS PePfSt (ed. by Borst (2006), 588–89) with Lect. comp. 4.5, 7 (Borst (2006), 598–
99, 601–02). On the early transmission of the AD argumenta, see Warntjes (2010b).
20
See the Appendix below.
21
Bede, De temporum ratione 11, 13, 45, 59, 61 (CCSL 123B, 312–15, 326–27,
420–22, 447, 451). On early medieval views of the Hebrew calendar, see also Wiesen-
bach (1986), 119–22; Nothaft (2012a); Nothaft (2014), 43–51; and the sources assem-
bled in Warntjes (2010a), 242–43. A roughly contemporary interpretation of the Dio-


 C. PHILIPP E. NOTHAFT

An attentive reader of De temporum ratione, such as Claudius, could

thus be tempted to equate lunar months in the Old Testament with lu-
nar months in the Easter computus and use this equation as a basis to
transfer biblical into Julian dates.
In principle, there was one very straightforward way of locating the
beginning of the biblical Hebrew year in the Julian calendar: since the
Easter full moon was supposed to represent the 14th day of Nisan, any
computist could find the latter’s current Julian date in his Easter table
and then count back 13 days to arrive at 1 Nisan. Claudius, however,
used a slightly more elaborate approach. In its rough outlines, it was
based on the 22nd chapter of De temporum ratione, where Bede presented
an ‘old formula’ for finding the age of the moon on any given day of the
year. The first step was to establish the lunar age on 1  January, which
was 9 in the first year of the cyclus decemnovenalis. For subsequent years,
it could be determined by adding the value of the ‘epact’, which corre-
sponded to the lunar age on 22 March. This value could be either looked
up in any standard Easter table or calculated by dividing the present year
(AM or AD) by 19 and multiplying the remainder by 11. The result of
this multiplication (subtracting multiples of 30) was the epact.22 Once
the lunar age of 1 January was known, the rest of the calculation pro-
ceeded like this:23

Si ergo uis scire hoc uel illo die quota sit luna, computa dies a principio
mensis Ianuarii usque in diem de quo inquiris et, cum scieris, adde aetatem
lunae quae fuit in kalendis Ianuariis, partire omnia per lviiii et si amplius
xxx remanserint, tolle triginta; et quod superest ipsa est luna diei quam
quaeris.

‘If you wish to know how old the Moon is on this or that day, count the
days from the beginning of January up to the day you want, and when
you know this, add in the age of the Moon on the kalends of January.
Divide the total by 59, and if more than 30 remain, subtract 30. What is
left over is the day of the Moon you are seeking.’

nysiac luna 14 as pascha Hebreorum is found in a unique addition to Isidore’s Chronica

maiora, datable to AD 800 and edited by Mommsen in MGH Auct. ant. 11, 491. I owe
this reference to Richard Landes.
22
Bede, De temporum ratione 20, 52, 57 (CCSL 123B, 346–49, 441, 446–47).
23
Bede, De temporum ratione 22 (CCSL 123B, 351; translation according to Wal-
lis (1999), 69). In this passage, Bede omits the subtraction of 1 for 1 January, which is
necessary because the count has to be exclusive, although he later includes it in his exam-
ples (tolle kalendas). See also Lib. ann. 43 (ed. by Borst (2006), 739).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

Claudius, on the other hand, sought the exact opposite—he started out
with lunar data and wanted to calculate the corresponding day in the Ju-
lian calendar. In order to do so, he had to reverse Bede’s principle. Know-
ing the lunar age on 1 January, he first added the days of the individual
lunations, which alternated between 30 and 29 days (save for the occa-
sional intervention of embolismic months, which always have 30 days),
up to the ‘Hebrew’ lunar date in question and then inferred the Julian
equivalent by counting the same number of days from 1 January. The lat-
ter could be done in accordance with a table for kalends, nones, and ides
of each month derived from the Latin Anatolius (De ratione paschali)
and found in the same chapter of De temporum ratione.24 This whole
approach, cumbersome as it may have been, is a testimony to Claudius’
originality, who, in order to reach his chronological aims, had to devise
formulae that could not simply be found in the existing handbooks. In-
deed, he deviated from Bede by adding 8 rather than 9 to the value of
the epact, thereby reproducing the lunar age on 31  December, which
he simply refers to as the ‘epact’.25 This was an expedient step, because
it circumvented the subtraction of 1, which became necessary in Bede’s
scheme due to the fact that 1 January was part of the counted interval.
As a matter of fact, his placing of the epact on the last day of the Julian
year indicates that Claudius learned the computistical trade not in Fran-
cia, but in his native Spain, as this placement can be shown to have been
a distinctive feature of early medieval Iberian computistics.26
For a closer look at Claudius’ approach, we can turn to the Exodus
from Egypt, which is the only date for which the Carolingian chronicler
outlined the steps of his calculation in any detail. According to Claudi-
us’ adaptation of Bede’s Vulgate chronology, the Exodus took place in
AM 2453. Since the AM era started with the second year of the cyclus
decemnovenalis, he first had to divide the year in question by 19 and in-
crease the remainder by one. In the case of the Exodus, the remainder
was two (2453modulo19=2), meaning that the Israelites had left Egypt
in the third year of the cycle. The epact for this year was 2×11=22, to
which Claudius added 8 days to arrive at his version of the epact, which
indicated the lunar age on the preceding 31 December. The result was
22+8=30, which, for computistical purposes, was equivalent to an epact

24
Bede, De temporum ratione 22 (CCSL 123B, 351–52); Anatolius, De ratione
paschali 10 (ed. by Mc Carthy and Breen (2003), 50).
25
Claudius of Turin, Brevis chronica (PL 104, 919C, 920A, 922D–923A, 925A).
26
See the Appendix in Mosshammer (forthcoming).


 C. PHILIPP E. NOTHAFT

of 0. This meant that 1 January fell on luna 1 and hence coincided with
the beginning of a ‘Hebrew’ lunar month. Following Bede, the month in
question could be identified with Tebeth, whereas Nisan was supposed
to correspond to the April lunation. In order to get from 1 January to
14 Nisan, Claudius thus had to count: 30+29+30+14=103  days. If
day 1 corresponds to 1 January, day 103 (in a common year) will fall on
13 April (31+28+31+13=103), which is indeed the date of the Easter
full moon in the third year of the Dionysiac cycle. For reasons unex-
plained, however, our chronicler counted only 102 days and thus chose
12 April, putting the Exodus (15 Nisan) on 13 April, the arrival in the
wilderness of Sinai (1 Siwan; cf. Exodus 19:1) on 27 May and the recep-
tion of the Law (4 Siwan; cf. Exodus 19:16) on 31 May.27
It is difficult to see what could have motivated such a lowering of the
lunar count by one day. The default position among early medieval com-
putists was that the lunations of the 19-year cycle alternated between
‘full’ months of 30 days and ‘hollow’ months of 29 days, such that the lu-
nations belonging to January, March, etc. would be full and those of Feb-
ruary, April, etc. would be counted hollow. This is the view expressed in
De temporum ratione, which makes it likely that Claudius, too, adhered
to it.28 It is worth noting, however, that according to this scheme the
lunation of April, which Bede associated with the month of Nisan, was
supposed to be a hollow month of 29 days. Yet elsewhere in De tempo-
rum ratione, the Northumbrian monk implied that the Hebrews began
their year with a full lunation, which could be taken to mean that the

27
Claudius, Brevis chronica (PL 104, 922–23): Iterum sume superiores annos 2453,
hos partire per decimam nonam partem […] adde unum et fiunt tria, et invenies quod ter-
tius fuerit annus cycli solaris, qui est decemnovalis: subtrahe unum, et remanent duo: ipse est
secundus annus epactae lunaris. Iterum dic: undecies bini fiunt 22, tantos enim dies habet in
duobus annis solaris cursus amplius lunari. His adde 8, et fiunt 30. Et quia annus embolis-
mus fuit, tanti exstiterunt in epacta. Iterum sume summam numeri 132, subtrahe triginta,
qui fuerunt in epacta, et remanent 102, et invenies quod secundum dies solares, pridie Idus
Aprilis feria secunda luna quarta decima exstiterit ipsum Pascha. Juxta vero dies lunares III
Kal. Aprilis fuit Neomenia, id est novae lunae principium. Secundum vero supputationem
nostram, quae fit per dies solares, Idus Apriles, luna quinta decima, alter die post Pascha,
inveniuntur exisse filii Israel de Ramesse, et venisse in Socoth et de Socoth VIII. Kal. Maii,
luna sexta decima in Phyairoth, et VI Kal. Junii, luna prima, qui est dies primus tertii men-
sis lunaris, venerunt ad montem Sinai, et acceperunt legem Deo loquente ad se per angelum
pridie Kalendas Junii, die quinquagesimo post agni occisionem, feria secunda, luna quarta,
quo die et Pascha in Aegypto celebraverunt. The passage in Roman type is garbled in Paris,
Bibliothèque nationale de France, Lat. 5001, on which Labbé’s edition was based. I have
emended it on the basis of Monza, Biblioteca Capitulare, c-9/69, 81va.
28
Bede, De temporum ratione 20 (CCSL 123B, 346–49).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

Bible used an opposing pattern of full and hollow lunations.29 In such a

scenario, the Hebrew lunar month corresponding to March would con-
sist of merely 29 days, leading to a displacement of the Easter full moon
in the Alexandrian 19-year cycle vis-à-vis the ‘Hebrew’ 14 Nisan. With
Heimo of Bamberg, who wrote his Consideratio annorum seculi et Chris-
ti Ihesu (also known as De decursu temporum) in AD 1135, we know of
at least one medieval computist who drew these conclusions and con-
sequentially differentiated between the ‘Hebrew’ and the ‘Christian’
value of the epact.30 Claudius’ calculation of the Exodus features the
same displacement of 14 Nisan by one day compared to the Dionysiac
cycle that was later advocated by Heimo. Since, however, all other lunar
calculations found in Claudius’ chronicle conform to the regular pattern
of lunations (see Appendix), it is far more likely that his Exodus date was
the result of an ordinary slip. Perhaps he mistakenly performed the sub-
traction of 1 for the kalends of January, which was necessary according
to Bede’s method, but should have been made superfluous by his use of
the ‘Iberian epact’.

Calculating the weekday

Besides converting the lunar dates of the Bible into dates in the Julian
calendar, Claudius was also interested in determining the corresponding
weekday. The common Julian year consists of 365 days and, since 365
equals 7×52+1, the weekday on a particular date of that calendar in-
creases from one year to another. If every Julian year were common, the
sequence of weekdays would recur after only 7 years. Due to the periodic
insertion of a leap-day, however, the increment is raised to two in every
fourth year. This leads to a sequence of 7×4=28 years, the so-called ‘solar
cycle’, which marks the recurrence of the weekday in the Julian calen-

29
Bede, De temporum ratione 11 (CCSL 123B, 312–15).
30
Heimo of Bamberg, De decursu temporum 1.2 (ed. by Weikmann (2004), 135):
Sciendum est autem, quia aliter se habet compotus Hebreorum, quo usi sunt antiqui pa-
triarche et prophetae, sed et evangelistae, et aliter se habet compotus Grecorum et Latino-
rum, quo nunc maxime utuntur Christiani. Hebrei quippe seriem temporum per lunares
annos estimabant, sed Greci et Latini per solares, et lunationes Aprilis, Iunii, Augusti et
caeterorum mensium, quos nos numeramus XXIX dierum, illi numerabant XXX, quos nos
XXX, illi XXIX, embolismos omnes XXX dierum sicut et nos preter ultimum; illi XXIX
dies tantum evenerant. Eisdem epactis, eisdem quoque concurrentibus usi sunt, quibus et nos
utimur, sed sedem epactarum, quam nos in XI Kal. Aprilis ponimus, ipsi posuerunt in XII
Kal. Aprilis.


 C. PHILIPP E. NOTHAFT

dar. In the Dionysiac Easter table, this cycle is represented by the ‘solar
epacts’ or ‘concurrents’, which were thought to designate the weekday
on 24 March. Curiously, Claudius did not make use of either the ‘solar
cycle’ or the ‘concurrents’, but instead chose to calculate the weekdays ‘in
his head’, availing himself of a method that worked independently from
any list or Easter table.
He started by dividing the present year of the world by 4, thereby ar-
riving at the total number of leap-days that had elapsed since the creation
of the world (18 March, AM 1). In the case of the Exodus, which took
place in AM 2453, 613 days had to be added (2453/4=613.25). A sec-
ond step prescribed the addition of six dies intercalares, yielding a total
sum of 3072, which he then divided by 7. The remainder of this division
indicated the weekday on 1 January, which in Claudius’ calculation for
the year of the Exodus was 6, meaning Friday (3072modulo7=6).31 In
order to understand the rationale behind this procedure, which Claudius
never bothered to explain, we have to recall that, according to his chron-
icle, the world began on a Sunday, 18 March 3952 BC. Since Sunday is
the first day of the week, it was possible to keep track of the year-by-year
increment of the weekday on 18 March by simply enumerating the years
of the world and then adding to this sum the number of leap-days that
had accrued since creation. By the year AM 4, for example, the weekday
on 18 March had shifted from 1/Sunday to 5/Thursday, because of the
leap-day added in this year. For sums greater than seven, it was of course
necessary first to subtract multiples of seven days and then turn to the
remainder.
The calculation just outlined only applies to the weekday on 18 March
or any other day of the Julian year equivalent to it. In order to adapt this
method to any day of the year, however, it was much more convenient
to first find the weekday on 1  January, which was identical to that of
18 March only in leap years. In common years, on the other hand, it dif-
fered by six days: in AM 1, for instance, 18 March was a Sunday, whereas
31
Claudius of Turin, Brevis chronica (PL 104, 922): Si vis hujus Paschae nosse tem-
pus, vel diem, vel feriam, aut lunam, sume annos ab initio mundi, qui sunt usque ad id
tempus 2453; hos partire per quartam partem, et dic: quater 600 fiunt 2400, et supersunt 53.
Iterum dic: quater deni, 40, et supersunt 13; quater terni, 12, et superest 1; et invenies quod
primus annus tunc exstiterit post bissextum. Sume igitur ipsos dies bissextiles, qui fuerunt
in toto retro tempore 613, hos adde ad superiores annos, et invenies summam eorum 3066:
his iterum adde sex dies, quos intercalares vocamus, et erunt 3072. Et ut feriam ipsius anni
invenire possis, partire supradictam summam, et dic septies 400 fiunt 2800, et supersunt 272.
Iterum dic septies 30 fiunt 210, et supersunt 62; et rursum septies octoni fiunt 56, et supersunt
sex; et invenies quod feria sexta secundum dies solares exstitit principium anni quam nos
more gentilium Kalendas Januarias nuncupamus.


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

the preceding 1 January would have fallen on a Monday. Accordingly,

in order to arrive at the weekday on 1  January in a common year, six
days had to be subtracted from the total weekday increment before it
could be divided by seven.32 This seems to be the rationale behind the
six dies intercalares, which Claudius mentions in his calculation of the
Exodus date. By adding rather than subtracting these six days, however,
he introduced an error of two days into his computation, which went
(2453+613+6)/7=438 R 6 (= Friday), when the correct result should
have been (2453+613–6)/7=437 R 1 (= Sunday). A comparison with
Claudius’ other calculations in the ‘reckoning section’ (see Appendix)
reveals that this was not just a momentary slip, but a systematic error,
one which pervaded his entire chronology.
It is difficult to say with any certainty whether this error was the
outcome of Claudius’ attempt to customize his own reckoning rules or
whether he was led astray by an existing formula, which prescribed the
same misguided addition of six dies intercalares. One example, kindly
suggested to me by Immo Warntjes, would be a parallel rule used for
reckoning with years AD in the Frankish Annalis libellus of AD  793,
which demanded the subtraction of 1.33 Since subtracting 1 and adding
6 are equivalent in calculations modulo 7, it is conceivable that Claudius
simply adopted this method for years AM without worrying about the
difference in weekdays between the two eras. As far as the terminology
of sex dies intercalares is concerned, the only comparable text known to
me occurs in London, British Library, Harley 3017, 186v, which was
copied between AD  862 and 864 at the monastery of St.-Benoît-sur-
Loire (Fleury).34 The passage in question was edited by Borst as part of
the Lectiones sive regula conputi, a Frankish collection of computistical
argumenta that was first redacted in 760:35

32
Alternatively, one could have added 1 rather than subtracting 6 (which is equiva-
lent in modulo7-calculations), as is done in Lect. comp. 3.6, MS Pf (Borst (2006), 589–
90). As Immo Warntjes points out to me, Claudius could have made his method apply
for all years (i.e. both common years and leap-years), provided that he always subtracted
1 from the sum of Anni Mundi before dividing them by 4 to obtain the leap-day incre-
ment. In this case, he would have based his calculations on a year beginning on 1 January
(which comes before the bissextus) rather than on 18 March (which comes after). My
preference for 18 March in the present discussion is based chiefly on Claudius’ discus-
sion of the date of Noah’s Flood (see n. 67 below), which indicates that he did not per-
form this subtraction.
33
Lib. ann. 20 (Borst (2006), 712–13).
34
See Borst (2006), 244–45, for a description of the manuscript.
35
Lect. comp. 8.12a (ed. by Borst (2006), 648).


 C. PHILIPP E. NOTHAFT

Quare iubet sex intercalares ad feriam adicere? Quia sexta feria Christus
conceptus et sexta feria crucifixus.

‘Why must one add six intercalary days to the weekday? Because Christ
was conceived and crucified on a Friday.’

The remark stands isolated among the other argumenta assembled in

Harley 3017 and thus fails to offer us an appropriate context for the
mentioned addition of six intercalary days. Indeed, it is difficult to see
how the weekday of the incarnation and crucifixion could provide any
justification for such a rule. A relation to Claudius’ own formula might
still be inferred from the unusual terminology of intercalares as opposed
to regulares, which is the term used in the argumenta that precede and
follow the passage in question. But it remains doubtful whether Claudi-
us had access to any such precedent by the time he wrote his chroni-
cle and it cannot be ruled out that the formula in question was added
to the bulk of the collection only after AD 814, which may even mean
that it was influenced by the chronicle of Claudius of Turin rather than
vice versa. In the absence of any further evidence, it may well be that
Claudius’ infelicitous formula for the calculation of the weekday was his
own invention. As we shall see, it is precisely this computational error on
Claudius’ part that provides the key to his dates for Christ’s crucifixion
and resurrection.

The early medieval ‘crisis’ of computistical chronography

For a better understanding of these dates in the context of Claudius’
chronicle, it will be worthwhile to take a step back and look at the wider
context and tradition of Christian chronological scholarship, since Late
Antiquity, which included a particular strand or modus operandi that I
would like to term ‘computistical chronography’. At the core of ‘compu-
tistical chronography’ lay the insight that the various Easter cycles late
antique Christians had developed in order to calculate future dates of the
Easter full moon could also be used retrospectively, that is to find out
when Passover or some other lunar date occurred in the past. Our earliest
example for an application of this principle is the so-called ‘statue of Hip-
polytus’, discovered near Rome in 1551, whose plinth shows a Greek in-
scription of a 112-year Easter cycle starting with the first year of the reign
of Alexander Severus (AD 222). In the margins of certain years of this


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

cycle, historical notes indicate the date of biblical Passovers, as they were
thought to have occurred in past cycles. Most significantly, the Passion of
Christ is marked next to the Easter full moon (14 Nisan) of the 32nd year
of the cycle (AD 253), thereby indicating that it took place on Friday,
25 March AD 29 (253–2×112=29).36 This date, which matched both the
Easter cycle and the historical data for the life of Jesus (AD 29 equals the
15th year of Tiberius, in which Jesus was baptized according to Luke 3:1),
soon became the dominant calendar tradition of the Latin West, as wit-
nessed by numerous sources from the third, fourth, and fifth centuries.37
As Venance Grumel has shown in his great handbook of Byzantine
chronology, late antique Christian scholars worked hard to create chron-
ological systems in which the key dates of the history of salvation—the
creation of the world as well as the incarnation, birth, death, and resur-
rection of Christ—would form a coherent match with the calendrical pa-
rameters provided by their Easter cycles.38 The two most influential sys-
tems of this kind were each based on a creation era—known respectively
as the Alexandrian and the Byzantine world era—whose starting year
was also the beginning of a 532-year Easter cycle. In addition, both sys-
tems were tied to a particular date for Christ’s Passion: 23 March AD 42
in the Alexandrian system, which influenced the Coptic and Ethiopic
Churches, and 23 March AD 31 in the Byzantine-Greek tradition.39
In the case of the Latin Church, a comparable system was proposed
in AD 457 by Victorius of Aquitaine, who had been commissioned by
the papal curia to reform the Roman Easter cycle. Victorius reacted by
producing a 532-year cycle, which had a hypothetical starting point in
5201 BC, the assumed year of the creation of the world. He also made
sure that the resulting Easter table would provide acceptable data for the
Passion of Christ, which he re-located to 26 March AD 28. Victorius’
Easter table, although it enjoyed a wide reception during the sixth and
seventh centuries, was eventually superseded by the work of Dionysius
Exiguus, whose Easter table was a more faithful representation of the
Alexandrian computus. In fact, the latter simply continued an existing
95-year Easter table that had presumably been commissioned by Bishop
Cyril of Alexandria a century earlier. One important modification Dio-

36
On the Easter table of Hippolytus, see Richard (1950); Mosshammer (2008),
121–25; Nothaft (2012), 38–52, with references to further literature.
37
On late antique dating traditions for the crucifixion, see Richard (1951), 32–42;
Lazzarato (1952), 349–423; Loi (1971).
38
Grumel (1958), 3–158. See also Grumel (1952).
39
See Nothaft (2012), 56–68, with references to further literature.


 C. PHILIPP E. NOTHAFT

nysius introduced to this table was his replacement of the year 248 of the
era of Diocletian, which designated the starting point of his continua-
tion, with the year 532 ‘from the incarnation of our Lord’ (ab incarna-
tione Domini nostri Iesu Christi).40 In order to grasp the consequences of
this decision, let us take a look at the Easter full moon dates or termini
paschales the Alexandrian 19-year cycle provides for the range of years
relevant for Christ’s death, here taken to be AD 27 to 37:

27 07 April (Mon)
28 27 March (Sat)
29 15 April (Fri)
30 04 April (Tue)
31 24 March (Sat)
32 12 April (Sat)
33 01 April (Wed)
34 21 March (Sun)
35 09 April (Sat)
36 29 March (Thu)
37 17 April (Wed)

At first glance it would appear that the acceptance of the Alexandrian

reckoning in the West made it possible to retain the widespread notion
according to which Jesus was crucified in AD 29. The Easter table shows
a full moon on Friday, 15 April in this particular year. Acceptance of this
date, it is true, would have required abandoning the typologically attrac-
tive idea that Jesus was crucified on 25 March, the traditional Roman
date of the vernal equinox, which was also often associated with the crea-
tion of the world and Christ’s incarnation. Since 15 April is the 14th day
of the moon, it would have also meant to side with the Gospel of John,
who indicated that Jesus was crucified on the day before Passover (14
Nisan), and hence against the synoptic evangelists, who made the cruci-
fixion coincide with the Passover feast (15 Nisan) and whose chronol-
ogy of the Passion was generally favoured by medieval writers.41 Yet the
true problem lay elsewhere: according to Luke (3:23), Jesus was ‘about
thirty years of age’ when he received baptism, which naturally meant

40
See Declercq (2002). On Victorius, see also Mosshammer (2008), 239–44. Edi-
tions of both Victorius and Dionysius are found in Krusch (1938).
41
On the preference for the synoptic chronology, see Prologus Theophili 4 (ed. by
Krusch (1880), 225); Epistola Proterii 2–3 (ed. by Krusch (1880), 272); Bede, De tempo-
rum ratione 47, 61 (CCSL 123B, 432, 452).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

that he cannot have been younger than 30 when he died. Since the natu-
ral implication of Dionysius Exiguus’ Annus Domini era was that Jesus
was born at the end of either 1 BC or AD 1, acceptance of this era made
a crucifixion in AD 29 virtually impossible to maintain.
Another potential candidate for the Passion year was AD  31, in
which the full moon is shown on Saturday, 24 March, suggesting a cru-
cifixion on 23 March. Although a crucifixion on luna 13 is not in full
conformity with the Gospels, it does make for an attractive Julian date,
because it implies that Jesus rose from the dead on the day of the old Ro-
man vernal equinox, 25 March. It is thus not surprising that the Passion
date on 23 March AD 31 enjoyed some acceptance in the Greek East
and eventually became the dominant tradition in Byzantine chronogra-
phy.42 It is conceivable that it might have come to play the same role in
the Latin West, once the Alexandrian Easter reckoning had established
itself there, but on close inspection it becomes clear that Dionysius’ im-
position of the AD-regime worked against such developments. While it
is true that, if Jesus was born at the end of 1 BC and crucified in spring
AD 31, he would have reached the minimum age of 30, chronographers
in Dionysius Exiguus’ time had already reached a consensus that this was
not a satisfying way of accounting for the Gospel data. Basing themselves
mainly on the narrative found in the Gospel of John, chroniclers such as
Eusebius of Caesarea had come to the conclusion that the public minis-
try of Jesus lasted not just a few months, but over three years, bringing
his age at crucifixion to roughly 33 years.43
Bede, among other writers, was influenced by this current of thought
and provided an extensive commentary on the Dionysiac era in the
47th chapter of De temporum ratione. Bede’s conclusion was firm: if the
Dionysiac era was a correct representation of Christ’s birth-date, then
the crucifixion should have taken place in AD 34, when Jesus was aged
33; and on 25 March, as the Latin fathers, including St Augustine, had
taught. Yet even a superficial look at the corresponding Easter tables
was bound to bring disillusionment: in AD 34, 25 March had been a
Thursday and the 18th day of the moon, and no other year in the vicinity
yielded a combination of astronomical and calendrical data that satis-
fied the parameters found in the Gospels.44 Needless to say, this para-

42
See Grumel (1958), 23, 28, 30, 91, 112, 121–24; Loi (1971), 60–62.
43
An excellent survey of patristic views of the duration of Christ’s public ministry
is provided in Ogg (1940).
44
Bede, De temporum ratione 47 (CCSL 123B, 427–33).


 C. PHILIPP E. NOTHAFT

dox seriously troubled later generations of computists. When a group of

Frankish ‘experts’ on Easter reckoning (compotiste) were interrogated in
AD 809, they confidently dated Christ’s Passion to 25 March, only to get
tangled up in chronological contradictions:45

Hinc eis iniunctum, ut annos Domini a praesenti sursum versus computan-

do usque ad primum perducerent et ab eo deorsum versus usque ad passio-
nem eius, inquirentes diem eiusdem passionis dominicae in VIII. Kalendas
Aprilis, luna quinta decima concordantem. Inspectis itaque maiorum cyc-
lis, cum invenire, quod eis mandatum fuerat, pre difficultate quadam ne-
quirent, responderunt non posse se convenientem in hoc repperire rationem.
Igitur quid sequi vellent, quaesitum est. Responsio: Auctoritatem patrum,
id est Augustini, Hieronimi, Dyonisii, Bedae, qui passionem Domini VIII.
Kalendas Aprilis predicant.

‘At this point they were told first to count back the years of the Lord from
the present to the first, and then from it forward to His Passion, finding
whether the day of His Passion agreed with [their answer of ] 25 March,
luna 15. And so, when they studied the traditional cycles and were unable
to find what had been commanded them because of an inherent difficulty,
they replied that they could not find a formula to make those calculations
agree. Then they were asked which calculation they wished to accept. Re-
sponse: The authority of the fathers, that is of Augustine, Jerome, Diony-
sius, Bede, who preached the Lord’s Passion on 25 March.’

Their response reveals the dilemma posed by a desire to conform with

evangelical and patristic authority, on one hand, and an acceptance of
the 532-year Easter cycle, on the other. The upshot of this conflict of
data was that the art of ‘computistical chronography’, which strove for
perfect coherence between historical chronology and the parameters of
the Easter cycle, was in a state of serious ‘crisis’ at the time when Claudi-
us composed his chronicle. It is this crisis that provides the background
for Claudius’ own attempt to calculate anew the date of the Passion.46

45
Cap. comp. 2 (ed. by Borst (2006), 1040–41). The present translation has been
modified from Jones (1963), 26, who also edits and discusses this text. Further notewor-
thy early medieval reactions to the same problem can be found in the late-seventh-centu-
ry text De comparatione epactarum Dionysii et Victorii, edited by Warntjes (2010a), 322–
26, and chapters 1.7 and 1.8 of the Seven-Book-Computus (Lib. calc.) of AD 809/12
(Borst (2006), 1119–27).
46
For later medieval attempts at a solution to this dilemma, see now Verbist
(2010). For further details on ‘computistical chronography’ and the history of the search
for the crucifixion date, see Nothaft (2012).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

Claudius’ date of the crucifixion

As we have seen earlier, Claudius of Turin dated the resurrection of Jesus
Christ to 23 March, thereby creating a chronological parallel to the day
of Adam’s creation. The clause secundum hanc nostram supputationem,
which directly precedes the relevant passage, clearly indicates that this
parallelism was arrived at by computistical means. Claudius does not ex-
pressly state the year in which he thought the resurrection to have taken
place, but from the letter-preface to Ado it appears that he dated it to
AM 3986=AD 34, following Bede’s opinion in chapter 47 of De tempo-
rum ratione.47 If we apply Claudius’ rule for calculating the weekday on
1 January to this year, we get (3986+996+6)modulo7=4, indicating that
1 January fell on a Wednesday and 18 March on a Tuesday, with the next
Sunday falling on 23 March. This is precisely the combination presup-
posed by Claudius’ own comments on the crucifixion and resurrection
of Jesus.48
Taken together, these considerations prove that Claudius dated the
Passion to 21 March AD 34. Since 21 March actually fell on a Sunday
in AD 34 and Jesus was crucified on a Friday, this date is chronologi-
cally impossible, but it is wholly consistent with the two-day error im-
plicit in Claudius’ formula for the weekday. As it turns out, this error
had the attractive side-effect of moving the Easter full moon, which fell
on 21 March in AD 34, from Sunday to Friday. This result brought his
Passion date into harmony with the Gospels, which prescribed a cruci-
fixion on Friday, luna 14 ( John) or luna 15 (synoptic evangelists). If he
proceeded by his usual method of calculation, Claudius was technically
forced to side with John, because 21 March was the date of the Easter full
moon (luna 14) in AD 34. Bede, by contrast, had clearly favoured the
synoptic chronology, writing that ‘no catholic may doubt that the Lord
mounted the Cross on Friday, on the 15th day of the Moon’.49 Some inter-
esting further light on this potentially controversial aspect of Claudius’
calculation is shed by a commentary on Matthew that is generally at-

47
See n. 16 above and Bede, De temporum ratione 47 (CCSL 123B, 430–31).
48
See n. 13 above.
49
Bede, De temporum ratione 47 (CCSL 123B, 432; translation according to Wal-
lis (1999), 128–29): Nam quod dominus xv luna feria sexta crucem ascenderit, et una
sabbatorum, id est die dominica, resurrexerit a mortuis, nulli licet dubitare catholico ne
legi, quae per agnum paschalem decima quarta die primi mensis ad vesperam immolari
praecipit, pariter et Evangelio, quod dominum eadem vespera tentum a Iudaeis et mane
sexta feria crucifixum ac sepultum, prima sabbati resurrexisse perhibet, videatur incredulus.


 C. PHILIPP E. NOTHAFT

tributed to Remigius of Auxerre (†AD 908).50 Chapter 351 of this vo-

luminous work contains an intriguing passage on the chronology of the
Passion and the Last Supper, which later also circulated separately as an
excerpt appended to the twelfth-century Gospel harmony of Zachary of
Besançon.51 I shall quote the relevant bits as found in an eleventh-centu-
ry copy of the complete Matthew commentary, contained in Florence,
Biblioteca Medicea Laurenziana, Plut. 20.22:52

Attamen sciendum quod si diligenter et congruenter anni computentur ab

initio mundi usque ad passionem domini, invenimus quum ipso anno quo
dominus passus est secundum annos lunares VIII idus Martii fuit neome-
nia, id est initium nove lune et secundum Hebreos ipsius anni principium,
duodecimo kalendarum Aprelium in ipso vernali aequinoctio, feria sexta,
fuit luna quartadecima et semis pascha Iudeorum.

‘Nevertheless, one should know that if the years from the beginning of
the world until the Lord’s Passion are reckoned carefully and appropri-
ately, we find that, since the new moon (that is, the beginning of the new
moon and the beginning of the year according to the Hebrews) in the
year in which the Lord suffered was on the 8th before the ides of March
[8 March] according to lunar years, [it follows that] the Passover of the
Jews was on 12th day before the kalends of March [21 March], the day of
the vernal equinox, Friday, with a lunar age of 14 and a half.’

After reflecting on the typological congruence between the slaughter-

ing of the paschal lamb on luna 14 and the crucifixion of Jesus, the ‘true
lamb of God’, on the same day, the author goes on to write:53

Sed forte querit aliquis quare ille preoccupavit comedere agnum. Cui di-
cendum est, quia non fuit sub lege ipse, voluit comedere agnum ut traderet
corporis et sanguis sui misteria discipulis. Videtur enim Iohannes huic op-
pinionem sentiens esse cum dicit Iudei autem ‘non introierunt in pretori-
um ut non contaminarentur sed manducarent pascha’ [18:28], quod enim

50
Tax (1991).
51
See, e.g., Graz, Universitätsbibliothek, 234 (saec. XIII), 243vb; Munich, Bayer-
ische Staatsbibliothek, Clm 4546 (saec. XIIex.), 268vb, both of which contain Zachary’s
Gospel harmony In unum ex quatuor. For an incomplete list of MSS of both this excerpt
and the entire commentary on Matthew, see Jeudy (1991), 467–71.
52
(Pseudo-)Remigius of Auxerre, Commentarium in Matthaeum 351 (Mt. 27:62;
Florence, Biblioteca Medicea Laurenziana, Plut. 20.22, 217r).
53
(Pseudo-)Remigius of Auxerre, Commentarium in Matthaeum 351 (Mt. 27:62;
Florence, Biblioteca Medicea Laurenziana, Plut. 20.22, 217r).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

duodecimo kalendarum crucifixus est, decimo resurrexit, quo die plasma-

tur legitur primus homo.

‘But perhaps someone asks why [ Jesus] anticipated the eating of the
Lamb. To him one should say that, since [ Jesus] was not under the Law,
he wanted to eat the Lamb [one day earlier], so he could transmit the
mysteries of his body and blood to the disciples. For it would seem that
John agrees with this opinion when he says that the Jews did not enter
the Praetorium, ‘lest they should be defiled, but that they might eat the
Passover’—namely that he was crucified on the 12th day before the kal-
ends [21 March] and resurrected on the 10th [23 March], on which day,
it is read, the first human was created.’

The affinities between this exposition and Claudius’ chronicle are palpa-
ble. Not only did Remigius (or whoever wrote the commentary) choose
the same unusual date for the crucifixion, but he drew the conclusion, in-
herent also to Claudius’ reckoning, that the corresponding lunar age was
14 rather than 15. As a result, it had to be conceded that Jesus did not
celebrate Passover together with the Jews, but instead anticipated the meal
by one day, eager to share bread and wine with the disciples before his
imminent death. His remarks also show how Claudius could have justi-
fied his chronological stance on scriptural grounds, i.e. on the basis of the
Gospel of John, without incurring the immediate suspicion of heterodoxy.
Although there are no straightforward verbal parallels between (Pseudo-)
Remigius and the pertinent passages in the chronicle of Claudius of Turin,
it is still likely that the latter lurks behind this exposition. Particularly strik-
ing is the reference to a calculation of the years since the beginning of the
world (si diligenter et congruenter anni computentur ab initio mundi usque
ad passionem domini), which seems out of place in a Gospel commentary
and points to the use of a chronographic source. Moreover, the commen-
tator stresses the parallelism between the calendar date of the resurrection
and the creation of Adam (23 March), which, as seen above, is also a cen-
tral feature of the corresponding discussion in Claudius’ chronicle.
More daring than the Johannine lunar age was Claudius’ choice
of Julian dates, given that a crucifixion on 21  March—unlike Bede’s
23  March or the more orthodox 25  March—was devoid of any prec-
edent or backing from patristic tradition.54 Nevertheless, Claudius may
have drawn some justification for his opinion from chapter  30 of De
54
Bede considered both 23 and 25 March in chapters 47 and 61 of De temporum
ratione, but only referred to 25 March in chapter 30 and in his Letter to Wicthed. By
contrast, in chapter 66 he made some circumspect remarks in favour of a crucifixion on


 C. PHILIPP E. NOTHAFT

temporum ratione, where the dates of the equinoxes and solstices in the
Julian calendar are discussed. Bede begins by detailing an old Roman
tradition, which put the four cardinal points of the year on the 8th day
before the kalends [i.e. the first day] of April, July, October, and January.
His comments on this tradition are worth quoting in full:55

Haec quidem gentiles, quibus non dissimilia de tempora etiam perplures

ecclesiae tradidere magistri, dicentes: viii kl. Apriles in aequinoctio verno
dominum conceptum et passum, eundem in solsitio brumali viii kl. Ianu-
arias natum; item beatum praecursorem et baptistam domini viii kl. Octo-
bres in aequinoctio autumnali conceptum, et in aestivo solstitio viii kl. Iuli-
as natum—addita insuper expositione quod auctorem lucis aeternae cum
cremento lucis temporariae concipi simul et nasci deceret, poenitentiae uero
praeconem quem oportebat minui cum inchoata minoratione lucis generari
pariter et concipi. Verum quiae, sicut in ratione paschali didicimus, aequi-
noctium vernale duodecimo kalendarum Aprilium die, cunctorum orien-
talium sententiis et maxime Aegyptiorum quos calculandi esse peritissimos
constat, specialiter adnotatur, caeteros quoque tres temporum articulos pu-
tamus aliquanto priusquam uulgaria scripta continent esse notandos.

