THE CHESTER BEATTY BIBLICAL PAPYRI DESCRIPTIONS AND TEXTS OF TWELVE MANUSCRIPTS ON PAPYRUS OF THE GREEK BIBLE FASCICULUS II THE GOSPELS AND ACTS BY FREDERIC G, KENYON LATE DIRECTOR AND PRINCIPAL LIBRARTAN OF THE BRITISH MUSEUM LONDON EMERY WALKER LIMITED 16 CLIFFORD’S INN 1933INTRODUCTION 1, DESCRIPTION HESTER BEATTY PAPYRUS LI is a codex which originally con- C tained the four Gospels and the Acts. When it first arrived in London it could only be described as a lump of papyrus, composed of a number of small leaves in very delicate condition, much frayed at the edges. Until these leaves had been carefully separated and mounted under glass by the skill of Dr. Ibscher, it was not possible to do more than ascertain that portions of some at least of the Gospels and of the Acts were included. When thus separated and mounted, it could be seen that they comprised a number of leaves, of which the upper margins were in some cases preserved, but mutilated at the bottom and on both sides, so that no line was complete. Subsequently a second parcel was found to contain portions of the Gospels of Luke and John, in a much better state of preservation. Although the lower portion of each leaf had gone, the complete width of the column of writing was preserved in six cases, while four more were nearly complete. Further, among them were two conjoint pairs of leaves, which are of material assistance in determining the original make-up of the volume. Finally (as mentioned in the General Introduction) some additional frag- ments of the papyrus, which had separately and subsequently been acquired by the University of Michigan, were most generously transferred to Mr. Beatty. These included several fragments of Luke (one amounting to nearly half a leaf) and one small scrap each of Mark and John, and constituted a material addition to the total text. From these materials it is possible, not only to describe the fragments that survive, but to ascertain with approximate accuracy the original form and contents of the manuscript. In its present condition, the manuscript consists of 30 leaves (60 pages) of papyrus, all more or less imperfect. The largest fragment, consisting of the conjoint pair of leaves containing Lk. x. 6—ix. 46, measures 15 in. in width and 7 in. in height. The smallest (apart from the scraps of Mk. xi and xii) measures 33% 1} in. The top of the column of writing is preserved in most cases, but not in all; ina few it is uncertain. The bottom of the column is lost in every case ; in the best-preserved leaves the loss amounts to about 5 lines ofvi INTRODUCTION text, besides the margin. In Lk. and Jn. either the whole width of the column is preserved, or one side of it, so that the distribution of the text between the lines is certain. In Mt.and Mk. and Acts it is not so; both sides of every column are damaged, so that the exact distribution of the text between the lines is not ascertainable. Although the top of the column of writing is generally preserved, the upper margin is not intact, and only in two cases has the page numeration escaped destruction. These two, however, are sufficient to give very valuable evidence as to the contents of the MS. The surviving numbers are 193 at the top of the page containing Acts xiv. 15-23, and 199 at the top of that containing Acts xvii. 9-17, which is the last page of the codex preserved. From these the contents of the MS. can be calculated, as shown in the following section. From these data it can be deduced that the codex originally consisted of approximately 220 leaves, measuring about 10% 8 in., with columns of writing about 74 in. high and 6+ in. in width, and containing usually 39 lines of text. The upper margin was at least 1} in., the lower presumably a little more. The inner margin, between the line of folding and the edge of the column of writing, is fin. in the two examples preserved; the outer margin may be presumed to have been about rin. The sheets of papyrus are so arranged that recto pages face recto and verso verso throughout the book. As shown in the General Introduction, this result can be arrived at in two ways. Either each sheet is folded separately into a conjoint pair of leaves, the successive pairs being then laid on one another and bound together; or a number of sheets are laid, before folding, one above the other, with the recto and verso sides alternately uppermost, and the whole folded together to form a quire of any size that may be selected. In the latter case the text will only be continuous over a pair of conjoint leaves in the innermost pair of the quire: in the former it will be continuous over each conjoint pair throughout the codex. The fortunate preservation of two successive conjoint pairs in Luke proves that the last-mentioned was the method employed here. Leaves 11 and 12 (according to the present numbering) contain the continuous text of Lk. x. 6—xi. 46 (with a few lines missing at the bottom), and leaves 13 and 14 contain the continuous text of xi. 50—xiii. 24. This proves that each shect was folded separately, or in other words that the quires were of two leaves only. These folded sheets or quires were then sewn together; the surviving examples show two pairs of holes for the threads, the pairs being 23 in. apart, and the holes of each pair 0°65 cm. from the line of folding. In one case a second pair of holes, slightly smaller, appears o'8 cm. above the upper pair. TheDESCRIPTION vil object is not clear, unless in this case it was thought necessary to reinforce the binding with a second thread. Whether the whole codex was then enclosed in a cover of leather or other material, there is no evidence to show; but such bindings have been found with early Coptic papyrus MSS. 2. CONTENTS The following is a list of the portions of the Gospels and Acts included in the MS. in its present state. Owing, however, to the condition of the papyrus, in which nearly every line is more or less mutilated, it will be understood that the mention of a verse does not imply that the whole of that verse is preserved. It may be that it is represented only by a few letters. Nothing but an inspec- tion of the printed text will show whether a given word or passage is extant in the MS. On the other hand, considerations of space will often show which of two or more alternative readings the MS. had, even when the whole passage is not preserved. Some use of such evidence is made in the apparatus criticus, and more minute scrutiny would probably establish some additional readings : but, in view of the character of the writing and the irregularity in the length of lines, caution is necessary in such calculations. The table of contents is as follows: Mt. xx. 24-32, xxi. 13-19, XXV. 41—xxvi. 3, 6-10, 19-33. Mk. iv. 36-40, v. 15-26, 38—vi. 3, 16-25, 36-50, vii. 3-15, 25—viii. 1, 10-26, 34—ix. 8, 18-31, xi. 27-33, xii. 1, 5-8, 13-19, 24-28. Lk. vi. 31-41, 45—Vvii. 17, ix. 26-41, 45—-x. 1, 6-22, 26—xi. 1, 6-25, 28-46, 50—xii. 12, 18-37, 42—xili. 1, 6-24, 29—xiv. 10, 17-33. Jn. x. 7-25, 31—xi. 10, 18-36, 43-57. Acts iv. 27-36, v. 10-20, 30-9, vi. 7—Vii. 2, 10-21, 32-41, 52—viii. 1, 14-25, 34—ix. 6, 16-27, 35—X. 2, 10-23, 31-41, Xi. 2-14, 24—xii. 5, 13-22, xiii. 6-16, 25-36, 46—xiv. 3, 15-23, xv. 2-7, 19-26, 38—xvi. 4, 15-21, 32-40, xvii. Q-17. That the codex included all the five books as a single unit is proved, not only by the identity of the hand throughout, but by the numeration of the pages, of which, as stated above, two instances (pp. 193 and 199) have fortu- nately survived. Calculation shows that a page of the papyrus is approximately equivalent to 36 lines of the Oxford text (ed. Souter, 1910). On this basis the four Gospels and Acts as far as ch. xvii. 8 (the bottom of the last complete page) would occupy 1974 pages, which is as close an approximation as could te bviii INTRODUCTION be desired to the 198 pages shown in the numeration of the papyrus. The space required for each book is approximately as follows : Matthew 493 pages Mark 305 Luke 503 4 John 38 os Acts 5° yy 218 The same result is arrived at by calculating the space required for the missing end of the manuscript. Page 198 ended with the end of Acts xvii. 8. From that to the end of the book the text occupies about 634 lines of Souter’s Oxford text, which is equivalent to just over 18 pages of the papyrus. The text would therefore have ended on p. 217; and as, in the two cases where the numeration is preserved, the odd numbers (193 and 199) are on the second pages of their leaves, there must have been somewhere (presumably at the beginning) one unnumbered blank page. This would give a total of 218 pages, or 109 leaves; but as an odd number of leaves is an impossibility (the codex being composed of sheets of papyrus, each of which when folded formed two leaves), we must assume a blank leaf at the end. The codex therefore con- sisted of 55 sheets, forming 110 leaves or 220 pages, one page at the beginning and two at the end being blank. With regard to the order of the books, the only evidence lies in the fact that Mark and Acts were closely associated in the papyrus as brought to England. This makes it probable that Mark stood last among the Gospels, as in the Freer MS. at Washington (W), where the order of the books is Matthew, John, Luke, Mark, the so-called Western order, which is found in the Codex Bezae and several MSS. of the Old Latin version. 