The archaeology of ancient cult: from foundation

deposits to religion in Roman Mithraism

Matthew M. McCarty, Mariana Egri and Aurel Rustoiu
Introduction
In the past two decades, the “archaeology of religion” has moved from the margins of
scholarship to the center, led by the growth of postprocessual archaeological hermeneu-
tics.1 Such theoretical frames — whether the materiality of religion, objects as agents, the
entanglement of humans and objects, or “thing theory” — demonstrate the centrality of
the physical world and its archaeological correlates to religion. They offer new ways of
posing questions about the construction of meanings for worshippers through materials.2
If the benefits of these approaches for placing archaeology at the heart of ancient
worshippers’ thought-worlds are clear, such “archaeologies of religion” also work counter-
productively, for they continue to isolate “religion” as a distinct object of ancient practice
and of modern enquiry. Yet as a host of scholarship on ancient religion has pointed out,
religious knowledge in the Roman world was not primarily predicated on internalized
beliefs or interpretative frameworks, but rather was mediated in a range of externalized,
social forms.3 The concept of “religion” being an interior state or consisting primarily of
propositional claims owes much to Enlightenment, Protestant thought; some have even
advocated abandoning the term “religion” altogether when dealing with antiquity, for
the term presupposes a domain of activity and thought that never had neat boundar-
ies in antiquity.4 Rather, religious knowledge in the Roman world was predicated upon
the performance of, and participation in, actions — in J. Scheid’s aphorism, “faire, c’est
croire”.5 It was primarily the manipulation of objects, images and environments — ritual
practice — that constituted the religious life of the Roman world. Instead of fixating upon
savoir penser, ways of thinking or the interpretative frameworks that worshippers might
apply to or derive from objects, it may be worth turning greater attention to savoir faire, to

1 Recent work mostly focused on prehistory and post-antiquity includes T. Insoll, Archaeology,
ritual, religion (New York 2004); id., The Oxford handbook of the archaeology of ritual and religion
(Oxford 2011); D. Barrowclough and C. Malone (edd.), Cult in context: reconsidering ritual in
archaeology (Oxford 2007); E. Kyriakidis, The archaeology of ritual (Los Angeles, CA 2007). For the
Mediterranean world in particular, see R. Raja and J. Rüpke (edd.), Blackwell companion to the
archaeology of religion in the ancient world (Oxford 2015).
2 The work edited by C. Moser and J. Knust, Ritual matters: material remains and ancient religion
(Ann Arbor, MI 2017), proposes to apply these frames to ancient religion, but it is telling that
the theoretical frames they propose are hardly taken up in any of the papers; instead, individual
chapters primarily offer reconstructions of rites.
3 E.g., J. Scheid, Quand faire, c’est croire: les rites sacrificiels des Romains (Paris 2005); J. Rüpke,
Religion of the Romans (Cambridge 2007) 6-15.
4 B. Nongbri, Before religion: a history of a modern concept (New Haven, CT 2013); cf. T. Masuzawa,
The invention of world religions, or, How European universalism was preserved in the language of
pluralism (Chicago, IL 2005), for the creation of “religion” as an academic category within
the context of 19th-c. colonialism. Note that it was precisely at this time, in the context of
encyclopedias, that the use of Mithraism and discussion of it as a religion, gained wide currency
in the Anglophone world especially: e.g., G. T. Bettany, The great Indian religions, being a popular
account of Brahmanism, Hinduism, Buddhism, and Zoroastrianism ... (London 1892).
5 Scheid (supra n.3).
© Journal of Roman Archaeology 32 (2019)
https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press
 280 M. M. McCarty, M. Egri and A. Rustoiu

the technical and craft knowledge that enabled such rites to happen ‘properly’ in the first
place. Along similar lines, C. Ando has stressed the practical and empirical nature of reli-
gious epistemologies in the Roman world: knowing how to deal with divine powers often
mattered more than particular propositional claims about those powers.6 Efficacy — how-
ever ancient worshippers might measure it — achieved through technical expertise and
the performance of rites was central to cult.
If religion in the Roman world was largely about social savoir faire, then it ought to be
treated like any other kind of craft or technology: areas and practices for which archaeol-
ogy has long had analytical tools. They are the same tools that have allowed analysis of
myriad other socio-cultural processes, their transmission, and the ways they were loaded
with significance. Before asking what sites, objects, or practices may have “meant” to, or
“acted” on, worshippers, we may first wish to ask questions about how the transmission
of these practices occurred and how it shaped them. We may, in other words, need a more
processual archaeology of cult.7 But what would such an approach look like in practice?
Close, contextual analysis of materials and site formation processes can allow the detailed
articulation of a ritualized chaîne opératoire that governed the performance of events deal-
ing with the divine. Most archaeological accounts of sanctuary sites and their rites stop
here, at what is, in essence, the starting point for anthropological accounts of religion: a
description of what happened, even if more ephemeral aspects of those practices (includ-
ing spoken words) are lost or must be reconstructed from the more usual, textual sources
for ancient ‘religion’. Moving beyond description of the unfolding of rites requires posing
additional questions that can shed light on the social dynamics of ancient cult. In the case
of ritual remains, the questions can be as simple as: How did people know what to do?
How was ritual knowledge transmitted through time and through space? Such questions
have long occupied anthropologists of religion, and such comparative work can offer fruit-
ful frameworks, if not explanatory mechanisms.
The cult of Mithras offers a unique opportunity to see how such an approach can work
in practice and how it can provide new insights on ancient cult.8 Mithraism is a good
test-case for two reasons. First, most accounts of the cult, even those that rely heavily
on material remains, focus on reconstructing Mithraism’s thought-content and its savoir
penser: beliefs, esoteric doxa, propositional claims, ‘star talk’.9 Such reconstruction hap-
pens through an hermeneutic process of decoding either individual symbols or entire

6 C. Ando, The matter of the gods: religion and the Roman empire (Berkeley, CA 2008) 13-15.
7 Cf. C. Renfrew, The archaeology of cult: the sanctuary at Phylokapi (London 1985).
8 Despite being modern coinages, terms like “Mithraism,” “Mithraic” and “Mithraist” provide
convenient shorthand to collect the body of things and people associated with worshipping
Mithras, and they will be maintained throughout this paper. For further discussion of these
problems, see R. Gordon, “Institutionalized religious options: Mithraism,” in J. Rüpke (ed.), A
companion to Roman religion (Oxford 2008) 392-405.
9 Such an approach became standard with F. Cumont, Les mystères de Mithra (Brussels 1913).
R. Beck, The religion of the Mithras cult in the Roman empire (Oxford 2006), explicitly rejected
“doctrine” from iconography as an antiquated approach, but through the lens of cognitive
science and ‘star talk’ he creates something that looks rather like Cumont’s mentalité; the
primary distinction is how Beck grounds the system of referents for signs in the cult. R. Turcan,
Mithra et le mithraïcisme (Paris 2004), read the cult in a Platonic light, heavy with allegory.
A. Mastrocinque, The mysteries of Mithras: a different account (Tübingen 2017), likewise focuses
primarily on symbolic, iconographic interpretation. D. Ulansey, The origins of the Mithraic
mysteries (New York 1989), focused on the astrological content by reading individual symbols.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 281

symbol-systems; that the cult encoded its content in metaphor or allegory is taken as a
given. Of course, it is not hard to see why accounts of Mithraism unfold in this manner,
given the source material from which they start. On the one hand, iconographically rich,
complex scenes of Mithras stabbing the bull are one of the most widespread features of
the cult. With details that seem neither to be iconic or narrative — including a veritable
zoo of animals that accompanies the act of Mithras stabbing the bull — such reliefs seem
to invite allegorical or symbolic readings.10 On the other hand, ancient textual mentions of
Mithraism seem to privilege similar symbolic understandings of the cult’s ‘content’. Even
accounts of the cult that rely largely on material or iconographic sources, beginning with
F. Cumont and continuing through R. Beck, take these texts as their point of departure.11
Most of these sources, however, are Christian polemic, in some cases written centuries
after the last archaeological evidence for Mithras-worship; sometimes contradictory and
unreliable, they are often given to allegorical framing.12 The other major text (itself citing
two earlier works devoted to Mithraism about which we know little) is Porphyry’s 3rd-c.
allegorical reading of Homer’s Odyssey, De antro nympharum. Instead of providing the key
to interpreting Mithraism, this work shoe-horns Mithras-worship into Porphyry’s wider
Neoplatonic arguments about the nature of the cosmos, quarrying the cult for evidence
of the cross-cultural recognition of Neoplatonic truths.13 Given such problems, the focus

10 Cf. J. Elsner, Art and the Roman viewer (Cambridge 1995), arguing that changing modes of
viewing images created new types of allegorical art.
11 Cumont (supra n.9). It is also telling that his Textes et monuments figurés relatifs aux mystères de
Mithra (Brussels 1899) began with literary testimonia. Beck (supra n.9) 16-17 explicitly takes
Porphyry as the “key” to Mithraism and its “intent”.
12 The main texts are collected in Cumont 1899 (supra n.11), and re-presented in A. S. Geden,
Select passages illustrating Mithraism, translated, with an introduction (London 1925). The heavily
filtered, sometimes willfully distorted, content of such polemics is difficult to evaluate: for
example, Pseudo-Nonnus describes initiation as involving 80 grades of hardship, in contrast
to the 7 grades that appear in Jerome and are accepted as “canonical” (Pseudo-Nonniani in IV
Orationes Gregorii Nazianzeni Commentarii 4.70; Jer., Ep. 107.2). The two major Christian sources
for Mithraism contemporary with the cult’s 2nd-c. floruit — Justin Martyr and Tertullian — are
both given to allegorical interpretation of sacred text and practice in such polemic contexts. While
Tertullian has often been seen as a non-allegorical interpreter of text, G. Dunn, “Tertullian’s
scriptural exegesis in De praescriptione haereticorum,” J. Early Christian Studies 14 (2006) 141-
55, demonstrates that he employs a range of interpretative tools in different polemic contexts,
using allegory primarily when his opponents did not; the symbolic reading of Mithraic rites,
then, implies that pagan (or heretical) opponents may have invoked the cult differently.
13 For Porphyry’s allegorical mode, see R. Lamberton, Homer the theologian: Neoplatonist allegorical
reading and the growth of the epic tradition (Berkeley, CA 1986); P. T. Struck, Birth of the symbol:
ancient readers at the limits of their texts (Princeton, NJ 2004). Indeed, De Antro has even been
read as its own polemic against Gnostic interpreters, and as a prescriptive manual for “proper”
modes of Neoplatonic interpretation: M. Edwards, “Porphyry’s ‘Cave of the Nymphs’ and the
Gnostic controversy,” Hermes 124 (1996) 88-100. For Porphyry and Neoplatonism in Mithraism,
see R. Turcan, Mithras Platonicus: recherches sur l’hellénisation philosophique de Mithra (Leiden
1975). For Porphyry as the ‘key’ to interpreting Mithraism, see Beck (supra n.9). Porphyry’s
cosmology does map neatly onto at least one mithraeum, that of Sette Sfere in Ostia (R. Gordon,
“The sacred geography of a mithraeum: the example of Sette Sfere,” JMithStud 1 [1976] 119-65),
but this may reflect the particular social and educational milieu in which it was built, rather
than a generalized principle. F. Coarelli, “Apuleio a Ostia?” DialArch 7 (1989) 27-42, followed
by R. Beck, “Apuleius the novelist, Apuleius the Ostian householder, and the Mithraeum of
the Seven Spheres: further explorations of an hypothesis of Filippo Coarelli,” in S. G. Wilson
and M. Desjardins (edd.), Text and artifact in the religions of Mediterranean antiquity (Waterloo,

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 282 M. M. McCarty, M. Egri and A. Rustoiu

on ways of thinking and interpreting engendered by these textual sources should not pre-
clude an approach that focuses on the cult as craft and that seeks to answer basic questions
about technical know-how and the actual practices within the cult: What was done in and
around a mithraeum? How often was this done? Who did these things? How standardized
were the practices across the empire?
The cult of Mithras was, in some ways, rather different than the forms of civic or state
cult for which practice-based models of religion were developed. Many of these differ-
ences were made quite explicit for worshippers, as means of establishing inversions and
distance from other forms of cult: Persian outfits and loan-words, sunken sanctuaries
rather than those raised on podia. Still, isolating Mithras-worship from its wider cultic
ecosystem — as something fundamentally different in kind from, say, cults of Mercury
or the emperor (which also preferred sanctuaries of similar designs and shared images)
— depends primarily on the problematic historiography that grouped Mithraism among
“Oriental cults” and teleological narratives that sought to explain the ultimate triumph
of Christianity.14 In the Roman empire, Mithraism and its material correlates developed
within, and depended upon, social structures and visual referents shared with the types of
public cults that J. Scheid examines.15 R. Gordon long ago demonstrated the ways Mithras-
cult was interwoven with the wider social structures of empire;16 more recently, D. Walsh
has shown the ways that societal changes in the 4th-5th c. led to the rapid decline in the
cult’s popularity — further evidence for how tightly Mithras-worship had been entangled
with wider Roman society.17 Recognizing that Mithras-worship was deeply integrated into
a practice-oriented world of cult craft encourages us to approach Mithraism from a similar
perspective.
The second reason Mithraism makes a good case-study is because the need to under-
stand the performance of ritual acts, and how the knowledge to do so was transmitted, is
particularly acute in this new elective cult. Transmission happened very rapidly (within
a century) and very widely (from Britain to Syria), in a manner strikingly different from
the majority of locally-oriented cults that were merely concerned with diachronic, inter-
generational transmission. In addition, most worshippers seem to have begun worshipping
Mithras as adults; unlike most techniques for dealing with the gods, Mithras-worshippers
were not socialized into Mithraism’s practices as children.18 This process of transmis-
sion has not been neglected in Mithraic studies, but it is generally viewed on the scale of
geographic diffusion from one region to another, or on the scale of institutions (imperial