‘This is what some of the pagans say; and very many of the Church’s teach-
ers recount things which are not dissimilar to these about time, saying that
our Lord was conceived and suffered on the 8th before the kalends of April
[25 March], at the spring equinox, and that he was born at the winter sol-
stice on the 8th before the kalends of January [25 December]. And again,
that the Lord’s blessed precursor and Baptist was conceived at the autumn
equinox on the 8th before the kalends of October [24 September] and born
at the summer solstice on the 8th before the kalends of July [24 June]. To
this they add the explanation that it was fitting that the Creator of eternal
light should be conceived and born along with the increase of temporal
light, and that the herald of penance, who must decrease, should be engen-
dered and born at a time when the light is diminishing. But because, as we
have learned in connection with the calculation of Easter, the judgment of
all men of the East (and especially of the Egyptians, who it is agreed, were
the most skilled in calculation) is in particular agreement that the spring
equinox is on the 12th before the kalends of April [21 March], we think
that the three other turning-points of the seasons ought to be observed a
little before [the date] given in the popular treatises.’

23 March. See Bede, De temporum ratione 30, 47, 61, 66 (CCSL 123B, 374, 432, 452,
464) and Epistola ad Wicthedum 12 (CCSL 123C, 642).
55
Bede, De temporum ratione 30 (CCSL 123B, 374; translation slightly modified
from Wallis (1999), 87–88).


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

The mentioned connection between the annual course of the sun and the
conception and birth of Jesus Christ and John the Baptist could be justi-
fied by recourse to John 3:30, where the Baptist is quoted as saying ‘he must
increase, but I must decrease’. This was sometimes interpreted as referring
to the change from increase to decrease of daylight, and vice versa, that
was associated with the two equinoxes.56 According to Bede, however, the
traditional dates to which the conception and birth of Jesus and John were
assigned did not really correspond to the true dates of the equinoxes and
solstices. Indeed, his words could be taken to imply that the four calendar
dates commemorating Jesus and John should be shifted by four days, to
the 12th day before the kalends of April, July, etc.57 Such a shift could also
be taken to affect the date of the Passion, which was generally assumed to
have taken place on the same calendar date as Christ’s conception (i.e. the
feast of the Annunciation on 25 March). One important authority for this
notion was St Augustine, who, in his Quaestiones Exodi, had claimed that
the typological link between the date of the conception and the Passion
was already hinted at in the book of Exodus (23:19: ‘Do not cook a young
goat in its mother’s milk’).58 Claudius was in all likelihood aware of this
idea, as he inserted a longer passage from Augustine’s Quaestiones into the
‘reckoning section’ of his Chronicle, and may have taken it very seriously.59
It is thus conceivable that he extended Bede’s criticism of the traditional
date of the vernal equinox to the date of the crucifixion, thereby finding an
important justification to move it from 25 March to 21 March.
In doing so, Claudius was—at least by his own lights—able to solve
the most nagging chronological problem Bede’s De temporum ratione

56
Augustine, De diversis quaestionibus 58.1 (ed. by Almut Mutzenbecher in CCSL
44A, 104); Augustine, Enarrationes in psalmos 132.11 (CCSL 40, 1934); Augustine,
In Iohannis Evangelium 14.5 (ed. by Radbod Willems in CCSL 36, 144); De solstitiis
et aequinoctiis (PL, Supplementum 1, 561); Caesarius of Arles, Sermones 216 (ed.  by
Germain Morin in CCSL 104, 859); Pseudo-Bede, De argumentis lunae (PL 90, 724);
Bede, In Lucae Evangelium expositio 1.1.24 (ed. by David Hurst in CCSL 120, 28). See
also Jones (1943), 366; McCluskey (1998), 27; Warntjes (2010a), 106–07, 150–53; and
Love (2007), 76–78, who shows that Bede used De solstitiis et aequinoctiis.
57
Such conclusions were indeed drawn at the beginning of the eleventh century by
Heriger of Lobbes, Epistola ad quendam Hugonem Monachum (PL 139, 1133–34). On
this work, see Verbist (2010), 15–33.
58
Augustine, Quaestiones Exodi 90 (ed.  by Johannes Fraipont in CCSL 33,
114–16). See also Augustine, De diversis quaestionibus 56 (CCSL 44A, 96); Augustine,
De Trinitate 4.5 (ed. by William J. Mountain in CCSL 50, 172); De solstitiis et aequinoc-
tiis (PL, Supplementum 1, 562–63).
59
Compare Claudius of Turin, Brevis chronica (PL 104, 922) with Augustine,
Quaestiones Exodi 47.6 (CCSL 33, 92).


 C. PHILIPP E. NOTHAFT

had posed to subsequent generations of computists: the problem of how

to find a satisfactory date of the crucifixion, one which would simulta-
neously fall on a Friday and a full moon. Bede himself had despaired of
finding an answer, as can be seen from his ironic remark that anyone who
managed to find the appropriate data for AD 34 should ‘give thanks to
God, for He has granted that you find what you were looking for, just as
He promised’.60 Claudius, close reader of Bede that he was, may well have
regarded his biblical calendar dates, which were the outcome of cumber-
some calculations, as God’s reward for an astute researcher.

Appendix: Claudius’ biblical calendar dates

1. The Flood year (AM 1656)
According to Genesis 7:11, the flood began on the 17th day of the second
month. Claudius mistakenly speaks of the 17th day of the first month
and equates it to Saturday, 5 April.61 The erroneous lunar date is also pre-
sent in the ‘display section’ of the chronicle, but not in Claudius’ com-
mentary on Genesis, where he correctly locates the flood’s beginning in
the second month.62 In the chronicle, Claudius’ mistake leads him to
perceive a chronological symmetry between the beginning of the flood
and Adam’s creation, which both allegedly happened on the 17th day of
the first lunar month.63 He also notes that AM 1656 was a leap year with

60
Bede, De temporum ratione 47 (CCSL 123B, 431–32; translation according to
Wallis (1999), 128): Et ideo circulis beati Dionysii apertis, si quingentesimum sexagesi-
mum sextum ab incarnatione domini contingens annum, quartam decimam lunam in eo
VIIII kal. Apr. quinta feria repereris, et diem paschae dominicum VI kal. Apr. luna decima
septima, age Deo gratias quia quod quaerebas, sicut ipse promisit, te invenire donavit.
61
Claudius of Turin, Brevis chronica (PL 104, 919): Et anno vitae illius 600 venit
diluvium secundum dies lunares, mense primo, septima decima die mensis; secundum so-
lares vero dies, Kalendis Aprilis [sic!], luna septima decima; quem Hebraei, qui menses non
a sole, sed a luna computant, septimum decimum diem vocaverunt primi mensis, septima
feria, quam nos sabbatum nuncupamus, ingressus est Noe in arcam, et hac ipsa die venit
diluvium. The Kalendis Aprilis (1 April) in Labbé’s edition has to be emended to nonas
Apriles (5 April), as correctly found in Madrid, Biblioteca Nacional, 9605, 113rb, and
Monza, Biblioteca Capitolare, c-9/69, 79vb.
62
Monza, Biblioteca Capitolare, c-9/69, 68v; Claudius of Turin, Commentarii in
Genesim (PL 50, 925C).
63
Claudius of Turin, Brevis chronica (PL 104, 919): Mirum omnipotentis Dei
judicium in humanum genus, quod per Adam luna septima decima accepit initium, in
non custodiendo Dei praeceptum septima decima in diluvio incurreret mortis supplicium.


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

an epact of 11 and that the new moon of the first Hebrew month was on
20 March. The year began on a Wednesday.64
Computistical check: Claudius’ claims with regards to the calendar
date and epact are in harmony with his computistical rules for translat-
ing biblical into Julian dates. AM 1656 is a Julian leap-year (1656mod-
ulo4=0) and the fourth year of the cyclus decemnovenalis (1656modu-
lo19=3; 3+1= 4), with an epact of 3 (=3×11–30), leading to a Claudian
‘epact’ of 11 (=3+8). Since the Claudian ‘epact’ designates the lunar age
on 31  December, it follows that 1  January is the 12th day of a 30-day
lunation. 1 Nisan (Nisan being equated with the April lunation) should
by implication be the 80th day of the leap year: 19+30+30+1=80.65 This
corresponds to 20 March (31+29+20=80). A new moon on 20 March
implies luna 17 on 5 April, as noted by Claudius. As for the weekday,
it has been established that AM 1656 is a Julian leap year, in which the
weekday on 1 January does not differ from that on 18 March. Accord-
ingly, the weekday on 1 January in the year of the flood should have been
a Thursday ((1656+414)modulo7=5). Claudius ignored this and calcu-
lated the weekday in his usual fashion, adding six dies intercalares to arrive
at Wednesday ((1656+414+6)modulo7=4). In a leap-year, the weekday
on 5 April differs from that on 1 January by 4 days (31+29+31+5=96;
–1, because the count has to be exclusive; 95modulo7=4). Since the year
AM 1656 started on a Thursday, 5 April should have accordingly fallen
on a Monday (5+4modulo7=2). Claudius, on the other hand, once
again ignored the bissextus and added only three days. Counting from
Wednesday, he arrived at Saturday.66

Cf. Claudius of Turin, Brevis chronica (PL 104, 917): Et ita ratione deducta X Kal. Aprilis
luna septima decima invenitur protoplastus Adam ex terrae limo esse formatus.
64
Claudius of Turin, Brevis chronica (PL 104, 919): Anno vitae Noe 600 impleti
sunt anni a conditione mundi 1656; eo enim anno quo diluvium fuit, annus bissextilis exsti-
tit: et si more nostro computarentur feriae vel Kalendae, essent ipso anno Kalendae mensis
Januarii feria quarta: et quia fuit annus cycli solaris 4, idcirco XI fuerunt in Epacta et XIII
Kalendas Aprilis fuit Neomenia, id est novae lunae principium, et ipsius anni secundum
menses lunares primi mensis initium, impletique sunt ipso anno anni Jubilaei a constitu-
tione mundi 33, cycli decemnovales 82 [read: 87], dies bissextiles 414.
65
Since AM  1656 is a Julian leap-year, the lunar bissextus is applied, increasing
the length of the lunation of February to 30; see Bede, De temporum ratione 41 (CCSL
123B, 405–07).
66
Immo Warntjes suggests that, rather than ignoring the bissextus, Claudius may
have calculated 1 January to fall on Tuesday (3+4=7). This would have been consistent
with the approach mentioned in n. 32 above, which subtracts 1 before calculating the
leap-day increment (1656+[(1656–1)/4]+6)modulo7=3=Tuesday). Yet both Labbé’s
edition, based on the Paris MS, and the other two MSS expressly mention feria IIII or


 C. PHILIPP E. NOTHAFT

2. Noah leaves the ark (AM 1657)

Noah left the ark in the following year on the 27th day of the second
month (Genesis 8:14). According to Claudius, this was a Sunday, 3 May,
exactly one year and 28 days after the flood’s inception. He also notes
that 1 January in AM 1657 fell on a Thursday, that the epact was 22 and
that the new moon of the first month fell on 9 March.67
Computistical check: AM  1657 is a Julian common year and the
fifth year of the cyclus decemnovenalis (1657modulo19=4; 4+1=5),
with a lunar epact of 14, leading to a Claudian ‘epact’ of 22. The new
moon of the first Hebrew month will hence be the 68th day of the
year (8+29+30+1=68), which is indeed 9  March (31+28+9=68).
The 27th day of the second month will fall 55 (=28+27) days later, on
3  May (22+30+3). According to Claudius’ weekday formula, 1  Janu-
ary AM 1657 was a Thursday ((1657+414+6)modulo7=5). Being the
123rd day of the Julian common year (31+28+31+30+3=123), 3 May
was indeed, according to Claudius’ calculation, bound to fall on a Sun-
day ((123–1)modulo7=3; (3+5)modulo7=1).

3. The Exodus from Egypt (AM 2453)

Claudius states that the 14th day of the first Hebrew month was 12 April,
30 March being the beginning of that month. The Exodus took place on
13 April, whereas the Law was received on Monday, 31 May, the fourth
day of the third month (Exodus 19:16).68
Computistical check: AM 2453 is a common year and the third year
of the cyclus decemnovenalis (2453modulo19=2; 2+1=3), with a lunar
epact of 22, leading to a Claudian ‘epact’ of 30. The new moon of the first
Hebrew month will thus be the 90th day of the year (30+29+30+1=90),
which is 31  March (31+28+31=90). Claudius, by contrast, notes
30 March, which makes the Passover / Easter full moon fall on 12 April.

quarta. See n. 64 above and MSS Monza, Biblioteca Capitolare, c-9/69, 80ra; Madrid,
Biblioteca Nacional, 9605, 113va.
67
Claudius of Turin, Brevis chronica (PL 104, 919–20): Secunda saeculi aetate,
prima hujus die, quae est decima septima dies mensis secundi, egressus est Noe de arca,
uxor ejus, et filii, et uxores filiorum ejus, die prima feria, quam nos propter resurrectionem
Domini Dominicam nuncupamus: quia, sicut supra jam dixi, si more nostro ipso tempore
computarentur feriae vel Kalendae, essent ipso anno, quo egressi sunt de arca, Kalendae
mensis Januarii, feria quinta anno primo post bissextum. Et quia fuit annus quintus cycli
solaris, qui est decemnovalis, et quartus epactalis, XXII fuerunt in Epacta, et fuit VII Idus
Martii mensis Neomenia, id est novae lunae principium. Et haec fuit secundum Hebraeos
primi mensis initium, et anni principium.
68
See n. 27 above.


 CLAUDIUS OF TURIN AND THE DATE OF CHRIST’S PASSION

The reception of the Law, on the fourth day of the third month, thus took
place on the 151st day of the year (30+29+30+29+30+4–1=151), which
is 31 May (31+28+31+30+31=151). AM 2453 began on a Friday, if we
follow Claudius’ reckoning rules ((2453+613+6)modulo7=6). 31 May
was bound to fall three weekdays later, on a Monday, being the 151st day
of the common year ((151–1)modulo7=3; (3+6)modulo7=2).

4. The year after the Exodus (AM 2454)

In the year after the Israelites’ Exodus from Egypt, on the first day of
the first month, the Tabernacle was erected (Exodus 40:2). According to
Claudius, this took place on Sunday, 20 March, whereas the following
Passover eve (14 Nisan) was celebrated on Saturday, 2 April.69
Computistical check: AM 2454 is a Julian common year and the fourth
year of the cyclus decemnovenalis (2454modulo19=3; 3+1=4) with a lu-
nar epact of 3, yielding a Claudian ‘epact’ of 11. The new moon of the
first Hebrew month fell on the 79th day of the year (19+29+30+1=79),
which is 20  March (31+28+20=79). As in the case of the flood year,
the Passover / Easter full moon thus fell on 2 April. As for the week-
day, 1 January AM 2454 was a Saturday according to Claudius’ formula
((2454+613+6)modulo7=0=7). 20  March or 1 Nisan occurred one
weekday later, on a Sunday ((79–1)modulo7=1), while 14 Nisan was a
Saturday.

5. The destruction of Solomon’s Temple (AM 3363)

Jerusalem was captured by the Babylonians ( Jeremiah 39:2) on the fifth
day of the fourth month, which Claudius equates to Saturday, 23 June.
The final destruction of the city ( Jeremiah 52:12) took place on the
tenth day of the following month, or 28  July. The Hebrew lunar year
began on 23 March.70

69
Claudius of Turin, Brevis chronica (PL 104, 923): Moyses anno vitae suae 80 educ-
tum ex Aegypto rexit populum Dei Israel in deserto annis 40, quorum primo anno taber-
naculum Domino construxit, et septem mensibus omne opus perficiens, mense primo, anni
secundi prima die mensis erexit, secundum dies solares ipso anno XIII Kal. Aprilis, feria
prima fuit Neomenia, id est novae lunae principium, et primi mensis initium: et Pascha
Hebraeorum feria septima, quarto Nonas Aprilis, luna quarta decima quam illi quartam
decimam primi mensis diem secundum dies lunares appellare consueverunt.
70
Monza, Biblioteca Capitolare, c-9/69, 82rb–va: Secundum dies lunares mense
quarto, quinta die mensis, iuxta solares vero dies VIIII kl. Iulias, luna quinta, feria sep-
tima, aperta est civitas […] mense quinto, decima die mensis lunaris, solaris vero V kl. Aug,
luna X […] destructi sunt muri civitatis. […] Unde si a me queratur per quid sciam hoc aut
quomodo potuit fieri ut talibus diebus vel comprehensa fuerit civitas vel incensa respondeo


 C. PHILIPP E. NOTHAFT

Computistical check: AM 3363 is a Julian common year and the first

year of the 19-year cycle (3363modulo19=0; 0+1=1), with a lunar epact
of 0, producing a Claudian ‘epact’ of 8. This leads to a new moon of the first
Hebrew month on the 82nd day of the common year (22+29+30+1=82),
which is 23 March (31+28+23=82). The fifth day of the fourth month is
bound to fall on the 174th day (22+29+30+29+30+29+5=174), which
is 23 June (31+28+31+30+31+23=174), and the tenth day of the fifth
month takes place on the 209th day (22+29+30+29+30+29+30+10=209),
which is 28 July (31+28+31+30+31+30+28=209). AM 3363 began on
a Monday according to Claudius’ formula ((3363+840+6)modulo7=2).
23 June, being the 174th day of the year, fell five weekdays later, on a Sat-
urday ((174–1)modulo7=5; 2+5=7).

quia ipso anno X kal. Aprilis extitit secundum dies lunares anni principium et primi mensis
initium. Colliguntur omnes anni ab initio mundi usque ad eversionem Hierosolymiae que
facta est per Nabuchodonosor […] 3363 anni […] dies bissextiles 840. See also Madrid, Bib-
lioteca Nacional, 9605, 115r. This passage is missing from Labbé’s edition.


 LISA CHEN OBRIST

WHAT A DIFFERENCE A DAY MAKES:

THE EIGHTH DAY OF THE WEEK IN
BOOK 10 OF HRABANUS MAURUS’
DE RERUM NATURIS

Abstract
Book 10 of Hrabanus Maurus’ De rerum naturis covers the topic of time and
how to calculate it. In the earliest witnesses and in all the printed editions, the
last chapter of the book, chapter  17, consists of one contiguous 80-word ex-
cerpt from Isidore of Seville’s Etymologiae (6.18) about the different meaning
of the Lord’s Day for the Christians and the Jews. The last sentence of this ex-
cerpt from Isidore makes mention of the Lord’s Day as the eighth day of the
week but provides no explanation for this manner of counting the days of the
week. However, in at least 18 later manuscripts, the earliest of which dates to
the twelfth century, chapter 17 is almost 500 words long and includes an ad-
ditional passage, after the excerpt from Etymologiae, which talks about why the
Lord’s Day is celebrated, what will happen on the last day when all believers will
be resurrected like Jesus, and the difference between the Christian and Jewish
understandings of eternal salvation. This paper identifies for the first time in
print the source for the bulk of this additional passage at the end of book 10. It
is taken from another work by Isidore, De ecclesiasticis officiis (1.24–25), a work
which has so far not been identified as a source anywhere else in the 22 books
of De rerum naturis.
Keywords
Hrabanus Maurus, De rerum naturis, De computo; Carolingian computus, me-
dieval calendars; eight-day week.

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 293-308
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114741
 LISA CHEN OBRIST

Introduction
My thesis project was a study, edition, and translation of book  10 of
Hrabanus Maurus’ De rerum naturis.1 I came to this topic when I was
introduced to the much larger project of a critical edition of the com-
plete De rerum naturis directed by Prof. William Schipper at Memorial
University in Newfoundland. He is at the moment completing the first
volume of the work to be published in the Corpus Christianorum Con­
tinuatio Medievalis series.2
As the title of my article suggests, I will be focusing on a particular
passage of book 10 that deals with the eighth day of the week. I was im-
mediately curious about this idea mainly because it was not an historical
aspect of the calendar that I was familiar with in the way that I was about
other notions of early medieval computus such as the leap year and the
difference between the lunar month and the modern western calendar
month. Since delving into the manuscripts, I have made some new dis-
coveries about the passage in book 10 concerning the eighth day of the
week that have not previously been discussed in print.
Hrabanus Maurus (c.AD  780–856) is most often associated with
the monastic foundation at Fulda where he received his early education
and became a monk.3 After being consecrated as deacon in AD 801, he
was sent to study under Alcuin at Tours. In AD 822, Hrabanus then
took on the role of abbot at Fulda which he fulfilled for the next twenty
years.4 Hrabanus was heavily influenced by his teacher and, in turn, be-

1
Time and how to calculate it: a study and edition with translation of Book Ten
of Hrabanus Maurus’ De rerum naturis (Defended April 2015, Supervisor: Prof. John
Magee), Centre for Medieval Studies, University of Toronto.
2
I would like to take this opportunity to thank Prof. William Schipper for his
guidance and support. It was only with his blessing that I was able to claim book 10 for
my own exploration and only with his generosity that I was able to access an abundance
of resources that made this project possible.
3
The 1150th anniversary of Hrabanus’ death in 2006 generated a large amount of
new scholarship; see, e.g., Archa verbi 4 (2007); Felten and Nichtweiss (2006); Haar-
länder (2006); Kotzur (2006). For helpful older surveys of Hrabanus’ place in the Caro-
lingian context, see Brunhölzl (1975), 325–40, 554–56, as well as the essays in Kottje
and Zimmermann (1982). For a comprehensive and recent survey of the history of the
monastery of Fulda, see Aris et al. (2004).
4
Born in Mainz in AD 784, raised at Fulda, where Bangulf sent him to the school
at Tours, Hrabanus added Maurus to his name at the wish of Alcuin who gave the nick-
name to his pupil. He returned for his second stay at Tours after AD 801. He was then
sent to Fulda until AD 822 where he served as abbot for twenty years. Then, in AD 847,
he became bishop at Mainz and died in AD 856. Always loyal to Louis the Pious, he also
had the favour of Louis the German and Lothair I, but with interruptions. According to


 WHAT A DIFFERENCE A DAY MAKES

came a renowned teacher himself, attracting the likes of Gottschalk of

Saxony, Walahfrid Strabo from Reichenau, and Servatus Lupus from
Ferrières to study under him at Fulda. His highly active political ca-
reer, as well as his reputation as an administrator and teacher, led to the
commissioning of many of his voluminous works. His writings include
several instruction books for clergy, biblical commentaries, acrostic po-
ems, letters he exchanged with scholars and politicians, and textbooks
for his students, including one wholly devoted to the study of compu-
tus.5
De rerum naturis, also known as De universo, is an encyclopedic
compilation in 22 books composed in the interim between Hrabanus’
20-year tenure as abbot of Fulda and his appointment to the archie-
piscopal seat at Mainz in AD 847.6 The main source and inspiration
for Hrabanus’ oeuvre was Isidore’s Etymologiae.7 Hrabanus, however,
expanded the scope of what Isidore had presented by adding an em-
phasis on the mystical significance of objects and he also re-arranged
the material according to a new hierarchy. De rerum naturis comprises
two sections: books 1 to 11 focus on theological matters (God, the Bi-

de Ghellinck (1939), 102–03, Hrabanus was a grammarian, teacher, moralizer, exegete,

and encyclopedist, poet, and all-round versifier and was highly regarded by his contem-
poraries and the elite. For details of Hrabanus’ career, see Schipper (1997b), 364–65. For
a history of Fulda and the influence of its abbots, see the recent work by Raaijmakers
(2012), especially 175–264 on the period of Hrabanus’ abbacy.
5
By far Hrabanus’ most popular work was the Liber de sanctae crucis which still
survives in 81 copies. For Hrabanus’ writings see Schipper (2007); Embach (2007); Fer-
rari (1999); Kottje (2012). For his De computo, see the edition by Wesley Stevens in
CCCM 44.
6
For a detailed description of the sequence of events leading to the writing and
publishing of Hrabanus’ De rerum naturis from Mainz and Fulda, see Schipper (1997b),
364–65. In a quote from the second prefatory letter to his friend Haymo of Halberstadt,
Hrabanus does reveal that he had been thinking about such a project since his student
days and that he finally had the opportunity to focus on it without outside distractions;
Hrabanus Maurus, Epistola 36 (ed. by Ernst Dümmler in MGH Epp. 5, 470–71): post­
quam me divina providentia ab exteriorum negotiorum cura absolvit, teque in pastoralis
cure officium sublimavit, cogitabam, quid tuae sanctitati gratum et utile in scribendo con­
ficere possem: quo haberes ob commemorationem in paucis breviter annotatum quod ante
in multorum codicum amplitudine, et facunda oratorum locutione dissertum copiose legisti.
According to Schipper (2007), 111–13, 104 n. 4, 113 n. 25, Hrabanus probably com-
posed the letters specifically for the edition he issued before AD 852 and likely com-
posed the letter to Haymo first, to accompany the copy of De rerum naturis he sent
to Haymo c.AD 846, and the second letter, to King Louis, perhaps in connection with
Hrabanus’ election and appointment as archbishop of Mainz in early AD 847.
7
Isidore’s Etymologiae are ed. by Lindsay (1911). For a more detailed description
of their contents, see Sanford (1949); Collison (1964), 34; Heyse (1969).


 LISA CHEN OBRIST

ble, the Church, etc.) and books 12 to 22 focus on secular affairs (the

world, nature, and aspects of daily life, etc.).8 Today, Hrabanus’ encyclo-
pedia survives in 35 complete and 24 fragmentary manuscripts, a few
of which are dated to the ninth century and possibly to his lifetime.9
The editio princeps of De rerum naturis was printed by Adolf Rusch in
1467.10 The Rusch edition was reprinted by George Colvener in his
collected edition of Hrabanus’ works in 1627 and again by Jacques-
Paul Migne in the series Patrologia Latina in 1851.11 To date, I  have
collated 18 manuscripts and the Rusch edition, and it is the edition I
produced for my dissertation that forms the basis for the textual analy-
sis in this article.
Book 10 discusses the phenomena caused by objects in the sky in-
stead of discussing the physical qualities of these objects, which are the
topic of book 9. The art and science of computus involves, among other
topics, the study of the motions of celestial objects, their actions and
their behaviours. The aim of computus is to devise tools and methods
to enable measurement of these activities. In other words: how do we
measure the passage of time, record past events, and predict future
occurrences?12
Hrabanus includes topics such as the division of time into ages,
years, months, weeks, days, hours and minutes; the parts of day and
night; the significance of feast days; and the importance of the sab-

8
This two-part division of the books is reflected in the early two-volume sets of
the work: the Reichenau copy (now Karlsruhe, Badische Landesbibliothek, Aug. perg.
96 and 68), the Worms copy (now Vienna, Österreichische Nationalbibliothek, 121),
and London, British Library, Harley 3092 (a tenth-century copy from Cusa). However,
according to Schipper (1997b), 363, ‘later two-volume copies, on the other hand, invari-
ably begin with Book 11 (de diversitate aquarum), suggesting that the earlier division was
original with Hrabanus’.
9
See the Appendix for a current list of known manuscripts as well as the works by
Schipper listed in the bibliography and Kottje (2012), 261.
10
Rusch (1467); see Scholderer (1939), 44.
11
Colvener (1627), i 51–272, repr. in PL 111, 1–680; there is a lacuna between the
words stirpe and Agareni in De rerum naturis 16.2 (PL 111, 438), which, according to
Schipper (2004), 2, corresponds precisely in its opening and closing words to a single leaf
in Rusch’s edition; the PL edition is unreliable because it contains a number of lacunae
and where Hrabanus quotes from Isidore, Migne corrected the text to correspond with
the Etymologiae; for more details, see Schipper (2004), 2.
12
For definitions of the subject of medieval computus see Warntjes (2010a),
XXXIII. A detailed discussion of the evolution of computistical terms can be found in
Wallis (1996), 383–87.


 WHAT A DIFFERENCE A DAY MAKES

bath and the Lord’s Day.13 Learning this type of basic information is
the first step in knowing how to deal with the challenges of reconciling
the Julian calendar, the liturgical calendar, and the lunar calendar. This
is a challenge that must be overcome in order to establish the dates of
the moveable feasts of the Christian liturgical year and it is of utmost
importance for knowing when to celebrate Easter. It reflects the basic
knowledge and understanding of similar topics discussed in De com­
puto14 and they, in fact, share some of the same excerpts from Isidore
and Bede.

Hrabanus’ sources for book 10 of his De rerum naturis

Hrabanus’ major sources for book 10 are Bede’s De temporum ratione15
(On the Reckoning of Time) and Expositio in sancti Iohannis evange­
lium16 (his Commentary on the Gospel of John), Cassiodorus’ Expo­
sitio in psalmos17 (Commentary on the Psalms), the Clavis Melitonis,18
and Isidore’s Etymologiae.19 The major studies which review the sources
­Hrabanus used in the De rerum naturis are Elisabeth Heyse’s and Di-
ane O. Le Berrurier’s dissertations and the works of William Schipper.
13
The chapters in book 10 cover the following topics: (1) times; (2) moments; (3)
hours; (4) days; (5) the parts of the day; (6) the night; (7) the seven parts of the night;
(8) darkness; (9) weeks; (10) months; (11) the seasons; (12) the year; (13) the saeculum;
(14) the six ages; (15) festal days; (16) the Sabbath; and (17) the Lord’s Day.
14
CCCM 44, 165–323.
15
Bede, De temporum ratione (ed. by Charles W. Jones in CCSL 123B). For a dis-
cussion of Bede’s sources in De temporum ratione see Jones (1943); Wallis (1999), lxxii–
lxxxv. For an analysis of resources available to Bede at Jarrow during the early eighth
century, see Ganz (2006), 100.
16
Bede, Expositio in sancti Iohannis evangelium (PL 92, 634–938).
17
Cassiodorus, Expositio in psalmus (ed. by Marcus Adriaen in CCSL 97–98).
18
The most popular form of the Clavis (Stegmüller (1940–80), iii 559–61
(no. 5574)) is edited by Pitra, (1852–58), ii 1–519; iii 1–307. Another version of the
Clavis (Stegmüller (1940–80), iii 561–62 (no. 5575)), considered by Pitra to be the ear-
liest, is edited by him in Pitra (1879–91), ii 6–127, with a subject-index on pp. 147–54.
For a summary of the current editions see Kaske (1988), 34–35. A new edition of the
Clavis Melitonis is being prepared by Toby Burrows for Corpus Christianorum; Burrows’
website has a bibliography: http://confluence.arts.uwa.edu.au/display/~tburrows/
Clavis+Melitonis+project.
19
Isidore, Etymologiae (ed. by Lindsay (1911)). See Barney (2006), 10–24, for a
discussion of how Isidore uses his sources. For a detailed discussion of the reception of
Isidore in and after Carolingian times, see Barney (2006), 24–28; Curtius (1953), 23;
Fontaine (1959), 863–88; Brehaut (1912), 15–34.


 LISA CHEN OBRIST

Heyse’s 1969 dissertation, under the direction of Bernhard Bischoff, re-

viewed the sources for all 22 books.20 Unfortunately, she based her study
on the text in the Patrologia Latina, which is not reliable.21 She was not
able to identify many passages that have now been identified by other
scholars and her manuscript contains many small errors.22 More recently,
William Schipper, in his preparations for the publication of an edition
of all 22 books, has published several articles and papers on the sources
Hrabanus used in De rerum naturis as well as some of his other exegetical
writings.23 After verifying Heyse’s findings, correcting a few and adding a
new source, I was able to compile the results in Table 1 below. Heyse only
identified six text sources for book 10: Bede’s De temporum ratione, Cas-
siodorus’ Expositio in psalmus, Clavis Melitonis, Isidore’s Etymologiae,
Bede’s Expositio in sancti Iohannis evangelium, and Gregory’s Moralia in
Job. As illustrated in Table 1, a further eight source texts by Ambrosias-
ter, Bede, Origen, and Paterius have now been identified for book 10.
Together, the 22 books of De rerum naturis are approximately
500,000 words.24 Book 10 has 9,639 words and it, therefore, makes up
approximately 2% of the encyclopedia. The unidentified passages in
book 10 equal 1,219 words. However, this count includes 385 words of
direct Biblical quotations. Therefore, the actual unidentified word count
is 834, making Hrabanus’ contribution to book 10 equal to c. 8.65% of
the content.

20
Heyse (1969); see an analysis of her work in Schipper (2004), 8–10.
21
Schipper (2004), 8.
22
For some reason, Heyse did not identify passages from chapter 4 of the Clavis
Melitonis (items 5, 25, 42–43, 67–68; Pitra (1879–91), ii 17–21) as being sources for
book 10. Perhaps this is because some of these were not copied verbatim but only para-
phrased, whereas the items from the Clavis Melitonis identified by Heyse were by and
large copied verbatim. But the items listed above are easily recognizable as the source
passages, with perhaps a few words changed or the verb tense altered but with always the
same biblical quotation used. These missed attributions could also be the result of typos
and unintentional copying mistakes in Heyse’s dissertation. This occurs, for instance,
with the excerpt Dies autem dominum significat in illa sententia evangelii: qui ambulat
in die non offendit, which Heyse identified as Clavis Melitonis 4.1 (Pitra (1879–91), ii
17) when it is actually 4.43 (Pitra (1879–91), ii 17). Another example of a mistake is
the identification of the source for chapters 16 and 17 as Etymologiae 16.1 instead of the
correct 16.18.
23
See Schipper’s works listed in the bibliography, in particular Schipper (2004);
Schipper (2008).
24
Schipper (2007), 103.


 Text Author Number of Chapters Number Percentage
Excerpts Containing of Words of book 10
Excerpts
De temporum ratione (CCSL 123B) Bede 11 1, 2, 3, 4, 6, 3820 39.63%
9, 10, 12, 14
Etymologiae (Lindsay (1911)) Isidore 16 1, 2, 3, 5, 7, 1414 14.67%
8, 11, 12, 13,
15, 16, 17
Clavis Melitonis (Pitra (1852–58), ii) Anonymous 27 1, 3, 4, 5, 6, 1278 13.26%
7, 8, 10, 11,
12
Expositio in psalmos (CCSL 97–98) Cassiodorus 22 1, 4, 5, 6, 7, 1226 12.72%
8, 11, 12, 13
Moralia in Job (CCSL 143) Gregory 3 6, 8, 10 163 1.69%


In sancti Marci evangelium expositio (CCSL 120) Bede 2 7, 11 125 1.30%
Liber de expositione veteris ac novi testamenti (PL 79) Paterius 2 8 88 0.91%
Commentarium in Isaiam prophetam (PL 24) Origen 2 10 70 0.73%
In sancti Iohannis evangelium expositio (PL 92) Bede 1 8 67 0.69%
WHAT A DIFFERENCE A DAY MAKES

Super epistolas catholicas expositio in primam epistolam s. Joannis Bede 1 3 54 0.56%

(CCSL 121)
In epistolam ad Romanos (CSEL 81) Ambrosiaster 1 3 40 0.41%
De divisionibus temporum liber (PL 90) ps-Bede 1 3 27 0.28%
Commentaria in epistolam B. Pauli ad Romanos (Caillau and Origen 1 4 26 0.27%
Guillon (1829))
Commentarii in epistolas sancti Pauli in primam epistolam ad Origen 1 2 22 0.23%
Corinthios (PL 30)

Table 1  Sources for book 10 of Hrabanus Maurus’ De rerum naturis by text.

 LISA CHEN OBRIST

The end of the last chapter of book 10 and its authorship

This, of course, could mean that these unidentified passages are the origi-
nal words of Hrabanus and this will factor into my analysis of the fol-
lowing excerpt that appears in some manuscripts at the end of the last
chapter of book 10, which discusses the Lord’s Day.  Of the 35 extant
manuscripts that contain book 10, I was able to access 30, either in situ
or through facsimiles. I was, therefore, able to verify the inclusion of this
addition in these manuscripts and found it present in 18 manuscripts,
that is, about half of the extant complete copies.25 These are the 393
words added to the end of chapter 17:26

Ipse est enim primus qui post septem reperitur octavus. Unde in ecclesiasten
ad duum testamentorum significationem datur illi septem et illi octo.27
Dominicum diem apostoli ideo religiosa sollempnitate sancxerunt quia
in eodem redemptor noster a mortuis resurrexit. Quique ideo dominicus
appellatur, ut in eo, a terrenis operibus vel mundi illecebris abstinentes
tantum divinis cultibus serviamus dantes scilicet diei huic honorem et re­
verentiam propter spem resurrectionis nostrae quam habemus in illo. Nam
sicut ipse dominus noster Iesus Christus redemptor et salvator noster tertia
die resurrexit a mortuis ita et nos resurrecturos in novissimo die speramus.
Unde etiam in dominico die stantes oramus quod est signum futurae re­
surrectionis. Hoc agit universa ecclesia quae in peregrinatione mortalitatis
inventa est expectans finem seculi quod in domini nostri Jesu Christi cor­
pore praemonstratum est qui est primogenitus a mortuis.28 Sabbatum au­
tem a priori populo in initio celebratum legimus ut figura esset in requiem.
Unde et sabbatum requies interpretatur. Apparet autem hunc diem etiam
in scripturis sanctis esse sollepnem. Ipse est enim dies primus seculi in ipso
formata sunt elementa mundi in ipso creati sunt angeli in ipso quoque a
mortuis resurrexit Christus in ipso de caelis super apostolos spiritus sanctus
descendit. In ipso credimus universam et inmensam multitudinem totius
mundi resurrecturam ad ipsum terribile dei iudicium quod caput est oc­
tavi et ab illa die octavum caelebrare sine cessatione sive iustos cum angelis
dei in perpetua exultatione sive reprobos cum diabolo et satelitibus eius in
perpetua dampnatione. Namque quod creavit angelos et a mortuis resur­

25
These 393 words appear at the end of chapter 17 after celebrari coepit in MSS Cc
Ck Cu E F G L Ob Oc Og On Or Pb Pv Ro Rs Rt Wo (for the abbreviations, see the Ap-
pendix). They are not included in the editio princeps by Rusch, nor in the 1627 edition
by Colvener, nor in PL 111, and nor in Heyse’s 1969 study.
26
Trans. by Thomas Knoebel and myself.
27
Ecclesiastes 11:2.
28
Colossians 1:18.


 WHAT A DIFFERENCE A DAY MAKES

rexit et spiritum sanctum super apostolos ad nostram corroborationem et

confirmationem misit. Nobis valet nil credere nisi fidem nostram bonis
perornemus et sanctis virtutibus carnisque in vitam aeternam credamus
resurrectionem. Nam nobis hoc apostolicum contestatur eloquium dicendo:
Fides sine operibus mortua est.29 Et dominus in aevangelio: Qui non cre-
diderit resurrectionem non habebit vitam aeternam.30 Manna in heremo
in eodem die de caelo data est primo. Sic enim dicit dominus: Sex diebus
collegetis manna in die autem sexto dupplum colligetis.31 Sexta enim
dies est parasceve quae ante sabbatum ponitur. Sabbatum autem septimus
dies quem sequitur dominicus in quo primum manna de caelo venit. Unde
intellegant iudei iam tunc praelatam esse iudaico sabbato dominicam nos­
tram. Iam tunc indicatum quod in sabbato ipsorum gratiam dei de caelo
ad eos nulla descenderit sed in dominica nostra in qua primum dominus
de caelo eam pluit.

‘For this day, the first to be found after the seven days, is the eighth.
Hence, even in Ecclesiastes its significance for the two Testaments
is stated: Divide your means seven ways for the one, or even eighth. The
apostles sanctified the Lord’s Day by religious solemnity, because on that
day our redeemer rose from the dead. Thus this day is called the Lord’s
so that on it, abstaining from earthly works and the allurements of the
world, we might serve him only in divine worship, giving honour and
reverence on this day, certainly, for the hope of our resurrection, which
we have in him. For just as our Lord, Saviour, and Redeemer Jesus Christ
himself rose from the dead on the third day, so also we hope that we
shall be raised on the last day. Hence also on the Lord’s Day we pray
standing because this is a sign of the future resurrection. The universal
church which is found in the pilgrimage of mortality does this, looking
forward at the end of time to what was first indicated in the body of our
Lord Jesus Christ, who is the first-born from the dead. We read that the
Sabbath was celebrated by the first people in the beginning so that their
bodily form might be at rest. Thus the Sabbath is interpreted as rest.
This day appears as solemn even in Sacred Scriptures. For this is the first
day of the period and on this day the elements of the world were formed,
on this day the angels were created, on this day also Christ rose from
the dead, on this day the Holy Spirit descended from heaven upon the
apostles. On this day, which is the beginning of the eighth period, we
believe that the universe in the multitude measure of the whole world
will be resurrected in the terrible judgment of God and from this day

29
James 2:26.
30
John 3:15–16.
31
Exodus 16:26–29.


 LISA CHEN OBRIST

on the octave, either the just celebrate without cessation with the angels
of God in perpetual exultation or the rejected celebrate with the devil
and his minions in perpetual damnation. For he created angels and rose
from the dead and he sent the holy spirit upon the apostles for our cor-
roboration and confirmation. Nothing is able to be believed by us unless
we greatly adorn our faith with good and holy virtues and believe in
eternal life and the resurrection of flesh. For it is witnessed by us in the
apostolic speech as he says: Faith without works is dead. Also the Lord in
the gospel says: He who will not believe in the resurrection will not have
eternal life. On this day in the desert manna from heaven was first given.
For thus says the Lord: Six days you shall gather it, on the sixth day he
gives you food for two days. Thus, the sixth day is the day of preparation
that is placed before the Sabbath. The Sabbath is the seventh day, which
is followed by the Day of the Lord on which the manna from heaven
first came. Consequently, the Jews may now understand that our Lord’s
Day is superior to the Jewish Sabbath. It is now clear that the grace of
God never descended on them from heaven on their Sabbath but rather
on our Lord’s Day, on which the Lord first rained down the manna.’

The excerpt starts off with an explanation of the eighth day of the week.32
The explanation, in this case, turns out to be quite simple. The eighth day
of the week is just the first day of the week.33 The first and the eighth day
of the week also happen to be the Lord’s Day, a day to commemorate the
significance of the life and death of Jesus. This system of counting the
weekdays is not unique, and in fact the opening of book 10 describes
another instance of inclusively counting the days of the week. That is:
32
For details about the history of the week and the different types (astrological, lu-
nar, civic 7, 8, 9, 10, and decan) and religious versions Jewish, Babylonians / Chaldeans,
Assyrian, Egyptian, Christian, etc., see: Richards (1998), 208–09, 266–75; Zerubavel
(1985), 6–22, 45–46, 63–65; Blackburn and Holford-Strevens (1999), 566–82.
33
As Wallis (1999), 278, notes, Bede also elaborated a theme found in Irish computi,
and ultimately derived from a hint in Isidore’s De natura rerum 3 (ed. by Fontaine (1960),
185), about the notion that the ‘week’ is not a univocal category; rather there are eight dif-
ferent kinds of weeks: ‘The message is that whereas human authority restricts the meaning
of time, the divine authority of Scripture expands and diversifies that meaning.’ The ninth
chapter in book 10 of Hrabanus’ De rerum naturis, the longest chapter at 1,661 words, is
composed of four excerpts from Bede’s De temporum ratione (from chapters 8–10; CCSL
123B, 299–312) and contains no original commentary from Hrabanus. It discusses the
eight types of scriptural weeks: the divine week of Creation; the conventional seven-day
week; Pentecost or the ‘feast of weeks’, i.e. the 50 days after Easter (7×7+1 days); the Jew-
ish seventh month (feast of the Atonement); the Sabbath of the land in the seventh year;
the Jubilee, the annual analogy of Pentecost (7×7+1 years); the ‘prophetic week’ of the
Old Testament; and the week of the World-Ages. Isidore also refers to the eighth day
of the week in Etymologiae 5.22 (Lindsay (1911), i), in which he underlines the cyclical
character of the week in which the eighth day is identified as the first.


 WHAT A DIFFERENCE A DAY MAKES

there are seven complete days in the span of a week but for the weeks
to exist in a cycle they must repeat themselves and thus the eighth day
of the week must exist and be at one and the same time the first and the
last day of the cycle. Without the concept of the eighth day, the week
would come to end and cease to be part of a cyclical sequence. This is just
another way of conceptualizing the seven-day week, which is still used
in many cultures today. The example given in chapter one of book 10
involves the Roman secular market week cycle known as the nundinae.34
The term literally means nine days and marks the repetition of market
events every nine days. That is: this cycle comprised eight days between
market days, with the market day, the first day of the cycle, also being the
ninth day.
The example of the nundinae is given within an explanation of the
different ways of conceptualizing time measurement. It is the ever-
present dichotomy of God and man, heaven and earth, of the human
and the divine, which is reflected in the organization of the 22 books
of the complete text. It is the problem of how to reconcile observations
of nature, what they see in the world around them, with what they read
about in the texts of divine authority. The nundinae is here given by
Hrabanus as an example of an entirely human and mundane construct
of time measurement. The eighth day of the week is also referred to in
chapter four in two sentences which Hrabanus borrows from the Clavis
Melitonis.35 This excerpt discusses the same instance of the Lord’s Day
being the first and the eighth day of the week as the additional passage at
the end of the last chapter.
Let me now return to his last chapter, chapter 17, of book 10.36 It
opens with Isidore’s passage on the different meanings of the Lord’s Day
for the Christians and the Jews in his Etymologiae, and this excerpt does
include mention in the last sentence of the Lord’s Day as the eighth
day of the week.37 The rest of the chapter continues on the topic of the
Lord’s Day being the eighth day and goes on to talk about why the Lord’s
34
Hrabanus Maurus, De rerum naturis 10.1 (PL 111, 285).
35
Clavis Melitonis 4 (Pitra (1879–91), ii 19): Dies octava resurrectionis dominicae
obtinet sacramentum. Item dies octava futuram omnium resurrectionem et diem iudicii,
cuius mysterio sexti psalmi titulatio praesignatur, hoc est pro octava.
36
Hrabanus Maurus, De rerum naturis 10.17 (PL 111, 308–09).
37
Isidore, Etymologiae 6.18.20–21 (Lindsay (1911), i): De dominico die: Domini­
cus dies proinde vocatur, quia in eo resurrectio domini nostri declarata est. Qui dies non
Iudaeis sed Christianis in resurrectione domini declaratus est, et ex illo habere coepit festivi­
tatem suam. Illis enim solum celebrandum sabbatum traditum est, quia erat ante requies
mortuorum; resurrectio autem nullius erat qui resurgens a mortuis non moriretur. Post­


 LISA CHEN OBRIST

Day is celebrated, what will happen on the last day when all believers will
be resurrected like Jesus, and the difference between the Christian and
Jewish understandings of eternal salvation. Therefore, the additional
text is a continuation and expansion of the discussion, resulting in the
extension of chapter 17 from about 80 words (from Isidore) to almost
500 words. This additional text not only concludes the chapter but also
concludes the entire book.

Source analysis
I recently discovered that more than half of this mystery passage is, not
surprisingly, taken from another work by Isidore of Seville, his De ec­
clesiasticis officiis (On the Ecclesiastical Offices). The text is taken from
book 1, chapters 24 (The Lord’s Day) and 25 (The Sabbath) and not in
sequential order.38 In fact, there are five separate excerpts from different
sections of chapters 24 and 25 and they are, in a way, spliced together.
Sources for the addition to Chapter 17 De dominico die are as follows:

Ipse est enim […] et illi octo: (22 words) Isidore, De ecclessiasticis officiis
(DEO) 1.25
Dominicum diem apostoli […] primogenitus a mortuis: (111 words) DEO
1.24
Sabbatum autem a priori  […] requies interpretatur: (19 words) DEO
1.25
Apparet autem hunc […] sanctus descendit: (43 words) DEO 1.25
In ipso credimus universam  […] habebit vitam aeternam: (106 words)
unidentified
Manna in heremo […] caelo eam pluit: (86 words) DEO 1.25

Isidore’s De ecclesiasticis officiis is not identified by Heyse or Schipper as a

source found in any of the 22 books of De rerum naturis. However, work
is still ongoing with the complete edition and this source will now have
to be kept in mind as a possibility for other, up until now, unidentified
passages. There is precedent for Hrabanus’ use of Isidore’s De ecclesiasticis
officiis before he set about writing his encyclopedia. In AD  819, Hra-
banus used Isidore’s text as a source for his De institutione clericorum. In
fact, in book 2, chapter 42 (De die dominica), Hrabanus used the exact

quam autem facta est talis resurrectio in corpore domini, ut preiret in capite quod corpus
ecclesiae speraret in fine, iam dies dominicus, id est octavus, qui et primus, celebrari coepit.
38
Isidore, De ecclesiasticis officiis 1.24–25 (ed. by Christopher M. Lawson in CCSL
113, 27–29).


 WHAT A DIFFERENCE A DAY MAKES

same excerpts that appear in the additional passage.39 Therefore, given

this precedent, it is possible that Hrabanus was responsible for adding
these excerpts to the end of book 10, chapter 17. As for the unidenti-
fied passage of 106 words that appears in the middle of this addition,
I am still in the process of searching other source texts, on the one hand
hoping to discover where else Hrabanus was mining his information,
but on the other hand wanting to restore and confirm the originality of
the words and give Hrabanus his due, which in much of the secondary
literature he has not received.40 In this case, I have no help from Heyse
to indicate the direction I might look in other than combing through
the works of the ‘usual suspects’ such as Isidore, Bede, Cassiodorus, and
Hrabanus’ own corpus. She based her source study on the text in the Pa­
trologia Latina, which goes back to the first edition by Rusch and which
does not include this additional text.

Manuscript evidence
However, there is always a spanner that gets thrown into the works and
here I must go back to the manuscript evidence. Of the 18 manuscripts
that I have so far confirmed to contain the additional text, none date
from the ninth century. In fact, the earliest manuscripts date from the
twelfth century, far from Hrabanus’ lifetime.41 I have used the adjective
‘additional’ a few times to describe the passage at the end of chapter 17 as
I believe we must entertain the possibility that it was added by someone
else and perhaps at a later date.42 It is not surprising that its first appear-

39
Hrabanus Maurus, De institutione clericorum 2.42 (ed. by Zimpel (1996), 397–
99); Isidore, De ecclesiasticis officiis 1.24–25 (CCSL 113, 27–29).
40
The De rerum naturis is not a mutilated, untidy mess of Isidore’s encyclopedia
as Collison (1964), 36, suggests, nor is it a ‘shameless plagiarism’ or poor imitation of
Isidore (de Ghellinck (1939), 103). In Collison’s esteem (Collison (2010), 272): ‘the
development of the encyclopedia during the next 500 years [i.e. after Isidore’s Etymolo­
giae], though of social interest, was undistinguished from the point of view of scholar-
ship. Rabanus Maurus (c. 776–856), one of the English scholar Alcuin’s favorite pupils,
compiled De Universo (‘On the Universe’), which, in spite of its being an unintelligent
plagiarism of St Isidore’s work, had a lasting popularity and influence throughout the
medieval period.’
41
Six of the 18 manuscripts date from the twelfth century: G Oc Or Pv Ro Rs (for
the abbreviations, see the Appendix).
42
There are two other additions to book 10 found in later manuscripts. In chap-
ter 1 (De temporibus; PL 111, 285–86), 54 words are included in the text after praecepit
in MSS Ck Cu E Ob On Or Og Ro Wo. The first 16 lines of chapter 1 are taken entirely
from Bede’s De temporum ratione 2 (De trimoda temporum ratione; CCSL 123, 274–75).
The text of these 16 lines is, in fact, two excerpts from De temporum ratione 2: the first


 LISA CHEN OBRIST

ance is in a twelfth-century manuscript during a time when Europe saw

a new wave in encyclopedism.43 From the number of surviving witnesses
and the early date of the first edition we can surmise that it was a popu-
lar and valued work. William Schipper has also shown that manuscript
copies from the twelfth century and later, particularly from England, are
highly annotated and obviously well-used.44

Conclusion
These possible explanations lead to questions regarding the status of
Hrabanus’ De rerum naturis. How much and in what way was his work
valued? Did later editors think they were fixing a mistake by completing
what they thought was an unfinished chapter?45 Could the additional
text be past marginal notes which were then inserted into the primary
text during the copying process before or during the twelfth century?

106 words of Bede’s chapter (Tempora igitur […] in lege praecepit) and the last 71 words
from Bede’s chapter (Porro natura duce […] et dies et annos). The middle part of De tem­
porum ratione 2, missing from Hrabanus’ chapter, is, in fact, the passage of 54 words
which appear in the text of the later manuscripts of book 10. In chapter 15 (De festivitati­
bus; PL 111, 307–08), 58 words are included in the text after sollemnitatis vestrae in MSS
C N St Va Vc (all the extant illustrated manuscripts). This additional passage discusses
the important dates of Jesus’ lifespan and are paraphrases of texts from Pseudo-Bede’s De
argumentis lunae (De solstitiis et aequinoctiis; PL 90, 724) and from Pseudo-Dionysius’
Argumentum XV (ed. by Krusch (1938), 79–80).
43
There are several significant aspects of western Europe in the twelfth century
which contributed to the flowering of a new encyclopedic movement. First, there was
the rediscovery of ancient medical, mathematical, and philosophical texts which arrived
via the Arabic world. Chief among these rediscovered texts was a corpus of Aristotelian
writings. Second, the rise of universities led to a new academic and intellectual culture
which eventually expressed itself in the thirteenth century through the humanist and
scholastic schools of thought. Third was the establishment of vast complete libraries in
universities, cathedrals, monasteries, and royal and civil courts. Lastly, new textual me-
diators, the academics, the intellectual royals, and the preachers from the new religious
orders, created new demands for encyclopedias with different content, formats, and
purposes. See Le Goff (1992), 37. For more details on twelfth- and thirteenth-century
encyclopedic texts and late medieval encyclopedism, see the works by Burrows (1987),
30–37; Collison (1964), 44–68; Grant (1999), 103–04; Kaske (1988), 41–49; Twomey
(1988), 186–99, 202–15.
44
Schipper (1997a), 21.
45
Chapters 15, 16, and 17 are very short and only include an excerpt from Isidore’s
Etymologiae similar to the opening structure of other chapters in the majority of the
22 books of De rerum naturis. It is very feasible that these chapters were unfinished and
that later editors strove to complete them (although I have so far found no evidence of
any additional passages associated with chapters 15 or 16).


 WHAT A DIFFERENCE A DAY MAKES

Was it considered a stable text? Or was it fair game for additions, in

some cases reflecting Hrabanus’ own method of compilation, organiza-
tion and analysis? Did a later scholar perhaps believe he was doing just
the same thing in imitating Hrabanus’ method? Our modern notions of
‘authorial originality’ and ‘plagiarism’ give this a negative spin. But was
it really such a bad thing to enhance the work, not because it was seen as
lacking or unworthy of reverence, but because it was seen as an impor-
tant and valued resource, worthy of attention and enhancements?46
So, it turns out that the eighth day of the week does not make that
much difference in the classical and medieval understanding of hebdom-
inal time measurement. It was not so unusual and foreign to readers of
Hrabanus’ work as it seemed to me when I first came across it. It is sim-
ply a different way of conceptualizing and counting the days of the week.
Different systems were successfully reconciled so that both theoretical
and literal textual explanations could co-exist with everyday experiential
reality. There is still, however, the question about the status of this addi-
tional 393-word text and what it means to our understanding of the sta-
tus of the entirety of De rerum naturis as well as who was responsible for
adding excerpts from another work by Isidore. It is precisely because of
these questions that this addition to the end of book 10 about the eighth
day of the week does make a significant difference to our understanding
of the reception and influence of Hrabanus’ encyclopedia.

Appendix: List of manuscripts containing book 10

of ­Hrabanus’ De natura rerum
A Karlsruhe, Badische Landesbibliothek, Aug. perg. 96 (Reichenau, saec.
ix/x)
An Angers, Bibliothèque municipale, 31 (olim 27) (La Baumette, saec. xv)
Ar Arras, Bibliothèque municipale, 506 (832) (St Vaast, saec. xi)
C Montecassino, Archivio dell’Abbazia, 132 EE (Montecassino, AD 1023)
Cc Cambridge, Corpus Christi College, 11 (Canterbury?, saec. xii/xiii)
Ck Cambridge, University Library, Dd 13.4 (Kirkstead, saec. xiii)
Cu Cambridge, University Library, Dd 1.30 (origin unknown, saec. xiii)
D Darmstadt, Hessische Landes- und Hochschulbibliothek, 416 (Heidel-
berg, 1467)
E El Escorial, Real Biblioteca, f. I.12 (origin unknown, saec. xiv)

Hathaway (1989).
46


 LISA CHEN OBRIST

F Florence, Biblioteca Medicea Laurenziana, Fondo S. Croce, Plut. XXXI

Sin. Cod. I (origin unknown, saec. xiv)
G Glasgow, University Library, Hunterian V.1.3 (366) (Chichester, saec. xiiex)
L Leiden, Universiteitsbibliotheek, VLF 5 (probably Italian, saec. xiv)
M Molfetta, Biblioteca del Seminario, 5.7.V (3) (origin unknown, saec. xiv)
Ma Madrid, Biblioteca Nacional, 12370 (origin unknown, saec. xiii)
N Naples, Biblioteca Nazionale, Borbon. V.C.46 (origin unknown, saec. xii)
Ob Oxford, Bodleian Library, Laud misc. 746 (Oxford, saec. xiii)
Oc Oxford, St. John’s College, 88 (Chichester, saec. xii)
Og Oxford, Trinity College, 64 (Gloucester, saec. xiii)
On Oxford, New College, 159 (origin unknown, saec. xiv)
Or Oxford, St. John’s College, 5 (Reading, saec. xiiex)
Pb Paris, Bibliothèque nationale de France, Lat. 2420 (origin unknown, saec.
xiv)
Pc Paris, Bibliothèque nationale de France, Lat. 7608 (origin unknown, saec.
xiii)
Pg Paris, Bibliothèque nationale de France, Lat. 11684 (Saint-Germain-des-
Prés, saec. xii)
Pr Paris, Bibliothèque nationale de France, Lat. 13411 (Saint-Germain-des-
Prés, saec. xii)
Pv Paris, Bibliothèque nationale de France, Lat. 16879 (Clairvaux, saec. xii)
Ro London, British Library, Royal 12.G.xiv (St Albans, saec. xii)
Rs Rheims, Bibliothèque municipale, 441 (origin unknown, saec. xii)
Rt Rheims, Bibliothèque municipale, 442 (St Denis, saec. xv)
St Stuttgart, Württembergische Landesbibliothek, Theol. et philos. 2ø45
(perhaps Konstanz, AD 1457)
V Venice, Biblioteca nazionale di S.  Marco, Lat. II 56 (origin unknown,
saec. xiv)
Va Vatican, Biblioteca Apostolica, Pal. lat. 291 (central or southern Germa-
ny, AD 1425)
Vc Vatican, Biblioteca Apostolica, Reg. lat. 391 (probably Italian, saec. xiv/
xv)
Wo Worcester, Cathedral Library, F.21 (origin unknown, saec. xii/xiii)
Z Zürich, Zentralbibliothek, Car. C 97 (263) (origin unknown, saec. xv)


 DÁIBHÍ Ó CRÓINÍN

ARCHBISHOP JAMES USSHER

(1581–1656)
AND THE HISTORY OF THE EASTER
CONTROVERSY

Abstract
Archbishop James Ussher is probably best known for his reckoning of the date
of the creation of the world (at the beginning of the night preceeding 23 Octo-
ber 4004 BC). However, his calculations were all based on a meticulous study of
the Old Testament and other early Christian and non-Christian chronographi-
cal writings. This paper announces the discovery of a previously-unnoticed
Oxford manuscript that lists the impressive array of patristic and post-patristic
writings on the subject of the early Easter controversy that he accumulated for
his researches.
Keywords
Annales veteris testamenti, Annales novi testamenti, Annals of the World, Cum-
mian’s Letter; Columba, Columbanus, Gregory the Great, Jacques Sirmond,
Jerome, James Ussher; computus, Easter controversy.

Introduction
James Ussher—described by a contemporary as ‘A Miracle of Learning’1—
was born in Dublin on 4 January 1581.2 A nephew of Richard Stanihurst
(1547–1618), author of a history of the reign of Henry VIII of England

1
The words of William Selden, as quoted by Ussher’s biographer, Bernard (1656),
8–9: Vir […] ad miraculum doctus. I am grateful to Katharine Simms and Bernard Mee-
han for confirming the source.
2
See Ford (2007), which supersedes the earlier biography by Buick-Knox (1967).
I have not seen Cunningham (2007).

Late Antique Calendrical Thought and its Reception in the Early Middle Ages, ed.  by
Immo Warntjes and Dáibhí Ó Cróinín, Turnhout, Brepols, 2017 (Studia Traditionis
Theologiae, 26), pp. 309-351
© BREPOLSHPUBLISHERSDOI 10.1484/M.STT-EB.5.114742
 DÁIBHÍ Ó CRÓINÍN

and a controversial ‘Description of Ireland’ (De rebus in ­Hibernia gestis),3

he was one of the first students in Trinity College Dublin (founded in
1591, opened in 1592), which he entered in 1594. That he evinced an
early interest in Bible studies and in biblical chronology is clear from a
letter (dated before 1600, the year in which Ussher received his MA) that
he wrote to Stanihurst while still a student, in which he stated:4

‘The principal part of my study at this time is employed in perusing the

writings of the fathers, and observing out of them the doctrine of the
ancient Church; wherein I find it very necessary that the reader should
be thoroughly informed touching his authors, what time they lived, and
what works are truly, and what falsely attributed to them.’

The letter shows clearly that ‘even as a young man he had a very high-
ly developed interest in the past and a very critical approach to evidence
dealing with it’.5 That early promise was to be fulfilled in a series of pub-
lications that came from his pen following his appointment as Professor
of Theological Controversies in Trinity College in 1607. Several of these
represent milestones also in the evolution of technical chronology and
in the history of the Easter controversy.
One of the earliest of Ussher’s publications already demonstrates his
particular preoccupation with matters to do with the paschal contro-
versy in the early and medieval churches. His Discourse of the Religion
professed by the Ancient Irish (1631) devoted two whole chapters to
‘the controversy which the Britons, Picts and Irish, maintained against
the Church of Rome, touching the celebration of Easter’.6 Kenney re-
marked that ‘Ussher could be charmingly frank in regard to his histori-
cal methods’,7 as the following quote illustrates:8

3
Barry and Morgan (2013). The original Latin text was published by Plantin at
Leiden in 1584. On Stanihurst, see Lennon (1981).
4
It is the very first of Ussher’s letters published in Elrington (1847), xv 3.
5
Mc Carthy (1997), 73.
6
Ussher (1631), 90–114 (chapters 9 and 10), the quote at 90–91. The title of the
work is often given (e.g., in Kenney (1929), 47 n. 181) as A Discourse of the Religion ancient­
ly professed by the Irish and British; my own copy of the book has the title given above. Ken-
ney (1929), 47 n. 181, points out that an earlier work of Ussher’s appeared (in partial form)
in 1609: Of the Original and First Institution of Corbes, Herenaches, and Termon Lands,
but that this was first published in full only in Vallancey (1770), 175–203. I am grateful to
Dr Siobhán Fitzpatrick (Royal Irish Academy) for checking the Vallancey reference for me.
7
Kenney (1929), 47 n. 181.
8
A note ‘To the Reader’ appended to the Discourse, in Elrington (1847), iv 376–77.


 ARCHBISHOP JAMES USSHER (1581–1656)

‘Although my principal intention in this discourse was to produce such

evidences as might shew the agreement that was betwixt our ancestors
and us [i.e., those of his own persuasion] in matters of religion, and to
leave the instances which might be alleged for the contrary to them unto
whom the maintaining of that part did properly belong; yet, I have upon
occasion touched upon that part also, and brought to light some things
which I met withal in such hidden antiquities, as in all likelihood would
not have come unto their notice without my discovery [He then refers
to the manuscripts in the collection of Sir Robert Cotton, which are
indicated in his list of sources, in order that others may consult them]
my intention herein being to deal fairly, and not to desire the concealing
of any thing that may tend to the true discovery of the state of former
times, whether it may seem to make for me or against me.’

The ‘public’ Ussher, however, was known for his fervent and outspoken
anti-Catholic rhetoric. On 8 September 1622, for instance, in the pres-
ence of the then new Lord Deputy (Falkland), he preached a notori-
ous sermon in Dublin in which he demanded severe repressive measures
against the papists,9 while in 1626–27 he strongly opposed the grant-
ing of toleration to Irish Catholics. There is a striking contrast between
these public statements of Ussher’s and his actions in the field of scholar-
ship, where he maintained friendly relations with contemporary Irish
Franciscan and Jesuit scholars, such as Luke Wadding and Brendan O
Conor, and Stephen White, secular clergy such as David Rothe (Roman
Catholic bishop of Ossory),10 and laymen, such as Conall Mageoghan,
‘for whose religious ideals he felt the greatest repugnance’.11 That this
respect was mutual is borne out by remarks in the Commentarius Ri­
nucinnianus (a report on the activities in Ireland of the papal emissary,
GianBattista Rinuccini during the years  1645–48, compiled between
1661 and 1666, by staunchly Catholic writers) to the effect that there
was hardly anything in Ussher’s Antiquitates (1639) that stood outside
the Catholic faith, save the reference to King Charles as Defensor fidei.12
9
Elrington (1847), i 58–60. Ussher subsequently maintained, however, that some
of the remarks in that sermon had been exaggerated; see his letter of 16 October 1622,
in Elrington (1847), xv 180–82; see now Boran (2015), i 254–55. I owe this observation
to Prof. John McCafferty, University College Dublin.
10
See O’Sullivan (1994–95), 7–49; Grosjean (1959), 158.
11
Kenney (1929), 48.
12
Commentarius Rinucinnianus (ed.  by Kavanagh, O’Ferrall, and O’Connell
(1932–49), i 242–44): haud pauca sunt quae jam animum iam Catholicum valde redo­
lent, nihilque (quod sciam) ibi habetur haeresim aperte spirans praeter solum ‘Fidei Defen­
soris’ titulum. I owe the reference to Leofranc Holford-Strevens.


 DÁIBHÍ Ó CRÓINÍN

The tone of Ussher’s Discourse is polemical but not unfair,13 and al-
ready attests to his wide familiarity with sources relating to the early his-
tory of the churches in these islands. His reference to evidence that he
had found in manuscripts belonging to Sir Robert Cotton very likely
relates to the next of his publications, best-known to historians of the
Easter controversy as the editio princeps of the famous paschal letter of
Cummian, which Ussher published from the only known copy.14 Ken-
ney described the letter as ‘the only important controversial document
written in Ireland regarding the paschal question that we still possess’,
but reported the manuscript as having been lost in the disastrous 1731
fire that destroyed a significant part of the Cotton Library.15 However,
the indefatigable Italian scholar, Mario Esposito, pointed out that the
manuscript was not, in fact, lost—though it was singed around the edges
in the fire (with some slight loss of text)—and he published some read-
ings from it. Esposito remarked that ‘Ussher appears to have made a care-
ful copy of the MS, and his edition may be relied upon’.16
The 1988 edition of Cummian’s letter by Maura Walsh and Dáibhí
Ó Cróinín revealed the full extent of both the biblical and the patristic
sources drawn on by its author;17 perusal of Ussher’s editio princeps dem-
onstrated his profound familiarity with many of those sources already,
bearing out what he had said to his uncle in that letter of his student
years. It is equally clear, however, that Ussher (in 1631–32, at any rate)
was not (yet) equally familiar with the computistical literature from the
early centuries of the church. On the other hand, by the time he came
to publish his massive Britannicarum ecclesiarum antiquitates,18 which
showed how thoroughly learned and well-read he had become on the
subject of the Easter controversy in the Irish and British churches in the
sixth and seventh centuries, Ussher did draw explicitly on the older pa-

13
Ussher (1613) opens with the words Romanam bestiam!
14
Cummian’s letter De controversia paschali was first edited by Ussher (1632), 24–
35, from the sole manuscript London, British Library, Cotton Vitellius A XII, fols 79r–
83r. See now the critical edition and translation in Walsh and Ó Cróinín (1988); see also
Walsh (2013).
15
Kenney (1929), 220–21 n. 57: ‘Taken by Ussher from a MS of the Cotton col-
lection, now lost’. Kenney described the text, for some reason, as ‘evidently much cor-
rupted’.
16
Esposito (1930), 240–45; citation at 242, a verdict endorsed in the modern re-
edition, Walsh and Ó Cróinín (1988), 53.
17
Walsh and Ó Cróinín (1988), 225–26, list all the non-biblical sources used, over
forty in total.
18
Ussher (1639).


 ARCHBISHOP JAMES USSHER (1581–1656)

tristic and post-patristic literature on the paschal debates of earlier cen-

turies.19 And when it comes to his great two-volume compendium of
biblical chronology, the Annales veteris testamenti and the Annales novi
testamenti, it is clear that his researches into the subject had advanced by
leaps and bounds.20
There is evidence of this already in a letter (dated 12 Nov. 1639), ad-
dressed by him to one of his many continental scholarly contacts, Chris-
tian Rave (Christianus Ravius, then based in Constantinople). Ussher
sent Ravius a ‘shopping-list’, as follows:21

Libri quos pro me requiri velim, hi sunt: Vetus Testamentum Syriacum,

non ex Hebraeo factum (illud enim jam habeo) sed ex Graeco versum, atque
obelis et asteriscis distinctum. Polycarpi et Ignatii Epistolae Syriace conver­
sae. Eusebii (non Historia Ecclesiastica quae passim prostat) Chronicum
Graecum, vel etiam Syriace versum. Si quid etiam versionum Symmachi,
Aquilae et Theodotionis reperiri possit. Julii Africani Chronicon. Hegessip­
pi Historia Ecclesiastica. Clementis Alexandrini Hypotyposeôn libri, et de
Paschate libellus Anatolii. Aniani et Panodori computi Paschales. Georgii
Syncelli Graecum Chronicon. Apollodori Graecum Chronicon. Phlegon de
Olympiadibus. Diodori Siculi, Polybii, Dionysii Halicarnassei, Dionis Coc­
ceiani libri illi, qui apud nos desiderantur. Hipparchi Astronomica, Graecè.
Hos libros omnes sollicitè vestiges velim quaqua transibis, et si quos repe­
rias, diligenter in adversariis notes locos ubi extant, et nomina eiorum in
quorum manibus sunt, itidemque pretium quo eos divendere velint, ut
et nomina nostratium mercatorum in iisdem locis commorantium, ut sic
postea, quando ad nos reversus fueris, accersere eos, si pretium placuerit,
possimus. Quod si et alii probi authores Graeci aut Syriaci, praeter supra­
dictos, in manus tuas inciderint, qui apud nos desiderantur, poteris et circa
illos eâdem uti diligentiâ.

‘The books that I would like you to seek out for me are these: a Syriac
Old Testament, not translated from the Hebrew (I have that already),
but from the Greek, marked with asterisks and obelisks; the letters of
Polycarp and Ignatius translated into Syriac; the Greek Chronicle of

19
This is the case also with his Discourse, in those two chapters (9 and 10; Ussher
(1631), 90–110) where he discusses the insular Easter controversy. Ussher cites Bede’s
Historia ecclesiastica at length, and various papal letters quoted by him, but no earlier
writers on the paschal question are mentioned.
20
Ussher (1650); Ussher (1654).
21
See now Boran (2015), ii 802–03. This edition supersedes the collection of
Ussher’s letters in Elrington (1847–64), xv and xvi; it also includes previously unpub-
lished letters of Ussher’s. The translation is mine.


 DÁIBHÍ Ó CRÓINÍN

Eusebius (not the Ecclesiastical History, which I have to hand), or else

the Syriac version; and if you can find some version of Symmachus, Aq-
uila and Theodotion; the Chronicle of Julius Africanus; Hegessipus’s
Ecclesiastical History; Clement of Alexandria’s Hypotoposeon books and
Anatolius’s book On Easter; Anianus and Panodorus’s books on Easter
computation; George Syncellus’s chronicle in Greek; Apollodorus’
chronicle in Greek; Phlegon on the Olympiads; the books of Diodorus
Sicillus, Polybius, Dionysius Halicarnassus and of Dion Cocceianus that
we are lacking, and Hipparchus’ Astronomy, in Greek.
Seek out traces of all these works wherever you go, and if you find some
of them, note carefully where they are and the names of those in whose
possession they are, and the prices at which they are willing to sell them,
so that we may notify our suppliers in those places and when you have
returned we can subsequently obtain them, if the price is right. And if
you should come into possession of other authors in Greek or Syriac,
besides those that we have listed above, and that we lack, you may nego-
tiate for them as well.’