ON 2000) 551-57, in fact attributes the reversed order of Mars and Venus on the benches to
Apuleius’ works and his personal connection to the block. This may suggest the rôle of personal
elaboration and forms of exegesis when decorating a mithraeum.
14 For the problems of Cumont’s “Oriental cults” as a historiographic category, see the papers in
C. Bonnet, V. Pirenne-Delforge and D. Praet (edd.), Les religions orientales dans le monde grec et
romain: cent ans après Cumont (1906-2006) (Rome 2009).
15 For Mithraic differentiation while belonging to the same religious system, see V. Huet, “Reliefs
mithriaques et reliefs romains ‘traditionnels’: essai de confrontation,” in Bonnet et al. (supra
n.14) 233-56.
16 R. L. Gordon, “Mithraism and Roman society: social factors in the explanation of religious
change in the Roman Empire,” Religion 2.2 (1972) 92-121.
17 D. Walsh, The cult of Mithras in late antiquity (Leiden 2018).
18 F. Prescendi, “Children and the transmission of religious knowledge,” in V. Dasen and
T. Spaeth (edd.), Children, memory, and family identity in Roman culture (Oxford 2010) 73-93.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 283

administration, soldiery, families) where this diffusion might take place.19 On a smaller
scale, the picture of how individuals learned to worship Mithras rests on analogy and
probability rather than on ancient evidence: Gordon has posited traveling Mithraic “col-
porteurs”, comparable to the Evangelical tract-peddlers of the 19th c.;20 alternatively, he
has posited a simple model of contact diffusion, with individuals moving to a new place
and bringing whatever knowledge of the cult they had with them, to form the basis of a
new community.21 Fine-grained archaeological analysis can offer the chance to test such
hypotheses about the transmission of savoir faire.
Such questions about practices and the knowl-
edge of how to enact them are the sorts of questions
answerable from archaeological analysis, and from
close, contextual analysis of finds, botanical and
faunal assemblages within mithraea.22 They gov-
erned our design of the Apulum Mithraeum III
Project, a mithraeum within the canabae (later pro-
moted to a municipium) outside the legionary fort
at Apulum in Dacia (fig. 1). The project examines
the micro-archaeology of cult practice. Preliminary
results from three seasons demonstrate the ways
that a focus on the archaeology of cult technolo-
gies can aid understanding of Mithras-worship.
Our approach allows for four key arguments to be
developed. First, a structured deposit in Mithraeum
III provides evidence for a foundation rite that
commemorated a shared meal, and allows descrip-
tion of the ritualized behaviors that resulted in
the deposit. Second, this type of deposit, which is
common in a range of mithraea across the empire,
offers concrete evidence for the commensurability
of complex Mithraic ritual practice through time
and space; it may be possible to speak of networks
of shared practice rather than either “Mithraism” or
Fig. 1. Main settlement areas of Apulum
cellular cults of Mithras. Third, building on work in (authors, after A. Diaconescu, in I. Haynes
the anthropology of religion, such commonalities in and W. S. Hanson [edd.], Roman Dacia
practice imply specialized ritual training, given the [JRA Suppl. 56, 2009] 107, fig. 4.13).

19 Diffusion models: Cumont (supra n.9) 33-69; R. Merkelbach, Mithras (Königstein 1984) 146-53;
M. Clauss, Cultores Mithrae (Stuttgart 1992) 253-58. Institutional vectors for the spread of cult:
Cumont (supra n.9); M. Clauss, The Roman cult of Mithras: the god and his mysteries (Edinburgh
2000) 33-38; R. Gordon, “The Roman army and the cult of Mithras: a critical view,” in C. Wolff
and Y. Le Bohec (edd.), L’armée romaine et la religion sous le Haut Empire romain (Lyon 2009)
379-450.
20 R. Gordon, “Who worshipped Mithras?” JRA 7 (1994) 459-74, followed by Clauss 2000 (supra
n.19) 37; Beck (supra n.9) 53. I. Tóth’s notion, in “A Dacian ‘apostle’ of the cult of Mithras?”
Specimina Nova 8 (1992) 153-60, of a Mithraic “apostle” has been solidly rejected. See further
below.
21 Gordon (supra n.8) 400.
22 An approach represented by M. Martens and G. de Boe (edd.), Roman Mithraism: the evidence of
the small finds (Brussels 2004).

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 284 M. M. McCarty, M. Egri and A. Rustoiu

infrequency of such foundation rites. Fourth, in the case of Mithraeum III this expertise
comes not from the person paying for the sanctuary, but rather from a separate, perhaps
mobile, ritual expert. Distinct networks of cultic know-how intersected at the founding
of the sanctuary. Although few archaeological contexts related to the use of the building
survive, a structured deposit set beneath the floor of the nave allows for a detailed under-
standing of the dénouement of the foundation rites.

The archaeology of cult: Apulum’s Mithraeum III

Mithraeum III was so named because a range of Mithraic material had been unearthed in
Apulum from the 18th c. onwards.23 The number of mithraea that once stood in the Colonia
Aurelia Apulense and the Municipium Septimium Apulense (where Mithraeum III is situated)
has been estimated at anywhere from 3 to 6.24 None of this material has archaeological
context; at least one concentration comes from the Colonia, and the largest concentration
comes from a farmer’s garden — the “Oancea mithraeum”, which more probably repre-
sents a lime kiln or other collection of culled statuary, given its diversity.25 The only clues
for dating this material come from onomastics in the inscriptions and from the sculptural
style, both problematic criteria; the third quarter of the 2nd c. A.D. is a rough terminus post
quem.26 Given at least two distinct findspots for Mithraic material found in Apulum prior
to 2008, there were at least two mithraea in the twin settlement areas; the discovery in
2008 of a new mithraeum brings the minimum number to three. It was discovered during
rescue excavations undertaken by the Muzeul Naţional al Unirii, Alba Iulia, when a local
developer sought to turn a private house and its backyard garden into a hotel. The site sits
within the canabae that developed around Dacia’s main legionary fort, c.400 m southwest
of the fort walls. Over two weeks, the team from the museum exposed rubble founda-
tions of a building that contained three broken inscriptions dedicated to Mithraic deities.27
Coupled with the building’s layout, these allowed identification of a mithraeum. Then,
given the economic uncertainties of the time, the construction project was halted and the
site backfilled.
From 2013 to 2016, the Apulum Mithraeum III Project, under the direction of the pres-
ent authors, documented the remains uncovered in 2008 and continued excavation of the
building and its surroundings using single-context methods practiced by the Museum
of London. In addition to the 2008 excavation which had removed any Roman use and
destruction layers across much of the sanctuary, two major post-Roman interventions
affected preservation and the possibility of reconstructing practices. First, in the 11th c.,

23 CIMRM nos. 1935-2005.

24 Cumont 1899 (supra n.11) 251. A. Rusu-Pescaru and D. Alicu, Templele romane din Dacia (Deva
2000) 137-38, and J. R. Carbó García, Los cultos orientales en la Dacia romana (Salamanca 2010)
131, suggest 4 mithraea, while C. Szabó, “The cult of Mithras in Apulum: communities and
individuals,” in L. Zerbini (ed.), Culti e religiosità nelle province danubiane (Ferrara 2015) 410,
suggests 5-6 mithraea.
25 Vermaseren (CIMRM 2, 277) recounts his unsuccessful expedition to locate the Oancea property.
For the finds, see V. Christescu, “Nouveaux monuments d’Apulum,” Dacia 3-4 (1927-32) 620-25;
C. Daicoviciu, “Neue Mitteilungen aus Dazien,” Dacia 7-8 (1937-40) 308-9; CIMRM nos. 1953-
67. In Inscriptions d’Apulum (Paris 2001) pl. 3, 30, I. Piso places the mithraeum near the corner of
Bulevardul Încoronării and Strada Izvorului, but the evidence for this positioning is lacking.
26 G. Sicoe, Die mithräischen Steindenkmäler aus Dakien (Cluj-Napoca 2014) 111-28.
27 M. E. Egri, M. M. McCarty, A. Rustoiu and C. Inel, “A new Mithraic community at Apulum
(Alba Iulia, Romania),” ZPE 205 (2018) 268-76.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 285

Fig. 2. Aerial photograph of site of Apulum Mithraeum III, with post-Roman interventions highlighted (Apulum
Mithraeum III Project).
a semi-sunken mediaeval house was built at the SW corner of the mithraeum, cutting
through the Roman foundations (fig. 2). Its builders quarried the mithraeum for construc-
tion materials. Ceramic roof-tiles were collected from the Roman ruin and used to create a
paved ‘patio’ outside the house at 25-30 cm above Roman ground level; the house’s central
oven was made of rubble packing that included part of a limestone column shaft, a cornice
block, and a well-head from earlier Roman structures (perhaps the mithraeum, but possi-
bly other buildings).28 Stones had clearly been robbed from the mithraeum’s foundations,
probably in the same period. The smaller stone altars in the mithraeum’s nave had been
smashed, while a larger altar was discovered tipped forward, face down, with portions of
its back roughly sawn off, presumably for re-use after the entire altar proved too heavy to
move. It is uncertain whether the altar was toppled in the collapse of the mithraeum, as
part of mediaeval quarrying efforts, or was ritually ‘decommissioned’ in a manner similar
to the altar in the Mithraeum at Inveresk, perhaps as part of the Roman ‘withdrawal’ from
Dacia in the mid-3rd c.29 The 11th-c. robbing left intact little evidence for the mithraeum’s
ancient use. Second, at some point in the post-Roman period the natural geology caused a
transverse slump, opening a sinkhole at the SE corner of the mithraeum (fig. 2). The corner
of the foundation slipped downwards, leaving a gaping pit that was used as a rubbish tip
until the early 18th c., when a Vauban-style fort was built over the Roman/mediaeval fort

28 Patio: Context 1273. Hearth: Context 1196.

29 Inveresk: F. Hunter, M. Henig, E. Sauer and J. Gooder, “Mithras in Scotland: a mithraeum
at Inveresk (East Lothian),” Britannia 47 (2016) 119-68. The date, nature, and impact of the
‘withdrawal’ — long accepted as being a formal military abandonment in 274/275, based
on SHA, Aurel. 39 — has recently been challenged by the observation that new coins cease
entering Dacia in the 250s: C. Găzdac, Monetary circulation in Dacia and the provinces from the
middle and lower Danube from Trajan to Constantine I (AD 106-337) (Cluj-Napoca 2010); see also
D. Ruscu, “L’abandon de la Dacie romaine dans les sources littéraires (II),” Acta Musei Napocensis
37 (2000) 265-75, and C. Oprean, “Dacia, Pannonia and Raetia in the time of Gallienus: a com-
parative approach,” in Actes du XIVème Congrès UISPP, Université de Liège, 2001 (BAR S1302;
Oxford) 13-18, for a critical comparison of literary, epigraphic and archaeological evidence.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 286 M. M. McCarty, M. Egri and A. Rustoiu

Fig. 3. Plan of Apulum Mithraeum III

(Apulum Mithraeum III Project).

of Alba Iulia, its outermost edge placed 100 m from the mithraeum; the spoil from excavat-
ing the new fort’s moat was dumped over the area of the mithraeum, sealing the Roman
and mediaeval remains. Again, the sinkhole meant that little of the Roman-period use of
the area was preserved intact.
Nonetheless, the mithraeum’s layout and date can be reconstructed from available data
(fig. 3). It sat on rubble foundations that rose about 1 Roman foot (28 cm) above ground-
level as a dwarf wall to protect a timber-and-mud superstructure. It was oriented E–W,
with the main entrance set at the NE corner where a doorsill was robbed out. That entrance
led into a small antechamber, subdivided by a wall on a lighter foundation on the S side
probably to create a small storage space, akin to those at other mithraea.30 From the ante-
chamber, one stepped down into a nave flanked by two raised benches whose floor level
can be reconstructed just below the outside ground level. The walls of the benches use
larger stones than do the outer walls, creating a sense of descent into a rocky cave. At the
W end of the nave, a mortar step, once paved with robbed stone or ceramic slabs (imprints
survive), leads up to a recessed niche. That niche presumably displayed a scene of Mithras
stabbing the bull, probably in relief given the fragment of a relief plaque discovered near-
by.31 The main altar, dedicated to Sol Invictus Mitrhas (sic), sat at the W end of the nave.
One of the fragmentary inscriptions found in 2008 was a statue-base dedicated to the
torchbearer Cautopates. It exactly matches a damaged statue-base whose inscribed face
was removed after antiquity but which was found tipped forward on the dirt floor of the

30 E.g., Riegel, Martigny and Bieshiem: A. Schatzmann, “Möglichkeiten und Grenzen einer funk-
tionellen Topographie von Mithrasheiligtümern,” in Martens and de Boe (supra n.22) 13-14.
31 Egri et al. (supra n.27).

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 287

nave along the N bench. The damaged base probably supported a statue of Cautes, the
two torchbearers being displayed towards the W end next to the benches. In other words,
the mithraeum was laid out in a fairly ‘canonical’ manner; with a pair of benches 7.5 m
long accommodating 20-25 reclining Mithraists, it was on a par with mithraea across the
empire.32 The building was designed and scaled to suggest that its anticipated usage was
for a community of ‘standard’ size.
The mithraeum’s rubble foundations were set into a layer of dirt 20 cm deep that was
spread evenly to cover the remains of earlier, lighter, post-built timber structures. The foun-
dations also sit over an earlier rubbish pit that was filled in prior to construction. Débris in
the rubbish pit included several fragments
of Lézoux black-slipped and sigillata ware
and southern Pannonian thin-walled bea-
kers that provide a terminus post quem of
A.D. 150-170, suggesting that we have the
earliest evidence for Mithraic cult in Dacia,
and one of the earliest mithraea in central
Europe.33
Unfortunately, the mediaeval and rescue
interventions make it nearly impossible to
speak about the ritual use of the building in
the Roman period. One structured deposit,
however, allows for a more detailed account
of ritual practice. In the floor of the nave, a
small box built of ceramic tiles cut through
the leveling layer for the building’s con-
struction (fig. 4). It contained a mix of ash,
charcoal, burned vegetal remains, and ani-
mal bone.34 The top of the box was exposed
in 2008 but it was excavated fully only in Fig. 4. The tile box deposit in the nave, after
excavation (Apulum Mithraeum III Project).