This is a veritable Who’s-Who of authors that wrote on the subject of

the paschal controversy.22 Some of them are already cited in the Anti­
quitates, but how did Ussher acquire his knowledge of them? The sourc-
es of Ussher’s manuscript library have been well researched.23 In the area
22
One of the most sensational of Ussher’s scholarly publications was his famous
edition of the letters of Ignatius and Polycarp; Elrington (1847), vii 87–267. The list
of desiderata ranges well beyond authors who wrote on the paschal controversy, but of
those cited by him (leaving aside the versio Symmachi, Aquilae et Theodotionis of the
Bible) many are discussed by Mosshammer (2008) in the following places: Eusebius of
Caesarea (Chronicon): 110–13, 267, 289, 325, 329–30; Anatolius of Laodicea: 130–45
(for the Liber Anatolii, see 136–43); Julius Africanus: 328–29, 385–421; Hegesippus is
the Latin translation of Flavius Josephus, Jewish Antiquities: 321, 324, 330, 412–13, and
for manuscript transmission, Siegmund (1949), 102–07 (citing Ussher (1639), 379);
the computi paschales of Annianus and Panodorus: 28, 198, 200–01, 357–84, 421, 426;
George Syncellus: 359–84; Phlegon: 337, 395; Diodorus Siculus: 11, 14, 53, 396–97,
419; Dionysius of Halicarnassus: 11, 13, 14, 19, 267, 388, 396, 398–400. The Hypotopo­
seon libri is a lost work attributed to Clement of Alexandria cited by Eusebius, Historia
ecclesiastica 6.14.1; it is an expository work discussed by Photius of which only fragments
survive; see GCS 3, fragments 1–23; Ussher (1639), 161, refers to Photius for a text of
the letter of the Emperor Honorius concerning Pelagius and Caelestius (and again on
pp. 193, 420). Apollodorus of Athens’s Graecum Chronicum was an influential work in
ancient biography (e.g., Diogenes Laertius 1.37–38, on Thales) and is reconstructed by
Jacoby (1902) and Jacoby (1923–59), as item 244; for brief summaries, see Mossham-
mer (1979), 116–17, 158–60, 283–88. The Dionis Cocceiani libri is the History of the
Roman historian Cassius Dio, surnamed Cocceianus; Ussher apparently wanted the lost
books (qui apud nos desiderantur). Polybius is the Greek historian; Ussher (1639), 377,
cites him for his description of the Britannicas Insulas.
23
See O’Sullivan (1956); Meehan (1986).


 ARCHBISHOP JAMES USSHER (1581–1656)

of historical chronology, however, it will be seen in what follows that he

must have seen and used the two great early seventeenth-century stud-
ies of technical chronology by Dionysius Petavius (Denis Petau) and
Aegidius Bucherius (Giles Bouchier),24 as well as Henricus Canisius.25
He had also certainly read (and disapproved of ) the great Joseph Justus
Scaliger’s Opus de emendatione temporum, which appeared first in Paris
in 1583 and was republished in 1629.26
Just as he did with native Irish scholars, when working on Early
Irish history,27 so too Ussher drew on his extensive friendship network
of scholars, in Europe and beyond, to seek out texts of this kind when
engaged on his chronological researches.28 The measure of his success
in these endeavours is borne out by his famous reckoning of the date of
Creation (which he recalculated as the night preceding Sunday, 23 Oc-
tober, 4004 BC),29 the product of unrivalled knowledge of both bibli-
cal and secular (i.e., mostly Greek and Roman) sources. When he an-
nounced his conclusions in The Annals of the World,30 he established
himself firmly as the leading authority in Europe on the subject.

Ussher’s use of sources for his chronological studies

A new discovery now enables us, for the first time, to determine exactly
which were the texts that Ussher had accumulated in preparation for
this chronological work. Oxford, Bodleian Library, Smith 31 (a paper
manuscript) is a miscellaneous eighteenth-century collection; part 3,
pp. 70–73, contains various texts with the heading: Veterum Scripta Pas­
chalia per Jacobum Usserium Armachanum (‘A collection of ancient pas-
chal writings [compiled] by James Ussher of Armagh’), beginning Anno

24
Petavius (1627) and Bucherius (1634).
25
Canisius (1601–04). See Grafton (1975), 174. According to Mac Carthy (1892),
367 n. 4, there was an annotated copy of Bucherius in the National Library of Ireland in
his time; it has since disappeared.
26
Scaliger (1629). He may also have read Scaliger (1606). See Grafton (1985), 131.
27
See Gwynn (1967); Cunningham and Gillspie (2004–05); ref. from Prof. John
McCafferty.
28
See Boran (2009), 176–94, 228–36 (notes).
29
See Mc Carthy (1997), which offers a necessary corrective to many of the er-
roneous and misleading accounts of Ussher and his works in various encylopaedia and
online web-based sources.
30
Ussher (1658).


 DÁIBHÍ Ó CRÓINÍN

Domini 196. I. Acta Caesariensis synodi sub Theophilo habita sive epistola
Philippi de Pascha; dubia admodum fidei: pag. 469. edit. Bucherii, et in
apographo meo MS (‘I. AD 196. I. Acts of the Council of Caesarea [con-
vened] under Theophilus, otherwise known as the Letter of Philip on
the Pasch, of doubtful authenticity, however; p. 469 of Bucherius, also
in my own MS copy’).31 The collection ends: XXXV. De sollemnitatibus
et sabbatis et neomeniis legalibus Anonymi tractatus ex MS. Cottoniano
(‘XXXV. The anonymous tract On the Festivals and Sabbaths and New
Moons of the Old Testament from a Cotton MS’), and is signed: Exsc­
ripsi ex libro autographo Usserii, Nov. XXV. 1697. T.S. (‘I have transcribed
this from an autograph book of Ussher, 25 November 1697. T. S.’).32
Who is this ‘T.S.’? He must surely be Thomas Smith (1638–1710),
sometime librarian of the Cotton library who published a catalogue of
that collection.33 He was the author of a Life of Ussher, and also ed-
ited the letters of the great English antiquarian, William Camden. Some
(perhaps all?) of his manuscripts ended up in the Bodleian in Oxford;
amongst them appears to have been an autograph volume of Ussher’s.34
Ussher, because of his involvement in politics, and as a result of the
events leading up to the rebellion of 1640–41 in Ireland, was replaced
as chief agent of ecclesiastical policy and as confidant of the viceroy by
John Bramhall, bishop of Derry (an appointee of the new Lord Deputy,
Thomas Wentworth); he left Ireland soon afterwards.35 However, flight
did not save him from further involvement in political affairs, and he was
soon called on to play his part in the impending ecclesiastical revolution
in England. The Long Parliament had held its first session on 3 Novem-
ber 1640, and Strafford (previously Lord Deputy in Ireland) was com-
mitted to the Tower of London on a charge of treason on 25 November;
Archbishop Laud (an erstwhile correspondent of Ussher’s36) followed
him there on 1  March 1641. The formal trial of Strafford opened in
31
In the left margin of the p. 70 (partially obscured by the binding) are the words:
[d]egenui / a et Eusebio, perhaps to be read: De genuina et Eusebio [Historia ecclesiastica].
32
I am grateful to Dr David Howlett (Oxford) for checking the text of the note for
me.
33
Smith (1696); see Wright (1997).
34
I am grateful to Dr Elizabethanne Boran (Worth Library, Dublin) for this infor-
mation. Dr Thomas Roebuck (University of East Anglia) assures me that ‘Smith’s com-
ments like this are completely reliable: if he says he has copied it from Ussher’s autograph
manuscript, then it was copied from Ussher’s autograph manuscript’. I am indebted to
him for this and for much other information concerning Smith.
35
McCafferty (2007), 193–98, 208, 210.
36
See Ford (1991–92).


 ARCHBISHOP JAMES USSHER (1581–1656)

Westminster Hall on 22  March 1641; on 12  May his execution took
place on Tower Hill. Ussher was with him to the end, kneeling beside
him on the scaffold as he said his last prayers.37
Ussher had preached at the opening of the Irish parliament on
20 March 1640, but left for England soon after that, bringing with him
his many books. The vicissitudes that Ussher’s collection of books and
manuscripts endured during the opening years of the 1640s are rela-
tively well known.38 His library had miraculously escaped the flames of
Drogheda in 1641, but flight to England offered no relief; his collection
was sequestered by the English parliamentary forces in Chester in 1643,
and both Ussher and his books were to endure further discomfitures
in the years that followed. According to a contemporary biographer
(Parr), ‘many others of his loose Papers and Manuscripts […] were either
lost in his often forced removals, or fell into the hands of the Men of
those spoiling times, who had no regard to things of that Nature’.39 The
Chester incident had been particularly upsetting because Ussher was
present and witnessed the events; another of his biographers (Bernard)
remarked that ‘it did much grieve him’.40 The rebels tossed his papers to
the winds, and though he recovered some of his materials, ‘he lost two
Manuscripts of the History of the Waldenses, which he never got again’.41
Furthermore, during the hiatus between the purchase of Ussher’s library
for Trinity College Dublin and its eventual entry into the university in
1661, the collection was housed in Dublin Castle, and though many of
the manuscripts that were stolen during this period were later recovered,
‘it is not clear how many were lost’.42
It appears, then, that not all of Ussher’s books made their way back
to Ireland after the Restoration; Ussher himself never returned to his
homeland. The collection of texts concerning the paschal question that
he had accumulated may have been one victim of those turbulent times.
However that may be, Thomas Smith’s transcription of Ussher’s auto-
graph inventory provides a precious insight into the range and extent of
Ussher’s researches in the field, and also into his working methods, for
it reminds us again of what he wrote to his uncle, Richard Stanihurst,

37
Gwynn (1967), 265–66.
38
For what follows, see Boran (2009), 177.
39
Parr (1686), Preface; Boran (2009), 177 and 228 n. 8.
40
Bernard (1656), 101; Boran (2009), 228 n. 9.
41
Bernard (1656), 100; Boran (2009), 228 n. 9.
42
Boran (2009), 228 n. 9.


 DÁIBHÍ Ó CRÓINÍN

back in his student days, about the importance of dating and locating
ecclesiastical authors, but also of establishing the authenticity—or oth-
erwise—of the writings attributed to them.43 In general, it can be said
that Ussher’s classification of computistical authors and texts, and his
arrangement of them in what he perceived to be their correct chrono-
logical order, are very impressive, and even in the case of dubious texts,
his analysis is borne out by modern historiography.
This is well illustrated, for example, by reference to both the first and
the last entry in his catalogue. The first entry, sub anno 196, the text re-
ferred to by Ussher as the Acta Synodi, has been described by the doyen
of modern computistical scholars, the great Charles  W. Jones,44 as ‘a
work that has proved an enigma, the so-called De ordinatione feriarum
paschalium or the Acta Synodi of the bishops assembled at the Synod of
Caesarea’. It exists in no fewer than 36 manuscripts and four recensions,
whose discrepancies have never been reconciled.45 Because it was fre-
quently cited by Irish authors of the seventh century, and because many
of its manuscripts derive from Irish or English archetypes, it was believed
by some nineteenth- and early twentieth-century scholars to belong to
a group of texts described (wrongly) as ‘Irish paschal forgeries’.46 Bede
knew the text in its B-recension and quoted from it verbatim, believing
it to be a genuine writing of Bishop Theophilus of Caesarea.47 The in-
spiration for the forgery—if such it was—lay possibly in the statement
of Jerome:48 Theophilus, Caesareae Palaestinae […] episcopus, sub Severo
principe adversum eos, qui XIV. luna cum Iudaeis pascha faciebant, cum

43
I give the full text of the inventory in the Appendix.
44
Jones (1943), 87.
45
There is an edition in preparation by Leofranc Holford-Strevens.
46
The forgery idea goes back to van der Hagen, but was taken up with gusto by
Krusch (1880), 303–10, who edited this and several others of the ‘forgeries’. He reiter-
ated that conclusion in Krusch (1884). His views were echoed by Mac Carthy (1892),
367–68, and again in his introduction to the Annals of Ulster (Mac Carthy (1901),
cxv–cxvii). Writing on the sources of Cummian’s letter, e.g., Krusch (1884), 151, said:
‘Die darauf folgende Liste der Verfasser von Ostercyclen ist das bunteste Mixtum von
Echtem und Falschem—doch überwiegt das Letztere—, was mir je vorgekommen ist’;
Mac Carthy (1901), cxxxix–cxl is almost a translation: ‘The list of cycles is perhaps the
most ludicrous tissue of fact and fiction in existence’.
47
Bede, De temporum ratione 47 (ed. by Jones (1943), 267–67; commentary p. 87).
The work is also cited in the eighth-century Collectio canonum Hibernensis 31.17–18
(ed. by Wasserschleben (1885), 68).
48
Jerome, De viris illustribus 43, based on Eusebius, Historia ecclesiastica 5.23, 25.
Cited by von Harnack (1958), 1/v, 503. The genuine Theophilan epistle has not sur-
vived.


 ARCHBISHOP JAMES USSHER (1581–1656)

ceteris episcopis synodicam et valde utilem composuit epistulam. (‘Theo-

philus, bishop of Caesarea in Palestine, along with other bishops, dur-
ing the reign of Severus composed a very useful synodical letter against
those who celebrated Easter on luna 14 with the Jews.’) The technical
details of the Acta need not concern us here; suffice it to say that all four
recensions are essentially alike, though they differ in significant details.
Dom André Wilmart believed that the recension which he published
was composed in Africa;49 Jones believed that the African text was the
original one, which was then modified in Spain, in the late sixth century,
and was in turn turned into a third recension when it was introduced to
Ireland.50 Either way, the work was not Irish.
What is of particular importance here, however, is that some of the
best manuscripts of the Acta give it the alternative title, Epistola Philippi
de Pascha, which is the alternative title given to the text in Ussher’s Bodle-
ian inventory: Acta Caesariensis synodi sub Theophilo habita, sive Epistola
Philippi de Pascha (‘Acts of the Council of Caesarea [convened] under
Theophilus, otherwise known as the Letter of Philip on the Pasch’). The
confusion (if such it is) is unique to a family of early computistical man-
uscripts known as the Sirmond Group, so-called because the best rep-
resentative of the group (Oxford, Bodleian Library, Bodley 309—saec.
XI, from Vendôme), once belonged to the famous seventeenth-century
French Jesuit scholar, Jacques Sirmond.51 The Bodleian purchased the
codex in 1698 with the collection of Edward Bernard, who had appar-
ently acquired it soon after Sirmond’s death but before the dispersal of
his library, which had passed to the Jesuit college in Paris and was sold
by auction in 1764.52 The Sirmond Codex was used by both Petavius and
Bucherius (as deduced by Jones) for their great compendia On Time,53
but the manuscript seems to have disappeared soon after that; at least,
none of the eighteenth- and nineteenth-century authors (including

49
Wilmart (1933), from Vatican, Biblioteca Apostolica, Reg. lat. 39 (saec. IX),
which has no discernible Irish connections.
50
Jones (1943), 88.
51
For discussion of this curious phenomenon, see Jones (1937), 206–07.
52
It is very likely the manuscript described as Victorii Aquitani Chronicon Paschale
in Bernard (1697), 228 (as Krusch (1880), 210, pointed out). Jones (1937), 207 n. 1,
reported the Bodleian librarian, Craster, as having remarked (presumably in a personal
communication) that ‘with hardly an exception’, all of Bernard’s Latin manuscripts had
come to him from the library of Nicholas Heinsius.
53
Petavius (1627), i 583: Vetus Codex, qui est penes P. Sirmondum, quae est farrago
Computisticarum disputationum. The word farrago here does not have the perjorative
sense of the modern English term.


 DÁIBHÍ Ó CRÓINÍN

the great Mommsen) who published studies of technical and histori-

cal chronology appears to have used it. Hence, for example, the Dutch
computist, Johannes van der Hagen, in his two great studies, published
numerous corrections and conjectures without having seen the codex.54
Whatever about the title Acta Synodi, it seems almost certain, however,
that the attribution of the text to an otherwise unknown Philip,55 and
its description as an Epistola, when the text is, in fact, in dialogue form,
must have resulted from some earlier dislocation in the manuscript
transmission.
An identical situation seems to have arisen in the case of the text
numbered XXXII in Ussher’s list, which is likewise found in the Sir-
mond Codex (Oxford, Bodleian Library, Bodley 309, 82v–84r), where
the rubric reads Exemplum suggestionis boni sci. primice., clearly intended
to introduce the memorandum prepared by the Vatican curial officials,
Bonus and Boniface, in AD  526, to advise Pope John on the proper
method of calculating Easter.56 The text that immediately follows the
rubric in Oxford, Bodleian Library, Bodley 309, however, is De sol­
lemnitatibus (the final item in Ussher’s Catalogue, No. XXXV), which
we shall discuss presently. Oxford, Bodleian Library, Digby 63, 59r–v
transmits the full text of the Suggestio. As Jones remarked,57 ‘Digby 63
proves itself in other respects a reliable source of knowledge of Insular
usage in the seventh century’, and the fact that it preserves the Suggestio
intact proves that it had parted company from the tradition otherwise
represented by the Sirmond Group at an early date, since the Suggestio
no longer survives in the Sirmond Group of manuscripts. It must surely
be the case that, in a similar way, the Acta Synodi (Pseudo-Theophilus)
must have preceded a separate text (inc. De sollemnitatibus et sabbatis et
neomeniis) that was attributed to (perhaps also composed by) the other-
wise unidentified Philippus.58 Whether this Philippus should be identi-

54
van der Hagen (1734) and (1736). See Jones (1937), 205: ‘Many times he says,
as on p. 43 [of Observationes]: An Bucherius ita in suo MSto legerit, dubito; mihi potius
videtur correctio Bucheriana.’
55
Leofranc Holford-Strevens suggests a possible explanation for the occurrence of
the name in Polycrates, cited by Eusebius, Historia ecclesiastica 5.24.2.
56
Published by Krusch (1926). There are no grounds for Mac Carthy’s suggestion
((1901), cxlvii) that the text is a forgery.
57
Jones (1943), 103.
58
For yet another example of the same phenomenon—also in the Sirmond Group
of manuscripts—see Ó Cróinín (2003a), 211–12. In this instance, the text in Oxford,
Bodleian Library, Bodley 309, 98r–99r ends: Finit de exemplaris cosmographi. Incipit pro­
logus Theophili, where the reference is clearly to that Liber Cosmographorum that Ceol-


 ARCHBISHOP JAMES USSHER (1581–1656)

fied as the author of the De sollemnitatibus is a moot point, but one that
need not be discussed here.
Mention of the tract De sollemnitatibus brings us to that second ex-
ample of Ussher’s methodology mentioned previously. The text occurs
as the last one in his Catalogue (No. XXXV), where it is described as
De sollemnitatibus et sabatis et neomeniis legalibus Anonymi tractatus,
ex MS Cottoniano. The work has had a chequered career (to use Jones
words).59 The letter (disputatio has that meaning in our period) is a
short treatise on the subject of Jewish festal practices and their relevance
to Christian paschal observances.60 It is addressed to an unidentified
uenerabilis papa, whose prayers are requested at the end. If the Disputa­
tio ever had a superscription, it appears now to be lost, and with it the
names of both the author and the recipient.61 But wherever the work is
rubricated in manuscripts it is attributed to Jerome (Disputatio Sancti
[h]Ierononimi and the like). It was first published by Vallarsi from the
manuscript Vatican, Biblioteca Apostostolica, Vat. lat. 642, 89r–90v
(saec. XI) and from Vallarsi’s edition it was taken into Migne’s Patrologia
Latina.62 A separate edition was published by Jean-Baptiste Pitra from
the Paris and London manuscripts (the latter being the same Cotton
codex (London, British Library, Cotton Caligula A XV) that Ussher
referred to).63 In 1885, however, strangely unaware of its previous pub-

frid, Bede’s abbot at Wearmouth-Jarrow, is reported as having given to King Aldfrith of

Northumbria in return for eight hides of land; Bede, Historia abbatum 15 (ed. by Plum-
mer (1896), i 380, and now ed. and trans. by Wood and Grocock (2013), 58–59; see
also Wood and Grocock (2013), xxxvi–xlviii, 164–65). The Liber Cosmographorum had
clearly been removed from the original manuscript before the Sirmond Group of copies
was made; no copy of it survives. Wood and Grocock’s comment ((2013), 58 n. 188),
‘Unfortunately, nothing is known of the contents of this manuscript’, ignores all the lit-
erature about the Sirmond Codex. Alden Mosshammer writes (personal communica-
tion): ‘The text you cite from Bodley 309 has the word incipit, but that word is not in the
Geneva or Tour MSS. The words come at the end of the Prologue of Theophilus, not the
beginning. I think the text in the underlying exemplar must have read finit de exemplari
cosmographi Prologus Theophili. In that case, it might mean that the Prologus Theophili
had been copied from an exemplar belonging to a cosmographer—perhaps Ceolfrid?’.
59
Jones (1943), 108.
60
For what follows, see Ó Cróinín (1997).
61
Unless, of course (as was suggested tentatively above) the rubric in Oxford,
Bodleian Library, Bodley 309, 94v: Incipit epistola Philippi de pascha preserves the au-
thor’s name.
62
Vallarsi (1766), i 114–20 (Ep. 149) = PL 22, 1220–24. The text is also ed. by
Isidor Hilberg in CSEL 56, 357–63; the Maurists published it also in their edition of
Jerome’s works (Martinay and Pouget (1693–1706), i 1103), amongst the spuria.
63
Pitra (1962), i xi–xiv, 9–13, 565.


 DÁIBHÍ Ó CRÓINÍN

lication—and despite the fact that he himself had encountered the text
previously (before 1880) in Cologne, Dombibliothek, 83-II, 201r–203r,
and correctly identified it as a spurious letter of Jerome,64 when Krusch
found the text again in Paris, Bibliothèque nationale de France, Lat.
16361 (without a rubric), he faced it as though it were a new work, and
(on internal evidence!) published it as a previously-unknown letter of
Columbanus.65 Controversy ensued in the German journals, principally
between Krusch, Wilhelm Gundlach, and Otto Seebass, with Gundlach
supporting Krusch’s case and Seebass (an authority on penitential lit-
erature) denying the alleged parallels between the concluding paragraph
of De sollemnitatibus and the opening section of Columbanus’ (authen-
tic) Epistola 5.66 So confident was Gundlach of the letter’s authenticity
that he published it as Epistola 6 in the MGH edition of Columbanus’
letters (simply reproducing Krusch’s text without consulting any other
manuscripts).67 Cabrol and Leclercq published a separate edition, as an
anonymous pre-Nicene tract.68
Jones observed that there was nothing definite in the letter to pre-
clude Jerome’s authorship, ‘except that Jerome is not apt to have writ-
ten such a letter’. Quite what he meant by that is not altogether clear;
but, be that as it may, his further remark, ‘I seem to note a similarity in
the opening paragraph, with its mass of Biblical quotation that is only
vaguely relevant, to the letter of Cummian to Seghine’,69 was singularly
unfortunate, as the very studied and careful composition of Cummian’s
letter has been vindicated by modern scholarship.70
Unlike Cummian’s letter, however, De sollemnitatibus has no saluta-
tion, but concludes with the words ora pro me, uenerabilis papa (‘pray

64
Krusch (1880), 204.
65
Krusch (1885).
66
The controversial literature is listed in Ó Cróinín (1997), 50 n. 10. The most
recent edition is in Walker (1957), 198–206.
67
MGH Epp. 3, 154–56. Krusch subsequently retracted his earlier attribution to
Columbanus: Krusch (1902), 20 n. 1. The work is no longer regarded as by Columbanus.
68
Cabrol and Leclercq (1900–13), ii 71–73 (where the MSS used are not speci-
fied). The Disputatio was even considered as a possible Pelagian tract; see Ó Cróinín
(1997), 50–51 nn. 16 and 17.
69
Jones (1943), 109. But even Homer nodded in this instance: Jones failed to re-
alise ((1937), 217) that the item in the Sirmond Group (e.g., Oxford, Bodleian Library,
Bodley 309, 97r–v) that he described as ‘a short tract on the mystical significance of
Easter. Not published to my knowledge’, is, in fact, an abbreviated version of the Dispu­
tatio. He corrected the omission in Jones (1943), 107.
70
See Walsh and Ó Cróinín (1988), 46–51; Walsh (1987).


 ARCHBISHOP JAMES USSHER (1581–1656)

for me, venerable father’); the author describes himself as a peregrinus.

He also states that the letter was solicited (Haec autem et a te postulatae
sunt; ‘these things were requested by you’), and the implication is that the
writer was commissioned by some higher authority to provide answers to
questions concerning the Easter problem. The fact that Rome alone fig-
ures as an Apostolic See, and the use of the word papa, need not necessar-
ily mean that the letter was addressed to a pope; Cummian, after all, refers
to St Patrick as papa noster, and the term is frequently used of bishops in
the older literature.71 The evidence of the manuscripts speaks strongly for
an Irish origin for De sollemnitatibus. The Cotton manuscript referred to
by Ussher is dated AD 743 in this section, and thus offers a terminus post
quem non. The earliest references to the work, however, are in Irish texts,
and it is cited (under Jerome’s name), for instance, in the seventh-century
Irish computistical handbook, De ratione conputandi.72
One possible solution to the problem of date and authorship has
been suggested in the controversy around the date of Easter that led to
the famous Synod of Whitby in AD 664.73 The tone of the Disputatio
is forceful, even combative, reminiscent of Columbanus, whom Krusch
described as ‘homo uehemens feroxque natura’.74 There were countless
Irish peregrini in England and on the Continent in the seventh century.
One who comes to mind is the Irishman referred to by Bede as acerrimus
ueri Paschae defensor nomine Ronan (‘a most sharp-tongued defender of
the true Pasch called Rónán’), who clashed bitterly with Fínán, bishop
of Lindisfarne (†AD  661).75 This Rónán might very well have been
called upon by a bishop (papa) of the Irish Romanist party in order to
refute the arguments of those in the Irish churches who held out in their
support for older Irish practices in the matter of Easter reckoning. Cer-
tainly, its advocacy of Alexandrian Easter rules precludes authorship by
Columbanus or any other partisan of the ‘Irish’ 84-year Easter tables.76
Most of this controversy must have been unknown to Ussher, but the
fact that he dated De sollemnitatibus in the seventh century, and placed
it last in his catalogue, demonstrates, yet again, how acute was his judge-

71
Walsh and Ó Cróinín (1988), 85 and nn. 208–09.
72
De ratione conputandi 98 (ed. by Ó Cróinín (1988), 204). See also Ó Cróinín
(1982), 118–20.
73
Ó Cróinín (1997), 52–53.
74
Krusch in MGH SS rer. Merov. 4, 20. Coming from Krusch, the verdict has a
certain piquancy about it!
75
Bede, Historia ecclesiastica 3.25 (ed. by Plummer (1896), 181).
76
For De sollemnitatibus, see also Warntjes (2010a), LXVIII–LXIX n. 175.


 DÁIBHÍ Ó CRÓINÍN

ment of all these materials. Rather than place it in the fourth or fifth cen-
tury (or even earlier, on the basis, e.g., of the use in it of the term papa),
he judged it to be the latest in time of all the texts that he had assembled.
Modern scholarship would concur with him in that judgement.

Identifying Ussher’s computistical sources from Oxford,

Bodleian Library, Smith 31 (pt 3), pp. 70–73
In what follows I will try to identify and discuss briefly the other texts
listed by Ussher, in the sequence in which he listed them.
No. II: Victoris Papae epistolae tres paschales: Victor (AD 189–98) is
best known for having threatened to excommunicate the Christians of
Asia Minor in AD 192 for their continued adherence to the (earliest)
Christian practice of observing the fourteenth day of the lunar month as
the beginning of the paschal festival (otherwise known as Quartodeci-
manism). The matter had been simmering since the mid-second century,
following what Eusebius describes as ‘synods and assemblies of bishops
held on this account, and all [in the West], with one consent, through
mutual correspondence, drew up an ecclesiastical decree that the mys-
tery of the Lord’s Resurrection should be celebrated on no other but the
Lord’s Day’.77 In the year AD 154–55 Bishop Polycarp of Smyrna went
to Rome to consult with Pope Anicetus on the problem, but the two
agreed to differ;78 however, the pope conceded the administration of the
Eucharist to Polycarp as a mark of respect. Pope Victor, for his part, was
dissuaded from his confrontational views by the intervention of Bishop
Irenaeus of Lyon; ‘his retraction saved the Church from schism’.79

77
Eusebius, Historia ecclesiastica 5.23: ob quam causam conventus episcoporum et
concilia per singulas quasque provincias convocantur, porogatisque ad se invicem epistulis de
singulis quisque locis unum omnes ecclesiasticum dogma confirmat, ne liceat aliquando nisi
in die dominica, in qua dominus surrexit a mortuis, dominicum paschae celebrare myste­
rium et in hac sola solvendum esse paschale ieiunium; Jones (1943), 10.
78
‘It is quite possible that that dispute was not about when to celebrate Easter but
whether. Irenaeus’s letter contrasts observers with non-observers, not Quartodecimans
with Dominicalists. However, Rome certainly did so by Victor’s day’ (personal com-
ment from Leofranc Holford-Strevens).
79
Jones (1943), 10. Columbanus, Epistola 1.4 (Walker (1957), 6), was scathing
about Victor’s attempt to force Sunday-observance on the Church, saying that he intro-
duced it as an anti-Semitic measure: Cum Iudaeis facere Pascha non debemus. Dixit hoc
olim et Victor episcopus, sed nemo orientalium suum recepit commentum. For good general
discussion, see Mosshammer (2008), 46–47, 110–11; Blackburn and Holford-Strevens
(1999), 791–92.


 ARCHBISHOP JAMES USSHER (1581–1656)

The Liber Pontificalis reads: Hic fecit constitutum ad interrogationem

sacerdotum de circulo paschae ut dominico paschae, cum presbiteris et epis­
copis factam conlationem et arcessito Theophilo episcopo Alexandriae facta
congregatione, ut a .xiiii. luna primi mensis usque ad .xxi. die dominicum
custodiatur sanctum pascha.80 (‘Here it was decided, because of queries
from bishops concerning the paschal cycle, and following discussion with
the bishops and priests, at the urging of Theophilus, bishop of Alexan-
dria, that the holy Pasch should be observed on Sunday from luna 14 to
21 of the First Month.’) Jones pointed out that Theophilus of Alexandria
was here confused with Theophilus of Caesarea, Victor’s contemporary.81
Though Ussher refers to three ‘supposed’ letters of Victor (epistolae sup­
posititiae), none is thought authentic, but copies of all have survived. The
supposed letter of Victor to Theophilus is cited from Isidore Mercator’s
Collectio; presumably the Pseudo-Isidorian Decretals are meant.82 The let-
ter to Theodosius and the letters to Desiderius and Paracoda are printed in
Patrologia Graeca from Mansi.83 Mansi reprinted the letters to Desiderius
and Paracoda from Labbaeus, who in turn reprinted them from Baronius,
Baronius had them ex Floriac. Biblioth. monimentis.84 Ussher’s reference
ad calcem bibliothecae Floriacensis is to Floriacensis vetus bibliotheca, com-
piled by Jean du Bois (= Ioannes a Bosco).85 In the second part, dedicated
to the xyston laevum of the library, Victor’s letter to Bishop Desiderius
(known only from this letter), squeezed in between Bishops Iustus and
Dionysius, is on pp. 24–25.86 Ussher knew these texts from Bosco, whose
work he cites in several places in the Antiquitates.87 Ussher also cites a text
referred to in Indice librorum Graecorum MSS. Bibliothecae Cesareae Vien­
nensis, anno 1579. confecto (‘in the index of Greek manuscripts of the Impe-
rial Library of Vienne, published in 1579’) that he had used.88

80
Liber Pontificalis 15, s.n. ‘Victor’ (ed.  by Theodor Mommsen in MGH Gesta
pontificum Romanorum 1, 19); Jones (1943), 36–37 n. 4.
81
Jones (1943), 36–37 n. 4.
82
PL 130, 121–24. There is a newer edition available online from the MGH at
http://www.pseudoisidor.mgh.de. The letter is quoted in Gratian’s Decretum 3.22 (new-
ly ed. in Friedberg (1959), i 1359–60); reference from Alden Mosshammer.
83
PG 5, 1485–90.
84
Baronius (1588–1607), ii 198.
85
a Bosco (1605b).
86
Reference from Leofranc Holford-Strevens.
87
Ussher (1639), 442, 444, 536, 558, 676.
88
Ussher (1639), 420. This may be a reference to the separately paginated mono-
graph by a Bosco (1605a), which has a table of contents headed: Catalogus quae hoc


 DÁIBHÍ Ó CRÓINÍN

No. III: Hippolyti Portuensis episcopi Canon Paschalis: Here referenced

to Scaliger’s De emendatione temporum (‘lib. VII, p. 677’). Mosshammer
provides an excellent discussion of Hippolytus’ paschal table.89 The fa-
mous discovery, c. 1550, of the statue of a headless seated figure, near the
church of San Lorenzo fuori le Mura, in the suburbs of Rome, was one of
the sensations of nineteenth-century Christian archaeology.90 Inscribed
in Greek characters on a plinth at the rear of the chair is a list of writings;
on one side of the chair there is a 16-year table of paschal full moons,
while on the opposite side there is a 112-year list of dates for Easter Sun-
day. Although no name appears on the statue, it was subsequently identi-
fied with that Hippolytus to whom Eusebius ascribed a work De pascha.91
The statue now stands in the foyer of the Vatican Library. Ussher clearly
drew on the commentary in De doctrina temporum.92
No. IV: Anonymi Scriptum Paschale […] ad annum v Gordiani […]
perductum: This is a reference to the Pseudo-Cyprianic Computus of
AD 243 (that year being the fifth year of the Emperor Gordian). Ussher
describes the work as having been ex antiquissimo MSto Cottoniano de­
sumptum (‘taken from a very old Cotton MS’), and he had his text from
London, British Library, Cotton Caligula A XV, 97v–105v, of which
he made a transcript (though he refers also to Canisius, whose Anti­quae
Lectiones he cites93). He mentions the work in his Chronologia Sacra,
commenting on the reign of Jechonia, where he says that he was assigned
eleven years of rule in Paschali tractatu, quem ineditum habeo, anno quin­
to Gordiani, Ariano et Papo consulibus, id est, anno aerae Christianae 243
(‘in an unedited paschal tract that I have, under the year of the consuls
Gordianus, Anianus and Papus, i.e., in the Christian era [AD] 243’).94
The reference to the chronicle anno 13º Alexandri Imperatoris (Chris­
tianae epochae 234) scripti is to the Liber Generationis, which is a Latin
version of the Chronicle of Hippolytus.95 It is interesting that Ussher did
not speculate that the Computus of 243 was by Cyprian.

laevum xyston Floriacensis veteris bibliothecae regiae suo ambitu concludit (reference from
Alden Mosshammer).
89
Mosshammer (2008), 116–25.
90
See Frend (1996).
91
Eusebius, Historia ecclesiastica 6.22.
92
Bucherius (1634), 295–96.
93
Canisius (1601–04), ii 578–600.
94
Ussher (1660) (= Elrington (1847), xii 126); cf. Mosshammer (2008), 125–27
and 228–29, and his article in this volume (pp. 43–70 above).
95
See Mosshammer (2008), 328.


 ARCHBISHOP JAMES USSHER (1581–1656)

No. V: Dionysii Alexandrini epistola ad Basilidem Pentapolis: Eusebi-

us says that Dionysius of Alexandria sent a letter to the Roman Church,
but he does not quote from it, nor explicitly state its subject.96 Else-
where he reports the same Dionysius as having addressed a letter to his
fellow Eastern bishops (in AD 247), in which he set forth a canon of
Easters for eight years, which later came to be known as an octaëteris; this
octaëteris—in some form—lasted until the end of the fourth century.97
Ussher cites no source for this letter.
No. VI: Anatolii Laodiceni episcopi Canon Paschalis: the Anatolian
paschal canon.98 Cited already as an authority by Columbanus, in his
letter to Pope Gregory of AD 600,99 and by Cummian in his letter to
Ségéne, abbot of Iona (†AD 652),100 the tract of Anatolius was a ‘star
witness’ at the Synod of Whitby in AD 664. Because the work has sur-
vived in nine manuscripts, all of which are either Irish in origin or are
directly traceable to Irish exemplars, it was believed by Krusch and oth-
ers to be one of the infamous ‘Irish Paschal Forgeries’, and to have been
concocted in Ireland in the sixth century.101 ‘With his usual rhetorical
flourish, Bartholomew Mac Carthy (1901: cxviii), judged the text ‘peer-
less in the field of fabrication’’.102 Mc Carthy and Breen, however, argued
that the Latin tract preserved a genuine Anatolian text, thereby agree-
ing with Bucherius, who had first published the text.103 The most recent
commentator, however (Mosshammer), is not convinced by the argu-
ments for authorship by Anatolius (in this version, at any rate); Ussher,
for his part, describes the text in his list as dubia fidei, which certainly
suggests that he suspected its authenticity.