32 This assumes 50 cm (30 concurrent reclining participants) to 60 cm (25 concurrent reclining

participants), as per L. M. White, “The changing face of Mithraism at Ostia,” in D. Balch and
A. Weissenrieder (edd.), Contested spaces: houses and temples in Roman antiquity and the New
Testament (Tübingen 2012) 471-74; cf. A. Hensen, Mithras der Mysterienkult an Limes, Rhein und
Donau (Stuttgart 2013) 397.
33 For dating Mithraism in Dacia (terminus post quem of A.D. 168, one monument datable to 178-
180), see Sicoe (supra n.26) 111-24. At Novae (Moesia), an inscribed relief showing Cautes/
Cautopates was dedicated by Melichrisus, a slave of the conductor publici portorii ripae Thraciae,
whose activity can be dated around A.D. 100 (CIMRM no. 2269; cf. Merkelbach [supra n.19]
148). The first probably datable monument from the Pannonian provinces is an altar from
Carnuntum, dedicated by C. Sacidius Barbarus, centurion of legio XV Appollinaris (CIMRM
no. 1718), stationed there until A.D. 113. The terminus ante quem is slightly uncertain since
vexillations may have returned to Carnuntum during the Marcomannic Wars (Gordon [supra
n.19] 393). Otherwise, the earliest evidence for the cult in the Pannoniae comes from Poetovio’s
Mithraeum I, dated to c.150-160 through its connection to conductores of the portorium Illyrici:
CIMRM nos. 1487-1510.
34 The archaeobotanical analyses were conducted by B. Ciută (Universitatea “1 Decembrie 1918”,
Alba Iulia); the archaeozoological analysis by G. El-Susi (Institute of Archaeology and Art
History, Cluj-Napoca). We are grateful to both for sharing their results.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 288 M. M. McCarty, M. Egri and A. Rustoiu

2013 under controlled conditions. The contents represented a single, structured deposi-
tion, although there was evidence for intrusive species (parts of a mouse skeleton that
taphonomically appeared much later than the other bones). Two of the side tiles were slop-
ing inward, the result of compaction after the box had been set in place. As the tiles showed
no signs of direct flame, the deposit was not burned in the box; rather, the contents appear
to have been burned elsewhere, collected, and put into the box. In other words, the deposit
represents intentional burning of vegetal and animal materials, presumably as an offering,
followed by the curation of the remains.
Table 1
CARBONIZED SEEDS FROM THE TILE BOX DEPOSIT IN THE NAVE OF
APULUM MITHRAEUM III (analysis by B. Ciută).
Latin name No. of charred seeds Common name
Vitis vinifera 2 grape vine
Prunus spinosa 1 stone blackthorn
Galium aparine 16 cleavers
Convolvulus arvensis 24 field bindweed
Cf. Raphanus raphanistrum 1 silica wild radish
44 seeds

Fig. 5. Carbonized seeds from

the tile box deposit (Apulum
Mithraeum III Project).

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 289

Especially for the small seed remains, the deposit probably includes only that portion of
the original material burned that had been collected and placed in the box. The botanical
remains included 44 carbonized seeds (fig. 5; Table 1), all intentionally burned. The major-
ity were weeds associated with cereal cultivation (Convolvulus arvensis, Galium aparine,
Raphanus raphanistrum). Convolvulus arvensis, which grows around (and chokes) cereals, is
often harvested along with grain and survives the threshing process, like Galium aparine
(both types of large seed are usually separated from grain at later stages of processing
using coarse sieves or by hand).35 Their presence here (and the absence of any cereals) may
suggest that they were part of agricultural waste dried and used as tinder; alternatively,
they possibly represent burned grain products or wildflowers.36 The fruit seeds, grape and
sloe probably represent the burned offering of these fruits, one cultivated, the other gath-
ered; such fruit offerings are common in Roman ritual contexts of all kinds.37 The presence
of sloe may hint at the season
of the ritual — the fruit appears
in late summer (August) and
ripens through November (in
modern Transylvania sloe is har-
vested in October or November)
— but a somewhat earlier time
of year seems probable based
on the Galium and Convulvulus
seeds, both plants developing
their seeds from August to Sep-
tember/October. In short, the
botanical remains suggest
burned vegetal offerings of fruit
and perhaps also grain products.
The faunal assemblage seems
to represent a similar process
of burning foodstuffs, but also
included unburned table-scraps
(fig. 6). Of the non-intrusive
bones or fragments in the box,
almost all that can be identified Fig. 6. Selection of bones from the tile box deposit (Apulum
belong to either adult chickens Mithraeum III Project).

35 C. J. Stevens, “An investigation of agricultural consumption and production models for

prehistoric and Roman Britain,” Environmental Arch. 8 (2003) 61-76. The high proportion of
weed seeds also stands in contrast to other ritual contexts (especially funerary) in the Roman
provinces, where the grains offered were well-cleaned: B. Cooremans, “The Roman cemeteries
of Tienen and Tongeren: results from the archaeobotanical analysis of the cremation graves,”
Vegetation History & Archaeobotany 17 (2007) 3-13.
36 M. Robinson, “Domestic burnt offerings and sacrifices at Roman and pre-Roman Pompeii,
Italy,” Vegetation History & Archaeobotany 11 (2002) 97, suggests that Silene gallica seeds in
Pompeian offering deposits might represent burned flowers.
37 Robinson ibid.; B. Zach, “Vegetable offerings on the Roman sacrificial site in Mainz, Germany
— short report on the first results,” Vegetation History & Archaeobotany 11 (2002) 101-6; K. Reed,
L. Lodwick, T. Leleković and H. Vulić, “Exploring Roman ritual behaviours through plant
remains from Pannonia Inferior,” Environmental Arch. 24 (2019) 28-37.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 290 M. M. McCarty, M. Egri and A. Rustoiu

(Gallus dom., 14) or pig (Sus s. dom., 15). The MNI represented includes 3 piglets (aged
around 2 months), 2 pigs (aged 8-10 and 14-16 months), and 3 chickens.38 Almost all of the
chicken bones come from the meaty portions (thigh, wing, drumstick); they were expertly
butchered, and portions selected, before being offered. At least 3 chicken thighs and 4
wings were exposed to direct flame but at a low temperature, consistent with offering at a
small altar (rather than cremation on a high-temperature pyre). However, 4 wings, 1 thigh
and 1 drumstick show no signs of direct flame but were probably stewed. This suggests
that the meat portions underwent two processes: some was burned as an offering, some
may simply have been table waste. The same is true of the piglet bones. Two bones — a
metapodial (from the foot) and femur (from the leg) were burned; five — a parietal and
canine (from the head), 2 ribs, and 1 humerus (from the leg) — were unburned. This sug-
gests that the piglets underwent a different process prior to being offered; instead of being
butchered with the meatiest portions selected for offering/consumption, the whole piglet
made its way into the sanctuary, perhaps cooked, offered and/or consumed whole. The
age of the piglet also hints at the season of the ritual. As represented by the agronomists,
Roman practice favoured two annual farrowings, one in mid-summer (June/July), a second
in mid-winter (December/January).39 Faunal remains from the Tienen mithraeum seem to
confirm that a second farrowing might happen in the Roman world.40 At two months old,
the piglet in the box would fit well with the seasonality of the botanical sample, suggesting
that it may have been slaughtered in August/September.

From micro-archaeology to practice

The contents of the box represent a complex rite, involving burning foodstuffs and also
collecting unburned foodstuffs; any unburned vegetal remains are unlikely to have sur-
vived. The remains seem to represent a multi-dish banquet. The burned and unburned
portions do not come from significantly different parts of the animals, suggesting that the
foodstuffs selected for burning came from the same dishes as those consumed.
The difference in treatment of these common dishes — some burned, some unburned
— may hint at two different kinds of diners: divine diners who consumed their portion
through burning, and human diners who consumed their portion and whose scraps were
collected without burning. The ‘logic’ of burning portions for a deity to eat is well attested
in the classical world.41 The centrality of dining to Mithraism has long been recognized:
the space of a mithraeum is laid out as a triclinium, with Mithras in the position of honor;
images of Mithras reclining to dine with Sol appear not only as framing scenes, but in some
cases on the back side of reversible bull-stabbing reliefs that would allow the dining god

38 No beef bones were found, although this does not preclude the presence of beef; beef is more
commonly cooked and consumed off the bone, unlike pig and chicken: R. C. G. M. Lauwerier,
“Bird remains in Roman graves,” Archaeofauna 2 (1993) 79.
39 R. C. G. M. Lauwerier, “Pigs, piglets and determining the season of slaughtering,” JArchSci 10
(1983) 483-88.
40 A. Ervynck and K. Dobney, “A pig for all seasons? Approaches to the assessment of second
farrowing in archaeological pig populations,” Archaeofauna 11 (2002) 7-22.
41 W. Burkert, Homo necans: the anthropology of ancient Greek sacrificial ritual and myth (Berkeley,
CA 1983) 4-6; M. Beard, J. A. North and S. R. F. Price, Religions of Rome (Cambridge 1998)
35-36; J. Scheid, La religion des Romains (Paris 2007) 75-76; Rüpke (supra n.3) 140-49;
J. Scheid, “Roman animal sacrifice and the system of being,” in C. Faraone and F. Naiden (edd.),
Greek and Roman animal sacrifice (Cambridge 2012) 84-95.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 291

to be present in the sanctuary with his worshippers; and faunal remains from mithraea
often show evidence for meals.42 The fact that in Mithraeum III the god ‘ate’ the same cuts
of meat as his worshippers, rather than either prized cuts or inedible bits, is more unusual
in the context of ancient cult meals. Normally, distinct portions were reserved for most
deities, highlighting their exceptionality and distinctiveness from human worshippers.43
In such a context, Mithras is less a member of a hierarchical community and more of an
equal of his worshippers. Like a good host in a non-cultic Roman banquet, Mithras did
not reserve (or have reserved) the best food for himself, but shared special dishes with his
worshippers.44
And special these particular dishes were. Analyses undertaken at other mithraea have
demonstrated how the faunal assemblages are markedly different than those generated
from domestic waste.45 Mithraea consistently contain high percentages of chicken bone, a
meat that seems to have been consumed rarely in non-religious contexts. The predilection
for chicken may have to do with the bird’s association with the sun, for Sol plays a signifi-
cant rôle in the cult, and with the purported geographic and cultural origin of Mithraism
in Persia (chicken was called ‘the Persian bird’).46 Similarly, the suckling pig consumed at
our mithraeum seems to be a rare delicacy; the animal was slaughtered before it reached
full size. As the general profile of animal bones from several sites points to a beef-rich diet
in the province of Dacia, the foods consumed by the Mithraic community from which
remains were placed in the box set the meal apart from regular dining activities, if not from
other Mithraic banquets.47

42 Clauss 2000 (supra n.19) 108-13. For Mithraic dining, see J. P. Kane, “The Mithraic cult meal and
its Greek and Roman environment,” in J. Hinnells (ed.), Mithraic studies (Manchester 1975) 313-
51; A. Griffith, “Amicitia in the cult of Mithras: the setting and social functions of the Mithraic
cult meal,” in K. Mustakallio and C. Kroetzl (edd.), De amicitia: friendship and social networks in
antiquity and the Middle Ages (Rome 2010) 63-77.
43 Rüpke (supra n.3) 140-49 emphasizes the hierarchical nature of this division within Roman
sacrifice, building on J. Scheid, Romulus et ses frères : le collège des Frères Arvales (Rome 1990); cf.
Scheid 2012 (supra n.41).
44 The excavators of the Kempraten mithraeum plan to test this hypothesis by examining whether
there were distinctions in the faunal assemblages generated as waste from different positions
on the benches inside. For placement and hierarchy in Roman triclinia, see K. M. D. Dunbabin,
The Roman banquet: images of conviviality (Cambridge 2003) 36-46. In “The Roman convivium
and the idea of equality,” in O. Murray (ed.), Sympotica. A symposium on the symposion (Oxford
1990) 308-20, J. H. D’Arms argued for a rhetoric of equality in Roman banqueting that conceals
a firmly delineated hierarchy; similar tensions seem to be at play between shared food and
discrepant seating placement in the mithraeum.
45 A. Lentacker, A. Ervynck and W. Van Neer, “Gastronomy or religion? The animal remains
from the mithraeum at Tienen (Belgium),” in S. J. O’Day, A. Ervynck and W. Van Neer (edd.),
Behaviour behind bones: the zooarchaeology of ritual, religion, status and identity (Oxford 2003) 77-94;
C. Olive, “Le faune exhumée des mithraea de Martigny (Valais) et d’Orbe-Boscéaz (Vaud) en
Suisse,” in Martens and de Boe (supra n.22) 147-56.
46 A. Lentacker, A. Ervynck and W. Van Neer, “The symbolic meaning of the cock. The animal
remains from the mithraeum at Tienen (Belgium),” in Martens and de Boe (supra n.22) 57-80.
47 A. Gudea, Contribuţii la istoria economică a Daciei romane (Cluj-Napoca 2007). There is little
evidence in primary contexts for the waste from regular communal meals in Apulum’s
Mithraeum III, but the general faunal profile from the perimeter of the building points to similar
portions of meat consumed, as do the faunal remains from other mithraea. Similar observations
have been made at other mithraea: A. von den Driesch and N. Pöllath, “Tierknochen aus
dem Mithrastempel von Künzing, Lkr. Deggendorf,” in. K. Schmotz (ed.), Vorträge des 18.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 292 M. M. McCarty, M. Egri and A. Rustoiu