96
Eusebius, Historia ecclesiastica 6.46. For what follows, see Mosshammer (2008),
118–21.
97
Eusebius, Historia ecclesiastica 7.20–22: Dionysius Paschales illas quas habemus
epistolas tunc temporis conscripsit […] probans festum Paschae diem nonnisi post aequinoc­
tium vernum celebrari oportere, octo annorum canonem publicavit. For the belief that the
octaëteris lasted, in some form, until the end of the fourth century, see Jones (1943), 14
(who cites Krusch and Ideler, both referring to Epiphanius).
98
Most recently ed. and trans. by by Mc Carthy and Breen (2003).
99
Columbanus, Epistola 1.3 (Walker (1957), 2): Anatolius, mirae doctrinae vir.
100
Cummian, De controversia paschali ll. 210–12 (ed. and trans. by Walsh and Ó
Cróinín (1988), 84–87, with discussion at 32–35). Good discussion in Jones (1943),
82–85.
101
See Krusch (1880), 311–27. Mac Neill (1931), 450, suggested that the work
might have originated in post-Roman Wales.
102
Mosshammer (2008), 137. For what follows, see Mosshammer (2008), 136–45.
103
Bucherius (1634), 450–66.


 DÁIBHÍ Ó CRÓINÍN

No.  VII: Constantini M[agni] Imperatoris epistolae duae: Eusebius

stated that Constantine addressed a long letter to the churches in gen-
eral devoted entirely to the subject of the paschal observance.104 He also
reports that he had himself composed a tract ‘On the Festival of the Pas-
cha’, which he had dedicated to the emperor, and he quotes Constan-
tine’s letter of acknowledgement.105 Ussher’s references were presumably
to these passages in Eusebius; no such letters are now extant.
No. VIII: Latinorum Paschalis laterculus centum annorum: This is to
be identified with the 100-year list contained in the Chronograph of
AD  354.106 The term Latinorum was intended to distinguish that list
from the 100-year list of Theophilus (of Alexandria); Ussher’s reference
is to Bucherius.107
No. IX: Ambrosii Mediolanensis episcopi epistola: The divergence of
Alexandrian and Roman Easter tables during the early centuries of the
church caused considerable difficulties. In AD 387 Ambrose, bishop of
Milan, addressed a letter to the bishops of the province of Aemilia on the
subject of the Easter date for that year (25 April, the latest possible date,
according to Alexandrian principles, and four days beyond the limit set
by older Roman tables). Ambrose’s letter is clear testimony to the fact
that Alexandrian rules and the 19-year tables that embodied them were
known and used in the West before the end of the fourth century, but
that other tables too were in use at the same time. Krusch (and Eduard
Schwartz) dismissed the letter as spurious,108 but their flimsy case is more
than adequately dealt with by Michaela Zelzer. Zelzer has a wonderful
rebuke of Krusch’s modus operandi: ‘Das ist keine brauchbare Methode,
um mit Angaben, die nicht ins Konzept passen, fertig zu werden.’109 The
letter was published by Bucherius, where Ussher found it.110
Nicholaus Mulerius, mentioned by Ussher in this section, was an as-
tronomer, apparently still honoured at the Rijksuniversiteit in Gronin-
gen, ‘where the student sailing club has named its flagship after him’.111

104
Eusebius, Life of Constantine 3.7–18. See Mosshammer (2008), 51.
105
Eusebius, Life of Constantine 4.34–35. See Mosshammer (2008), 146.
106
See Mosshammer (2008), 213–16.
107
Bucherius (1634), 252–59 (with reference also to 264).
108
Krusch (1880), V; Schwartz (1905), 54–55.
109
Zelzer (1978), 190 n. 10.
110
Bucherius (1634), 474. The letter is now ed. by Michaela Zelzer in CSEL 82,
222–34.
111
Leofranc Holford-Strevens, in a personal communication. Alden Mosshammer
suggests, as an alternative, Nicola Milanesio, author of a work De Miracoli de Santissimo


 ARCHBISHOP JAMES USSHER (1581–1656)

No. X: Iohannis Chrysostomi presbyteri Antiocheni Sermo Paschalis:

The sermon in question is the famous Easter homily, Sermo cathecheticus
in Pascha (which is still read in Orthodox churches); the text is printed
in Patrologia Graeca (amongst the dubia).112 Ussher cites the eight-vol-
ume edition of Henry Savile.113
No.  XI: Pachomii Abbatis Laterculum Quadragesimale et Paschale:
Not a table, presumably, but the mnemonic verses, beginning Nonae
Aprelis norunt quinos, that are ubiquitous in medieval computistical
manuscripts, and allegedly delivered to the Egyptian monastic founder,
Pachomius, by an angel.114 Ussher cites neither Bucherius nor Petavius
for the verses (nor any other source), but he would have found them in
any of the Cotton manuscripts that he refers to.115
No.  XII: Theophili Alexandrini prologus in laterculum Paschalem
C. annorum: A table of 100 years is known to have been composed by
Bishop Theophilus of Alexandria (AD  385–412). This he sent with a
dedicatory letter to the Emperor Theodosius. This table, beginning
in AD 380, was accompanied by an appendix of quaestiones, or math-
ematical formulae, that could be used to calculate the date of Easter.
According to Jones,116 ‘the ‘quaestiones’ are undoubtedly our Prologue’,
and Bucherius published this Prologue from the Sirmond Codex in De
doctrina temporum,117 and Petavius in Opus (apparently also from the
Sirmond MS), who published the Greek text as well (from the Chroni­
con Paschale).118 Ussher seems to have used both. Krusch re-edited the
letter.119

Sacramento (Venice 1615), though he adds: ‘No relevance I can think of ’ (personal com-
munication). Perhaps Johannes Franciscus Balbus, author of a Tractatus Foecundus, et
Perutilis de Praescriptionibus (1564); but that is a treatise on Roman Law, with no obvi-
ous relevance to our subject.
112
PG 59, 721–24. Cf.  Floëri and Nautin (1957), 111–73, which addresses the
question, why Christmas and Epiphany are celebrated on fixed dates, but Easter is not.
Its attribution to Chrysostom is considered dubious.
113
Savile (1610–12), v 940.
114
Nonae Aprelis is ed.  by Karl Strecker in MGH Poetae 4,2.3, 670–71; see also
Howlett (2013), 115–19 (who argues for an Irish origin of the verses).
115
For the legend, see Jones (1943a); for background to the verses (and a correction
of Jones’s article), see Ó Cróinín (1982a).
116
Jones (1943), 30.
117
Bucherius (1634), 471–73.
118
Petavius (1627), ii 501–03.
119
Krusch (1880), 220–26.


 DÁIBHÍ Ó CRÓINÍN

No.  XIII: Chronographia cyclis 1820 annorum distincta ad Vitalem

quendam: Alden Mosshammer, in personal communication, writes as
follows:

‘This entry is a real puzzler. I can think of no Chronographia organized by

cyclis annorum except that of Annianus of Alexandria—one of the lost
works that Ussher asked for in the letter cited above. Syncellus says that
Annianus composed 11 cycles of 532 years. See Mosshammer (2008),
200. Could ‘1820’ be a copying or reading error for ‘532’? With an ep-
och of 25 March 5492 BC, the eleventh cycle ended in AD 360. Syncel-
lus also says that one or another of the monks Annianus and Panodorus
compiled 5904 years up to the time of Theophilus. 5904=AD 412, [the
year of the] death of Theophilus. I suppose if one of these fellows wrote
in the year 412, the other might have composed his work in 395. But I
don’t know how Ussher could have acquired such information. Who
was Thomas Alan? I can’t believe that a copy of Annianus actually sur-
vived and came into this man’s hands. And I don’t know any dedication
ad Vitalem. The only historical work I can think of that ended in 395 is
the Ecclesiastical History of Rufinus. But Rufinus does not use consular
dates, and he addressed his work to Chromatius, not Vitalis.’

According to Ussher’s note, this work he obtained ex MS.o Thomas

Alani. ‘William O’Sullivan has outlined the principal sources of Ussher’s
collection, and these included a large bank of manuscripts bought fol-
lowing the death of Thomas Allen, Henry Savile of Bank, [and] others
bought once John Dee’s library was released to the market’.120
One wonders what this Chronographia could have been, and what
happened to Thomas Allen’s manuscript copy of it.
No.  XIV: Theophili Alexandrini Paschalium epistolarum trium:
These three letters circulated amongst the letters of Jerome.121 It is not
clear from Ussher’s list which edition of Jerome’s letters he used.
No.  XV: Synesii Cyraenaei Ptolemaidis episcopi epistolia duo: These
are the letters numbered 9 and 13 in the corpus of Synesius of Cyrene.
Letter 9 to Theophilus thanks him for his paschal letters. Letter 13
to Peter announces the date for Easter as 19 Parmouthi [= 14 April].
Petavius published an edition in 1612.122

120
Boran (2009), 178; cf. Watson (1978), 300.
121
See Siegmund (1949), 134–35.
122
Reprinted in PG 66, 1346, 1350.


 ARCHBISHOP JAMES USSHER (1581–1656)

No.  XVI: Innocentii  I. epistola ad Aurelium Carthaginensem epis­

copum: Ussher quotes this letter in his Antiquitates as ‘Innocentius I. in
Epistola ad Aurelium Carthaginensem, de anni CCCCXIV. Paschata
verba faciens’:123

Cùm ante diem undecimum Calendarum Aprilium pène Luna decima­

sexta colligatur (nam quippiam minus est) itemque cùm ante diem quar­
tum Calendarum earundem veniat vigesima tertitia; existimavi undeci­
mo Calendarum memoratarum die festa Paschalia celebranda: quoniam
in vigesimâ tertiâ Lunâ nullum Pascha unquam, ante hoc Pascha, factum
esse cognoscimus.

‘Since the 16th moon occurs the day before the 11th Kalends of April
(and that is the earliest date), and since likewise the 23rd moon occurs
the day before the 4th Kalends [of that month], I have decided that the
paschal feast should be celebrated on the 11th Kalends, since we have
never before heard of a pasch on the 23rd moon.’

He cited a 1591 Roman edition of decretal letters, presumably Epis­

tolae decretales summorum pontificum;124 the letter is indeed on p. 62 of
that edition.
The reference to Crabb must be to Pierre Crabbe, who published the
letters of Innocent in a Conciliorum omnium, etc. (Cologne 1551).125
No. XVII: Cyrilli Alexandrini episcopi xxix Sermonum Paschalium
epilogi: These are the 29 letters (usually called homilies) that were is-
sued by Alexandria in advance of the paschal celebrations every year
and circulated to local and distant churches; they cover the years
AD  414–44.126 The epilogi are probably the final paragraph of each
letter, in which the dates of the impending Quadragesimae and Easter
were announced every year. There was an edition by Antonio Salmatia
(Antwerp 1618).
No.  XVIII: Ejusdem [Cyrilli] epistola ad Synodum Carthaginensem:
The clue here is Ussher’s additional remark: quam ad annum 607 refert
Petavius (‘which Petavius relates to the year 607’). The text is quoted at
length by Cummian, who cited it as a witness in favour of the Alexan-

123
Ussher (1639), 480. See Holford-Strevens (2011), 6.
124
de Aquino (1591).
125
See PL 20, 517–18, and Holford-Strevens (2011), 6.
126
They are reprinted in PG 77, 401–82.


 DÁIBHÍ Ó CRÓINÍN

drian reckoning.127 It also occurs in the seventh-century Irish computisti-

cal handbook, De ratione conputandi (but without the section concerning
Victorius’ cycle that occurs in Cummian’s citation).128 Petavius (1627) and
Bucherius (1634) published it from the Sirmond Codex; Petavius declared
it a forgery.129 Bartholomew Mac Carthy wrote:130 ‘In view of the fact that,
not to mention older writers, Ideler, De Rossi and Mommsen did not ac-
cept the conclusion of Petavius, the fraud was exposed anew by Krusch, in
a manner that obviates cavil.’ Krusch re-edited the text,131 using Petavius
and five manuscripts, and agreed with Petavius’s conclusion of forgery.
The technical data in the Epistola132 are incompatible with an 84-year
cycle, but are found in the Victorian tables at AD 691, a date too late,
and in the Alexandrian cycle at AD 607. Mac Carthy therefore conclud-
ed that the forgery was created in AD 606, in anticipation of an anomaly
that would occur in the following year, when the ‘Irish’ and Victorian
Easters would fall on 16 April, while the Alexandrian fell on 23 April.133
Jones remarked:134 ‘We know little enough of what was done anywhere
about Easter, but […] we can be sure of one thing—at no place in Ireland
was luna xxii ever observed, unless by chance (and even of this we have
no evidence) the mistake was made in trying to follow the Victorian
tables in conformity with Rome’. The historical circumstances would
match those described in Ireland in Cummian’s time.
Jones, however, suggested that the forgery might have been con-
cocted in Canterbury as a kind of pia fraus, to persuade a reluctant Irish
church to adopt Victorian tables: ‘If the letter was sent to the Irish from
Canterbury, some attempt may have been made to obscure its immedi-
ate source’. The tampering that was supposedly done to an original letter
of Cyril’s ‘could not, by the furthest stretch of the imagination, be con-
sidered fraud’. He therefore maintained that ‘another ‘Irish Forgery’ was
not a fraud and was not Irish’.135

127
Cummian, De controversia paschali ll. 223–43 (Walsh and Ó Cróinín (1988), 86–
89). For detailed discussion, see Jones (1943), 92–97; see also Grosjean (1946), 225–43.
128
De ratione conputandi 98 (ed. by Walsh and Ó Cróinín (1988), 203).
129
Petavius (1627), i 144–45; Bucherius (1634), 481–84.
130
Mac Carthy (1901), cxxxiv n. 4.
131
The Epistola Cyrilli is ed. by Krusch (1880), 344–49.
132
Epistola Cyrilli 3 (Krusch (1880), 345); cf. Krusch (1880), 101–09.
133
Mac Carthy (1901), cxxxv.
134
Jones (1943), 94.
135
Jones (1943), 95–96.


 ARCHBISHOP JAMES USSHER (1581–1656)

Jones pointed out,136 that another Epistola Cyrilli is found in Milan,

Biblioteca Ambrosiana, H 150 inf. (the so-called ‘Bobbio Computus’):137
‘but the rubric is mistaken; the text looks like a set of excerpts from Isi-
dore’.
Ussher believed (probably correctly) that the Easter tables in Isidore’s
Etymologiae were an extension of the Cyrillan (Alexandrian) tables, but
since no good modern edition of Isidore’s tables exists, it is not possible
to be definitive on the matter.138 He also mentions the tables of Diony-
sius Exiguus, which he gives as No. XXVIII in his list.
No.  XX: Paschasini Lilybetani episcopi ad Leonem Papam epistola:
Paschasinus was bishop of Lilybaeum (Marsala in Sicily) and subse-
quently served as legate for Pope Leo at the Council of Chalcedon in
AD 451.139 In anticipation of the problems that were expected to arise
regarding the date of Easter in AD 444, Leo wrote to him, asking his
opinion about the conflicting Roman and Alexandrian reckonings for
the coming year. In his reply Paschasinus advised that he thought the Al-
exandrians were correct in proposing to observe Easter on 23 April, luna
.xviii. He supported his case with a little story that ‘explained’ the mys-
tery by means of a wondrous miracle (cuius misterii miraculum tale est).
According to his deacon, Lollianus, there was a certain remote parish,
set in the most desolate part of the country among impassable moun-
tains and impenetrable forests. In the baptistery of that impoverished
church the font filled miraculously every year on Easter eve. In AD 417,
when the church observed the Roman date of Easter (on 25 March) the
font remained dry; however, it filled at the usual hour on 22  April—
Easter Sunday by the Alexandrian reckoning!140 Impressed, it seems, by
the power of miracle (if not by argument), Leo sanctioned the obser-
vation of Easter Sunday in AD 444 on 23 April (two days beyond the
limits allowed by Roman paschal tables). Ussher states that he used the
edition in Bucherius;141 Krusch lists other editions.142

136
Jones (1943), 97 n. 1.
137
Bobbio Computus 140 (PL 129, 1353–54). In fact, the passage is from Isidore,
Etymologiae 6.17; see Warntjes (2010a), LII n. 127.
138
See Krusch (1884), 117–19. Immo Warntjes has a paper on the subject in prepa-
ration.
139
For this and what follows, see Mosshammer (2008), 62–63, 66, 169, 193, 196,
204, 238.
140
Epistola Paschasini 2­–3 (ed. by Krusch (1880), 249–50).
141
Bucherius (1634), 75.
142
Krusch (1880), 246.


 DÁIBHÍ Ó CRÓINÍN

Jones wrote:143 ‘To the list of previous editions add Ussher, Britan­
niae [sic] Ecclesiarum Antiquitates, London, 1687 (p. 480) and Dublin,
1639 (p. 926)’. But these are citations, not editions of the entire text.
No. XXI: Leonis I. epistolae: Ussher names no source, but he presum-
ably used the Ballerini edition of Leo’s letters. Krusch re-edited the ones
relating to Easter.144
No. XXII: Proterii Alexandrini ad Leonem I.: Just as had happened
in AD 444, so also in AD 455 the Roman and Alexandrian paschal reck-
onings diverged in a serious way. In AD 453 Pope Leo, anticipating the
problems to come, wrote a letter to the Emperor Marcian, solliciting his
help; he sent similar letters to Constantinople (but not, apparently, to
Alexandria).145 No doubt the fact that Leo had studiously ignored him
was the reason why Proterius’ response to Marcian’s request for advice was
frosty in tone. Before the end of May AD 454, Leo had received a letter
from him outlining, in full detail, where the fault lay in the matter, mak-
ing it clear that the Alexandrian church was the mother of paschal calcula-
tions and its expertise in such matters should not be doubted. He recom-
mended the Easter date of 24 April for the coming year: Et nos enim et tota
Aegyptia regio atque oriens universus sic ipsum diem caelebraturi sumus, Deo
praestante (‘all Egypt and the East will observe this day, God willing!’),
and further rubbed Leo’s nose in it by calling attention to the fact that in
AD 550 (Deo praestante!) they would observe an even later Easter—on
25 April.146 As he had done in AD 444, Leo capitulated again.147
Ussher’s reference to having read Proterius’ letter Dionysio Exiguo
interprete refers to the fact that Proterius (deliberately) had sent his let-
ter to Leo in Greek; the Latin translation had been made by Dionysius
Exiguus; it was the version that circulated in the West, together with
Dionysius’ own table. The Greek original has not survived.
No.  XXIII: Hilari Romani archidiaconi ad Victorium Aquitanicum
epistola: The conflict of AD  455 between the Alexandrian and Roman
churches over the proper date of Easter in that year resulted directly in the
creation of the Easter table of Victorius of Aquitaine.148 It is very likely that

Jones (1943), 55 n. 3.

143

Krusch (1880), 251–65; cf. also Krusch (1880), 129–38.

144

145
For what follows, see Jones (1943), 56–61; cf. also Mosshammer (2008), 59, 65,
190–93, 203, 243.
146
Krusch (1880), 270–71.
147
Holford-Strevens (2011), 6–7.
148
For what follows, see Jones (1943), 61–68; Mosshammer (2008), 239–41.


 ARCHBISHOP JAMES USSHER (1581–1656)

Leo, following his discomfiture at the hands of Proterius of Alexandria,

decided to take action to ensure that no such embarrassment would trou-
ble the Roman bishops in future. His archdeacon, Hilarus, wrote to Vic-
torius, asking him to inquire at his leisure (per otium diligenter inquiras)
about the causes of the recurring problems, and hopefully to come up with
a solution. Hilarus may also have sent Victorius a copy of the current Al-
exandrian tables and asked him to compare their data with Roman tables.
The letter of Hilarus, Victorius’ Prologue, and his 532-year tables
were first published by Bucherius, De doctrina temporum, where Ussher
consulted them.149 Petavius had earlier published a part of the work.150
Both used the Sirmond Codex. Because Hilarus subsequently (AD 461)
succeeded Leo as pope, some manuscripts of the Victorian tables de-
scribe him as papa (rather than archidiaconus), thereby creating the im-
pression that the tables were sanctioned by Rome from the outset.
No. XXIV: Victorii Aquitanici ad Hilarum Rescriptum cum prologo
in periodum suam Paschalem annorum 532: Little is known of Victorius.
Gennadius of Marseilles (a younger contemporary), in his catalogue of
eminent ecclesiastical writers (De viris illustribus), described him as a
scrupulosus calculator, and Victorius may have enjoyed a reputation as
a mathematician.151 Mommsen identified him as the author of a Liber
calculi that survives in a number of manuscripts. It consists of multipli-
cation- and division-tables for the numbers 2–98, and a discussion of
fractions.152 Jones remarked:153 ‘The Calculus is not the product of an
intellectual giant; the mathematical ability shown there would not pre-
clude serious errors in Easter-calculation’.
Columbanus, in his letter to Gregory the Great (AD 600), poured
scorn on Victorius’ tables, saying that they had been scrutinised and
found wanting by the most expert Irish computistical scholars (Scias
namque nostris magistris et Hibernicis antiquis philosophis et sapientissimis
componendi calculi computariis Victorium non fuisse receptum; ‘For you
should know that Victorius has not been accepted by our teachers, by the
ancient scholars of Ireland, by the mathematicians most skilled in reck-
oning chronology’), and were deemed to be ‘more worthy of ridicule or

Bucherius (1634), 2–10.

149

Petavius (1627), ii 885–88.

150

151
Gennadius, De viris illustribus 88.
152
For information on the various editions of the work, see Jones (1939), 53. See
now Peden (2002).
153
Jones (1943), 62.


 DÁIBHÍ Ó CRÓINÍN

pity than of authority’ (magis risu vel venia dignum quam auctoritate).154
Victorius fared no better at the hands of Bishop Victor of Capua, in the
mid-sixth century, or of Bede, in the eighth.155 Modern authorities, how-
ever, are inclined to be more generous. Mosshammer, for example, has
written:156 ‘Victorius’s 532-year list of epacts, weekdays, and Easter Sun-
days represents both a technically accurate use of the 19-year cycle and
a reasonable compromise between Alexandrian and Roman traditions’.
He was working to a brief, above all to avoid 25 April as a date for Easter.
The Oxford Sirmond Codex that Bucherius and Petavius used does
not have alternative dates for some years in the Victorian tables; most
manuscripts omit these.157 Mommsen’s edition158 is unsatisfactory, and
was criticised by Mac Carthy and Jones.159 Krusch re-edited Victorius’
computistica.160
No. XXV: Metrodori cyclus annorum 532: I know nothing more of
Metrodorus than what is stated by Alden Mosshammer:161

‘As far as we know, it was Annianus who first put the 532-year period
into the service of Paschal calculations. One of the books that Photius
read included a 532-year cycle beginning in the first year of Diocletian
and attributed to one Metrodorus. Photius says (Bibl. cod. 115) that he
could not ascertain who this Metrodorus was. Antoine Pagi (1689: vi)
identified him with the philosopher of that name whom Jerome (p. 232
Helm) dates to AD 330. Of that Metrodorus, George Cedrenus writes
that he visited India during the reign of Constantine to study philoso-
phy with the Brahmins, and while there introduced the Indians to wa-
termills and baths. Metrodorus was sufficiently well known that Photius
could not have been ignorant of him. The 532-year Paschal table at-
tributed to him was probably a production of the sixth or early seventh
century, one among the many such tables to which the author of the
Chronicon Paschale refers (20.19–21.5).’

154
Columbanus, Epistola 1.4 (Walker (1957), 6–7).
155
Bede, De temporum ratione 51 (ed.  by Jones (1943), 270–73); Epistola ad
Wicthedum 8 (ed. by Jones (1943), 322–23); see Jones (1943), 74.
156
Mosshammer (2008), 243.
157
On this, see Warntjes (2010a), LXXXIV–LXXXV n. 228.
158
Victorius’ Cyclus and Prologus were ed. by Theodor Mommsen in MGH Auct.
ant. 9, 667–735.
159
Mac Carthy (1901), lxxxvii; Jones (1943), 61 n. 4; see also his lengthier discus-
sion in Jones (1934).
160
Krusch (1938), 1–87.
161
Mosshammer (2008), 199.


 ARCHBISHOP JAMES USSHER (1581–1656)

Photius speaks of ‘one Metrodorus’, of whom he knew no more than that

his 532-year paschal cycle was compiled ‘according to the accurate calcu-
lation of luna XIV’. Ussher adds a reference to ‘another Greek’s criticism
of him’ (cum alterius Graeci in eum censura), referring to the anonymous
eight-book ‘Third Discourse on the Holy Feast of the Pascha’, reviewed
immediately afterwards in Photius § 116, which makes numerous criti-
cisms of Metrodorus.
Ussher must have become aware of Metrodorus via an edition of
Photius (ex bibliotheca Photii § 115–16), since no other edition of the
work is known. Photius was Patriarch of Constantinople in the ninth
century, and published a catalogue of the library there.162
No.  XXVI: Germani episcopi  […] ad Hormisdam Romanam
Pontificem:163 The Suggestio referred to by Ussher was undoubtedly that
authored by Bishop Germanus, the deacons Felix and Dioscorus, and
the priest Blandus concerning the date of Easter in AD 520.164 After the
near-schism of the Symmachan papacy, Hormisdas’s reign marked a re-
turn to cordial relations with the East; the Suggestio of AD 520 may have
set in train the chain of events that led to Dionysius Exiguus’ composi-
tion of his Easter table.
No. XXVII: Johannis episcopi Constantinopolitanus Relatio ad eun­
dem: A reference to the letter that Bishop John of Constantinople ad-
dressed to Pope Hormisdas, announcing the date of Easter in AD 520.165
Instantis autem dominicae passionis numerum, qui vestris continetur lit­
teris, ex repositis apud nos diei paschalis annalibus recte habere significa­
mus. (‘The date of the current paschal Sunday, as it is contained in your
letter, we have given correctly from our records.’) Ussher clearly found it
in the same printed collection as the foregoing Suggestio.
No.  XXVIII: Dionysii Exigui epistola ad Petronium: In the year
AD  525, Dionysius Exiguus composed a 95-year table of paschal full
moons and Easter Sundays, covering the period that he (and we) num-
ber as AD (Anno Domini) 532 to 626, and addressed it, with a prefa-
tory letter, to a Bishop Petronius, otherwise unknown. Shortly after that
(and apparently in response to a request to explain his methodology),
Dionysius addressed a letter (No. XXXI in Ussher’s list; see below) to

162
Mosshammer (2008), 118.
163
For what follows, see especially Moshammer (2008), 59–71.
164
The Suggestio is ed. by Otto Günther in CSEL 35,2, 642; see Krusch (1926), 48
n. 1.
165
CSEL 35,2, ii 638; see Krusch (1926), 48 n. 2.


 DÁIBHÍ Ó CRÓINÍN

Boniface and Bonus, secretaries (primicerii) to Pope John, elaborating

on what he had said in the preface to his table. Various reasons have been
proposed to explain why Dionysius might have done such a thing; the
most likely one (suggested by Jones166) is that the current Alexandrian
(Cyrillan) table was due to expire in AD 531; someone may have ap-
proached Dionysius with a request to provide a new cycle, based on Al-
exandrian principles. Dionysius simply copied the last 19-year cycle of
that Cyrillan table (AD 513–31) and then extended it for 95 more years.
The only change he made to the rubrics of the Cyrillan table is that he
discarded the dating by the regnal year of Diocletian: ‘I did not wish to
preserve the memory of the impious persecutor in my cycles, but rather
chose to denote the times from the birth of Our Lord Jesus Christ’ (no­
luimus circulis nostris memoriam impii et persecutoris innectere, sed ma­
gis elegimus ab incarnatione domini nostri Iesu Christi annorum tempora
praenotare).167 From this simple act has come our Christian era (‘from
the Incarnation’).
The standard modern edition of Dionysius’ table and its accompany-
ing documents is by Bruno Krusch.168 However, of this Jones remarked:169
‘There is no real improvement in this edition, and in some ways a marked
retrogression’. Better by far is the 18th-century edition by Jan.170 ‘Jan’s edi-
torial judgments and suggestive notes are models for editors.’171
No. XXIX: Ejusdem cycli Paschales […] et continuatio Authoris, qui
anno 617 vixit: The paschal cycle referred to is Dionysius’ table discussed
above; the continuatio is the text that circulated anonymously in some
manuscripts, but that in others is ascribed to an unidentified Felix.172
The ‘Whitby Manuscript’ to which Ussher refers is Oxford, Bodleian
Library, Digby 63; it contains folios 70v–71r the text of the Continuatio,
under the heading Successor Dionisii. Whether Petavius did, in fact, use
Digby 63 is a moot point.173

166
Jones (1943), 68.
167
Dionysius Exiguus, Epistola ad Petronium (ed. by Krusch (1938), 64).
168
Krusch (1938), 59–87.
169
Jones (1943), 69 n. 6.
170
Jan (1718) (= PL 67, 19–23, 453–520).
171
Jones (1943), 68 n. 6.
172
On Felix, see Mosshammer (2008), 349–51; Jones (1943), 73–77; Cuppo
(2012).
173
There is a colour reproduction of Oxford, Bodleian Library, Digby 63, 6r, on the
front dust-jacket of Mosshammer (2008).


 ARCHBISHOP JAMES USSHER (1581–1656)

No. XXX: Aegytiorum regulae Paschales: To his 95-year table Dio-

nysius appended a set of (originally nine) rules (argumenta) for calen-
drical computations (e.g., how to determine the indiction, the year of a
lunar cycle, the year in a decennovenal cycle, and so on). The argumenta
offered a method of calculating the date of Easter, instead of simply
looking it up in a table.174 Though no such formulae have come down
from the Alexandrian church, there is no reason to doubt Dionysius’
statement that they represented ‘Egyptian learning’ (Necnon et argu­
menta Aegyptiorum sagacitate quaesita subdidimus; ‘In addition we have
subjoined formulae derived from Egyptian wisdom’).175 Jones rightly
remarked:176 ‘These argumenta of Dionysius, for which there is no ear-
lier extant parallel, stimulated western thought.’ They spawned a myriad
of mathematical algorithms, ranging over every conceivable (and some
inconceivable) problems,177 and set computists to wondering how the
mechanics of Easter tables, and related phenomena, actually worked.
They might be regarded as the beginnings of genuinely scientific math-
ematical speculation in the Middle Ages.
The reference in Ussher’s list to a continuation of the argumenta
down to the year AD 581 (cum additionibus pauculis, anno 581, ab alio
adjecti; ‘dated 581, with a number of additional years added by some-
one else’) relates to a set of the argumenta that is associated, in some
manuscripts, with Cassiodorus (No. XXXIII in Ussher’s list; see further
below). Ussher clearly recognised that these argumenta were additional
to Dionysius’ original set (of nine).
No. XXXI: Dionysii epistola ad primicerios notariorum: This is the
letter that Dionysius addressed to Boniface and Bonus, secretaries to
Pope John (AD 523–26), discussed above.
No.  XXXII: Bonifacii Primicerii (vel potius Boni Secundicerii) sug­
gestio: A memorandum prepared by the Vatican curial officials, Bonus
and Bonifatius, in AD 526, to advise Pope John on the proper method of
calculating Easter in that year. The Sirmond Codex (Oxford, Bodleian
Library, Bodley 309, 82v–84r), has the rubric: Exemplum suggestionis
boni sci. primice., clearly intended to introduce the work. However, the
text that immediately follows the rubric is the De sollemnitatibus (the

174
See now Warntjes (2010b).
175
Dionysius Exiguus, Epistola ad Petronium (Krusch (1938), 67).
176
Jones (1943), 71.
177
An argumentum in the Sirmond group of manuscripts reckoned the number of
momenta spent by Jonah in the belly of the whale (2,880)! See Ó Cróinín (1995), 212
n. 88.


 DÁIBHÍ Ó CRÓINÍN

final item in Ussher’s Catalogue, No.  XXXV below).178 The Suggestio

was published by Krusch from Oxford, Bodleian Library, Digby 63.179
Its purpose would have been to counsel Pope John as to the advisability
of accepting the Alexandrian (in this case, Dionysiac) reckoning.
No. XXXIII: Aurelii Cassiodori Senatoris computus Paschalis:180 It is
not necessary to enter here into the question of whether or not Cassi-
odorus was responsible for the Computus paschalis.
Ussher refers to a 1622 text of Cassiodorus. Presumably the Geneva
edition by Petrus and Iacobus Chouêt is meant.
No. XXXIV: Morini Alexandrini Disputatio: The sixth of ten com-
putistical texts cited as authorities by Cummian in his letter,181 the
work survives in eleven manuscripts, and is cited also in the De ratione
conputandi182 and the Munich Computus.183 Jones remarked:184 ‘The
evidence of the manuscripts points clearly to Ireland, and there is no re-
corded reference to it that is not probably Irish.’ In a footnote to that
comment he added that ‘the only references to it, so far as I can tell, are
the one in Cummian, and one in a work, De Bissexto, attributed errone-
ously to Alcuin, but actually Irish and pre-Bedan’.
Bucherius published the text from the Sirmond Codex, while du
Cange published it again in Chronicon paschale.185 Muratori published a
further edition from the Bobbio Computus, and Jan a partial one (from
Oxford, Bodleian Library, Digby 63, 79r–81r).186 Jones planned a new
edition, but stated that, having collated all the manuscripts, and ‘in spite
of access to many variants, I cannot understand the work’!187

178
See above, p. 10.
179
Krusch (1926).
180
The Computus paschalis of AD 562 is ed. by Lehmann (1912), 52–55. See Neu-
gebauer (1982); Warntjes (2010b), 43–44, 57, 67–68.
181
Cummian, De controversia paschali l. 213 (Walsh and Ó Cróinín (1988), 86–
87); see Walsh and Ó Cróinín (1988), 40–41.
182
De ratione conputandi 86 (Walsh and Ó Cróinín (1988), 194).
183
Munich Computus 11, 36, 38 (ed. by Warntjes (2010a), 42, 96, 102); see Warn-
tjes (2010a), LXIV–LXV.
184
Jones (1943), 97.
185
du Cange (1688), 480–81 (Appendix 23; from Paris, Bibliothèque nationale de
France, Lat. 4860, 150r–v).
186
Muratori (1697–1713), iii = PL 129, 1357–58 (Bobbio Computus 146 = Mi-
lan, Biblioteca Ambrosiana, H 150 inf., 80r); Jan (1718) = PL 67, 460.
187
Jones (1943), 97 n. 2.


 ARCHBISHOP JAMES USSHER (1581–1656)

The purpose of the text is exceedingly obscure and has defied expla-
nation by the best modern computists (it is not mentioned by Moss-
hammer). It appears to favour the Victorian reckoning (Passion on luna
14 and Resurrection on luna 16), and it is cited to that effect by the
author of De ratione conputandi. A passage at the end of the tract has
been interpreted as providing a date of composition in either AD 604
or AD 632, and Jones even suggested that this section might have been
penned by Cummian.188 But since the premises on which these deduc-
tions are made are, in fact, groundless, and since the passage concerned
does not, I believe, belong to the original tract, the arguments for dating
it to the early seventh century (in Ireland?) fall to the ground.189
Eric Graff has argued persuasively for an early fourth-century Galli-
can origin (suggesting Marinus, bishop of Arles, as author, c.AD 313).190
Ussher does not cite either Bucherius or Petavius for this text, but seems
to imply that he had it directly from an unidentified manuscript. He
could have found it in the Cotton collection (London, British Library,
Cotton Caligula A XV, 82v–83v).
No. XXXV: De sollemnitatibus et sabbatis et neomeniis: This is the text
we discussed above (pp. 320–24). Though he gives no date for this work
(as he had done for every previous text in the list, with the notable excep-
tion of Morinus, De ratione paschali, No. XXXIV above), it is striking
that Ussher placed it last in his list, seeming to indicate that he believed
it to be a seventh-century composition, rather than a patristic or post-
patristic one. In this he has been vindicated by modern scholarship.

Conclusion
What is most impressive about Ussher’s list, looking at it from start to
finish, is the way in which it demonstrates, to an extraordinary degree,
the full extent of the paschal controversy in its chronological range, and
the astonishing grasp that Ussher had of the sources for that controversy.
We have seen how his interest in the subject must have been well devel-

188
Jones (1943), 97–98 n. 2.
189
Walsh and Ó Cróinín (1988), 41.
190
Graff (2010), with an edition 141–42. See earlier discussion in Cordoliani
(1945–46), 28–34; Grosjean (1946), 225–39. See also Strobel (1984), 120, who argued
for a Gallican origin, suggesting the sixth-century Marius, bishop of Avenches, as possi-
ble author. Marius is well known as the author of a chronicle, ed. in PL 72, 793–802 and
now by Favrod (1991).


 DÁIBHÍ Ó CRÓINÍN

oped already when he published his Discourse of the Religion professed by

the Ancient Irish (1613), and again with his Veterum epistolarum Hiber­
nicarum sylloge (1632), and how that initial interest had developed into
a thorough mastery of the materials already by the time he came to pub-
lish his Britannicarum Ecclesiarum Antiquitates (1639). In the Sylloge he
published the editio princeps of Cummian’s letter—the foundation-stone
of all later discussion of the Easter controversy in Ireland; the Antiqui­
tates demonstrates how his researches in the field had come to take in the
entire sweep of patristic, post-patristic, and early medieval computistical
writings.
There can be little doubt that Ussher’s initiation into the subject
had come from his acquisition of the two greatest monuments of sev-
enteenth-century technical chronology, Petavius’s Opus de doctrina tem­
porum and Bucherius’s De doctrina temporum, as well as the six volumes
of Canisius’s Antiquae lectiones. But he had clearly moved beyond the
initial stage of merely gathering source material for the history of the
early Irish church. The evidence of his correspondence with other schol-
ars, at home and abroad, shows that he was anxious to set the early Irish
experience against the wider backdrop of the universal Christian church,
and for that purpose he needed to accumulate as large a corpus of origi-
nal documentary sources as he could lay his hands on. He may not have
had in mind—initially, at any rate—the creation of a universal system of
world history and chronology, such as the later Enlightenment scholars
had derided in Scaliger,191 but there can be no doubt that he was piqued
by the challenge of trying to reduce the seeming chaos of earlier centu-
ries to some kind of strict chronological order.
Scaliger had compiled the seven volumes of his De emendatione
temporum deliberately, as he said himself, ‘to stir up the lethargy of our
chronologers, who seem all to have sworn never to tell the truth’.192 An-

191
Anthony Grafton (1975), 157, cites a passage from Giambattista Vico, The New
Science, as follows: ‘These two marvellous geniuses, Joseph Justus Scaliger and Denis Pe-
tau, with their stupendous erudition […] failed to begin their doctrine at the beginning
of their subject matter. For they began with the astronomical year, which […] was un-
heard of among the nations for a thousand years, and in any case could have assured them
only of conjunctions and oppositions of constellations and planets in the heavens, and
not of any of the things that had happened here on earth nor of their sequence […] And
on this account their work has shed little light on the beginnings or on the continuation
of universal history.’
192
Scaliger and Heinsius (1627), 90: Habeo in manibus opus affectum de Emen­
datione temporum in quo veterum excito omnium nostrorum Chronologorum, qui omnes
ὁμοθυμαδòν in unum id jurasse videntur, ut verum numquam dicant; Grafton (1975), 158
n. 9.


 ARCHBISHOP JAMES USSHER (1581–1656)

thony Grafton has described Petavius as ‘clearer-headed than Scaliger’,

while Alden Mosshammer described him as ‘Scaliger’s most intelligent
imitator’.193 But the comparison is neither fair nor valid, as Petavius, in
fact, published the first five books of his Opus de doctrina temporum to
refute Scaliger’s work,194 in the process dismissing Scaliger as a ‘halluci-
nator’ and a ‘mendacious inventor’.195 Ussher does not engage in any such
polemic in his own publications (apart from the occasional blast against
perceived Catholic errors!); that was to come only in his Annales. For
the purpose of reconstructing the history of the paschal controversy in
the early centuries of the church, however, Ussher stuck strictly to the
texts. We have every reason to be thankful that he did so.

Appendix: Oxford, Bodleian Library, Smith 31 (pt 3),

pp. 70–73
(3)  Veterum Scripta Paschalia per Jacobum Usserium Armachanum |p. 70

Anno I. Acta Caesariensis synodi sub Theophilo habita, sive

Domini epistola Philippi de Pascha dubia admodum fidei:
196. pag. 469. edit. Bucherii, et in apographo meo MS.196
196. II. Victoris Papae epistolae tres paschales supposititiae:
una ad Theophilum, quae in Isidori Mercatoris
Collectione habetur: Duae ad Viennenses episcopos
Desiderium (al. Dionysium) et Paracodam: ad
calcem bibliothecae Floriacensis Johannis à Bosco,
pag. 24–26.
222. III. Hippolyti Portuensis episcopi Canon Paschalis
ἐκκεαδεκαετηρικὰ Graecè a Scaligero lib. vii. de
emendatione temporum p. 677, Latinè ab Aegidio
Bucherio, pag. 295–96 editus et commentario
illustratus.

193
Grafton (1975), 174; Mosshammer (2008), 10.
194
Mosshammer (2008), 10.
195
See Conley (2004), 176, cited in Mosshammer (2008), 10, who describes Petavi-
us’s Opus as ‘a far superior work’. Scaliger had attracted equal vituperation from John
Donne, as Grafton (1975), 160 and n. 18, remarks.
196
[d]egenui | a et Eusebio added in left margin.


 DÁIBHÍ Ó CRÓINÍN

243. IV. Anonymi Scriptum Paschale ad hekkadecateridae

rationem accommodatum et ad annum v Gordiani:
(consulatumque Arriani et Papi perductum: cujus
initium [Multo quidem tempore]197 ex antiquissimo
MSº Cottoniano desumptum: cujus chronologia
conferenda cum chronico anno 13º Alexandri
Imperatoris (Christianae epochae 243) scripti. tomi
secundi. Antiquae lectionis Henrici Canisii: (et
videndum anne idem utriusque operis author)
[2]60. V. Dionysii Alexandrini epistola ad Basilidem Pentapo-
lis episcopum, de Dominicae Resurrectionis
tempore, et Paschalis jejunii termino. Graeco-Latinè.
[26]6. VI. Anatolii Laodiceni episcopi Canon Paschalis:
pag. 439 (edit. Bucherii) dubia fidei.
[3]25. VII. Constantini M. Imperatoris epistolae duae: una ad
Alexandrinum, alia ad Eusebium Caesariensem.
Graeco-Latinè.
[357]. VIII. /Latinorum/ Paschalis laterculus centum annorum.
editus a Bucherio Paschalis laterci [struck out].
pag. 252–59 (cum 256. et 264.) |p. 71)
[357]. IX. Ambrosii Mediolanensis episcopi epistola (83a) ad
episcopos per Aemiliam constitutos: recensita a
Bucherio pag. 474. Nicolao Mulerio, et Francisco
Balbo.
387. X. Johannis Chrysostomi presbyteri (corr. <episcopi)
Antiocheni. Sermo Paschalis, tomo v. edit. Saviliana
p. 940.
388. XI. Pachomii Abbatis Laterculum Quadragesimale et
Paschale quod ab Angelo accepisse dicitur.
388. XII. Theophili Alexandrini prologus in laterculum
Paschalem C. annorum (iuxta Alexandrinos ad
Theodosium Seniorem Aug. /Latinè/ editus a
Bucherio p. 471. Graecè et Latinè a Petavio. tom. 2
pag [].
395. XIII. Chronographia cyclis 1820 annorum distincta ad
Vitalem quendam, in Olybrii et Probini consulatu
scripta ex MS.º Thomas Alani.
401. XIV. Theophili Alexandrini Paschalium epistolarum
trium epilogi: (ex operibus Hieronymi.)

197
Square brackets added later.


 ARCHBISHOP JAMES USSHER (1581–1656)

412. XV. Synesii Cyraenaei Ptolemaidis episcopi epistolia

duo: 9 et 13, ad Theophilum Alexandrinum et
Petrum Presbyterum.
414. XVI. Innocentii I. epistola ad Aurelium Carthaginensem,
episcopum: edit. Romana. p. 62. emendatius verò à
P. Crabb edita: tom. 1. Concil. fol. 272 et Bucherio
p. 48[0?]. (collat. cum pag. 158–59. de tempore,
quo scripta est)
414. XVII. Cyrilli Alexandrini episcopi, xxix Sermonum
Paschalium epilogi: ab anno 413. ad 442.
419. XVIII. Ejusdem epistola ad Synodum Carthaginiensem
cum adjectâ lacinia, quam ad annum 607. refert
Petavius. tom. 1. pag. 221 et 2. p. 884. 893. 894 sed
Cyrilli asserere conatur Bucherius p. 119 et 484.) et
appendice aliâ in Cottoniano MS.º eidem subjectâ:
cujus initium, oportuno tempore Dominus Jesus.
|p. 72)
437. XIX Ejusdem Cyrilli prologus in cyclum Paschalem
XCV annorum: (a Petavio editus. tom. 2. p. [] et
Bucherio p. 481.) Quintus illius cyclus ex Dionysio
Exiguo infra habetur descriptus: quibus priorum 4
ex Isidori Hispalensis libris Originum adjici etiam
possunt: (confirmati ex Epilogis Sermonum
Paschalium, de quibus supra ann. 414.
443. XX. Paschasini Lilybetani episcopi ad Leonem Papam
epistola MS.198 et edit. a Bucherio p. 75.
453. XXI. Leonis I. epistolae.
454. XXII. Proterii Alexandrini ad Leonem I. de ratione
Paschali epistola Dionysio Exiguo interprete.
457. XXIII. Hilari Romani archidiaconi ad Victorium Aquitani-
cum epistola: edita a Bucherio. p. 1.
457. XXIV. Victorii Aquitanici ad Hilarum Rescriptum: cum
prologo in periodum suam Paschalem annorum
532. edit. a Bucherio a pag. 2. ad 10.
520. XXV.199 Metrodori (?) cyclus annorum 532. cum alterius
Graeci in eum censura; ex bibliothecâ Photii
§ 115–16.

198
The abbreviation ‘MS’ appears to have been added.
199
XXV corr. <XXIV.


 DÁIBHÍ Ó CRÓINÍN

520. XXVI. Germani episcopi. Felicis et Dioscori Diaconorum,

et Blandi. Presbyteri suggestio ad Hormisdam
Romanum Pontificem; tom. 1. Decretal. epist. edit.
Rom. p. 530.
520. XXVII. Johannis episcopi Constantinopolitanus Relatio ad
eundem: ibid. pp. 529–30.
525. /XXVIII./ Dionysii Exigui epistola ad Petronium. ex MS. et
edit. Petavii et Bucherii p. 485.
XXIX. Ejusdem cycli Paschales: quibus praeponendus
titulus ex codice Regio apud Petavium. pp. 889–90.
et subicienda continuatio Authoris, qui anno 617
vixit, ex MS Whitbeiensi vide Petavi. tom. 2. p. 404.
XXX. Aegyptiorum regulae Paschales ad annum 525 ab
eo | p. 73) adaptatae. cum additionibus pauculis,
anno 581. ab alio adjecti.
526. XXXI. Ejusdem Dionysii epistola ad primicerios notario-
rum. ex MS. et edit. Bucherii. p. 489.
526. XXXII. Bonifacii Primicerii (vel potius Boni Secundicerii)
notariorum Suggestio ad Jo[h]annem Papam200 de
ratione Paschali. MS.
562. XXXIII. Aurelii Cassiodori Senatoris computus Paschalis.
pag. 1313 operum, edit. Anno 1622.
XXXIV. Morini Alexandrini disputatio de ratione Paschali.
MS.
XXXV. De sollemnitatibus et sabbatis et neomeniis
legalibus Anonymi tractatus, ex MS. Cottoniane.

Exscripsi ex libro autographo Adversariorum J. Usserii. Nov. xxv.

1697. T[homas] S[mith].

Timothei computus201
Graeco-Latino v. tom. 3. Petavii.

200
Add. in r. marg.: sed de hoc consule in a later hand.
201
This list of titles looks like an afterthought by Ussher. The first may possibly refer
to the text called Liber Timothei episcopi de Pascha discussed by Siegmund (1949), 64–65
n. 2. Siegmund suggests it may be identical with the Codex sancti Augendi that Krusch
(1880), 210 (with reference to Bucherius (1634), 494), mentions as having seen. He re-
fers to Montpellier, Bibliothèque de la Faculté de Médecine, 157, which has this text,
together with others of the Alexandrian school, including three letters of Theophilus
(as translated by Jerome), an unidentified B. Gregorii papae de eadem sollemnitate, and
excerpts from Eusebius, Historia ecclesiastica 6–7 concerning Easter. It also has a copy of
Ceolfrid’s letter to the Pictish king, Nechtan (though the name is garbled: ep. Goffridi)
and excerpts from Bede’s Historia ecclesiastica; see Jones (1943), 160.


 ARCHBISHOP JAMES USSHER (1581–1656)

Pii Papae I. epistola.202

Victoris Papae epistolae tres.203
Caesariensis Synodi Acta.204
Ejusdem fragmenta ex Eusebio.205
Hippolyti canon.206

Fols 10v–40v of the Montpellier MS have a series of letters, introduced by the

words Incipiunt epistolae paschales Theophili Alexandrinae urbis episcopi ad totius Egyp­
tii episcopos, followed by the three letters: 1) inc. Sollemnitatis augustae sermo diuinus
(10v–22v); 2) inc. Christum Ihesum Dominum glorie fratres karissimi rursum consona
uoce (22v–31v), and 3) inc. Nunc quoque Dei uiua sapientia (31v–40v).
I cannot account for the additional reference: Graeco-Latino v. tom. 3. Petavii;
Petavius was published in 2 volumes.
202
Alden Mosshammer writes (personal communication): ‘It is apparently a me-
dieval ‘urban myth’ that Pius I issued a decree that Easter should be observed only on
a Sunday. An early version of the story is in Bede’s Chronicle at 4113 (Wallis, p. 202)
Scaliger debunked it, Animadversiones in Chronologica Eusebii [Leiden 1606] 219. The
letter is among those forged in the Pseudo-Isidorian decretals (see PL 130, 111). The
text of the alleged letter is quoted in Gratian’s Decretum (composed about 1150), De
Consecratione: Distinctio iii, cap. XXI as follows: Cap. XXI. Non nisi die dominico sanc­
tum Pasca celebretur. Item Pius Papa, IX. a Petro Apostolo, in primo suorum decretalium.
Nosse uos uolumus, quod Pasca Domini die dominico annuis temporibus sit celebrandum.
Istis enim temporibus Hermes doctor fidei et scripturae inter nos effulsit. Et licet nos idem
Pasca predicto die celebraremus, tamen quia quidam, inde dubitabant ad corroborandas,
animas eorum eidem Hermae angelus Domini in habitu pastoris apparuit, et precepit ei,
ut Pasca Domini die dominico ab omnibus celebraretur. Unde et nos apostolica auctoritate
institiumus, omnes eadem seruare debere, quia et nos eadem seruamus, non debetis a capite
quoquo modo discedere. This can be found in col. 1358 of Corpus Iuris Canonici, 1, new
ed. by Aemilius Friedburg [Leipzig 1879; repr. Graz 1959].’
203
For the three letters of Pope Victor, see above, No. II.
204
This must be the Acta Synodi, above No. I.
205
Perhaps a reference to the specific text in Eusebius, Historia ecclesiastica 5.25,
where he quotes from the end of a letter sent by those present at the Council of Caesarea.
206
Presumably another reference to the Hippolytan table, No. III above.


 DÁIBHÍ Ó CRÓINÍN

Plate 1 Oxford, Bodleian Library, MS. Smith 31, p. 70. The opening page of Thomas Smith’s
list of Scripta Paschalia (items I–VIII) copied in 1697 from a lost MS. of Archbishop James
Ussher’s.


 ARCHBISHOP JAMES USSHER (1581–1656)

Plate 2 Oxford, Bodleian Library, MS. Smith 31, p. 71. Ussher’s items IX–XVIII.


 DÁIBHÍ Ó CRÓINÍN

Plate 3 Oxford, Bodleian Library, MS. Smith 31, p. 72. Ussher’s items XIX–XXX.


 ARCHBISHOP JAMES USSHER (1581–1656)

Plate 4 Oxford, Bodleian Library, MS. Smith 31, p. 73. End of Ussher’s list (items XXXI–
XXXV), with dating-clause.


 BIBLIOGRAPHY

a Bosco, J. (1605a) Antiquae Sanctae ac Senatoriae Viennae Allobrogum Gallico-

rum Sacrae et profanae plurimae Antiquitates, Lyon.
— (1605b) Floriacensis vetus bibliotheca, Benedictina, Sancta, Apostolica, Pon-
tificia, Caesarea, Regia, Franco-Gallica, Lyon.
Allen, M. I. (1998) ‘The Chronicle of Claudius of Turin,’ in A. C. Murray, Af-
ter Rome’s fall: narrators and sources of early medieval history, Toronto,
288–319.
— (2003) ‘Universal history 300–1000: origins and western developments,’ in
D. M. Deliyannis, Historiography in the Middle Ages, Leiden, 17–42.
Alturo, J. (1994a) ‘La escritura visigótica de origen transpirenaico: una approxi-
mación a sus particularidades,’ Hispania sacra 46, 33–64.
— (1994b) ‘El glosario del manuscrito en excritura visigótica Paris, BN, lat.
609,’ Euphrosyne 22, 185–200.
Alvarez Miranda, D. J. (1928) Antiphonarium mozarabicum de la catedral de
León, León.
Antolin, P. G. (1910–23) Catalogo de los códices latinos de la real biblioteca del
Escorial, 5 vols, Madrid.
Arbeiter, A. and S. Noack-Haley (1999) Christliche Denkmäler des frühen Mit-
telalters vom 8. bis 11. Jahrhundert, Mainz.
Aris, M.-A. et al. (2004) ‘Fulda, St Salvator,’ in F. Jürgensmeier, F. L. Büll, and
R. Schwerdtfeger, Germania Benedictina, vol. 7: Die benediktinischen
Mönchs- und Nönnenklöster in Hessen, St Ottilien, 213–434.
Aubineau, M. (1966) ‘Les 318 serviteurs d’Abraham (Gen., XIV, 14) et le nom-
bre des pères au Concile de Nicée (325),’ Revue d’histoire ecclésiastique
61, 5–43.
Barley, M. and R. P. C. Hanson (1968) Christianity in Britain, 300–700, Leicester.
Barlow, T. (1660) Jacobi Usserii Armachani Chronologia sacra, Oxford.
Barney, S. A. (2006) The Etymologies of Isidore of Seville, Cambridge.
Baronius, C. (1588–1607) Annales ecclesiastici, 12 vols, Rome.


 Bibliography

Barry, J. and H. Morgan (2013) Great deeds in Ireland: Richard Stanihurst’s De

Rebus in Hibernia Gestis, Cork.
Bayless, M. (1998) ‘The Collectanea and medieval dialogues and riddles,’ in
M. Bayless and M. Lapidge, Collectanea pseudo-Bedae, Dublin, 13–24.
Bellagente, E. (1999) ‘La Chronica de sex aetatibus di Claudio vescovo di To-
rino,’ Aevum 73, 237–46.
Bernard, N. (1656) The life and death of the most reverend and learned father of
our Church, Dr James Usher, late Archbishop of Armagh and Primate
of all Ireland, London.
— (1697) Catalogus Manuscriptorum Bibliothecarum Angliae, Oxford.
Bernoulli, C. A. (1895) Hieronymus und Gennadius, De viris inlustribus, Leip-
zig.
Bieler, L. (1949) The life and legend of St Patrick: problems of modern scholarship,
Dublin.
— (1952) Libri epistolarum Sancti Patricii episcopi, Dublin.
— (1979) The Patrician texts in the Book of Armagh, Dublin.
Binchy, D. (1962) ‘Patrick and his biographers, ancient and modern,’ Studia
Hibernica 2, 7–173.
Bisagni, J. and I. Warntjes (2008) ‘The Early Old Irish material in the newly
discovered Computus Einsidlensis (c.AD 700),’ Ériu 58, 77–105.
Bischoff, B. (1966–81) Mittelalterliche Studien: ausgewählte Aufsätze zur
Schriftkunde und Literaturgeschichte, 3 vols, Stuttgart.
Bischoff, B. (1990) Latin palaeography: Antiquity and the Middle Ages, trans. by
D. Ó Cróinín and D. Ganz, Cambridge.
Bitel, L. (1990) Isle of saints: monastic settlement and Christian community in
early Ireland, London.
Blackburn, B. and L.  Holford-Strevens (1999) The Oxford companion to the
year: an exploration of calendar customs and time-reckoning, Oxford.
Blake, M. (2009) Aelfric’s De temporibus anni, Cambridge.
Bonnaz, Y. (1987) Chroniques asturiennes, Paris.
Boran, E. (2009) ‘Ussher and the collection of manuscripts in Early Modern
Europe,’ in J. Harris and K. Sidwell, Making Ireland Roman: Irish Neo-
Latin writers and the Republic of Letters, Cork, 176–94.
— (2015) The correspondence of James Ussher, 1600–1656, 3 vols, Dublin.
Borst, A. (1993) ‘Alkuin und die Enzyklopädie von 809,’ in P. L. Butzer and
D. Lohrmann, Science in western and eastern civilization in Carolin-
gian times, Basel, 53–75.
— (1998) Die karolingische Kalenderreform, Hannover.
— (2001) Der karolingische Reichskalender und seine Überlieferung bis ins 12.
Jh., 3 vols, Hannover.
— (2006) Schriften zur Komputistik im Frankenreich von 721 bis 818, 3  vols,
Hannover.
Bouchier, G. (1655) Belgium Romanum ecclesiasticum et civile, Liège.


 Bibliography

Boulhol, P. (2002) Claude de Turin: un évêque iconoclaste dans l’occident caro-

lingien, Paris.
Brehaut, E. (1912) An encyclopedist of the Dark Ages: Isidore of Seville, New
York.
Brent, A. (1995) Hippolytus and the Roman church in the third century: com-
munities in tension before the emergence of a monarch-bishop, Leiden.
Bronisch, A. P. (2005a) Die Judengesetzgebung im katholischen Westgotenreich
von Toledo, Hannover.
— (2005b) ‘Die westgotische Reichsideologie und ihre Weiterentwicklung im
Reich von Asturien,’ in F.-R. Erkens, Das frühmittelalterliche König-
tum, Berlin, 161–89.
Brou, L. and J. Vives (1959) Antifonario visigótico mozárabe de la catedral de
León, Barcelona.
Brown, M. P. (1996) The Book of Cerne: prayer, patronage, and power in ninth-
century England, Toronto.
Bruin, F. (1977) ‘The first visibility of the lunar crescent,’ Vistas in Astronomy 21,
331–58.
Brunhölzl, F. (1975) Geschichte der lateinischen Literatur des Mittelalters, vol. 1:
von Cassiodor bis zum Ausklang der karolingischen Erneuerung, Munich.
Bucherius, A. (1634) De doctrina temporum commentarius in Victorium Aqui-
tanum, Antwerp.
Buick-Knox, R. (1967) James Ussher, archbishop of Armagh, Cardiff.
Bullough, D. A. (1960) ‘‘Christe fave votis’,’ Scriptorium 14, 346–48.
Burgess, R. W. and M. Kulikowski (2013) Mosaics of time, vol. 1: Historical in-
troduction to the chronicle genre from its origins to the high Middle Ages,
Turnhout.
Burrows, T. (1987) ‘Holy information: a new look at Raban Maur’s De naturis
rerum,’ Parergon: Bulletin of the Australian and New Zealand Associa-
tion for Medieval and Renaissance Studies n.s. 5, 28–37.
Bury, J. B. (1905) The life of St Patrick and his place in history, London, repr.
1971.
Byrne, F. J. (1967) ‘Seventh-century documents,’ Irish Ecclesiastical Record 108,
164–82.
Cabrol, F. and H. Leclercq (1900–13) Reliquiae liturgicae vetustissimae ex ss.
patrum necnon scriptorum ecclesiasticorum monumentis selectae, 2 vols,
Paris.
Caillau, A. B. and M. N. S. Guillon (1829) Origines commentariorum in epis-
tolam sancti pauli ad romanos (interprete rufino), Paris.
Calder, G. (1917) Auraicept na n-éces: the scholars’ primer: being the texts of the
Ogham tract from the Book of Ballymote and the Yellow Book of Le-
can, and the text of the Trefhocul from the Book of Leinster, Edinburgh,
2nd ed. Dublin 1995.


 Bibliography

Camón Aznar, J. (1960) ‘La miniatura española en el siglo X,’ Spanische

Forschungen der Görres-Gesellschaft 16, 16–36.
— (1975) ‘El arte en los Beatos y el códice de Gerona,’ in Beati in Apocalypsin
Libri duodecim: Codex Gerundensis, Madrid, 109–69.
Camps i Sòria, J. (2000) Catalunya en la época carolingia: arte y cultura antes del
románico (siglos IX y X), Barcelona.
Canisius, H. (1601–04) Antiquae lectiones, seu antiqua monumenta ad histo-
riam mediae aetatis illustrandam, 6 vols, Ingolstadt.
Capelli, A. (1962) The elements of abbreviation in medieval Latin paleography,
trans. by D. Haiman and R. Kay, Lawrence.
Carney, J. (1958) ‘A new chronology of the Saint’s life,’ in Ryan (1958), 24–37.
— (1961) The problem of St Patrick, Dublin, repr. 1973.
Charles-Edwards, T. M. (2000) Early Christian Ireland, Cambridge.
Claude, D. (1970) Geschichte der Westgoten, Stuttgart.
— (1971) Adel, Kirche und Königtum im Westgotenreich, Sigmaringen.
Colgrave, B. and R. A. B. Mynors (1969) Bede’s Ecclesiastical history of the Eng-
lish people, repr. 1991, Oxford.
Collins, R. (1995) Early medieval Spain: unity in diversity, 400–1000, 2nd ed.,
New York.
— (2004) Visigothic Spain 409−711, Malden.
Collison, R. (1964) Encyclopedias: their history throughout the ages, New York.
— (2010) ‘History of encyclopaedias,’ Encyclopaedia Britannica, 15th ed., xviii
271–77.
Colson, F. H. (1926) The week: an essay on the origin and development of the
seven-day cycle, Cambridge.
Colvener, G. (1627) Magnentii Hrabani Mauri […] opera, quae reperiri po-
tuerunt, omnia, 6 vols, Cologne.
Conley, T. (2004) ‘Vituperation in early seventeenth-century historical studies,’
Rhetorica 22, 169–82.
Contreni, J. J. (2005) ‘Bede’s scientific works in the Carolingian age,’ in S. Leb-
ecq, M. Perrin, and O. Szerwiniack, Bède le Vénérable entre tradition et
postérité, Villeneuve d’Ascq, 247–59.
Cordoliani, A. (1942) ‘Études du comput: un texte de comput espagnol du
VIIIe (?) siècle,’ Bibliothèque de l’école des chartes 103, 65–68.
— (1945–46) ‘Les computistes insulaires et les écrits pseudo-alexandrins,’ Bib-
liothèque de l’école des chartes 106, 5–34.
— (1951) ‘Los tratados y figuras de los códices ‘Aemilianensis’ y ‘Vigilianensis’
y el tratato de cómputo y calendaria antiguo de la Iglesia de España,’
Revista bibliográfica y documental: archivo genereal de erudición his-
pánica 5, 117–52.
— (1956) ‘Textes du comput espagnol du VIe siècle: encore le problème des
traités de comput de Martin de Braga,’ Revista de archivos, bibliotecas
y museos 62, 685–97.


 Bibliography

— (1961) ‘Contribution à la littérature du comput ecclésiastique au moyen âge,’

Studi medievali, Serie 3 2, 169–208.
Corning, C. (2006) The Celtic and Roman traditions: conflict and consensus in
the early medieval church, New York.
Cunningham, B. and R. Gillspie (2004–05) ‘James Ussher and his Irish manu-
scripts,’ Studia Hibernica 3, 81–99.
Cunningham, J. (2007) James Ussher and John Bramhall: the theology and poli-
tics of two Irish ecclesiastics of the seventeenth century, Aldershot.
Cuppo, L. (1998) ‘Roma magistra historiae: the year 680 as caput saeculi in Cas.
641,’ in J. Hamesse, Roma magistra mundi: Itineraria culturae medi-
evalis, Turnhout, 121–36.
— (2011) ‘Felix of Squillace and the Dionysiac computus I: Bobbio and North-
ern Italy (MS Ambrosiana H 150 inf.),’ in Warntjes and Ó Cróinín
(2011), 110–36.
— (2017) ‘Text and context: the annotations in Verona, Biblioteca Capitolare
XXII (20),’ in M. Teeuwen and I. van Renswoude, The annotated book
in the early Middle Ages: practices of reading and writing, Turnhout.
Curtius, E. R. (1953) European literature and the Latin Middle Ages, trans. by
W. R. Trask, London.
Cutler, A. (1965) ‘The ninth-century Spanish martyrs’ movement and the ori-
gins of western Christian missions to the Muslims,’ Muslim World 55,
321–39.
de Abadal y de Vinyals, R. (1949) La batalla del adopcionismo en la desinte-
gración da la Iglesia visigoda: discurso leído en la recepción, Barcelona.
de Aquino, A. (1591) Epistolae decretales summorum pontificum, 3 vols, Rome.
Declercq, G. (2000) Anno Domini: the origins of the Christian era, Turnhout.
— (2000a) Anno Domini: les origines de l’ère chrétienne, Turnhout.
— (2002) ‘Dionysius Exiguus and the introduction of the Christian era,’ Sacris
Erudiri 41, 165–246.
de Ferreras, J. (1700–27) Synopsis historica chronologica de España, 16 vols, Ma-
drid.
de Ghellinck, J. (1939) Littérature latine au moyen âge, Paris, repr. Hildesheim
1969.
de la Fuente, V. (1855–59) Historia eclesiástica de España o adiciones a la historia
general de la iglesia, 4 vols, Barcelona.
de Palol, P. (1986) El tapís de la creació de la catedral de Girona, Barcelona.
de Palol, P. und G. Ripoll (1999) Die Goten: Geschichte und Kunst in Westeu­
ropa, Augsburg.
de Paor, L. (1993) Saint Patrick’s world, Dublin.
de Paor, M. (1998) Patrick, the pilgrim apostle of Ireland, Dublin.
de Paor, M. and L. de Paor (1964) Early Christian Ireland, London.
de Rossi, G. B. (1857) Inscriptiones christianae urbis Romae septimo saeculo an-
tiquiores: volumen primum, prolegomena, Rome.


 Bibliography

Deswarte, T. (2013) ‘Polygraphisme et mixité graphique: note sur les additions

d’Arias (1060–70) dans l’Antiphonaire de León,’ Territorio, sociedad y
poder 8, 67–84.
Díaz y Díaz, M. C. (1972) Liber de ordine creaturarum: un anónimo Irlandés del
siglo VII, Santiago de Compostela.
— (1983) Códices visigóticos de la monarquia leonesa, León.
— (1990) ‘Mis experiencias en el campo de las abreviaturas visigóticos,’ in Las
abreviaturas en la enseñanza medieval y la transmisión del saber, Bar-
celona, 305–11.
— (2007) ‘Some incidental notes on music manuscripts,’ in S. Zapke, Hispa-
nia vetus: manuscritos litúrgico-musicales de los orígenes visigóticos a la
transición franco-romana (siglos IX–XII), Bilbao, 93–112.
Dorfbauer, L. (2013) ‘Der Genesiskommentar des Claudius von Turin, der
pseudoaugustinische Dialogus questionum und das wisigotische In-
texuimus,’ Revue d’histoire des textes, n.s. 8, 269–306.
Dubler, C.  E. (1962) ‘Voraussetzungen der Epoche des Beatus von Liébana,’
in J. Marqués Casanovas et al., Sancti Beati a Liébana in apocalypsin
codex gerundensis, Olten, 8–26.
du Cange, C. (1688) Paschalion seu chronicon paschale: a mundo condito ad
Heraclii imperatoris annum vicesimum, Paris.
Duchesne, L. (1886–92) Le Liber pontificalis: texte, introduction et commen-
taire, 2 vols, Paris.
Dümmler, E. (1895) ‘Über Leben und Lehre des Bischofs Claudius von Turin,’
Sitzungsberichte der Königlich Preussischen Akademie der Wissenschaf-
ten zu Berlin 23, 427–43.
Dumville, D. N. (1983) ‘Motes and beams: two insular computistical manu-
scripts,’ Peritia 2, 248–56.
— (1993) ‘Bishop Palladius’s computus,’ in D.  N. Dumville, Saint Patrick –
A.D. 493–1993, Woodbridge, 85–88.
Duncan, D. E. (1998) The calendar: the 5000-year struggle to align the clock and
the heavens – and what happened to the missing ten days, London.
Elrington, C. L. R. (1847–64) The whole works of the Most Rev. James Ussher,
Lord Archbishop of Armagh, and Primate of all Ireland, 17 vols, Dublin.
Embach, M. (2007) Die Kreuzesschrift des Hrabanus Maurus De laudibus sanc-
tae crucis, Trier.
Englisch, B. (2002) Zeiterfassung und Kalenderprogrammatik in der frühen
Karolingerzeit: das Kalendarium der Hs. Köln DB 83–2 und die Syno­
de von Soissons 744, Sigmaringen.
— (2010) ‘Karolingische Reformkalender und die Fixierung der christlichen
Zeitrechnung,’ in Warntjes and Ó Cróinín (2010), 238–58.
— (2011) ‘Ostern zwischen Arianismus und Katholizismus: zur Komputistik
in den Reichen der Westgoten im 6. und 7. Jh.,’ in Warntjes and Ó
Cróinín (2011), 76–109.


 Bibliography

Ernst, J. (1896) ‘Analekten,’ Zeitschrift für katholische Theologie 20, 360–62.

— (1898) ‘Zur handschriftlichen Überlieferung des Liber de Rebaptismate,’
Zeitschrift für katholische Theologie 22, 179–80.
Espenak, F. (2007) Six millennium catalog of phases of the moon, http://eclipse.
gsfc.nasa.gov/phase/phasecat.html.
Esposito, M. (1930) ‘Notes on Latin learning and literature in medieval Ireland:
1,’ Hermathena 20, 225–60, repr. in M.  Esposito, Latin learning in
medieval Ireland, ed. by M. Lapidge, London 1988, article IV.
Etchingham, C. (1999) Church organisation in Ireland A.D. 650 to 1000, May-
nooth.
Faber, E. (2014) Von Ulfila bis Rekkared: Die Goten und ihr Christentum, Stutt-
gart.
Fabricius, J. A. (1716–18) S Hippolyti episcopi et martyris Opera, 2 vols, Ham-
burg.
Favrod, J. (1991) La Chronique de Marius d’Avenches (455–581): texte, traduc-
tion et commentaire, Lausanne.
Fell, J. (1700) Sancti Cæcilii Cypriani opera, Oxford 1682, 3rd ed. Paris.
Felten, F.  J. and B.  Nichtweiss (2006) Hrabanus Maurus: Gelehrter Abt von
Fulda und Erzbischof von Mainz, Mainz.
Férotin, M. (1904) Le Liber ordinum: en usage dans l’église wisigothique et mo-
zarabe d’Espagne du Ve au XIe siècle, Paris, repr. Rome 1996.
Ferrari, M. C. (1973) ‘Note su Claudio di Torino ‘Episcopus ab ecclesia damna-
tus’,’ Italia medioevale e umanistica 16, 291–308.
— (1999) Il Liber sanctae crucis di Rabano Mauro: testo, immagine, contesto,
Bern.
Floëri, F. and P. Nautin (1957) Homélies pascales 3: une homélie anatolienne sur
la date de Pâques en l’an 387, Paris.
Floriano, A. C. (1949–51) Diplomática española del período astur: estudio de las
fuentes documentales del reino de Asturias (718–910), 2 vols, Oviedo.
Fontaine, J. (1959) Isidore de Seville et la culture classique dans l’Espagne Wisig-
othique, Paris.
— (1960) Isidore de Seville: Traité de la nature, Paris, repr. 2002.
— (1973) L’art préroman hispanique, Paris.
Ford, A. (1991–92) ‘Correspondence between Archbishops Ussher and Laud,’
Archivium Hibernicum 46, 5–21.
— (2007) James Ussher: theology, history, and politics in Early-Modern Ireland
and England, Oxford.
Fotheringham, J. K. (1922) ‘The Easter Calendar and the Slavonic Enoch,’ Jour-
nal of Theological Studies 23, 49–56.
Franke, F.-R. (1958) ‘Die freiwilligen Märtyrer von Cordova und das Verhält-
nis der Mozaraber zum Islam: nach den Schriften des Speraindeo,
Eulogius und Alvar,’ Spanische Forschungen der Görres-Gesellschaft 13,
1–170.


 Bibliography

Frederiksen, P. (1992) ‘Tyconius and Augustine on the Apocalypse,’ in R. K.