The community participating in this meal was also not tiny. The diversity of foods on
offer suggests a banquet fit for a group. The amount of meat consumed hints at more than
a few participants. With at least 3 chickens and 5 piglets sacrificed, it is possible to estimate
the weight of the meat they provided. Assuming a weight of around 1.5 kg of meat per
chicken, and 6-25 kg per piglet, based on their ages, the animals represented in the box
furnished at least 34 kg (and up to 133.5 kg) of meat.48 Even with one piglet — whose head-
to-toe remains imply use of the entire carcass — at least 6 kg of meat was processed and
consumed: enough to feed a substantial group of diners, even if only the portions found
within the box were consumed on-site.
Beyond the communal aspects of the meal, the collection in the box of scraps from both
divine and human diners reflects a second feature of this ritual: the deposit is part of a foun-
dation rite. The box appears to have been set below the floor level of the nave, probably to
be sealed when the floor was laid. It was cut into the levelling layer that prepared the area
for construction of the mithraeum. As the 2008 excavation removed soil down to below the
top of the vertical tiles of the box, the stratigraphic relationship between floor level and
box must be inferred from other, preserved, forms of evidence. The tipped (Cautes?) altar
nearby and the mortar step at the W end of the nave offer evidence for the original level
of the floor (fig. 7). The tipped altar sat on a layer of compacted earth that extended up to
242.44 m asl. The mortar bed-
ding for pavers sat at 242.40 m
asl (fig. 7 at A), but the top of
the pavers themselves would
have been several cm higher,
probably around the level of
the mortar bedding for the
next step at about 242.46 m
asl.49 As preserved, the high-
est edge of the tiles extended
Fig. 7. Diagram showing heights of features in the nave. A: Mortar
bedding for paved steps to cult niche (ht. of lower bedding 242.4 m asl; up to 242.34 m asl, or 6-11 cm
height of step 242.47 m asl). B: Foundation/dwarf wall separating nave lower than the reconstructed
and cult niche (ave. ht. 242.7 m asl). C: Tile box (max. ht. 242.34 m floor level of the mithraeum;
asl). D: Leveling/fill layer in nave, truncated by 2008 excavations (ave. the lowest probable level of
ht. 242.4 m asl). E: Proposed lowest level of flooring (242.4 m asl). the floor is shown in fig. 7 at
E. Even if another tile had originally covered the top of the box, adding an extra 4-5 cm of
thickness, this would be below the dirt floor of the nave. Comparable deposits in mithraea
(see below) similarly sat at or below floor level. All of this suggests that the tile box was
set under the floor of the sanctuary, which means that this deposit was buried before the
building was fully completed and opened for use: the banquet and the curation of its
remains were part of a foundation or dedication rite. Once the floor was laid, the deposit
was inaccessible; this foundation was a one-time event.

Niederbayerischen Archäologentages (Rahden 2000) 145-62 (Künzing); Lentacker et al. (supra n.45)
(Tienen).
48 For the weight of Roman chickens (1.5-2.0 kg, on a par with modern examples), see J. Peters,
Römische Tierhaltung und Tierzucht (Rahden 1998) 222-26. For the weight of the piglets, I am
grateful to G. El-Susi, whose analysis is forthcoming.
49 Context 1111.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 293

A number of distinct rites in antiquity, with different aims, often get grouped under
the heading of “foundation deposit” or “building offering”. Beyond well-studied Near
Eastern and Greek examples, in the Roman world these include everything from shafts
cut and filled with animal remains that were common in the towns of Roman Britain, to
the burial of sacrificed animals at the entrances to buildings while construction was tak-
ing place, to coins or other objects set under thresholds, to curated tableware buried under
hearths.50 Such offerings, which seem to vary widely across the empire, are often explained
as representing the continuity of earlier, micro-regional traditions; for example, under-
hearth deposits are primarily confined to the Valais region, shaft burials are common in
Britain, and other animal deposits are found primarily on the empire’s northern fringes.51
Yet “building offering” and “foundation deposit” are both chronological in that they place
the act in time relative to construction of a building. Such typologies do little to explain the
logic or purpose behind the banquet and the curation of its remains at Apulum. As with all
special deposits, greater analytical precision is necessary to understand what the deposit
was doing here.52
One path forward is to consider the function of such deposits. Looking at Near Eastern
and Archaic Greek deposits, G. Hunt suggested trying to separate different aims behind
depositional behaviour.53 Some building deposits appear to be protective, filled with
apotropaic materials such as epigraphic imprecations or set at liminal locations includ-
ing beneath thresholds;54 others are sanctifying, creating a sacred space either by using

50 In general: V. Lambrinoudakis, “Consecration / consécration / Konsecration / consacrazione,”

ThesCRA 3 (2005) 304-46; I. Beilke-Vogt, Das Opfer im archäologischen Befund: Studien zu den sog.
Bauopfern, kultischen Niederlegungen und Bestattungen in ur- und frühgeschichtlichen Siedlungen
Norddeutschlands und Dänemarks (Rahden 2007). Near Eastern foundation deposits: C. Ambos,
Mesopotamische Baurituale aus dem 1. Jahrtausend v.Chr. (Dresden 2004). For discussion in the
Greek world, see S. Weikart, Griechische Bauopferrituale: Intention und Konvention von rituellen
Handlungen im griechischen Bauwesen (Berlin 2002); G. Hunt, Foundation rituals and the culture
of building in ancient Greece (Ph.D. diss., Univ. of N Carolina, Chapel Hill 2006). There is no
comparable, overarching synthesis for the Roman world. For Romano-British shaft-burials,
see P. Woodward and A. Woodward, “Dedicating the town: urban foundation deposits in
Roman Britain,” World Arch. 36 (2004) 68-86. For horses buried at the entrances to buildings at
the Roman farm at Druten: R. C. G. M. Lauwerier, “The economic and non-economic animal:
Roman depositions and offerings,” in S. J. O’Day, A. Ervynck and W. Van Neer (edd.), Behaviour
behind bones: the zooarchaeology of ritual, religion, status and identity (Oxford 2003) 69. For hearths,
see F. Wiblé, Martigny-la-romaine (Martigny 2008) 123-24.
51 For continuities from local Iron Age traditions as explanatory category, see F. A. Gerritsen, Local
identities: landscape and community in the late prehistoric Meuse-Demer-Scheldt region (Amsterdam
2003) 63-65; M. Groot, Animals in ritual and economy in a Roman frontier community: excavations
in Tiel-Passewaaij (Amsterdam 2008) 113; M. Fulford, “Links with the past: pervasive ‘ritual’
behaviour in Roman Britain,” Britannia 32 (2011) 199-218; J. Magness, “The ‘foundation deposit’
from the Dura Europos synagogue reconsidered,” in B. D. Wescoat and R. G. Ousterhout (edd.),
Architecture of the sacred (Cambridge 2012) 231-47. For localized patterning, cf. D. Schmid,
“Bauopfer in Augusta Raurica: zu kultischen Deponierungen im häuslichen Bereich,” in
C. Ebnöther and R. Schatzmann (edd.), Oleum non perdidit. Festschrift für Stefanie Martin-Kilcher
zu ihrem 65. Geburtstag (Basel 2010) 285-93.
52 Cf. R. Osborne, “Hoards, votives, offerings: the archaeology of the dedicated object,” World
Arch. 36 (2004) 1-10.
53 Hunt (supra n.50).
54 Hunt ibid. 175; cf. S. Ellis, “‘Papsukkal’ figures beneath the daises of Mesopotamian temples,”
Revue d’Assyriologie 61 (1967) 60-61 for Mesopotamia.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 294 M. M. McCarty, M. Egri and A. Rustoiu

numinous materials or establishing a logic to the space.55 This seems to be the case with
the special deposits set under the floor and within the benches of the mithraeum at Mun-
delsheim, where half of a bovine skull was set in each of the benches.56 Although the two
halves did not definitely come from the same animal, their placement suggests that they
worked to articulate a logic for the space, equating the right/left side of the mithraeum
with those sides of the animal. Such deposits could be commemorative, marking a par-
ticular event integral to the foundation of the building (e.g., inauguration rites). Such
categories need not be exclusive; a gladius included within the barracks at Vindonissa may
have been both protective and symbolic.57 Distinguishing among these perceived func-
tions based on materials and placement allows a more nuanced understanding of ritual
behaviors and their material traces. In the case of our tile box, the commemorative rôle of
the deposit is clear. The depositor(s) collected remains from an event (a banquet), rather
than selecting more symbolic offerings, even if the portion of the remains curated for inclu-
sion suggest that they metonymously pointed to the entire event. The box sits in the center
of the nave, towards the W end, rather than in a liminal place; it does not appear primarily
to be protective.
Commemorating a specific meal (perhaps an inaugural meal) suggests that communal,
small-group dining played an important practical and conceptual rôle in the cult. Noting
that such dining occurred within the cult and could be ritualized in significant ways is
hardly novel; rather, the box at Apulum sheds new light on the ways such an event could
be treated as central and foundational within a Mithraic community. Instead of esoteric,
allegorical ‘meanings’, this banqueting event assumes both chronological and conceptual
precedence in the creation of the mithraeum. The community itself, the group of diners,
also takes on central significance, and it existed prior to completion of the mithraeum.58
Reconstructing archaeologically this act of foundation allows a different perspective on
the cult.
A word should be said about the timing of the event in August/September, given the
association in scholarship of Mithraic cult with astronomical phenomena.59 Cited as proof
of Mithraists observing important astronomical moments such as solstices and equinoxes
are, first, a celebration of deceased members at Virunum that seems to have been timed
to coincide with the summer solstice, according to which a ritual tied to non-calendrical
events (people dying) was made to fit astronomical time;60 and second, a large-scale
banquet perhaps connected to renovation of the mithraeum at Tienen that involved the
slaughter and consumption of hundreds of animals whose age at slaughter has suggested

55 Hunt ibid. 175-76.

56 D. Planck, “Ein römisches Mithräum bei Mundelsheim, Kreis Ludwigsburg,” Archäologische
Ausgrabungen in Baden-Württemberg 1989/90, 177-83. Cf. R. Beck, “Ritual, myth, doctrine, and
initiation in the mysteries of Mithras: new evidence from a cult vessel,” JRS 90 (2000) 169, n.102.
57 Although Schmid (supra n.51) 286 sees the sword as ritually deposited, others prefer a more
workaday explanation: E. Deschler-Erb, “Vindonissa: Ein Gladius mit reliefverzierter Scheide
und Gürtelteilen aus dem Legionslager,” Jb. Pro Vindonissa 1996-97, 13-31, sees the burial as a
personal hiding place within the barracks.
58 Pace Beck (supra n.56) 105: “no mithraeum, no mysteries of Mithras”. Cf. N. M. Swerdlow, “On
the cosmical mysteries of Mithras,” CPh 86 (1991) 48-63.
59 E.g., Ulansey (supra n.9); Beck (supra n.56); Mastrocinque (supra n.9).
60 R. Beck, “‘Qui mortalitatis causa convenerunt’: the meeting of the Virunum mithraists on June
26, A.D. 184,” Phoenix 52 (1998) 335-44.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 295

a possible correlation with the summer solstice61 (ongoing re-analysis of the evidence,
however, makes this connection less probable). While the date of inauguration of our
mithraeum might be stretched to the autumnal equinox (around September 23), it seems
more probable that the rite occurred slightly earlier.62 That is, the mithraeum’s foundation
rite was not correlated with an astronomical event or freighted with astrological signifi-
cance. The Apulum community did not commemorate the stars; they commemorated a
banquet shared with the god and an event held for their group.

Fig. 8. Mithraea with evidence for boxed or potted deposits in the floor of the nave (del. M. M. McCarty on
basemap by esri, USGS, NOAA).

Founding mithraea through time and space

The deposit at Apulum has close parallels in other mithraea across the empire. Acknowl-
edging the wide spread of a particular foundation ritual provides a more complete
understanding of how religious knowledge (savoir faire) circulated within the cult. A range
of special deposits commemorating particular events have been recognized in mithraea
from Tienen to Dura-Europos (fig. 8).63 First catalogued by M. Martens, such deposits hint

61 M. Martens, “The mithraeum at Tienen (Belgium): small finds and what they can tell us,” in
Martens and de Boe (supra n.22) 43-44.
62 Similar timing for rites is suggested in the mithraeum at Aquincum: I. Vörös, “A Tribunus Lati-
claviusok háza az Aquincumi 2-3. sz.-i legio-táborban. A mithraeum állatcsont-leletei,” Buda-
pest Régiségei 28 (1991) 127-32.
63 Martens (supra n.61); A. Hensen, “Templa et spelaea Mithrae. Unity and diversity in the
topography, architecture and design of sanctuaries in the cult of Mithras,” in J. F. Quack and
C. Witschel (edd.), Entangled worlds: religious confluences between East and West in the Roman
Empire (Tübingen 2017) 384-412.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 296 M. M. McCarty, M. Egri and A. Rustoiu

Table 2
Mithraea with evidence for boxed deposits in the floor of nave
Mithraeum Date (A.D.) Province Context Construction Contents
Carrawburgh c.200 Britannia Military Stone slabs Ash, charcoal, chicken
Phase I bones
Carrawburgh c.240 Britannia Military Stone slabs Ash, charcoal, pinecone
Phase IIA
Carrawburgh c.300 Britannia Military Stone slabs Beaker, ash, charcoal,
Phase III domestic fowl bones
Dura-Europos c.169 Syria Military Tile slabs Ash, charcoal, chicken
bones
Krefeld mid 2nd c. Germania Periurban Stone slabs Unknown
Inferior
Tienen c.200 Germania Periurban Tile slabs Ash
Inferior
Orbe-Bosceaz c.200 Gallia Villa Wooden box Burned & unburned
chicken/piglet bones,
fragments of 2 cups
Mariana c.150 Sardinia & Periurban Brick Ash, unknown
Corsica
London c.200 Britannia Urban Wooden box Organic matter
Apulum c.175 Dacia Urban Tile slabs Ash, charcoal, chicken
and piglet bones
Bornheim- c.175 Germania Villa Tile slabs Charcoal, applique of
Sechtem I Inferior snake
Bornheim- c.400 Germania Villa Stone box Charcoal, applique
Sechtem III Inferior of Cautes, coin of
Valentinian, cup, boar
tusk, slag
Mainz 2nd c. Germania Urban Tile slabs Cookpot, unknown
Superior
Martigny c.180 Alpes Periurban Pot Pot, burned and
Poeninae unburned chicken,
piglet, and ovicaprine
bones
Mundelsheim Late 2nd/ Germania Periurban Pot Pot, chicken and piglet
early 3rd c. Superior bones; one of several
buried deposits
at a shared depositional behaviour.64 Yet while Martens wished to see all structured depos-
its in mithraea as participating in a common spectrum of ritually dealing with rubbish (a
feature shared by other, non-Mithraic cults), a subset of these deposits shares an even closer
relationship.65 All consist of under- or at-floor-level boxes, c.1 Roman foot square, created
using either tiles or stone slabs, situated in the center of the nave and containing similar
types of materials (Table 2; fig. 9). In most cases these boxes and their contents were noted

64 M. Martens, “Re-thinking sacred ‘rubbish’: the ritual deposits of the temple of Mithras at Tienen
(Belgium),” JRA 17 (2004) 333-53.
65 For examples in non-Mithraic cult, cf. A. Schäfer and M. Witteyer (edd.), Rituelle Deponierungen
in Heiligtümern der hellenistisch-römischen Welt (Mainz 2013).