Emmerson and B. McGinn, The apocalypse in the Middle Ages, Ithaka,
20–37.
Frend, W. H. C. (1996) The archaeology of Early Christianity: a history, London.
Friedberg, A. (1959) Corpus Iuris Canonici, 2 vols, 2nd ed., Graz.
Ganz, D. (2006) ‘Anglo-Saxon England,’ in E. Leedham-Green and T. Webber,
The Cambridge history of the libraries in Britain and Ireland, 3 vols,
Cambridge, i 91–108.
García Avilés, A. (2001) El tiempo y los astros: arte, ciencia y religión en la alta
edad media, Murcia.
García Lobo, V. (1979) ‘El Beato de San Miguel de Escalada,’ Archivos Leoneses
33, 205–70.
García Toraño, P. (1986) Historia de el reino de Asturias, Oviedo.
García Volta, G. (1979) Die Westgoten: Aufbruch und Untergang eines Ger-
manenvolkes, Berg.
Gil, I. (1973) Corpus scriptorum muzarabicorum, 2 vols, Madrid.
Gil Fernández, J. (1985) Crónicas asturianas, Oviedo.
Görke, W. (2011) Datum und Kalender: von der Antike bis zur Gegenwart, Hei-
delberg.
Gómez-Moreno, M. (1919) Iglesias mozárabes: arte español de los siglos IX a XI,
Madrid.
Gómez Pallarès, J. (1986) Estudis sobre el Computus Cottonianus (Diss. Barce-
lona 1986), www.tdx.cbuc.es/handle/10803/5248.
— (1988–89) ‘Los textos latinos de cómputo de los mss. Paris, Bibl. Nat., NAL,
2169 y León, Bibl. de la Catedral, N. 8: una edición,’ Analecta Sacra
Tarraconensia 61–62, 373–410, repr. in Gómez Pallarès (1999), 63–92.
— (1989) ‘El Computus Cottonianus en los mss. Londres, B. M., Cotton Ca-
ligula A XV; Paris, B. N.‚ NAL 2169 y León, Archivo de la Catedral,
N. 8: un nuevo enfoque de la cuestión,’ Actas del VIII Congresso Espa-
ñol de Estudios Clásicos, vol. 3, Madrid, 501–06, repr. in Gómez Pal-
larès (1999), 57–62.
— (1999) Studia chronologica: estudios sobre manuscritos latinos de cómputo,
Madrid.
Gómez-Ruiz, R. (2007) Mozarabs, Hispanics and the cross, New York.
Gorman, M. (1997) ‘The Commentary on Genesis of Claudius of Turin and
biblical studies under Louis the Pious,’ Speculum 72, 279–329.
Grumel, V. (1952) ‘Les premières ères mondiales,’ Revue des études byzantines
10, 93–108.
— (1958) La Chronologie, Paris.
Graff, E. (2010) ‘The recension of two Sirmond texts: Disputatio Morini and De
divisionibus temporum,’ in Warntjes and Ó Crónín (2010), 112–42.
Grafton, A. (1975) ‘Joseph Scaliger and historical chronology: the rise and fall
of a discipline,’ History and Theory 14, 156–85.


 Bibliography

— (1985) ‘From De die natali to De emendatione temporum: the origins and set-
tings of Scaliger’s chronology,’ Journal of the Warburg and Courtauld
Institutes 48, 100–43.
Grosjean, P. (1946) ‘Recherches sur les debuts de la controverse pascale chez les
Celtes,’ Analecta Bollandiana 64, 200–44.
— (1959) ‘Notes sur quelques sources des Antiquitates de Jacques Ussher,’ Ana-
lecta Bollandiana 77, 154–87.
Grotefend, H. (1991) Taschenbuch der Zeitrechnung, 13th ed., Hannover.
Guarducci, M. (1977) ‘La statua di ‘sant’Ippolito’,’ in Ricerche su Ippolito, Rome,
17–30.
Gwynn, A. (1967) ‘Archbishop Ussher and Father Brendan O Conor,’ in Fran-
ciscan Fathers, Father Luke Wadding: commemorative volume, Dub-
lin, 263–83.
Haarländer, S. (2006) Hrabanus Maurus zum Kennenlernen: Ein Lesebuch mit
einer Einführung in sein Leben und Werk, Mainz.
Hanson, R. P. C. (1968) Saint Patrick: his origins and career, Oxford.
Hathaway, N. (1989) ‘Compilatio: from plagiarism to compiling,’ Viator 20,
19–44.
Heil, J. (1997) ‘Claudius von Turin – eine Fallstudie zur Geschichte der
Karolingerzeit,’ Zeitschrift für Geschichtswissenschaft 45, 385–412.
— (2000) ‘‘Nos nescientes de hoc velle manere’—‘We wish to remain ignorant
about this’: timeless end, or: approaches to reconceptualising escha-
tology after A.D. 800 (A.M. 6000),’ Traditio 55, 73–103.
Heil, W. (1965) ‘Der Adoptianismus, Alkuin und Spanien,’ in B. Bischoff, Karl
der Große: Lebenswerk und Nachleben 2: Das geistige Leben, Düssel-
dorf, 95–155.
Henel, H. (1942) Aelfric’s De temporibus anni, London.
Herbers, K. (2006) Geschichte Spaniens im Mittelalter, Stuttgart.
Herwig, W. (2005) Gotische Studien: Volk und Herrschaft im frühen Mittelalter,
München.
Heyse, E. (1969) Hrabanus Maurus’ Enzyklopädie ‘De rerum naturis’: Untersu-
chungen zu den Quellen und zur Methode der Kompilation, Munich.
Holford-Strevens, L. (2005) The history of time: a very short introduction, Ox-
ford.
— (2008) ‘Paschal lunar calendars up to Bede,’ Peritia 20, 165–208.
— (2011) ‘Church politics and the computus: from Milan to the ends of the
earth,’ in Warntjes and Ó Cróinín (2011), 1–20.
Holl, K. (1915–33) Epiphanius: Ancoratus und Panarion, 3 vols, Lepizig.
Howlett, D. (1994) Liber epistolarum Sancti Patricii episcopi: the book of letters
of Saint Patrick the Bishop, Dublin.
— (1995a) The Celtic Latin tradition of biblical style, Dublin.
— (1995b) ‘Five experiments in textual reconstruction and analysis,’ Peritia 9,
1–50.


 Bibliography

— (1996a) ‘Rubisca: an edition, translation, and commentary,’ Peritia 10, 71–

90.
— (1996b) ‘Seven studies in seventh-century texts,’ Peritia 10, 1–70.
— (1997) British books in biblical style, Dublin.
— (1998a) Cambro-Latin compositions, their competence and craftsmanship,
Dublin.
— (1998b) ‘Synodus prima Sancti Patricii: an exercise in textual reconstruction,’
Peritia 12, 238–53.
— (2002) ‘A miracle of Maedóc,’ Peritia 16, 85–93.
— (2005) Insular inscriptions, Dublin.
— (2006) ‘Gematria, number and name in Anglo-Norman,’ French Studies Bul-
letin 60, 90–92.
— (2008a) ‘On the new edition of Anatolius’ De ratione paschali,’ Peritia 20,
135–53.
— (2008b) ‘Insular inscriptions and the problem of coincidence: a reply,’ Cam-
brian Medieval Celtic Studies 56, 75–96.
— (2010a) ‘Architecture, music, and time in Wulfstan’s verse,’ in M. Henig and
N.  Ramsay, Intersections: the archaeology and history of Christianity
in England, 400–1200: papers in honour of Martin Biddle and Birthe
Kjølbye-Biddle, Oxford, 179–200.
— (2010b) ‘Computus in Hiberno-Latin literature,’ in Warntjes and Ó Cróinín
(2010), 259–323.
— (2010c) ‘Hiberno-Latin poems on the Eusebian canons,’ Peritia 21, 162–71.
— (2010d) ‘Two mathematical poets,’ Peritia 21, 151–57.
— (2011–12) ‘The Old-Irish hymn ‘Brigit bé bithmaith’,’ Peritia 22–23, 182–87.
— (2013) ‘Music and the stars in early Irish compositions,’ in M.  Kelly and
C.  Doherty, Music and the stars: mathematics in medieval Ireland,
Dublin, 111–28.
Huber, W. (1969) Passa und Ostern: Untersuchungen zur Osterfeier der alten
Kirche, Berlin.
Hughes, K. (1963) ‘Irish monks and learning,’ in Los monjes y los estudios, Pob-
let, 61–86, repr. in K.  Hughes, Church and society in Ireland, A.D.
400–1200, ed. by D. N. Dumville, London 1987, article XII.
— (1966) The church in early Irish society, London.
— (1972) Early Christian Ireland: introduction to the sources, London.
Huglo, M. (1954) ‘Christe fave votis,’ Scriptorium 8, 108–11.
Hultsch, F. (1864–66) Metrologicorum scriptorum reliquiae, 2 vols, Leipzig.
Instituto de Historia Eclesiástica Padre Enrique Flórez Madrid (1953) Antifon-
ario visigótica mozárabe de la catedral de León, Madrid.
Jacoby, F. (1902) Apollodors Chronik, Berlin.
— (1923–59) Die Fragmente der griechischen Historiker, 15 vols, Berlin.
Jan, J. W. (1718) Historia cycli Dionysiani cum argumentis paschalibus, Witten-
berg.


 Bibliography

Jenkinson, F. J. H. (1908) The Hisperica famina, Cambridge.

Jeudy, C. (1991) ‘Remigii autissiodorensis opera (Clavis),’ in D.  Iogna-Prat,
C. Jeudy, and G. Lobrichon, L’école carolingienne d’Auxerre: de Mure-
thach à Remi, 830–908, Paris, 457–500.
Jones, C. W. (1934) ‘The Victorian and Dionysiac paschal tables in the West,’
Speculum 9, 408­–21, repr. in C.  W. Jones, Bede, the schools and the
computus, ed. by W. M. Stevens, Aldershot 1994, article VIII.
— (1937) ‘The ‘lost’ Sirmond manuscript of Bede’s computus,’ English Histori-
cal Review 52, 204–19, repr. in C. W. Jones, Bede, the schools and the
computus, ed. by W. M. Stevens, Aldershot 1994, article X.
— (1939) Bedae pseudepigrapha: scientific writings falsely attributed to Bede,
Ithaca.
— (1943) Bedae opera de temporibus, Cambridge.
— (1943a) ‘A legend of St Pachomius,’ Speculum 18, 198–210, repr. in C. W.
Jones, Bede, the schools and the computus, ed. by W. M. Stevens, Alder-
shot 1994, article VII.
— (1963) ‘An early medieval licensing examination,’ History of Education Quar-
terly 3, 19–29.
— (1976) ‘Bede’s place in medieval schools,’ in G.  Bonner, Famulus Christi:
essays in commemoration of the thirteenth centenary of the birth of the
Venerable Bede, London, 261–85.
Kaske, R. E. (1988) Medieval Christian literary imagery: a guide to interpreta-
tion, Toronto.
Kautzsch, E. and A. E. Cowley (1985) Gesenius’ Hebrew grammar, 2nd ed., Oxford.
Kavanagh, S., R. O’Ferrall, and R. O’Connell (1932–49) Commentarius Rinuc-
cinianus, de sedis apostolicae legatione ad foederatos Hiberniae Catholi-
cos per annos 1645–1649, 6 vols, Dublin.
Kenney, J. F. (1929) The sources for the early history of Ireland, vol. 1: ecclesiastical.
An introduction and guide, New York, 2nd ed. Dublin 1966, repr. 1993.
Kottje, R. (2012) Verzeichnis der Handschriften mit den Werken des Hrabanus
Maurus, Hannover.
Kottje, R. and H.  Zimmermann (1982) Hrabanus Maurus, Lehrer, Abt und
Bischof, Mainz.
Kotzur, H.-J. (2006) Hrabanus Maurus: Auf den Spuren eines karolingischen
Gelehrten, Mainz.
Krusch, B. (1880) Studien zur christlich-mittelalterlichen Chronologie: Der
84jährige Ostercyclus und seine Quellen, Leipzig.
— (1884) ‘Die Einführung des griechischen Paschalritus im Abendlande,’ Neues
Archiv der Gesellschaft für ältere deutsche Geschichtskunde 9, 99–169.
— (1885) ‘Chronologisches aus Handschriften,’ Neues Archiv der Gesellschaft
für ältere deutsche Geschichtskunde 10, 84–89.
— (1910) ‘Das älteste fränkische Lehrbuch der dionysianischen Zeitrechnung,’
in Mélanges offerts à M. Émile Chatelain, Paris, 232–42.


 Bibliography

— (1926) ‘Ein Bericht der päpstlichen Kanzlei an Papst Johannes I. von 526
und die Oxforder HS. Digby 63 von 814,’ in A. Brackmann, Papsttum
und Kaisertum: Forschungen zur politischen Geschichte und Geisteskul-
tur des Mittelalters, Munich, 48–58.
— (1938) ‘Studien zur christlich-mittelalterlichen Chronologie: Die Entstehung
unserer heutigen Zeitrechnung,’ Abhandlungen der Preußischen Aka­
demie der Wissenschaften, Jahrgang 1937, phil.-hist. Klasse, Nr. 8, Berlin.
Kuypers, A. B. (1902) The prayer book of Aedeluald the bishop, commonly called
the Book of Cerne, Cambridge.
Labbé, P. (1657) Novae bibliothecae manuscriptorum librorum, 2 vols, Paris.
Landes, R. (1988) ‘Lest the millennium be fulfilled: apocalyptic expectations
and the pattern of Western chronography 100–800 CE,’ in W. Ver-
beke, D. Verhelst, and A. Welkenhuysen, The use and abuse of eschatol-
ogy in the Middle Ages, Leuven, 137–211.
Lazzarato, D. (1952) Chronologia Christi seu discordantium fontium concordan-
tia ad juris normam, Naples.
Le Berrurier, D. O. (1978) The pictorial sources of mythological and scientific il-
lustrations in Hrabanus Maurus’ De rerum naturis, New York.
Lehmann, P. (1912) ‘Cassiodorstudien II: die Datierung der Institutiones und
des Computus paschalis,’ Philologus 71, 278–99, repr. in Lehmann
(1959–62), ii 41–55.
Lehmann, P. (1959–62) Erforschung des Mittelalters, 5 vols, Stuttgart.
Leibniz, G. W. (1698) Accessiones historicae, Leipzig.
Lejbowicz, M. (2006) ‘Des tables pascales aux tables astronomiques et retour:
formation et réception du comput patristique,’ http://methodos.re-
vues.org/538.
— (2010) ‘Les Pâques baptismales d’Augustin d’Hippone, une étape contournée
dans l’unification des practiques computistes latines,’ in Warntjes and
Ó Cróinín (2010), 1–39.
Lennon, C. (1981) Richard Stanihurst the Dubliner, 1547–1618: a biography
with a Stanihurst text ‘On Ireland’s past’, Dublin.
Levison, W. (1946) England and the Continent in the eighth century, Oxford.
Lindsay, W. M. (1910) Early Irish minuscule script, Oxford.
— (1911) Isidori Hispalensis episcopi Etymologiarum sive Originum libri XX,
Oxford.
— (1915) Notae latinae: an account of abbreviation in Latin manuscripts of the
early minuscule period (c. 700–850), Cambridge.
Löfstedt, B. (2003) Virgilius Maro Grammaticus opera omnia, Leipzig.
Loi, V. (1971) ‘Il 25 Marzo data pasquale e la cronologia giovannea della pas-
sione in età patristica,’ Ephemerides liturgicae 85, 48–69.
Love, R. (2007) ‘Bede and John Chrysostom,’ Journal of Medieval Latin 17,
72–86.


 Bibliography

Ludlow, F. M. (2010) The utility of the Irish Annals as a source for the reconstruc-
tion of climate, 2 vols, unpubl. Ph.D. thesis Trinity College, Dublin.
Ludlow, F. et  al. (2013) ‘Medieval Irish chronicles reveal persistent volcanic
forcing of severe winter cold events,’ Environmental Research Letters
8, 024035.
Lumper, P. G. (1795) Historia theologico-critica de vita, scriptis, atque doctrina
sanctorum patrum, aliorumque scriptorum ecclesiasticorum trium pri-
morum saeculorum ex virorum doctissimorum literariis monumentis
collecta: Pars XI: Complectens praeter Novatiani et Cornelii, potissi-
mum S. Cypriani vitam, scripta et fidei doctrinam, Vienna.
Mac Carthy, B. (1892) The codex Palatino-Vaticanus no. 830, Dublin.
(1901) Annala Uladh: The Annals of Ulster, vol.  4: introduction and sources,
Dublin, repr. 1999.
Mac Neill, E. (1931) ‘Beginnings of Latin culture in Ireland,’ Studies 20, 39–48,
449–60.
Mac Neill, J. (1934) St Patrick, apostle of Ireland, London.
Mahler, E. (1916) Handbuch der jüdischen Chronologie, Leipzig.
Manitius, M. (1911) Geschichte der lateinischen Literatur des Mittelalters. Erster
Band: von Justinian bis zur Mitte des zehnten Jahrhunderts, Munich.
Mansi, J.  D. (1902–23) Sacrorum conciliorum nova et amplissima collectio,
55 vols, Paris, repr. Graz 1960–62.
Martinay, J. and A.  Pouget (1693–1706) Hieronymus (S.  Eusebius): Opera,
emendata studio et opera monachorum S. Benedicti, 5 vols, Paris.
Martínez Fuertes, B. (1936) Montes y Peñalba: ensayo histórico-artistico, Zaragoza.
Massons i Rabassa, E. (2002–03) ‘La iconografía del diable al claustre de Santa
Maria de l’Estany,’ Lambard. Estudis d’art medieval 15, 11–38.
Mattei, P. (2001) ‘Remarques sur la tradition textuelle (manuscrite et imprimée)
du De rebaptismate,’ Studia Patristica 36, 35–45.
Matter, E. A. (1985) ‘Theological freedom in the Carolingian age: the case of
Claudius of Turin,’ in G. Makdisi, D. Sourdel, and J. Sourdel-Thom-
ine, La notion de liberté au moyen âge: Islam, Byzance, Occident, Paris,
51–60.
— (1992) ‘The apocalypse in early medieval exegesis,’ in R. K. Emmerson and
B. McGinn, The apocalypse in the Middle Ages, Ithaka, 38–50.
McCafferty, J. (2007) The reconstruction of the Church of Ireland: Bishop Bram-
hall and the Laudian reforms, 1633–1641, Cambridge.
Mc Carthy, D. P. (1993) ‘Easter principles and a fifth-century lunar cycle used
in the British Isles,’ Journal for the History of Astronomy 24, 204–24.
— (1994) ‘The origin of the latercus paschal cycle of the insular Celtic church-
es,’ Cambrian Medieval Celtic Studies 28, 25–49.
— (1997) ‘The biblical chronology of James Ussher,’ Irish Astronomical Journal
24, 73–82.


 Bibliography

— (1999–2005) ‘Chronological synchronisation of the Irish Annals,’ www.

irish-annals.cs.tcd.ie.
— (2008) The Irish Annals: their genesis, evolution and history, Dublin.
— (2010) ‘Bede’s primary source for the Vulgate chronology in his chronicles in
De temporibus and De temporum ratione,’ in Warntjes and Ó Cróinín
(2010), 159–89.
— (2011) ‘On the arrival of the latercus in Ireland,’ in Warntjes and Ó Cróinín
(2011), 48–75.
— (2017) ‘Analysing and restoring the chronology of the Irish Annals,’ in
R. Kenna, M. Mac Carron, and P. Mac Carron, Maths meets myths:
complexity-science approaches to myths, folktales, chronicles and histo-
ries, New York, 177–94.
Mc  Carthy, D.  P. and A.  Breen (2003) The ante-Nicene Christian Pasch: De
ratione paschali – the paschal tract of Anatolius, bishop of Laodicea,
Dublin.
McCluskey, S.  C. (1998) Astronomies and cultures in early medieval Europe,
Cambridge.
McGinn, B. (1998) Visions of the end: apocalyptic traditions in the Middle Ages,
New York.
McKitterick, R. (2004) History and memory in the Carolingian world, Cam-
bridge.
— (2006) Perceptions of the past in the early Middle Ages, Notre Dame.
Meehan, B. (1986) ‘The manuscript collection of James Ussher,’ in P. Fox, Treas-
ures of the Library, Trinity College, Dublin, Dublin, 97–110.
Meissner, J. (1929) The Celtic church in England after the Synod of Whitby, London.
Millares Carlo, A. (1954) ‘Textes et figures de l’antiphonaire de León,’ Archivos
leonenses 8, 258–87.
— (1961) ‘Manuscritos visigóticos: notas bibliográficas,’ Hispania sacra 14,
337–444.
Monceaux, P. (1901–23) Histoire littéraire de l’Afrique chrétienne depuis les orig-
ines jusqu’à l’invasion arabe, 7 vols, Paris.
Monferrer Sala, J.  P. (2004) ‘Mitografía hagiomartirial: de nuevo sobre los
supuestos mártires cordobeses del siglo IX,’ in M. Fierro, De muerte
violenta: política, religión y violencia en Al-Andalus, Madrid, 415–50.
Mosshammmer, A. A. (1979) The Chronicle of Eusebius, Lewisburg.
— (2008) The Easter computus and the origins of the Christian era, Oxford.
— (2013) ‘The Praefatio (Prologus) sancti Cyrilli de Paschate and the 437-year
(not 418!) paschal list attributed to Theophilus,’ Vigiliae Christianae
67, 49–78.
— (forthcoming) ‘Three computistical texts—Expositio Bissexti, Quo Tem-
pore Initium Mundi, and Ratio Lunae, with an appendix on Spanish
epacts’, in I. Warntjes and D. Ó Cróinín, Proceedings of the 4th Galway
Computus Conference.


 Bibliography

Mostert, M. (1989) The library of Fleury: a provisional list of manuscripts, Hil-

versum.
Muratori, L. A. (1697–1713) Anecdota, quae ex Ambrosianae bibliothecae codi-
cibus, 4 vols, Milan.
Mynors, R. A. B. (1937) Cassiodori Senatoris Institutiones, Oxford.
Nagel, H. (1998) Karl der Große und die theologischen Herausforderungen seiner
Zeit: Zur Wechselwirkung zwischen Theologie und Politik im Zeitalter
des großen Frankenherrschers, Frankfurt.
Nestle, E. (1901) Die Kirchengeschichte des Eusebius aus dem Syrischen übersetzt,
Leipzig.
Neugebauer, O. (1979) Ethiopic astronomy and computus, Wien.
— (1982) ‘On the Computus paschalis of ‘Cassiodorus’,’ Centaurus 25, 292–302.
Neuss, W. (1931) Die Apokalypse des hl. Johannes in der altspanischen und alt­
christlichen Bibel-Illustration: Das Problem der Beatus-Handschriften,
2 vols, Münster.
Nogara, B. (1929) Codices Vaticani Latini: Codices 1461–2059, Vatican.
Nothaft, C. P. E. (2012) Dating the Passion: the life of Jesus and the emergence of
scientific chronology (200–1600), Leiden.
— (2012a) ‘Between crucifixion and calendar reform: medieval Christian per-
ceptions of the Jewish lunisolar calendar,’ in J. Ben-Dov, W. Horowitz,
and J. Steele, Living the lunar calendar, Oxford, 259–67.
— (2014) Medieval Latin Christian texts on the Jewish calendar: a study with five
editions and translations, Leiden.
O’Brien, J.  M. (2011) ‘Locating authorities in Carolingian debates on image
veneration: the case of Agobard of Lyon’s De picturis et imaginibus,’
Journal of Theological Studies, n.s. 62, 176–206.
Obrist, B. (2002) ‘Les manuscrits du De cursu stellarum de Grégoire de Tours
et le ms Laon, Bibliothèque municipale 422,’ Scriptorium 56, 335–45.
Ó Cróinín, D. (1982) ‘A seventh-century Irish computus from the circle of
Cummianus,’ Proceedings of the Royal Irish Academy 82C, 405–30,
repr. in Ó Cróinín (2003), 99–130.
— (1982a) ‘Mo Sinu maccu Min and the computus at Bangor,’ Peritia 1, 281–
95, repr. in Ó Cróinín (2003), 35–48.
— (1983) ‘The Irish provenance of Bede’s computus,’ Peritia 2, 229–47, repr. in
Ó Cróinín (2003), 173–90.
— (1983a) ‘Sticks and stones: a reply,’ Peritia 2, 257–60.
— (1985) ‘‘New heresy for old’: Pelagianism in Ireland and the papal letter of
640,’ Speculum 60, 505–16, repr. in Ó Cróinín (2003), 87–98.
— (1986) ‘New light on Palladius,’ Peritia 5, 276–83, repr. in Ó Cróinín (2003),
28–34.
— (1995) Early medieval Ireland 400–1200, London.


 Bibliography

— (1997) ‘The computistical writings of Columbanus,’ in M. Lapidge, Colum-

banus: studies on the Latin writings, Woodbridge, 264–70, repr. in Ó
Cróinín (2003), 48–55.
— (2000) ‘Who was Palladius ‘First Bishop of the Irish’?,’ Peritia 14, 205–37.
— (2003) Early Irish history and chronology, Dublin.
— (2003a) ‘Bede’s Irish computus,’ in Ó Cróinín (2003), 201–12.
— (2010) ‘The continuity of Irish computistical tradition,’ in Warntjes and Ó
Cróinín (2010), 324–47.
Ogg, G. (1940) The chronology of the public ministry of Jesus, Cambridge.
— (1954) ‘The tabella appended to the Pseudo-Cyprianic De pascha computus
in the Codex Remensis,’ Vigiliae Christianae 8, 134–44.
— (1955) The Pseudo-Cyprianic De pascha computus, London.
Ohashi, M. (2010) ‘The Annus Domini and the sexta aetas: problems in the
transmission of Bede’s De temporibus,’ in Warntjes and Ó Cróinín
(2010), 190–203.
O’Loughlin, T. (2005) Discovering Saint Patrick, London.
Omont, H. (1883) ‘Interrogationes de fide catholica ( Joca monachorum),’ Bib-
liothèque de l’école des chartes 44, 58–71.
O’Rahilly, T. F. (1942) The two Patricks, Dublin.
— (1946) Early Irish history and mythology, Dublin, repr. 1971.
Orchard, N. (2002) The Leofric Missal, 2 vols, London.
Orlandis, J. (2003) Historia del reino visigodo español, 2nd ed. Madrid.
Orlandis, J. and D. Ramos-Lissón (1981) Die Synoden auf der Iberischen Halb­
insel bis zum Einbruch des Islam (711), München.
Oroz Reta, J., M.-A. Marcos Casquero, and M. C. Díaz y Díaz (1993–94) San
Isidoro de Sevilla: Etimologías, 2 vols, 2nd repr., Madrid.
O’Sullivan, W. (1956) ‘Ussher as a collector of manuscripts,’ Hermathena 88,
34–58.
— (1994–95) ‘Correspondence of David Rothe and James Ussher, 1619–23,’
Collectanea Hibernica 36–37, 7–49.
Oulton, J. E. L. (1957) ‘The epistle of Cummian De controversia paschali,’ Stu-
dia Patristica 1, 128–33.
Palmer, J.  T. (2011) ‘Computus after the paschal controversy of AD  740,’ in
Warntjes and Ó Cróinín (2011), 213–41.
— (2011a) ‘Calculating time and the end of time in the Carolingian world,
c. 740–820,’ English Historical Review 126, 1307–31.
— (2014) The apocalypse in the early Middle Ages, Cambridge.
Papandrea, J. L. (2008) The Trinitarian theology of Novatian of Rome: a study in
third-century orthodoxy, Lewiston.
Parker, R. A. (1950) The calenders of ancient Egypt, Chicago.
Parr, R. (1686) The life of the most reverend father in God, James Usher, late Lord
Arch-Bishop of Armagh, primate and metropolitan of all Ireland, London.


 Bibliography

Peden, A. M. (2003) Abbo of Fleury and Ramsey: Commentary on the Calculus
of Victorius of Aquitaine, Oxford.
Petavius, D. (1627) De doctrina temporum, 2 vols, Paris.
Petitmengin, P. (1974) ‘Notes sur des manuscrits patristiques latins. II: Un
‘Cyprien’ de Cluny et la lettre apocryphe du pape Corneille,’ Revue
des études augustiniennes 20, 15–35.
— (1975) ‘Un monument controversé: le ‘Saint Cyprien’ de Baluze et Dom
Maran (1726),’ Revue d’histoire des textes 5, 97–136.
Pitra, J. B. (1852–58) Spicilegium Solesmense, 4 vols, Paris, 2 vols 2nd ed. Graz
1962.
— (1879–91) Analecta sacra Spicilegio Solesmensi parata, 9 vols, Paris.
Planta, J. (1802) A catalogue of the manuscripts in the Cottonian Library depos-
ited in the British Museum, London.
Plummer, C. (1896) Venerabilis Baedae opera historica, 2 vols, Oxford.
Pochoshajew, I. (2007) Die Märtyrer von Córdoba: Christen im muslimischen
Spanien des 9. Jahrhunderts, Frankfurt.
Pohl, W. and P. Herold (2002) Vom Nutzen des Schreibens: soziales Gedächtnis,
Herrschaft und Besitz im Mittelalter, Wien.
Polara, G. (1979) Virgilio Marone Grammatico Epitomi ed Epistole, Naples.
Preuschen, E. (1902) Eusebius Kirchengeschichte, Buch VI und VII, aus dem Ar-
menischen übersetzt, Leipzig.
Pritchett, W. K. (1982) ‘The calendar of the gibbous moon,’ Zeitschrift für Pa-
pyrologie und Epigraphik 49, 243–66.
Raaijmakers, J. (2012) The making of the monastic community of Fulda, c. 744–
900, Cambridge.
Rambaud-Buhot, J. (1968) ‘La critique des faux dans l’ancien droit canonique,’
Bibliothèque de l’école des chartes 126, 5–62.
Rankin, S. (forthcoming) ‘Si … memoraliter decantare consuescas: Melodies for
computistical verse,’ Anglo-Saxon 2.
Ribémont, B. (2001) Les origines des encylcopédies médiévales: d’Isidore de Séville
aux Carolingiens, Paris.
Richard, M. (1950) ‘Comput et chronographie chez saint Hippolyte,’ Mélanges
de science religieuse 7, 237–68, repr. in M.  Richard, Opera minora,
3 vols, Paris 1976–77, article 19.
— (1951) ‘Comput et chronographie chez saint Hippolyte (II),’ Mélanges de
science religieuse 8, 19–50, repr. in M. Richard, Opera minora, 3 vols,
Paris 1976–77, article 19.
Richards, E. G. (1998) Mapping time: the calendar and its history, Oxford.
Rusch, A. (1467) Hrabani Mauri, De sermonum proprietate, sive, Opus de uni-
verso, Strasbourg.
Ryan, J. (1958) Saint Patrick, Dublin.
Sánchez-Albornoz, C. (1972–75) Orígenes de la nación española, 3 vols, Oviedo.


 Bibliography

Sánchez Velasco, J. (2000) Elementos arquitectónicos de época visigoda en el Mu-

seo Arqueológico de Córdoba, Córdoba.
Sanford, E. M. (1949) ‘Famous Latin encyclopedias,’ Classical Journal 44, 462–
67.
Savile, H. (1610–12) Tou en hagiois patros hemon Ioannou Archiepiskopou Kon-
stantinoupoleos tou Chrysostomou ton heuriskomenon, 8 vols, Eton.
Scaliger, J. J. (1595) Hippolyti episcopi Canon pascalis, Leiden.
— (1606) Animadversiones in Chronologica Eusebii, Leiden.
— (1627) Epistolae omnes quae reperiri potuerunt, nunc primum collectae ac edi-
tae, ed. by D. Heinsius, Leiden.
— (1629) Opus de emendatione temporum, 2nd ed. Geneva.
Schäferdiek, K. (1967) Die Kirche in den Reichen der Westgoten und Suewen
bis zur Errichtung der westgotischen katholischen Staatskirche, Berlin.
— (1969­–70) ‘Der adoptianische Streit in Rahmen der spanischen Kirchen-
geschichte,’ Zeitschrift für Kirchengeschichte 80, 291–311; 81, 1–16,
repr. in K.  Schäferdiek, Schwellenzeit: Beiträge zur Geschichte des
Christentums in Spätantike und Mittelalter, ed.  by W.  A. Löhr and
H. C. Brennecke, Berlin 1996, 381–416.
Schipper, W. (1989) ‘Rabanus Maurus, De rerum naturis: a provisional check
list of manuscripts,’ Manuscripta 33, 109–18.
— (1997a) ‘Annotated copies of Hrabanus Maurus’s De rerum naturis,’ English
Manuscript Studies: 1100–1700, n.s. 6, 1–23.
— (1997b) ‘The earliest manuscripts of Hrabanus Maurus’ De rerum naturis
(Karlsruhe, Badische Landesbibliothek, MS Aug. 68 and Vienna, Ös-
terreichische Nationalbibliothek MS 121),’ in P. Binkley, Pre-modern
encyclopaedic texts, Leiden, 363–77.
— (2004) ‘Rabanus Maurus and his sources,’ in A. A. McDonald and M. W.
Twomey, Schooling and society: the ordering and reordering of knowl-
edge in the western Middle Ages, Leuven, 1–21.
— (2007) ‘Montecassino 132 and the early transmission of Hrabanus’ De rerum
naturis,’ Archa Verbi 4, 103–26
— (2008) ‘Textual varieties in manuscript margins,’ in R. H. Bremmer and S. L.
Keefer, Signs of the edge: space, text and margins in medieval manu-
scripts, Leuven, 25–54.
Schlag, H. E. (2008) Ein Tag zuviel: aus der Geschichte des Kalenders, Gerns-
bach.
Schlunk, H. and T.  Hauschild (1978) Hispania antiqua: die Denkmäler der
frühchristlichen und westgotischen Zeit, Mainz.
Schmid, J. (1904) Die Osterfestberechnung auf den britischen Inseln vom Anfang
des vierten bis zum Ende des achten Jahrhunderts, Regensburg.
— (1907) Die Osterfestberechnung in der abendländischen Kirche vom 1.
Allgemeinen Konzil von Nicäa bis zum Ende des VIII. Jahrhunderts,
Freiburg.


 Bibliography

Scholderer, V. (1939) ‘Adolf Rusch and the earliest Roman types,’ The Library,
3rd series 20, 43–50.
Schwartz, E. (1905) ‘Christliche und jüdische Ostertafeln,’ Abhandlungen der
Königlichen Gesellschaft der Wissenschaften zu Göttingen, philologisch-
historische Klasse, Band 8, Nr. 6, Berlin.
Schwartz, E. and T. Mommsen (1903–09) Eusebius Werke, 2. Band: Die Kirch-
engeschichte, 3 vols, Leipzig.
Sénac, P. (2000) La frontière et les hommes (VIIIe–XIIIe siècle), le peuplement
musulman au nord de l’Ebre et les débuts de la reconquête aragonaise,
Paris.
Siegmund, A. (1949) Die Überlieferung der griechischen christlichen Literatur in
der lateinischen Kirche bis zum zwölften Jahrhundert, Munich.
Silva-Tarouca, C. (1931) ‘Nuovi studi sulle antiche lettere dei papi,’ Gregori-
anum 12, 3–56.
Siniscalco, P. (1978) ‘Le età del mondo in Beda,’ Romanobarbarica 3, 297–331.
Smart, W. M. (1958) Foundations of astronomy, London.
Smith, T. (1696) A catalogue of the manuscripts in the Cottonian Library, Lon-
don.
Smyth, M. (1996) Understanding the universe in seventh-century Ireland, Wood-
bridge.
Souter, A. (1949) A glossary of later Latin to 600 A.D., Oxford.
Springsfeld, K. (2002) Alkuins Einfluß auf die Komputistik zur Zeit Karls des
Großen, Stuttgart.
Stegmüller, F. (1940–80) Repertorium biblicum medii aevi, 11 vols, Madrid.
Stern, S. (2001) Calendar and community: a history of the Jewish calendar, Ox-
ford.
— (2012) Calendars in Antiquity: empires, states, and societies, Oxford.
Strobel, A. (1977) Ursprung und Geschichte des frühchristlichen Osterkalenders,
Berlin.
— (1984) Texte zur Geschichte des frühchristlichen Osterkalenders, Münster.
Suchier, W. (1955) Das mittellateinische Gespräch Adrian und Epictetus nebst
verwandten Texten (Joca monachorum), Tübingen.
Tax, P.  W. (1991) ‘Remigius of Auxerre’s Psalm Commentary and the Mat-
thew Commentary attributed to him: questions of authenticity,’
in D.  Iogna-Prat, C.  Jeudy, and G.  Lobrichon, L’école carolingienne
d’Auxerre: de Murethach à Remi, 830–908, Paris, 413–24.
Thomas, C. (1981) Christianity in Roman Britain to AD 500, London.
Thompson, E. A. (1985) Who was Saint Patrick?, Woodbridge.
Thorndike, L. (1954) ‘Computus,’ Speculum 29, 223–38.
Tihon, Α. (1978) Le ‘petit commentaire’ de Théon d’Alexandrie aux tables faciles
de Ptolémée, Vatican.
Traube, L. (1909–20) Vorlesungen und Abhandlungen, 3 vols, ed. by P. Lehmann
et al., Munich, repr. 1965.


 Bibliography

Troncarelli, F. (1988) ‘Il consolato dell’Anticristo,’ Studi medievali 30, 567–92.

— (2014) ‘Osservazioni sul Reginense latino 2077,’ Scriptorium 68, 79–101.
Twomey, M.  W. (1988) ‘Appendix: medieval encyclopedias,’ in R.  E. Kaske,
Medieval Christian literary imagery: a guide to interpretation, Toron-
to, 182–215.
Ussher, J. (1613) Gravissimae quaestionis de Christianarum ecclesiarum succes-
sione et statu historica explicatio, London, repr. in Elrington (1847–
64), ii 1–414.
— (1631) A discourse of the religion anciently professed by the Irish and British,
London, repr. in Elrington (1847–64), iv 235–381.
— (1632) Veterum epistolarum Hibernicarum sylloge, Dublin, repr. in Elrington
(1847–64), iv 432–44.
— (1639) Britannicarum ecclesiarum antiquitates, Dublin, repr. in Elrington
(1847–64), v–vi.
— (1650) Annales veteris testamenti, Dublin, repr. in Elrington (1847–64), viii–x.
— (1654) Annales novi testamenti, Dublin, repr. in Elrington (1847–64), xi 1–176.
— (1658) The annals of the world deduced from the origin of time, and continued
to the beginning of the Emperor Vespasian’s reign, London.
— (1660) Chronologia sacra, Oxford, repr. in Elrington (1847–64), xi 475­–xii 144.
Valdeon Baruque, J. (1994) ‘Evolución historica del reinado de Alfonso  III,’
in F. J. Fernández Conde, La epoca de Alfonso III y San Salvador de
Valdediós, Oviedo, 19–26.
Vallancey, C. J. (1770–1804) Collectanea de rebus Hibernicis, 6 vols, Dublin.
Vallarsi, D. (1766–72) Hieronymi opera omnia, 2nd ed., 11 vols, Venice.
van Banning, J. (1997) ‘Claudius von Turin als eine extreme Konsequenz des
Konzils von Frankfurt,’ in R. Berndt, Das Frankfurter Konzil von 794,
2 vols, Mainz, ii 731–49.
van der Hagen, J. (1734) Observationes in veterum patrum et pontificum prologos
et epistolas paschales aliosque antiquos de ratione paschali, Amsterdam.
— (1736) Dissertationes de cyclis paschalibus, Amsterdam.
Verbist, P. (2010) Duelling with the past: medieval authors and the problem of the
Christian era, c. 990–1135, Turnhout.
Vidal Álvarez, S. (2000) ‘Las artes del objecto en Cataluña del final del reino
visigodo al románico (siglos VIII–X): el tesoro altomedieval de Sant
Pere de Rodes,’ in Camps i Sòria (2000), 255–58.
Vion, A. (1665) Veterum epistolarum Hibernicarum collectionem, Paris.
Vives, J. et al. (1963) Concilios visigóticos e hispano-romanos, Madrid.
Vives Gatell, J. and Á. Fábrega Grau (1949) ‘Calendarios hispánicos anteriores al
siglo XII,’ Hispania sacra 2, 119–46, 339–80.
von den Brincken, A.-D. (1957) Studien zur lateinischen Weltchronistik bis in
das Zeitalter Ottos von Freising, Düsseldorf.
Vones, L. (1993) Geschichte der iberischen Halbinsel im Mittelalter 711–1480:
Reiche – Kronen – Regionen, Sigmaringen.


 Bibliography

von Euw, A. (1998) ‘Kompendium der Zeitrechnung, Naturlehre und Him-

melskunde,’ in J. M. Plotzek and U. Surmann, Glaube und Wissen im
Mittelalter: die Kölner Dombibliothek, München, 136–56.
von Harnack, A. (1893–1904) Geschichte der altchristlichen Litteratur bis Euse-
bius, 2 vols, Leipzig, 2nd ed. 1958.
Walker, G. S. M. (1957) Sancti Columbani opera, Dublin, repr. 1970, 1997.
Wallis, F. (1996) ‘Chronology and systems of dating,’ in G. Rigg and F. Man-
tello, Medieval Latin studies: an introduction and bibliographic guide,
Washington, DC, 383–87.
— (1999) Bede: The Reckoning of Time, Liverpool.
Walsh, M. (1987) ‘Some remarks on Cummian’s paschal letter and the Com-
mentary on Mark ascribed to Cummian,’ in P.  Ní Chatháin and
M. Richter, Irland und die Christenheit / Ireland and Christendom: Bi-
belstudien und Mission / The Bible and the Missions, Stuttgart, 216–29.
— (2013) ‘Cummian’s letter: science and heresy in seventh-century Ireland,’ in
M. Kelly and C. Doherty, Music and the stars: mathematics in medi-
eval Ireland, Dublin, 99–110.
Walsh, M. and D. Ó Cróinín (1988) Cummian’s letter De controversia paschali
together with the related Irish tract De ratione conputandi, Toronto.
Warntjes, I. (2005) ‘A newly discovered Irish computus: Computus Einsidlensis,’
Peritia 19, 61–64.
— (2010a) The Munich Computus: text and translation. Irish computistics be-
tween Isidore of Seville and the Venerable Bede and its reception in Caro-
lingian times, Stuttgart.
— (2010b) ‘The Argumenta of Dionysius Exiguus and their early recensions,’ in
Warntjes and Ó Cróinín (2010), 40–111.
— (2010c) ‘A newly discovered prologue of AD 699 to the Easter table of Vic-
torius of Aquitaine in an unknown Sirmond manuscript,’ Peritia 21,
255–84.
— (2011) ‘The Computus Cottonianus of AD 689: a computistical formulary
written for Willibrord’s Frisian mission,’ in Warntjes and Ó Cróinín
(2011), 173–212.
— (2012) ‘Köln als naturwissenschaftliches Zentrum in der Karolingerzeit: Die
frühmittelalterliche Kölner Schule und der Beginn der fränkischen
Komputistik,’ in H. Finger and H. Horst, Mittelalterliche Handschrif-
ten der Kölner Dombibliothek, Viertes Symposion, Köln, 41–96.
— (2015) ‘Victorius vs Dionysius: the Irish Easter controversy of AD 689,’ in
P. Moran and I. Warntjes, Early medieval Ireland and Europe: chronol-
ogy, contacts, scholarship, Turnhout, 33–97.
Warntjes, I. and D. Ó Cróinín (2010) Computus and its cultural context in the
Latin West, AD 300–1200, Turnhout.
Warntjes, I. and D. Ó Cróinín (2011) The Easter controversy of Late Antiquity
and the early Middle Ages: its manuscripts, texts, and tables, Turnhout.


 Bibliography

Wasserschleben, H. (1885) Die irische Kanonensammlung, 2nd ed., Leipzig, repr.

Aalen 1966.
Watson, A. G. (1978) ‘Thomas Allen of Oxford and his manuscripts,’ in M. B.
Parkes and A. G. Watson, Medieval scribes, manuscripts and libraries:
essays presented to N.R. Ker, London, 279–314, repr. in A. G. Watson,
Medieval manuscripts in post-medieval England, Aldershot 2004, ar-
ticle XIII.
Weikmann, H.  M. (2004) Heimo von Bamberg: De decursu temporum, Han-
nover.
Wemple, S. F. (1974) ‘Claudius of Turin’s organic metaphor or the Carolingian
doctrine of corporations,’ Speculum 49, 222–37.
Wiesenbach, J. (1986) Sigebert von Gembloux: Liber decennalis, Weimar.
Williams, H. (1912) Christianity in early Britain, Oxford.
Williams, J. (1991) ‘Die Geschichte des Morgan-Beatus,’ in J. Williams et al.,
Beatus-Apokalypse der Pierpont Morgan Library: ein Hauptwerk der
spanischen Buchmalerei des 10. Jahrhunderts, Stuttgart, 17–30.
— (1994–2003) The illustrated Beatus: a corpus of the illustrations of the Com-
mentary on the Apocalypse, 5 vols, London.
Willis, J. (1994) Ambrosii Theodosii Macrobii Saturnalia, 2 vols, Stuttgart.
Wilmanns, W. (1872) ‘Ein Fragebüchlein aus dem neunten Jahrhundert,’
Zeitschrift für deutsches Alterthum 15, 166–80.
Wilmart, A. (1933) ‘Reg. Lat. 39: un nouveau texte du faux concile de Césarée
sur le comput pascal,’ in A. Wilmart, Analecta Reginensia. Extraits des
manuscrits latins de la Reine Christine conservés au Vatican = Studi e
Testi 59, Vatican, 19–27.
— (1937–45) Codices Reginenses Latini, 2 vols, Vatican.
Wilson, H. A. (1896) The Missal of Robert of Jumièges, London.
Wolf, K. B. (1999) Conquerors and chroniclers of early medieval Spain, 2nd ed.,
Liverpool.
Wood, I. and C. Grocock (2013) Abbots of Wearmouth and Jarrow, Oxford.
Wormald, F. (1934) English Kalendars before A. D. 1100, London.
Wright, C. D. and R. Wright (2004) ‘Additions to the Bobbio Missal: De dies
malus and Joca monachorum (fols 6r–8v),’ in Y. Hen and R. Meens,
The Bobbio Missal: liturgy and religious culture in Merovingian Gaul,
Cambridge, 79–139.
Wright, C. J. (1997) Sir Robert Cotton as collector: essays on an early Stuart cour-
tier and his legacy, London.
Zahn, T. (1884) Supplementum Clementium, Erlangen.
Zapke, S. (2007) Hispania vetus: musical-liturgical manuscripts: from visigothic
origins to the franco-roman transition, 9th–12th centuries, Bilbao.
Zelzer, M. (1978) ‘Zum Osterfestbrief des heiligen Ambrosius und zur römis-
chen Osterfestberechnung des 4. Jahrhunderts,’ Wiener Studien 91,
187–204.


 Bibliography

Zerubavel, E. (1985) The seven day circle: the history and meaning of the week,
New York.
Zimmer, H. (1901) ‘Keltische Kirche in Britannien und Irland,’ Realen-
cyklopädie für protestantische Theologie und Kirche 10, 204–43.
Zimmer, H. (1902) The Celtic Church in Britain and Ireland, trans. by A. Meyer
from Zimmer (1901), London.


 INDEX

Biblical citations

Genesis: Ecclesiastes:
1: 271 11:2: 300–01
1:11: 154 46:5: 259
1:14: 269
2:3: 174 Jeremiah:
7:11: 288 39:2: 291
8:14: 290 52:12: 291
14:14: 54
45:10–47:27: 124 Matthew:
27:62: 284–85
Exodus:
12:2: 124, 148 Luke:
12:2–18: 107 3:1: 279
12:18: 101–02, 124 3:23: 280
16:26–29: 301–02 23:44–45: 187, 238
19:1: 274
19:16: 274, 290 John:
23:19: 287 3:15–16: 301–02
40:2: 291 3:30: 287
11:9: 238
Joshua:
10:12: 264 Romans:
10:13: 259, 264 9:24: 227
10:14: 259, 264
1 Corinthians:
2 Kings: 9:27: 227
23:23: 54
 Index

Colossians: Apocalypse:
1:18: 300–01 6:12–17: 189
12:1–8: 189
James:
2:26: 301–02 1 Enoch:
72:8–32: 235

Sources

Abbo of Fleury: 4, 27–28 Ambrose of Milan:

De sacramentis:
Acta synodi Caesarea: 153–56, 5.21–23: 174
158–59, 166, 316, 318–21, 347 Epistolae extra collectionem:
1: 150 13 (AD 386): 19, 328

Admonitio generalis of AD 789: 11 Ambrosiaster:

In epistolam ad Romanos: 299
Ado of Vienne:
Chronicon in aetates sex Anatolius Latinus:
divisum: 269 De ratione paschali: 7, 18, 38,
71–93, 102–03, 125, 219,
Ælfric: 314, 327
De temporibus anni: 5, 9, 29 10: 273
62–63: 261 11: 235
13: 252–53, 257
Ailerán the Wise: 14: 121–22
Canon Euangeliorum: 228
Aratus Latinus: 24–25
Alcuin: 11
Arg. Aquens. of AD 816: 14
Pseudo-Alcuin:
De bissexto: 250–51, 340 Arn Serm.: 22

Aldhelm: Augustine: 104

De virginitate: Contra Faustum:
11: 226 16.6: 174
De civitate Dei:
Alexander de Villa Dei: 22.30: 174
Massa compoti: 5 De diversis quaestionibus:
56: 287
58.1: 287


 Index

De Genesi ad litteram: 1: 223

1.1: 250 2: 305–06
2.13: 155 3: 237–38, 254
2.14: 244–46 4: 237–38
4.24: 250 6: 232, 238, 269
De Genesi contra Manicheos: 11: 232–33, 271, 274–75
1.4.7: 239 12: 230
De haeresibus: 147 13: 230, 271
De trinitate: 16: 236
4.4: 233, 245–47 20: 272, 274–75
4.5: 287 21: 234
Enarrationes in pslamos: 22: 272–73
132.11: 287 24: 237
Epistola ad Ianuarium: 103 30: 285–86
In Iohannis Evangelium: 36: 236
14.5: 287 38: 232–33, 237, 248, 252
Questiones Exodi: 39: 233, 245–47, 251–52
90: 287 40: 232
47.6: 287 41: 232, 235, 289
45: 271
Augustinus Hibernicus: see De 47: 144–45, 270, 280–81,
mirabilibus sacrae scripturae 283, 285–88, 318
51: 139, 336
Auraicept na n-éces: 223 52: 272
57: 272
Pseudo-Balbus: 59: 271
Liber de asse minutisque eius 61: 271, 280, 285–86
portiunculis: 65: 144–46
7: 237 66 (Chronica maiora): 266,
269–70, 285–86
Beatus of Liébana: 188–90, 193–4, 67: 266
198 Epistola as Wicthedum:
8: 336
Bede: 6, 10, 21–22, 74, 154 12: 286
De natura rerum: 9 Explicatio Apocalypseos:
De temporibus: 9, 132 9: 152
6: 230 17: 151–52
9: 236 Expositio in Iohannis
10: 232 evangelium: 297–99
22 (Chronica minora): 266 Expositio in Lucae evangelium:
De temporum ratione: 9, 132, 1.1.24: 287
297–99, 302 Expositio in Marci evangelium:
praef.: 266 299


 Index

Expositio in primam epistolam Cap. comp. of AD 809: 14

Iohannis: 299 2: 282
Historia abbatum:
15: 320–21 Cassiodorus:
Historia ecclesiastica: 17 Complexiones in Apocalypsi:
2.4: 140 13: 167–68
2.19: 128–29 Computus of AD 562: 4, 6, 20,
3.25: 166, 323 139, 339–40
3.29: 163–65 Expositio in psalmos: 297–99
5.21: 157, 346 Institutiones:
2.praefatio.1: 168
Pseudo-Bede: 1.23: 139
De argumentis lunae: 259, 287, 1.23.2: 142
306 1.23.3: 142–43
1.29.3: 168
Benchuir bona regula: 227 1.3.1: 168–69

Bobbio Computus (see also Milan, Censorinus:

Biblioteca Ambrosiana, H 150 De die natali liber:
inf.): 18.3–4: 240–41
33: 240 18.12: 240–41
35: 240
39–43: 240, 247 Christum peto: 226–28
42: 259
43: 257 Chronicon paschale: 329, 336, 340
44: 251
140: 333 Chronograph of AD 354: 48, 52,
146: 340 328

Bobbio Missal: 260 Claudius of Turin:

Brevis chronica: 38–39, 265–92
Brigit bé bith maith: 227
Claudius Ser.: 267–70
Byrhtferth:
Enchiridion: 9, 29 Clavis Melitonis: 297–99
4: 298, 303
Pseudo-Byrhtferth:
Glossae: 29 Clemens Scottus:
Problemata arithmetica: 228
Caesarius of Arles:
Sermones: Cölner Prolog: see Prologus Paschae
216: 287
Collatio compoti Romani et Arabici:
Calcidius: 23 34


 Index

Collectio Canonum Hibernensis: 137 Computus Carthaginensis of AD

21: 133 455: 15–16
25: 133 1.6: 51
29: 133 2.4–5: 51
31: 318
37: 133 Computus Digbeanus of AD 675:
39: 133 243, 258–59
40: 133
42: 133 Computus Einsidlensis: 20, 132,
45: 133 264
52: 133 97: 234
56: 133 99: 253–55
105: 240, 251
Columbanus: 154 106: 234
Epistolae: 107: 233, 248–49
1: 16–17 108: 233, 240, 244, 248–49,
1.3: 327 257, 259
1.4: 115, 144, 150–51, 324,
335–36 Computus Pariensis of AD 754: 21
1.5: 140
2: 16–17 Computus Petri: 36

Pseudo-Columbanus: Computus Rhenanus of AD 776: 23

De saltu lunae: 32
De sollemnitatibus: 316, Concilium Bracarense:
320–24, 339–41 2.9: 183

Comp. Col. of AD 805: 137 Concilium Ovetense: 196

2.2: 237
3.2: 247, 249 Concilium Toletatum:
3.3: 232, 239, 247, 251 4.5: 183
4.8: 122–23 4.17: 189
5.7: 259 7: 194–95
9.17: 184
Computus of AD 243 (Pseudo- 10: 194–95
Cyprian, De pascha computus): 11: 194–95
8, 14, 38, 43–70, 167, 326 13.18: 184, 195–96
19: 237–38 14: 194–95
15: 194–95
Computus of AD 757: 23 16: 194–95

Computus of AD 789: 23 Crónica albeldense: 196


 Index

Cummian: De duobus testibus: 148

De controversia paschali: 10,
38, 94–137, 154, 312, 318, De mirabilibus scripturae: 3
322–23 Prologue: 263
ll. 6–17: 95–96 1.6: 260
ll. 20–26: 107 1.7: 237
ll. 32–41: 107–08 2.4: 259, 261–63
ll. 47–51: 98
l. 65: 107 De ordine creaturarum: 226
ll. 71–78: 107 9.4–6: 237
ll. 80–85: 107–08
ll. 86–89: 166 De pascha computus: see Computus
ll. 92–93: 126 of AD 243
ll. 93–94: 101
ll. 99–100: 109 De ratione conputandi: 10, 20, 132,
l. 110: 131 218, 264
ll. 114–18: 97 21: 237
ll. 123–29: 102–03, 107 26: 250
ll. 170–72: 130–31 28: 236
ll. 180–86: 102–03, 107, 35: 234
109–10 39: 233
ll. 204–23: 100–01, 104, 46: 236
107, 124, 157, 219, 327, 50: 253–54
340 51: 253–54
ll. 223–29: 107, 109–11, 235 52: 231, 243
ll. 223–43: 241, 332 53: 236–37, 240
ll. 259–70: 125–27 55: 244–45
l. 283: 97 56: 233, 245, 247
ll. 284–85: 109 57: 232
58: 263
Cunestabulus: 35 86: 340–41
98: 323, 332
Pseudo-Cyprian:
De pascha computus: see De ratione temporum (Annus solis):
Computus of AD 243 212–28

De argumentis lunae: see Pseudo- De sollemnitatibus: see Pseudo-

Bede Columbanus

De comparatione epactarum Dionysii Dial. Burg. of AD 727: 14

et Victorii of AD 689: 21, 282 12: 247
14: 247
De divisionibus temporum: 242–43,
249, 299 Dial. Langob. of c.AD 750: 14


 Index

Dial. Neustr. of AD 737: 14 Epist. Rat. of AD 809: 14

27: 247
27a: 233 Epistola Cyrilli: 158–60, 241,
28: 255–56 331–32
28a: 234, 252, 255–56, 263 3: 332

Dicuil: Epistola Pascasini: 111, 158–59,

Liber de astronomia: 3, 21 333–34
2–3: 333
Dionysius Exiguus: 2, 12, 19–20,
27–28, 38, 50, 74, 97, 102, 104, Epistola Proterii: 158–59, 334
127–28, 138–81, 183–86, 197, 2–3: 280
266, 279–81, 334, 337–39
Argumenta: 20, 142–43, 158, Eterius of Osma:
339 Adversus Elipandum libri duo:
Epistola ad Petronium 193
(Prologus): 55, 126, 148,
158–60, 166–68, 337–39 Eugipius:
Epistola ad Bonifatium et Excerpta ex operibus
Bonum: 160, 236, 337–39 S. Augustini:
118: 103
Pseudo-Dionysius:
Argumentum XV: 306 Eusebius of Caesarea:
Argumentum XVI: 243–45, Chronicon: 266, 281, 3314
247, 257–59, 261, 264 Historia ecclesiastica:
5.23: 318, 324
Disputatio Chori et Praetextati: 38, 5.24: 320
230–31, 236 5.25: 318, 347
6–7: 346
Disputatio Morini: 18, 48–49, 104, 6.20: 52
158, 340–41 6.22: 51, 53, 326
ll. 61–67: 102 6.48: 327
7.20–22: 327
Elipandus: Vita Constantini:
Epistola ad Migetium: 193 3.5: 145
3.7–18: 328
Ennius: 4.34–35: 328
Annales:
291: 173 Felix of Squillace: 19–20, 139–40,
144–46, 156–76, 338
Epiphanius:
Panarion: 50 Gennadius:
De viris illustribus: 147


 Index

88: 335 2: 51

Gerland the Computist: 7, 33, 36 Hilarus:

Epistola ad Victorium: 334–35
Gildas:
De excidio Brittaniae: 226 Hilary of Poitiers:
De Synodis:
Gregory of Tours: 86: 55
De cursu stellarum: 256–57 Tractatus super psalmos:
Historia Francorum: 135.7: 261
5.17: 185
10.31: 197 Hippolytus: 7–8, 43, 48, 51–53,
55, 98, 115, 278–79, 326
Gregory the Great:
Epistola ad Leandrum: Hrabanus Maurus:
5: 153 De computo: 10, 295, 297
Epistolae: De institutione clericorum:
11.41: 152 2.42: 304–05
11.48: 152 De rerum naturis: 39
11.51: 152 10: 293–308
14.2: 174 10.1: 303, 305–06
Moralia in Job: 298–99 10.15: 306
Registrum epistolarum: 10.17: 300–08
2.37: 147 11: 296
8.34: 147 16.2: 296
Epistolae:
Helperic: 36: 295
Liber de computo: 3, 26–27 Liber de sanctae cruces: 295

Heimo of Bamberg: In primo certe canone: 227–28

De decursu temporum: 8–9, 28
1.2: 275 Innocent I:
Epistola ad Aurelium
Heriger of Lobbes: Carthaginensem: 331
Epistola ad quondam Hugonem
Monachum: 287 Irish annals: 10, 15, 106, 128, 218

Hermannus Contractus of Isidore of Seville:

­Reichenau: 12–13, 31–37 Chronica: 271–72
1.235–39: 197
Quintus Julius Hilarianus: De ecclesiasticis officiis: 39
Expositum de die paschae et 1.24–25: 304–08
mensis: De natura rerum:


 Index

1: 241 Leo I: 334

3: 167, 302
6: 231–32, 236, 240 Leo monachus:
Etymologiae: 19, 295–99 Epistola of AD 627: 16–17
5.22: 302 4: 192
5.36: 198 6: 192
5.39: 197
6.17: 192, 231–32, 239–40, Leofric Missal: 260–61
333
6.18: 298, 303–04 Lib. ann. of AD 793: 13, 23, 26
6.36: 236 20: 277

Jerome: Lib. calc. of AD 818: 13, 23, 26

Commentarii in Isaiam: 1.7: 282
9.28.21: 259 1.8: 282
De viris illustribus: 147
43: 318 Lib. comp. of AD 809/12: 13, 23,
70: 50 25–26
In Aggaeum: 107 1.15: 268

Joca monachorum: 260 Liber pontificalis: 152, 164

15: 325
Julian of Toledo:
De comprabatione aetatis sextae: Liber quaestionum: 107–08
3.34: 197–8
Macrobius: 23
Laidcenn mac Baíth Bannaig: Saturnalia:
Lorica: 226 1.13.9: 236
1.14.3: 236
laterculus of Augustalis: 15 1.14.4: 236
1.14.6: 230, 236
latercus: 3, 10, 15–16, 97–98, 101, 1.15.1: 233
103, 122–23, 126, 129–31,
133–6 Marianus Scottus: 8, 28

Lect. comp. of AD 760/92: 23 Martianus Capella: 23–24

3.4: 271 De nuptiis Philologiae et
3.5: 271 Mercurii:
3.6: 277 2.220: 173
4.5: 271
4.7: 271
8.12a: 277–78


 Index

Martin of Braga: 7, 185 Commentarii in primum

epistolam sancti Pauli ad
Merobaudes: Corinthios: 299
Carmina: Commentarium in Isaiam
4.5: 173 prophetam: 299

Missale mixtum secundum regulam Ovid:

B. Isidori dictum Mozarabes: 249 Metamorphoseon libri:
13.334: 173
Mozarabic Chronicle of AD 754: Tristia:
194 4.2.55: 173

Muirchú: Pacificus of Verona: 7

Vita Patricii: 105–06, 123–25,
131, 134 Pascasinus:
Epistola: see Epistola Pascasini
Munich Computus: 16, 20, 32,
137, 264 Paschale Campanum: 147, 151–52,
6: 254 169–70
7: 250
8: 250 Paterius:
11: 340 Expositio in veteris ac novi
21: 236 testamenti: 299
29: 234
36: 340 Patrick:
38: 340 Confessio: 106, 125
39: 253
41: 232, 233, 239–40, 243, Paulinus of Nola:
249–50, 252 Carmina:
44: 48, 250 27.499: 173
58: 235
62: 235 Pliny: 13, 23
66: 261–62
Prol. Aquit. of AD 721: 14, 156
Nonae Aprilis: 102, 227–28, 329
Prologus Cyrilli: 18, 102, 104, 241
Notker of St Gall: 8: 236
Quatuor questionibus compoti:
31–32 Prologus Paschae (Cölner Prolog):
149, 170
Origen: 4: 155
Commentarii in epistolam
sancti Pauli ad Romanos:
299


 Index

Prologus Theophili: 18, 102, 104, Ser. nov.: 268

329
4: 280 Sigebert of Gembloux:
Liber Decennalis: 8–9
Prosper of Aquitaine:
Chronicon: 139 Statius:
Silvae:
Proterius: 2.7.20: 173
Epistola: see Epistola Proterii
Suggestio Boni: 140, 159–62, 320,
Quaest. Austr. of AD 764: 14 339–40
2.9: 239–40, 243, 259
Supputatio Romana: 2, 16, 98, 115,
Quaest. Langob. of c.AD 780: 14 119, 149–50

Ratio pascae: 146–53 Synesius of Cyrene:

Epistolae:
Ratio solis vel lunae cursus atque 9: 330
bissexti: 241–42 13: 330

Ratio spere Dionisii de circula mag- Synod of Merida (AD 666):

no paschae: 140–46, 153–56, 6: 184
177–81
Synod of Zaragossa (AD 691):
Reichskalender: 13 194–95
c12: 158 2: 184

Reinher of Paderborn: Tertullian:

Computus emendatus: 5, 35 Adversus Iudaeos:
8.18: 48
(Pseudo-) Remigius of Auxerre:
Commentarium in Matthaeum: Theon:
351: 284–85 Lesser Commentary on the
Handy Tables: 50
Robert of Jumièges:
Missal: 261 Thomas Scottus:
Bis duo nam nivei praesunt et
Roger of Hereford: 35–37 quinque nigelli: 228

Sampirus of Astorga: Tibullus:

Chronicon regum Legionis: 196 Carmina:
2.51: 173


 Index

Tírechán: Prologus: 335–36

Collectanea: 105–06, 123–24, 4: 52, 55
131 5: 52
7: 48
Tractatus Adthanasi: 120–21 9: 235
11: 111–12
Vegetius: 12: 112
Epitoma de re military: 173
Virgilius Maro:
Vergil: Epitomae:
Eclogae: 11: 256
4.10: 173
Walahfried Strabo: 10, 25
Versus de annis a principio: 218
Walcher of Malvern: 33–34
Victorian prologue of AD 699: 21
Willibrord:
Victorius of Aquitaine: 2, 12, Martyrologium: 173
16–18
Calculus: 257, 335 Zachary of Besançon:
Cyclus: 15, 38, 48, 98, 102, 104– In unum ex quattuor: 284
06, 111–20, 125, 134–35,
139–40, 142, 165, 194–5, Zeitz table of AD 447: 15–16
235, 279–80, 334–36

Manuscripts

Angers, Bibliothèque municipale, Bern, Burgerbibliothek, 645:

31 (olim 27): 307 50v: 247
72r–74v: 153–56
Arras, Bibliothèque municipale,
506 (832): 307 Bremen, Universitätsbibliothek,
msc 0041: 192
Berlin, Staatsbibliothek, Phillipps
1830: Cambridge, Corpus Christi Col-
1v–2r: 166 lege, 11: 307

Berlin, Staatsbibliothek, Phillipps Cambridge, University Library, Dd

1833: 27 1.30: 307

Bern, Burgerbibliothek, 408: 164


 Index

Cambridge, University Library, Dd 805r: 173

13.4: 307
Florence, Biblioteca Medicea Lau-
Cambridge, University Library, renziana, Fondo S. Croce, Plut.
L.1.1.10: XXXI Sin. Cod. I: 308
21r: 225
87v: 225 Glasgow, University Library,
Hunterian V.1.3 (366): 308
Cologne, Dombibliothek, 83–II:
156, 186 Graz, Universitätsbibliothek, 234:
15r–36v: 26 243v: 284
141r: 152
184r–185v: 192 Karlsruhe, Badische Landesbiblio-
193v–197r: 158, 170 thek, Aug. perg. 68: 296
201r–203r: 322
215r: 241 Karlsruhe, Badische Landesbiblio-
thek, Aug. perg. 96: 296, 307
Darmstadt, Hessische Landes- und
Hochschulbibliothek, 416: 307 Karlsruhe, Badische Landesbiblio-
thek, Aug. perg. 195: 175
Einsiedeln, Stiftsbibliothek, 321 42r: 171
(647):
82–125: see Computus Laon, Bibliothèque municipale,
Einsidlensis 422: 26
137–45: 235
Leiden, Universiteitsbibliotheek,
El Escorial, Real Biblioteca, d.I.1: VLF 5: 308
200–01
Leiden, Universiteitsbibliotheek,
El Escorial, Real Biblioteca, d.I.2: VLQ 69:
200–01 7r–47v: 226

El Escorial, Real Biblioteca, f.I.12: León, Archivo de la Catedral, 8:

307 19, 186, 200–11
1v: 202–03
Épinal, Bibliothèque municipale, 5v–9v: 203
72: 226 9v–19v: 203
19v: 204–11
Erfurt, Universitätsbibliothek, 20r–27v: 203
Amplonianus fol. 42: 226 25v: 204

Florence, Biblioteca Medicea Lau-

renziana, Amiatino 1:


 Index

London, British Library, Cotton Milan, Biblioteca Ambrosiana, F

Caligula A XV: 47, 186, 321, 60 sup.:
323 61r: 223
73r–80r: 47
77r: 160–61 Milan, Biblioteca Ambrosiana, H
82v–83v: 49, 341 150 inf.:
97v–105v: 43–45, 47–50, 4v: 241
60–61, 326 50r–v: 156, 167
107r: 47 80r: 340
108r–v: 139, 169
London, British Library, Cotton
Tiberius A III: 30 Molfetta, Biblioteca del Seminario,
5.7.V (3): 308
London, British Library, Cotton
Vitellius A XII: Montecassino, Archivio
79r–83r: 95, 312 dell’Abbazia, 132 EE: 307
81v: 99
Montecassino, Archivio
London, British Library, Harley dell’Abbazia, 150: 162, 173
3017:
186v: 277–78 Montpellier, Bibliothèque de la
Faculté de Médecine, 157: 346
London, British Library, Harley 10v–40v: 347
3092: 296
Monza, Biblioteca Capitolare,
London, British Library, Royal c–9/69:
12.G.xiv: 308 66r–83v: 267
68v: 288
Lucca, Biblioteca Capitolare Fe- 79r–v: 268–69, 288
liniana, 490: 15 80r: 289–90
81v: 274
Madrid, Biblioteca Nacional, 3307: 82r–v: 291–92
25 115v–116r: 267

Madrid, Biblioteca Nacional, 9605: Munich, Bayerische Staatsbiblio-

104r–112r: 267 thek, Clm 4546:
112v–117v: 267 268v: 284
113r: 269, 288
113v: 289–90 Munich, Bayerische Staatsbiblio-
115r: 291–92 thek, Clm 14725:
23v: 161
Madrid, Biblioteca Nacional,
12370: 308 Naples, Biblioteca Nazionale,
Borbon. V.C.46: 308


 Index

New York, Pierpont Morgan Oxford, New College, 159: 308

Library, 644:
112r: 189–90 Oxford, St. John’s College, 5: 308

Oxford, Bodleian Libray, Auct. D. Oxford, St John’s College, 17: 9, 31

2. 14: 173
Oxford, St. John’s College, 88: 308
Oxford, Bodleian Libray, Bodley
309: 120, 131–32, 319–20, Oxford, Trinity College, 64: 308
329, 332, 336
64v–73v: 242–43 Padua, Biblioteca Antoniana, I
82v–84r: 320, 339–40 27: 15
97r–v: 322 74v–75r: 235
98r–99r: 320–21
101r: 241 Paris, Bibliothèque nationale de
103v: 236 France, Lat. 609: 19
104r: 236 54r–55v: 185
106r: 120–21
Paris, Bibliothèque nationale de
Oxford, Bodleian Library, Digby France, Lat. 2420: 308
63: 38, 157–76
6r: 338 Paris, Bibliothèque nationale de
20v: 157 France, Lat. 4860:
49v–68v: 170 148v–149v: 167
59r–v: 162, 163–64, 320 150r–v: 340
60r: 160–61
63r–67r: 157, 166, 169 Paris, Bibliothèque nationale de
70v–71r: 139, 157, 168, 338 France, Lat. 4871:
72v–79r: see Computus 110r–v: 151
Digbeanus
79r–81r: 340 Paris, Bibliothèque nationale de
80v–87v: 170 France, Lat. 5001:
1r–8v: 267, 274
Oxford, Bodleian Library, Junius 1r: 269
11: 170–71, 175
Paris, Bibliothèque nationale de
Oxford, Bodleian Library, Laud France, Lat. 5239:
misc. 746: 308 4r–6r: 167

Oxford, Bodleian Library, Smith Paris, Bibliothèque nationale de

31 (pt 3): France, Lat. 5543:
70–73: 39, 309–51 7r–9r: 167


 Index

Paris, Bibliothèque nationale de Reims, Bibliothèque municipale,

France, Lat. 7608: 308 441: 308

Paris, Bibliothèque nationale de Reims, Bibliothèque municipale,

France, Lat. 9561: 226 442: 308

Paris, Bibliothèque nationale de Rome, Biblioteca Angelica, 1413: 8

France, Lat. 10837: 173
St Gall, Stiftsbibliothek, 161:
Paris, Bibliothèque nationale de 44: 245
France, Lat. 11684: 308
St Gall, Stiftsbibliothek, 878: 25
Paris, Bibliothèque nationale de
France, Lat. 11777: St Gall, Stiftsbibliothek, 904:
244v: 46 171–72

Paris, Bibliothèque nationale de St Petersburg, Nacionál’naja Bib-

France, Lat. 13411: 308 liotéka, Q.v.I.6:
1r: 144
Paris, Bibliothèque nationale de
France, Lat. 16879: 308 Stuttgart, Württembergische
Landesbibliothek, Theol. et
Paris, Bibliothèque nationale de philos. 2ø45: 308
France, Nouv. acq. lat. 1613:
12v: 161 Troyes, Bibliothèque municipale,
604: 169
Paris, Bibliothèque nationale de
France, Nouv. acq. lat. 1615: Vatican, Biblioteca Apostolica, Pal.
155r: 161 lat. 291: 308

Paris, Bibliothèque nationale de Vatican, Biblioteca Apostolica, Pal.

France, Nouv. acq. lat. 2169: lat. 824:
2v: 186 1v: 168
5v–9v: 186
Vatican, Biblioteca Apostolica,
Paris, Bibliothèque nationale de Reg. lat. 39: 319
France, Nouv. acq. lat. 16361:
322 Vatican, Biblioteca Apostolica,
275r: 161 Reg. lat. 123: 26, 156

Reims, Bibliothèque municipale, Vatican, Biblioteca Apostolica,

298 (now destroyed): 156 Reg. lat. 324: 46, 60, 63–65


 Index

Vatican, Biblioteca Apostolica, 68r–69r: 140–46, 153–56,

Reg. lat. 391: 308 177–81

Vatican, Biblioteca Apostolica, Venice, Biblioteca nazionale di S.

Reg. lat. 586: Marco, Lat. II 56: 308
108r–111v: 166
125r–v: 245–46 Verona, Biblioteca Capitolare,
XXII (20):
Vatican, Biblioteca Apostolica, 93r: 162
Reg. lat. 755:
1v–3v: 166 Vienna, Österreichische National-
Vatican, Biblioteca Apostolica, bibliothek, 121: 296
Reg. lat. 2077: 139–40, 147
78r–v: 148, 152 Vienna, Österreichische National-
79r–81r: 146–56 bibliothek, Ser. Nov. 37: 256
96v–98v: 147, 165, 169–70
99r: 152 Wolfenbüttel, Herzog-August-
101r: 172 Bibliothek, Cod. Guelf. 532
Helmst. (579):
Vatican, Biblioteca Apostolica, Vat. 85r–v: 151
lat. 642:
89r–90v: 321 Worcester, Cathedral Library, F.21:
308
Vatican, Biblioteca Apostolica, Vat.
lat. 1548: 139–44 Zürich, Zentralbibliothek, Car. C
51r–75v: 141 97 (263): 308