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 297

Fig. 9. Plans of
mithraea with
boxed deposits
in the floor of the
nave.
A: Carrawburgh,
Phase I.
B: Carrawburgh,
Phase II.
C: Carrawburgh,
Phase III (all
after Richmond
and Gillam [infra
n.75] figs. 1-3).
D: Krefeld (after
Reichmann [infra
n.67] fig. 1.
E: Tienen (after
Martens [infra
n.68] fig. 154).
F: Orbe-Boscéaz
(after Martin-
Pruvot [infra
n.69] fig. 4).
G: Bornheim-
Sechtem (after
Ulbert [infra n.80]
fig. 1).
H: Mariana (after
P. Chapon,
unpublished).
I: London,
Phase IIa (after
Shepherd [infra
n.72] fig. 97)
(del. M. M.
McCarty).
in passing or suffered later interventions. At Dura-Europos, a tile box was set into the floor,
just in front of the main altar, of a mithraeum founded around A.D. 169; the box contained
ash and “small” (probably bird) bones.66 At Krefeld, a small box built of stone slabs was
set into the floor of the mithraeum close to one end of the mid-2nd c. timber structure; its
contents were not analysed, but a sketch made at the time suggests that the top of the box
extended slightly above the packed earth of the mithraeum’s floor.67 At Tienen, a tile box
containing ash collected elsewhere sat in the nave of the early 3rd-c. mithraeum; no bone
was reported.68 At Orbe-Boscéaz, a late 2nd- or early 3rd-c. mithraeum had a box (probably

66 H. Pearson, “Architecture and history,” in M. Rostovtzeff, F. E. Brown and C. B. Welles (edd.),

The excavations at Dura-Europos conducted by Yale University and the French Academy of Inscriptions
and Letters. Preliminary report of the seventh and eighth seasons of work (New Haven, CT 1939) 70.
67 C. Reichmann, “Das Mithräum,” in R. Pirling (ed.), Das römisch-fränkische Gräberfeld von Krefeld-
Gellep, 1975-1982 (Stuttgart 1997) 21-24. We are grateful to the author for sharing the drawing.
68 M. Martens, Life and culture in the Roman small town of Tienen (Ph.D. diss., Vrije Universiteit

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 298 M. M. McCarty, M. Egri and A. Rustoiu

of wood) set in the center of its nave; it contained fragments of two cups as well as a mix
of burned and unburned chicken and piglet bones — certainly curated banquet remains.69
The Mainz mithraeum is more problematic as the site was not excavated scientifically but
by treasure-hunters during construction of an office building. Later accounts based on
the memory of those present allowed some degree of reconstruction of the building and
positioning of its finds; these included a tile box in the nave that contained a Flavian-era
cookpot, but the dating of the site is deeply problematic.70 This deposit, however, was
distinct from the later intentional deposition of the Wetterau-ware krater with appliqués
related to the cult that has garnered more attention.71 Most recently, the mithraeum at
Mariana has a brick-built box set under the floor of the nave, but the heavy rains and high
water-table at the time of the rescue excavations meant that the contents were not isolated
and analysed. At London, the Phase IIA mithraeum, built c.A.D. 280, had a box in the center
of the nave of Floor 3 that contained “dark soil”.72 The practice of placing organic remains
in a box under the floor of a mithraeum’s nave was thus a shared one, ours being simply
the most closely analysed deposit of this type.
Minor variations in the rite are also visible. Under the floor of the Martigny mithraeum
(late 2nd c.), a small pot was placed containing ash and burned and unburned bones from
chickens, piglets and ovicaprines.73 The excavator interpreted the contents as remains of a
meal. Although the container differs from the other examples, the function of the offering
— commemorating a meal — and its placement in the nave are consistent. A similar pot
with piglet and chicken bones was set under the floor of the Mundelsheim mithraeum, but
there it was also part of a wider symbolic use of buried deposits in different areas, includ-
ing the two parts of bovine skull buried in the benches.74 The foundation rite was charged
with extra symbolic weight in a manner not paralleled elsewhere.
Two further mithraea have multiple deposits each seemingly linked to rebuilding or
renovation. The Carrawburgh mithraeum, excavated in the 1950s, remains one of the most
completely published.75 The excavators recognized three phases: a first construction of the
late 2nd or early 3rd c., an enlargement a generation later, and a final rebuilding in late 3rd
or early 4th c.76 Within the nave, two boxes contained burned remains: the first, labelled
a “hearth”, contained charcoal and 3 chicken bones; the second, labelled a “bunker”,
contained ash and a carbonized pinecone, interpreted as being burned for its aromatic
qualities.77 Both were ascribed to Phase 1. The notion of under-floor, sealed deposits does

Amsterdam 2012) 262-64.

69 C. Martin-Pruvot, La villa gallo-romaine d’Orbe-Bosceaz: rapport sur les campagnes de fouille 1996-
1997: le Mithraeum, le chäteau d’eau (Lausanne 2000) 32 and 110.
70 I. Huld-Zetsche, Der Mithraskult in Mainz und das Mithräum am Ballplatz (Mainz 2008) 3 and
81-82.
71 For the Mainz vessel, see Beck (supra n.56); I. Huld-Zetsche, “Der Mainzer Krater mit den
sieben Figuren,” in Martens and de Boe (supra n.22) 213-28.
72 J. D. Shepherd, The Temple of Mithras, London (London 1998) 75.
73 F. Wiblé, “Les petits objets du mithraeum de Martigny,” in Martens and de Boe (supra n.22)
138-39; Wiblé (supra n.50) 153-54.
74 Planck (supra n.56) 90.
75 I. A. Richmond and J. P. Gillam, “The temple of Mithras at Carrawburgh,” Arch.Ael. 29 (1951) 1-92;
I. Richmond, “The cult of Mithras and its temple at Carrawburgh,” in R. L. S. Bruce-Mitford
(ed.), Recent archaeological excavations in Britain (London 1956) 65-86.
76 Richmond and Gillam ibid. 62-80.
77 Ibid. 7-8.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 299

not seem to have occurred to the excavators. A photograph suggests a slightly different
sequence: the top of the “bunker” appears to sit at or just below the level of the gravel floor
of Phase 1, while the “hearth” sits higher, beneath the compacted yellow gravel of Phase
2A.78 Given that the two features were both under the Phase 2A floor, it is not surprising
that the excavators associated them with Phase 1; instead, it seems more probable that the
“hearth” was set in place as part of the Phase 2 renovations, and subsequently covered.
With the final reconstruction, a different sort of structured deposit was set in the founda-
tion trench for the footings of three altars: it consisted of another “box” made of upright
slabs, with a Castor-ware beaker inside that was filled with ash, pinecone and domestic
fowl bones.79 In each major rebuilding the deposits seem to have changed location around
the space of the nave, but were relatively consistent in what each structure contained, even
if the Phase 3 deposit was slightly more elaborate. This scenario of three re-foundation
rites, each with its own box set in the nave near the far end, is made more probable by
comparison with the sequence identified at Bornheim-Sechtem. The excavation uncovered
remains of a late 4th-c. mithraeum that seemed to overlie the fragmentary remains of a
2nd-c. one (and possibly an intermediate building). The date and identification of the ear-
lier building is based on the stylistic dating of re-used Mithraic sculpture.80 In Phase I,
a tile box set under the floor contained charcoal and a lead-glazed appliqué of a snake;
in Phase III, a stone box set above and adjacent to the tile box contained a fragment of a
statue, a lead-glazed appliqué of Cautes, a boar tusk, a beaker, slag, and a coin of Valen-
tinian.81 The lead-glazed appliqués belong to the same vessel, as does a third, showing a
lion, that was placed in a structured deposit under the apse of the Phase III building.82 The
excavator suggested that during the Phase III construction the builders uncovered the tile
box and placed the appliqué there before re-sealing the deposit and creating a new, similar
deposit.83 Here, both the original foundation and the re-foundation include similar deposi-
tional rituals. To these re-foundation deposits we might add another feature in the London
mithraeum. An unusual base whose bottom was buried in the nave close to the apse of the
Phase III temple (late 3rd/early 4th c.) included a small recess, inside which a greyware jar
had been set that contained grey soil and chicken bones.84 This, however, might represent
a slightly different rite, more akin to the jar-burials at Mundelsheim.
From these deposits several common points emerge. First, there are two chronologi-
cal clusterings. Almost all of the deposits occur in mithraea founded in the late 2nd or
early 3rd c. (also the period when most mithraea seem to appear). But the second cluster,
in the late 3rd/4th c., seems to suggest a different pattern: at Carrawburgh III, Bornheim-
Sechtem III and London III, the foundation rites that led to the deposits seem to be more
complex and involve a greater variety of materials. Similarly, when the Mainz mithraeum

78 Ibid. pl. 2B.

79 Ibid. 35-6.
80 C. Ulbert, “Das Mithraeum von Bornheim-Sechtem bei Bonn,” in Martens and de Boe (supra
n.22) 81-88.; J.-C. Wulfmeier, “Ton, Steine, Scherben — Skulpturen und Reliefkeramiken aus
dem Mithraeum von Bornheim-Sechtem,” ibid. 89-94; C. Ulbert, J.-C. Wulfmeier and I. Huld-
Zetsche, “Ritual deposits of Mithraic cult-vessels: new evidence from Sechtem and Mainz,” JRA
17 (2004) 354-70.
81 Ulbert ibid. 86.
82 Confirmed by Neutron Activation Analysis: Ulbert et al. (supra n.80) 362-64.
83 Ibid. 362.
84 Shepherd (supra n.72) 84.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 300 M. M. McCarty, M. Egri and A. Rustoiu

was (probably) rebuilt at a much higher level, a new deposit for the custom-made Wetterau
krater with appliqués was created. In other words, an earlier and simpler rite that seems to
have commemorated foundation meals shared by Mithras and his worshippers was elabo-
rated over time to include a wider range of materials and vessels.
The geographic and social distribution of these deposits is also noteworthy. They appear
in almost all regions where mithraea are attested and in ones frequented by different
types of communities, from soldiers (Carrawburgh, Dura) to urban populations (Mainz,
Mariana), to customs administrators (Apulum), and to the residents of villas (Bornheim-
Sechtem, Orbe-Boscéaz). Despite recent efforts to break Mithraism down into local or
regional “strains”, this rite extends from Britain to Syria. (The absence of these deposits
in mainland Italy, despite a large number of excavated mithraea, especially at Rome and
Ostia, is equally striking, although most Italian mithraea had mosaic floors which were not
excavated or lifted.) In only a few cases can the presence of such a box be ruled out entirely
through excavations that went below the floor level and through the absence of later inter-
ventions that would have destroyed such a deposit.85
This set of deposits does not reflect the exact replication of a fixed liturgy. Each shows
some variation and a degree of elaboration, sometimes in different directions, by including
slightly different types of material. Yet each follows a common schema, an adapted ‘recipe’
that demonstrates the same core components: meal remains, set in a container, with that
container set beneath the nave’s floor. This type of structured deposit, set beneath the floor
of the sanctuary in a box (or vessel), is also almost unique to mithraea; it is not simply
part of a widespread Roman ritual habit.86 Despite increasing scholarly interest in “sacred
rubbish” and ritual deposits in Roman sanctuaries, there is only one potential parallel for
underfloor tile- or stone-box deposits.87 Set near the altar of the late 3rd-c. rebuilding of
the “Triangular Temple” at Verulamium, a structure pockmarked by various structured
deposits, was a tile box containing a bull skull.88 Rather than meal remains, however, this
seems to be a more symbolic deposit. The identification of the deity venerated is uncer-
tain, although on the basis of pine seeds R. E. M. Wheeler attributed the temple to Cybele;
pinecones, however, enjoyed a wider ritual use in Roman Britain as a form of incense,
including in Mithraism.89 Still, offerings of meal remains in boxes set below the floor seem
to be unique to mithraea. The significance of a foundation rite in sanctuaries to Mithras
should not be understated. A number of accounts have even challenged the notion of a rela-
tively coherent “Mithraism” as being the figment of the imagination of modern scholars

85 For example, no trace was found at Künzing (K. Schmotz, “Der Mithrastempel von Künzing,”
Vorträge des 18. Niederbayerischen Archäologentages [Leidorf 2000] 111-43), despite complete and
careful excavation. Yet even other recent excavations leave the question open; at Inveresk, as the
‘decommissioning’ of the mithraeum involved heavy interventions in the nave near the altars
(where one would expect to find such a deposit), the original presence of such a deposit cannot
be ruled out: Hunter et al. (supra n.29).
86 There are parallels in much earlier (Bronze Age) periods: Weikart (supra n.50) 33 notes stone
boxes placed under the floors of buildings but not in exactly the same relative locations.
87 Cf. A. Steiner, “Kultgruben im Tempelbezirk der heliopolitanischen Gottheiten in Carnuntum?
Ein Vorbericht,” in M. Meyer and V. Gassner (edd.), Standortbestimmung. Akten des 12. Öster-
reichischen Archäologentages in Wien (Vienna 2010) 297-304.
88 R. E. M. Wheeler and T. V. Wheeler, Verulamium: a Belgic and two Roman cities (Oxford 1936)
117-19.
89 L. A. Lodwick, “Evergreen plants in Roman Britain and beyond: movement, meaning and
materiality,” Britannia 48 (2017) 135-73.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 301

trying to build a synthetic picture from fragmentary testimonia.90 Yet while commonali-
ties in the general pottery and faunal assemblages in recently excavated mithraea point
to similar genres of activities taking place, this is the first clear attestation of a common,
complex, ritualized act. Even the rite most frequently discussed in Mithraic studies, ini-
tiation, has largely been assumed to be shared across Mithraic communities.91 Now, for
the first time, we see clear and concrete evidence not of diverse local cults of Mithras, but
of a relatively stable and shared ritual system that stretched across the empire. Still, not
every Mithraic community shared these rites. Instead of either “Mithraism” as a coherent
religious system, or autonomous Mithraic “cells” (i.e., largely independent re-creations
of Mithras-worship that engaged little with one another), we find evidence for networks
of shared practice. While diffusionist models of the cult tend to privilege geographically-
distinct strains of Mithras-worship, where physical proximity is a key determinant, the
broad spread (and the equally broad absence) of common inaugural rites suggest that the
cult must be conceptualized in terms of different forms of connectivity.
The deposits also raise another important question: how did a Mithraist learn the proper
way to inaugurate a mithraeum? Unique to mithraea and widespread, but not present in
all Mithraic sanctuaries, these foundation deposits offer a rare chance to understand how
religious knowledge (savoir faire) may have circulated within the cult. When combined
with anthropological work on the transmission of ritual knowledge, such deposits shed
important light on the cult of Mithras.

Anthropological moves from practice to savoir faire

The maintenance and transmission of any cultural practice through time and space
— whether tool-making or rituals for dealing with capricious deities — requires struc-
tures that enable those practices to be shared and learned. The nature of transmitting
complex cult acts has long been the focus of anthropologists of religion, especially within
socio-cultural contexts where such knowledge is not predicated upon sacred texts. These
anthropological studies point to 5 possible means by which a rite could be performed and
re-performed, transmitted through time and space:
1) most basically, by becoming familiar through repetition, as habit;92
2) by being semantically recomposed of stock elements that are themselves familiar from
repetition;93

90 E.g., P. Adrych et al., Images of Mithra (Oxford 2017); R. Gordon, “Projects, performance and
charisma: managing small religious groups in the Roman empire,” in G. Petridou, R. Gordon
and J. Rüpke (edd.), Beyond priesthood: religious entrepreneurs and innovators in the Roman Empire
(Berlin 2017) 275-314.
91 R.Turcan, “Hiérarchie sacerdotal et astrologie dans les mystères de Mithra,” in R. Gyselen
(ed.), La science des cieux: sages, mages et astrologues (Leuven 1999) 249-61, problematizes this
assumption.
92 On such approaches in the anthropology of religion, and an attempt to ground them in cognitive
science, see H. Whitehouse, “Implicit and explicit knowledge in the domain of ritual,” in
I. Pyysiäinen and V. Anttoen (edd.), Current approaches in the cognitive science of religion (London
2002) 133-52.
93 In The archetypal actions of ritual: a theory of ritual illustrated by the Jain rite of worship (Oxford),
C. Humphrey and J. Laidlaw argue from ethnographic observation of Jain rites that such discrete
building blocks form the basic units of all ritualized action, and are learned independently of
any meanings freighted upon a whole sequence.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 302 M. M. McCarty, M. Egri and A. Rustoiu

3) by being recorded in external media (like prescriptive texts);

4) by being tightly coordinated with other symbols in the cult so as to make the rite seem
like an ‘obvious’ extension and manifestation of those symbols and able to be recalled
through connections;94
5) by the formal training of a priestly ritual expert.95
Testing such models against evidence for the cult of Mithras suggests that the rites
resulting in foundation deposits set in boxes imply the fifth option, and an important rôle
for trained priests in the cult. Further, such priests were probably travelling experts, which
would offer the strongest evidence for the oft-assumed Mithraic “colporteur”.
In general terms, most cult craft in antiquity was transmitted through a combination
of repetition and reconstitution from stock elements. The lack of formal religious or rit-
ual education in the Roman world has long been noted.96 Most types of ritual act were
learned by participation or by watching; this process of religious socialization began in
childhood.97 There were, of course, exceptions, cases where more formal instruction may
have trained ritual experts. For example, Statius (Silv. 5.3) describes his father actively
teaching ritual experts. Similarly, those expected to execute particular rituals according
to the strict formal guidelines of past performance might undergo periods of ‘apprentice-
ship’; this is how Dionysius of Halicarnassus (Ant. Rom. 2.67) interprets the 30-year term
of a Vestal Virgin, the first decade being devoted to watching, the second to doing, and the
final one to teaching. Still, as a general rule, the savoir faire of Roman religion was learned
through experience and repetition.
In such a participatory learning system, the frequency with which rites, or their constit-
uent steps, are experienced is the prime determinant of how well the practice is learned.98
The more a process is experienced, the better it is remembered: a kind of mental ‘script’
is created and reinforced with each subsequent performance. High frequency of perfor-
mance (or witness) of such rites allows their replication in stable forms, although minor
innovations are possible, driven by discrete circumstances or by the personal flourishes of
a given officiant.99 Public and domestic sacrifices were periodic and frequent, and often
built from similar ‘building blocks’, formal steps that could be recombined for slightly

94 Fredrik Barth’s work on Baktaman initiation rites (Ritual and knowledge among the Baktaman
of New Guinea [New Haven, CT 1975] and Cosmologies in the making: a generative approach to
cultural variation in inner New Guinea [Cambridge 1987]), which are particularly comparable to
Mithraic rites, points to the ways complex ritual scripts could be re-created from participants’
experience, their memory, and a handful of external aides-mémoire.
95 V. Turner, “Religious specialists,” in D. Sills (ed.), International encyclopedia of the social sciences
(New York 1972) 437-44, in attempting to create a taxonomy of religious specialists, reserves the
title “priest” for such a technician.
96 Beard et al. (supra n.41) 27; Rüpke (supra n.3) 10-11.
97 Prescendi (supra n.18).
98 H. Whitehouse, “Memorable religions: transmission, codification and change in divergent Mela-
nesian contexts,” Man n.s. 27 (1992) 777-97; R. N. McCauley and E. T. Lawson, Bringing ritual
to mind: psychological foundations of cultural forms (Cambridge 2002) 48-50; R. F. Ellen, Nuaulu
religious practices: the frequency and reproduction of rituals in Moluccan society (Leiden 2012).
99 R. Firth, Rank and religion in Tikopia: a study of Polynesian paganism and conversion to Christianity
(London 1970) 233-60, distinguishes between “circumstantial”, “systemic”, and “personal”
variation in the performance of rites, demonstrating that even in simple rites there is wide
divergence between individual performances.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 303

different contexts.100 More individual engagement with the gods — for example, the pro-
cess of entreaty, vow and sacrifice/dedication — was occasional rather than calendrical,
but frequent enough that formulas could be learned and reproduced in different contexts,
as the host of “votum solvit” inscriptions attests. Occasional and infrequent rites could be
re-composed of stock elements that were both frequent and familiar, or they could rely
on external aides-mémoire. The performance of the Saecular Games offers a good example
of this latter process; Augustus’ version ‘invented’ these rites by combining well-known
forms of procession, sacrifice and spectacle, but then provided an authoritative epigraphic
account that helped to shape future iterations of these rites.101
At first glance, this model of frequent participation seems incompatible with Mithraic
cult, where one of the primary rites, the initiatory “mysteries”, were seemingly ‘one-offs’
for individual initiates. Even if we assume that an individual underwent multiple initia-
tions over his life (one for each grade), ancient authors suggest that each of these initiations
involved different types of practices and symbols.102 Indeed, recent accounts of Mithraic
cult, drawing on the cognitive science of religion, argue that the cult operated differently
and depended upon a distinct system of episodic memory: the heightened emotion of ini-
tiatory “rites of terror” seared them into an initiate’s memory as if by flash photography,
allowing the initiate to recall them accurately and reproduce them.103
Yet recourse to such an alternative explanation may be unnecessary, especially given
the issues that empirical psychological research have raised with the theories of ‘flash-
bulb memory’ espoused by such accounts.104 Barth was at first puzzled by the degree to
which Baktaman initiation rites were replicated in very stable forms over his years of field-
work, given the lack of textual script for their dénouement and communities’ reliance on a
few experts whose memory and mortality repeatedly jeopardized faithful reproduction.105
While Barth argues that the re-creation of a ritual order depended in part on external aides-
mémoire, the other key practice that allowed the maintenance of communal ritual expertise
was the way that recent initiates were involved in subsequent initiations. Even if each
person underwent his own initiation only once, he would be a witness or a participant

100 For this principle in ritual generally, see Ellen (supra n.98). For the formal recomposition of
ritual practices, see R. N. Lawson and E. T. McCauley, Rethinking religion (Cambridge 1990)
56-59. In Roman religion, the phenomenon has rarely been studied explicitly, thanks to the
ingrained division between “public” and “private” religion. Robinson (supra n.36) 98 notes
similar practices in public/domestic rites at Pompeii. J. Scheid, “Sacrifices for gods and
ancestors,” in J. Rüpke (ed.), Blackwell companion to Roman religion (Oxford 2007) 263-71, suggests
the commensurability of these two realms.
101 B. Schnegg-Köhler, Die augusteischen Säkularspiele (Munich 2002) 156-85.
102 Porph., De Ant. 15; Ambrosiaster, Quaest. 113.11. Discussion of rites: Clauss 2000 (supra n.19)
102-5.
103 R. Beck, “Educating a Mithraist,” in B. Bøgh (ed.), Conversion and initiation in antiquity (Bern
2014) 247-54; L. H. Martin, The mind of Mithraists: historical and cognitive studies in the Roman cult
of Mithras (New York 2015) 34-36.
104 W Hirst and E. A. Phelps, “Flashbulb memories,” Current Directions in Psychological Science 25
(2016) 36-41, rejecting the distinctiveness of “flashbulb memory”. Cf. W. Hirst et al., “A ten-year
follow-up of a study of memory for the attack of September 11, 2001: flashbulb memories and
memories for flashbulb events,” J. Experimental Psychology: General 144 (2015) 604-23: examining
recollections of the September 11 attacks, researchers found high levels of inconsistency
persisted in memories; inaccuracies and inconsistencies were most likely to be ‘corrected’ (that
is, standardized and shared) when reinforced by repetition in other media or conversation.
105 Barth 1987 (supra n.94).

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 304 M. M. McCarty, M. Egri and A. Rustoiu

in the initiations of others: that is, there was a high degree of participatory frequency in
the rite. Of course, simply positing possibilities from parallels is not enough; fortunately,
there is more direct evidence for this more frequent participation in Mithraism. Painted
scenes along the benches of the Capua mithraeum have long been interpreted as show-
ing initiatory rites, many involving a nude initiate undergoing various ordeals.106 In each
scene, the initiate is accompanied by a larger figure who seems to be directing the rites (an
expert); yet a third figure, clothed in a short tunic and of medium size, also stands beside
the initiate. Frequently identified as a “mystagogue”, the figure offers evidence for the
involvement of participants beyond the initiate and the expert.107 Initiation rites in Mithra-
ism were probably learned and replicated much like other rituals in the Roman world: by
experience and repetition.
Even foundation rites within houses and other kinds of sanctuary, which were single
events within the life of a given structure, may nevertheless have been learned and repro-
duced through repetition. Many of these are distinctly regional (for example, confined to
Britain, or even to micro-regions therein, or particular civitates). In such localized settings,
and where similar founding rites were applied to multiple structures with relative fre-
quency, it is easy to see how particular ‘ways of doing’ might move and circulate among
family members, neighbours, and social groups in regular contact, akin to other cultural
practices. In these cases, such foundation rites might fall into the occasional but repeated
pattern of other rites: over one individual’s lifetime, he or she might observe the building
and inauguration rite of multiple homes or hearths within the community.
Yet foundation rites of the kind attested by the below-nave deposits in mithraea are
clearly different in frequency than these other genres of ritual practice. They are founda-
tional, happening once in the life of a mithraeum structure. When these rites do happen
multiple times in the same location, with rebuildings and refoundations (Carrawburgh,
Bornheim-Sechtem, perhaps London), such events are separated by a generation or more,
so that it is unlikely that the same individual would have been present to participate in
multiple foundation rituals there. Mithraic foundation rites are also distinct from other
types of foundation rites in the Roman world. Given the spatial spread of these deposits,
and even accounting for mobility on the part of many Mithras-worshippers, it is similarly
unlikely that a regular worshipper would be present at more than one (possibly two) foun-
dations of different mithraea, which was not enough to learn the pattern.108 Thus, Mithraic
foundation rites were not learned by frequent participation: a different model must be
sought.
In one sense, the communal meal with Mithras that formed the first archaeologically-
visible step in inaugurating a mithraeum was a building block for the more complex
foundation rite.109 Still, this meal offers no apparent guidance for collecting and burying
remains in the nave. Nor does this second step have an obvious parallel in other sorts of

106 M. J. Vermaseren, The mithraeum at S. Maria Capua Vetere (Leiden 1971).

107 R. Gordon, “The Mithraic body: the example of the Capua Mithraeum,” in G. Casadio and
P. Johnston (edd.), Mystic cults in Magna Graecia (Austin, TX 2009) 290-313.
108 For the mobility of Mithraists, see S. Price, “Religious mobility in the Roman empire,” JRS 102
(2012) 7-11.
109 On Mithraic dining, see Lentacker (supra n.46). For learning by repetition in these cult meals,
see L. Martin, The minds of Mithraists: historical and cognitive studies in the Roman cult of Mithras
(London 2015) 47–51.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 305

foundation rites, especially not in a way that would be shared across the vast area in which
the rite occurs. It is unlikely that recomposition from common, ritual building blocks pro-
vides an explanation for the similarities in the rite. Similarly improbable seem to be portable
media that externalize instructions for the rite and that could be transferred from place to
place. Although such ritual handbooks are attested for certain experts — those classed
as magicians, whose circulating recipes survive in the PGM —, this seems less likely in
Mithraism. There is no surviving evidence for such textual supports, nor are there indices
that might indirectly suggest wide use of such texts. The two papyrus texts sometimes
linked with the cult of Mithras have nothing to do with Mithraism.110 Although R. Beck has
argued that the fragments of verses and hymns attested in ancient authors and the dipinti
of the Santa Prisca mithraeum are evidence for ritual texts in Mithraism, these are probably
specific to single communities. The hymn at Santa Prisca especially seems to be the work of
an amateur poet whose meter is slightly off, rather than a polished and widely circulated
‘hymnal’. Such fragments do imply the composition and use of texts by Mithraic groups,
but in a localized and internal way.111 Next, there is scant evidence for the circulation of
portable objects bearing Mithraic iconography, and few objects specific to mithraea that
were not manufactured locally, whether bull-slaying reliefs carved from locally available
materials or cult vessels commissioned from the nearest centers of pottery production.112

110 The evidence for textual transmission in Mithraism is slim. The so-called Mithras Liturgy, with
its regionally-specific details in the ritual prescriptions, must stem from a syncretistic community
in Egypt who probably simply borrows Mithras as a sun deity. In Eine Mithrasliturgie (Leipzig
1910), A. Dieterich claimed the papyrus was part of Mithras-cult; he was followed by H. D.
Betz, The “Mithras Liturgy.” Text, translation, and commentary (Tübingen 2003). F. Cumont, “Un
livre nouveau sur la liturgie païenne,” Revue de l’Instruction publique en Belgique 47 (1904) 1-10,
rejected this association, a view argued recently by R. Gordon, “Probably not Mithras,” CR
55 (2006) 99-100. The “Mithraic Catechism” (see W. Brashear, A Mithraic catechism from Egypt
(P. Berol. 21196) [Vienna 1992]) has been more widely accepted (e.g., Beck [supra n.9] 63) as
genuinely Mithraic, but that attribution stems primarily from the appearance of “lions”, which
also appear throughout Egypt in “Gnostic” texts.
111 Beck (supra n.103) argues for the teaching of Mithraic “principles” using esoteric texts, based on
dipinti at S. Prisca and on Firm. Mat., Err. 5.2, where Maternus cites a line of Greek hexameter
attributed to Mithras-worshippers’ “prophet” (sicut propheta eius tradidit nobis dicens). Still, the
use of Greek hexameter as a widespread feature in a predominantly Latinate cult would be
surprising.
112 R. Gordon, “Small and miniature reproductions of the Mithras icon,” in Martens and de Boe
(supra n.22) 259-84. Small-scale Danubian plaques, given their unique style, are the easiest
objects to observe moving, and they spread quite widely: Heddernheim I and III (CIMRM
1084, 1128), Siscia (CIMRM 1475), Cologne (CIMRM 1019; less certainly Danubian), Friedberg
(CIMRM 1054), Rome (see E. Lissi-Caronna, Il mitreo dei Castra Peregrinorum [Rome 1986] pl.
XVI), Lauriacum (CIMRM 1422), Eisenstadt (CIMRM 1648), Vindobona (CIMRM 1650), Alcsút
(CIMRM 1740); Linz (CIMRM 1415); and Caesarea Maritima (see R. J. Bull et al., The Mithraeum
at Caesarea Maritima [Boston, MA 2017]). The use of such small, movable plaques may represent
a particular mode of worship that S. Hijmans (“Material matters: object, authorship, and
audience in the arts of Rome’s empire,” in S. Alcock, M. Egri and J. F. D. Frakes [edd.], Beyond
boundaries: connecting visual cultures in the provinces of ancient Rome [Los Angeles, CA 2015]
84-103) connects to the cult of the Danubian riders. For imported reliefs in Dacia, see Sicoe
(supra n.26) 90-93 (at communities such as Potaissa, the small imported relief seems to have no
bearing on shaping the iconography or composition of the large-scale relief in the mithraeum).
On the Mainz vessel, see Huld-Zetsche (supra n.70). Local production of the custom-made
vessel at Bornheim-Sechtem: Ulbert et al. (supra n.80). Local production of the Pons Aeni bull-
stabbing vase: J. Garbsch, “Das Mithraeum von Pons Aeni,” BayVgBl 50 (1985) 402. The central

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 306 M. M. McCarty, M. Egri and A. Rustoiu

This might suggest that cult paraphernalia in general, including texts, did not regularly
circulate. In addition, minor variations in the attested foundation deposits — the use of
pots rather than boxes, differences in how far towards the apse the deposit is set in the
nave — hint at a degree of improvisation rather than the rigidity of a prescriptive, textual
recipe for the rite. Taken together, these indices for the lack of shared text and of objects
that circulated among Mithraic communities suggests that ritual handbooks are unlikely.
Finally, while the circulation and stability of many features of Mithraic visual and material
culture can be attributed to the relatively tight (and redundant) symbol-system of the cult,
foundation rites are not among these. R. Beck and L. Dirven have both demonstrated how
aspects of Mithraic imagery and practice serve mutually to reinforce one another as basic
conceptual axioms, thereby promoting stability of these core propositions.113 Yet the foun-
dation deposits had very few links to a wider Mithraic myth or symbol-system: their form
is not stored externally in other cultural products such that one could deduce its form and
reconstruct it from images, myth, or even other rituals. This type of foundation deposit is
not implicated in the narrative of Mithraic myth or in the dense web of symbols that might
be memorable thanks to the ways they mutually reinforce one another.
With these four alternatives judged unlikely, a remaining option to explain how
common foundation rites came to be adopted in a range of mithraea becomes the most
probable: that the deposits represent the presence of trained ritual experts at the moment
of a mithraeum’s foundation. The nature of that training might have been abstract —
teaching participants what they would do if they were ever to found a mithraeum. It is
nearly certain that most worshippers of Mithras did receive some sort of education in
things Mithraic as they ascended through various initiatory grades; most accounts argue
that that these teachings involved deeper understandings of esoteric postulates, savoir
penser.114 Further evidence that education in initiation did not routinely involve formal
ritual training comes from the way the priestly title sacerdos is distinguished from grades of
initiation.115 Occasional epigraphic pairings like pater et sacerdos presume that these posi-
tions could be separable; notably, though, no other grade besides pater is ever adjoined to
sacerdos in a title, suggesting that the ritual expertise of a sacerdos was reserved for those at
the highest levels of the cult.116 It is improbable that most Mithras-worshippers learned, in

Gaulish terra sigillata cup dedicated at Angers, commissioned from Lézoux by a worshipper
from Ambiens, represents a unique case of a moving object and a moving worshipper: M. Molin,
J. Brodeur and M. Mortreau, “Les inscriptions du mithraeum d’Angers/Iuliomagus (Maine-et-
Loire): nouvelles données sur le culte de Mithra,” Gallia 72 (2015) 424-26.
113 Beck (supra n.9) 5-6 (repeated, axiomatic “sacred postulates”); L. Dirven, “The mithraeum as
tableau vivant: a preliminary study of ritual performance and emotional involvement in ancient
mystery cults,” Religion in the Roman Empire 1 (2015) 20-50.
114 Although M. Clauss (“Die sieben Grade des Mithras-Kultes,” ZPE 82 [1990] 183-94; id. 1992
[supra n.19]) argues that the grades are actually priestly positions into which only a small
number of worshippers sought initiation, R. Gordon (“Ritual and hierarchy in the mysteries of
Mithras,” ARYS 4 [2001] 245-73) demonstrates the unlikelihood of this idea, based primarily on
the evidence for frequent initiations in the Dura-Europos graffiti. Education: Beck (supra n.103).
115 Clauss 1992 (supra n.19) 43, argues that terms like pater, antistes and sacerdos are synonyms.
F. Mitthof, “Der Vorstand der Kultgemeinden des Mithras,” Klio 74 (1992) 275-90, argues for
problems with such a view.
116 Mitthof ibid. argues that sacerdotes represent an even higher step in the cult hierarchy; while this
may be unlikely (and titles rarely occur beyond Italy: Turcan [supra n.91]), the fact that terms
related to ritual expertise and the prestige it brings are included within the cult suggest the
importance of craft knowledge. Cf. Gordon (supra n.114) 248-49.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 307

an abstract way, the rites involved in founding a mithraeum; this sort of technical knowl-
edge was reserved for a very few.
That leaves a small group of experts trained with the skills to found a mithraeum. And
with limited numbers capable of founding (or teaching others to found) mithraea, it seems
most probable that the rapid spread of this rite across a large geographic expanse is attrib-
utable to mobile ritual experts. Such experts would imply that more went into founding
a mithraeum than the ‘contact-contagion’ model of regular Mithraists moving with what-
ever level of knowledge they already had.
There is substantial evidence for mobile ritual experts in the Roman world (mostly
freelance technicians like diviners, astrologers and ‘magicians’, but also more institution-
alized figures). Within the cult of Mithras, there is some epigraphic evidence that could
be read to support the presence of such figures.117 There are examples not only of mobile
worshippers who ‘joined’ different Mithraic communities or paid for new temples, but of
ritual expertise being sought from outside a community. The “album” of members of a
Mithraic community at Virunum seems to record, in different hands, the changing annual
composition of the community between A.D. 183 and 201.118 When 5 members, including
two named as patres, died within a year (183-184), a new member, Trebius Alfius, labeled
as a pater, joined the community. Through careful analysis of the palaeography and layout
of the inscription, G. Piccottini argued that Alfius must have been a pater from elsewhere,
brought in by the Virunum community to fill their leadership lacuna.119 Even if onomas-
tics suggest that this Mithraic pater was not coming from too far away (he was probably
based within Noricum, given that his name suggests he was a freedman member of the
familia of Marcus Trebius Alfius, lessee of the rights to iron mining in the province), he
was still potentially an outside expert brought in.120 Similarly, in the “Mithraeum Fagan”
at Ostia, the pater Valerius Heracles made a dedication with two sacerdotes, but only the
pater appears in other texts from the mithraeum, which led F. Mitthof to suggest that the
sacerdotes were ‘guest’ dedicators here, brought in for a specific cult act.121 Leaving aside
the problems with differentiating the titles sacerdos and pater within Mithraic communities
(the terms are sometimes interchangeable, and much of the evidence seems to represent a
particular set of status concerns in central Italy), it seems that on occasion Mithraic com-
munities might seek expert knowledge (ritual or esoteric) from beyond their immediate
members,122 even if in both of these cases the experts may have been sought from close by.
The specialized savoir faire implied by foundation deposits across a vast geographic area,
however, hints that such experts might also travel considerably farther.
The lack of epigraphic attestations of such traveling figures has, of course, led to their
dismissal from accounts of the spread of Mithraism: epigraphy, however, cannot offer a

117 For wandering experts in general, see H. Wendt, At the temple gates: the religion of freelance
experts in the Roman empire (Oxford 2016); cf. also Gordon (supra n.90). While drawing parallels
between Mithraism and Christianity is always a fraught endeavor, mobile experts circulating
among Christ-focused communities are best attested there, from the 1st c. A.D. on: G. Thiessen,
Sociology of Early Palestinian Christianity (Philadelphia, PA 1978), saw Christianity as a cult of
wandering charismatics (sensu Weber).
118 G. Piccottini, Mithrastempel in Virunum (Klagenfurt 1994).
119 Ibid. 34-35.
120 CIL III 4788.
121 CIMRM 313; Mitthof (supra n.115) 287.
122 Note that Valerius Heracles elsewhere in the same Mithraeum (CIMRM 311) is called a sacerdos.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 308 M. M. McCarty, M. Egri and A. Rustoiu

total picture of the internal dynamics of a cult. Given that euergetism and its correlating
epigraphic behavior were tied to particular forms of social prestige, it may not be sur-
prising that anyone whose position depended on technical skill, rather than on making
expenditures, would be nearly invisible in inscriptions. While epigraphically such figures
may be hard to detect, archaeology of these foundation deposits may provide the stron-
gest evidence for the possibility of such figures in the cult. Their relatively shared training
precludes the notion of them being wholly autonomous peddlers of cult, able to elaborate
freely on the rites and to move in new and different directions.123 Of course, further work
is needed on the extent to which ritual-craft expertise can be found in the archaeology of
mithraea. But the epigraphic dossier at Apulum’s Mithraeum III happens to permit a more
detailed picture of the micro-dynamics between ritual expertise and mithraeum founda-
tion, and points to the need to distinguish between those who paid for a Mithraic spelaeum
and the ritual experts who inaugurated it.

Who was the ritual expert? Epigraphy and the community at Apulum
The three inscribed monuments (and presumably the fourth damaged base) from the
mithraeum were all dedicated to Mithraic deities — Sol Invictus Mithras, Cautopates, and
Transitus Dei — by the same man, Vitalis, who identifies himself as an arcarius (treasurer).
His single name implies that he was a slave. Although no surviving inscription documents
who paid for the mithraeum structure itself, given that Vitalis provided the most impor-
tant furnishings and monuments (main altar, depictions of Cautes and Cautopates), we
may surmise that he paid for the structure itself: in that sense, he will have been a core
member, even the founder, of the mithraeum. Yet analysis of his position and comparable
monuments suggest that he was not the ritual expert responsible for creating the founda-
tion deposit.
The inscriptions offer a probable scenario for the social context in which Vitalis learned
about the cult: as an administrator within the publicum portorium Illyrici, the imperial
customs zone that encircled the central provinces (the Pannoniae and Moesiae, Dalmatia,
Dacia).124 Although arcarii are attested in many milieux — from civic treasurers, to trea-
surers for institutions (like public baths), to those of private individuals —, three things
suggest that Vitalis was involved in the customs system.125 First, an office of the portorium
Illyrici at Apulum is attested by other inscriptions, which opens the possibility that a slave
arcarius at Apulum might work there.126 Second, portorium slaves (including arcarii) are
prolific dedicants of monuments in provinces where that institution was active, implying
that the customs system offered a venue where slave administrators could amass sufficient
wealth to make monumental dedications127 (arcarii in other institutions tend to enter the
epigraphic record through epitaphs and seem not to have been in a socio-economic position

123 Pace Gordon (supra n.90).

124 For more detailed arguments about Vitalis and his connection to the portorium and its particular
forms of Mithras-worship, see M. M. McCarty, M. Egri and A. Rustoiu, “Connected communities
in Roman Mithraism: a view from Apulum Mithraeum III,” Phoenix 71 (2017 [2018]) 370-92.
125 Arcarii in general: M. Silvestrini, “Gli arcarii delle città,” MEFRA 115 (2005) 541-54.
126 I. Piso and V. Moga, “Un bureau du Publicum Portorium Illyrici a Apulum,” Acta Mus. Napocensis
35 (1998) 105-8.
127 S. J. de Laet, Portorium (Brugge 1949); P. Kritzinger, “Contrascriptores im römischen Zollwesen,”
in B. Takmer, E. N. Akdoğu Arca and N. Gökalp Özdil (edd.), Vir doctus anatolicus: studies in
memory of Sencer Şahin (Istanbul 2016) 567-89.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 309

to make dedications).
Third, the fact that
Vitalis makes a dedi-
cation to Transitus
Dei suggests that he
was both involved in
the customs system
and acquired some of
his savoir faire within
Mithraic communities
from that institution.
Dedications to Tran-
situs are extremely
rare, coming from a
small group of central
European sites (fig.
10); their close connec-
tion is indicated by the Fig. 10. Sites with dedications to Transitus Dei (del. M. M. McCarty, on
basemap by esri, USGS, NOAA).
way these dedications
share an unusual form in Latin, Transitu rather than Transitui.128 The most notable dedica-
tion to Transitus, on the base of a statue of Mithras carrying the bull (it is unclear whether
the statue is of Transitus or simply to Transitus), was dedicated at Poetovio in Mithraeum I,
one whose community was comprised almost exclusively of portorium personnel.129 Indeed,
Poetovio seems to have served as a kind of headquarters for the portorium, with admin-
istrators coming there and being sent out to customs-houses on the edges of this group
of provinces. All other surviving dedications to Transitus come from sites that either cer-
tainly (because of epigraphic attestations) had portorium offices or that probably had such
based on their location at transit-points on the edge of a customs zone: Savaria, Brigetio,
Malvesia, Carnuntum and Gorsium.130 In other words, there seems to be a close connection
between dedications to Transitus and the customs system.131 That Vitalis makes a dedica-

128 We are grateful to the anonymous reviewer for this suggestion. The homogeneity of this form
stands in contrast to the variety of ways Mithras is invoked: at Apulum, as Mitrhae; at Poetovio,
as Mithrae.
129 E. Will, “Les fidèles de Mithra à Poetovio,” in Adriatica praehistorica et antiqua (Zagreb 1970)
633-38.
130 Savaria: P. Kiss, “Mithras-Altäre aus Savaria,” in I. Lazar (ed.), Religion in public and private
sphere. Acta 4th int. colloquium ‘The autonomous towns of Noricum and Pannonia’ (Koper 2011) 183-
91; see de Laet (supra n.127) 222, for the customs post there.
Brigetio: CIMRM 1737; de Laet ibid. 194, for customs.
Malvesia: CIMRM 1900; the town was an important node (the seat of a beneficiarius consularis)
and at the crossing over the Drina between Moesia and Dalmatia.
Carnuntum: CIMRM 1722; as a major entry point into Pannonia, and with the next-closest
customs station along this stretch of the provincial border found at Brigetio, it seems unlikely
that the portorium was not collected.
Gorsium: CIMRM 1811; the site sits at the crossing of major roads connected to other customs
sites (Brigetio and Aquincum), as well as on the border between the Pannoniae, but it may be the
least probable candidate for a portorium station in this list.
131 In “Das lokale System der mithraischen Personifikationen im Gebiet von Poetovio,” Arch.
Vestnik 28 (1977) 385-92, I. Tóth argues for the centrality of Poetovio in creating a distinct strain

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 310 M. M. McCarty, M. Egri and A. Rustoiu

tion to Transitus suggests two things: first, that he was an arcarius within this customs
system; second, that he learned how one might worship Mithras — that is, by making a
dedication to Transitus Dei— within this milieu.
To note that the customs system was a vector for the transmission of Mithraic worship
is hardly novel.132 The office seems to have created a close professional group and social
network that allowed for the recruitment of colleagues into the cult. Portorium personnel
were also mobile; they circulated among different customs stations for both longer and
shorter terms. Vitalis will have come into contact with Mithras through colleagues in the
customs administration, perhaps even at Poetovio; it was in this milieu that he was per-
suaded both to worship the god and eventually perhaps to invest in a new mithraeum. He
learned a particular manner of honouring Mithras as or through Transitus.
Yet the tile-box foundation rites do not seem to have played a rôle at other mithraea
tied to the portorium, which implies that such foundation rites were learned elsewhere
rather than in this cultic milieu. Most dedications to Transitus, and others made by porto-
rium personnel, are stray finds, without archaeological context. Two sites, however, clearly
demonstrate the absence of such foundation deposits. The first is Poetovio’s Mithraeum I.
Although excavated early on, in 1898, it was well documented for the period.133 Subsequent
consolidation and reconstruction for display meant that sub-floor levels were excavated;
no foundation box is documented and it is nearly certain that no such foundation rite was
ever enacted here. The other excavated mithraeum related to the portorium, at Ad Enum
near Pons Aeni, similarly shows no sign of a foundation deposit. Despite careful excava-
tion in 1978, there remain problems with the dating and phasing of the building.134 Two
large (1.8 x 1.4 m, 3 x 1.8 m) pits seemed to cut into the nave’s mortar surface; the excavator
suggested the mortar bedding was entirely covered with wooden boards.135 Although the
pits contained much material — pottery and building material, stone (some of it carved)
fragments, glass, mortar, charcoal, iron nails —, the excavator argued against them being
structured deposits, seeing them instead as features that could be used in ritualized pag-
eants.136 The coin finds also precluded their use as sealed foundation deposits: Pit 1 near
the altar contained 69 coins dating from Claudius Gothicus (268-270) to Arcadius (388-
402), while Pit 2 in the middle of the nave contained coins 103 coins similarly running from
Probus (280) to Arcadius.137 Since the mithraeum seems to have been founded in the mid-
to late 2nd c., these pits seem to have been accessible in the final phase of the mithraeum:
they were not sealed deposits. The positions of the coins within the pits are not recorded
but, given the general composition of the material and the size of the pits, they more likely

of Mithraism that favored such personifications which spread via the customs system; cf. also
P. Beskow, “The portorium and the mysteries of Mithras,” J. Mith. Stud. 3 (1980) 1-18.
132 E.g., Cumont (supra n.9) 62-63; Merkelbach (supra n.19) 167-74; Beskow, ibid.
133 W. Gurlitt, “Vorbericht über Ausgrabungen in Pettau,” Jh. Öst. Arch. Inst. Wien 1899, 87-96; id.,
“Ausgrabungen auf der Stätte der Römerstadt Poetovio,” Mitt. K.K. Central-Comm. zur Erfor-
schung und Erhaltung der Baudenkmale 1900, 91-96.
134 Garbsch (supra n.112) 428-35 proposes a late 1st-c. A.D. date; B. Steidl, “Neues zu den Inschriften
aus dem Mithraeum von Mühlthal am Inn. Pons Aeni, Ad Enum und die statio Enensis des
publicum portorium Illyrici,” BayVorgBl 73 (2008) 82, convincingly suggests a mid-2nd c. date
based on the limited (residual) 1st-c. material and evidence for foundation of the vicus.
135 Garbsch ibid. 364.
136 Ibid. 360.
137 Ibid. 391.

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 From foundation deposits to religion in Roman Mithraism 311

served a different purpose than foundation boxes. As at Poetovio, the Mithras-worshippers

at Ad Enum seem not to have participated in the foundation rites attested at other sites.
The lack of foundation boxes in mithraea associated with portorium personnel suggests
that foundation rites like those at our mithraeum were not part of the practices shared in
this milieu. If there was a strain of Mithraism unique to portorium personnel that involved
distinct practices such as honoring Transitus, this strain did not include an inaugural meal,
curation, and deposition of its remains under the nave. This means that, while Vitalis
learned about the cult and how to worship Mithras in the milieu of the portorium, he prob-
ably did not pick up the foundation rite in this same milieu. This savoir faire had to come
from somewhere — or rather, someone — else. Vitalis and the small group of diners at
the banquet would have invited a ritual expert drawn from a different Mithraic milieu
to assist in founding the new mithraeum. From this, we can draw two conclusions. First,
when speaking of the ‘founder’ of a mithraeum, there may have been a distinction between
the person paying for construction/furnishings and the one with the expertise to conduct
foundation rites. This need for specific expertise stands in contrast to most other ‘private’
(non-civic) cults in the Roman world; paying for a temple on private land normally was
enough to allow the builder to name himself as priest, in an act that allowed him to acquire
significant amounts of social capital via a priestly position that required no great ritual
knowledge.138 Second, the distinction at Apulum between mithraeum-builder or patron (a
customs agent) and a ritual expert who had ‘learned’ the cult outside the customs system
points to the likelihood of traveling ritual experts.
In these twin sources of savoir faire, one rooted in the portorium system’s strain of
Mithras-cult, one rooted in traveling ritual expertise, we also begin to see the ways that
practices in individual Mithraic communities might appear to converge or diverge from
one another. Such communities were not simply founded as replications of whatever
knowledge their builder had previously acquired (with all of the potential for innovation
in this act); nor were their activities wholly dependent upon replication of prescriptive
ritual schemata in the minds of traveling, trained experts. Focusing on the micro-dynamics
of ritual ‘know-how’ in one mithraeum offers insights into the range of hybrid processes
that drove diversification and standardization within the cult.

Conclusion
On the historical level, it is possible to combine an archaeologically-derived description
of foundation rites at Apulum’s Mithraeum III with models and questions drawn from
the anthropology of religion to tell the micro-historical story of how one mithraeum came
to be. In the third quarter of the 2nd c., a treasurer for the Illyrian customs system named
Vitalis was posted to the customs bureau at Apulum. He had evidently learned of Mithras,
and had learned how to worship Mithras — including dedicating to Transitus —, from
colleagues in the customs milieu, perhaps even at their main office in Poetovio. At his new
posting Vitalis evidently decided to establish a mithraeum. Doing so, however, may have

138 The best example may be P. Perelius Hedulus, who names himself sacerdos perpetuus after
dedicating a temple to the Gens Augusta (ILAfr 353; J. B. Rives, Religion and authority in Roman
Carthage [Oxford 1995] 55-57). Within Dacia, this phenomenon is also attested (IDR III.2.20);
it can perhaps also be seen in the elevation of a figure to sacerdos of Jupiter Dolichenus for
rebuilding a temple at Apulum (IDR III.5.1.217).

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press

 312 M. M. McCarty, M. Egri and A. Rustoiu

required ritual expertise that he did not possess, with the result that a ritual expert was
sought from elsewhere, someone who had learned how to perform rare foundation rites
outside the customs-system community. That expert would have conducted inaugural
rites with Vitalis and a group of other worshippers: they dined with the god, consuming
special foods like chicken, and then commemorated the occasion by collecting remains
from the meal and burying them in a specially-made box in the nave of the mithraeum
while it was still under construction.
These small-scale dynamics, emerging from a focus on religious savoir faire, offer new
perspectives on the cult in general, supporting some hypotheses about the cult while chal-
lenging other assumptions about how it worked. The foundation rite puts emphasis on
the social (rather than the symbolic) aspects of the cult. The homogeneity of foundation
rites in a series of mithraea points not only to a degree of ritual standardization previ-
ously assumed, but also offers evidence for mobile, trained ritual experts in the cult. The
distinction between these experts and the epigraphically documented individuals who
bankrolled the construction of the sanctuaries also suggests that we may need to think
differently about the dynamics between standardization and diversification within the
cult, seeing them in relation to different forms of connectivity and seeing the reproduc-
tion of the cult in new locations as an act of hybridization. Yet more than historical details
about the cult, these observations show the potential of the processual micro-archaeology
of ancient cult, when combined with anthropological and archaeological models of
knowledge-transmission, to generate new accounts of religious life in antiquity. In short,
we should focus attention on building ancient religion from the ground up, from individ-
ual deposits and archaeological contexts, through the engagement of models of cultural
and cultic transmission.
[email protected] University of British Columbia, Vancouver
[email protected] Institute of Archaeology, Romanian Academy
[email protected] Institute of Archaeology, Romanian Academy

https://doi.org/10.1017/S1047759419000151 Published online by Cambridge University Press