3. WRITING The codex is written throughout in a small and very clear hand. The letters are approximately square in formation, i.e. are about equal in height and width, as opposed to the greater width which is characteristic of Ptolemaic hands, and the greater height characteristic of the Byzantine period. They have, however, a decided slope to the right, as opposed to the uprightness generally found in Roman hands of the first two centuries. Letters such as « and o have rounded backs, the tops rather flattened and brought well over.WRITING ix ais generally composed of two strokes forming an acute angle on the left, and a slightly curved stroke drawn across the opening of the angle. » is well curved and rather broad. The top stroke of ¢ is distinct from the rest of the letter, while the middle stroke, or rather point, is joined to the lower stroke by acurve. wis well rounded, the central upright lower than the two outer ones. There is no excessive length in v, ¢, and y. The other letters have the simple, unexaggerated form which is characteristic of good Roman hands. The writing is very correct, and though without calligraphic pretensions, is the work of a competent scribe. Abbreviations are confined to the usual nomina sacra, and are early in type. The following are, it is believed, all that occur : 85, Ov, Ti, Oa = Beds, Oedv, Beod, Bed. ] = lnaods (TS twice only), iyvody, inood (voc, and dat.). KS, Ke (OF kv), KV, KY, KO = KUpios, KUpte, KUpLov, Kupiov, KupiC. Wa, TH, Wi, WA = wvedpa, mveiparos, medpart, Tveipara. WP, TPa, TPS OF TS, Wp, TS = waryp, warépa, warpds, warpi, warépes. DY = oravpde WD = vids, vidy, vid. pireen. Xpavovs (apparently) = xpuorriavods (Acts xi. 26). The abbreviation ™ is very unusual, but is found in two Oxyrhynchus papyri (1079 and 1224), which are assigned by the editors to the late third or fourth century. 7p also is rare and 73 appears to be unknown. The only other abbreviation is a stroke above a vowel to represent » at the end of a line (e.g. odd = ddov), The rough breathing is used occasionally, chiefly with the article and rela- tives. A filling-mark (>) occurs rarely at the end of a line. Jota adscript is used habitually after y and w, but not after a. v ededxvorrexdy is general. Punctuation (a high point) is used occasionally by the first hand. A later hand has added more, in heavy dots or strokes above the line at the end of a clause. In one place (Acts xv. 41—xvi. 1) proper names, both personal and topographical, are marked by dots both before and after, above the line. The diaeresis is used generally over initial « or v, but sometimes over v it is replaced by a horizontal line (%). Where the two letters occur together, as in wos, three dots are placed above them.x INTRODUCTION The following spellings are noteworthy : Barca (Jn. xix) carSounaioy ByOcadav (once Badoadav) caovn’ (regularly) dvvda (= W88as, Acts ix. 38) go\opwvos in Gospels, -wvros in Acts. poons 4. DATE The dating of the MS., which naturally is of great importance, has to be determined solely on palaeographical considerations. It is plain that the hand, as described above, comes between the small, square, upright hands of the first two centuries, and the larger, rougher hands of the Byzantine period. The individual forms of letters are early, with the simplicity characteristic of the Roman period. The curves of « and o, and the absence of exaggeration in v and ¢, are also signs of relatively early date. Dating by forms of individual letters is, however, often misleading ; the general appearance of the writing is a safer guide. The hand is in the sequence between that which is to be found, for example, in the Herodas papyrus of the first or second century, and that of the magical papyri which are generally assigned to the fourth. It has none of the characteristics of the vellum uncial hands of the fourth and later centuries. It has some resemblance to the Freer or Washington codex of the Gospels, but appears to be earlier. Its sloping character is in accordance with the general practice of the third century. The date which I should assign to it is the third century, and the first half of it rather than the second. In a matter of such importance it would be unsatisfactory to rely on a single opinion: but precisely the same estimate was independently formed by papyrologists of the experience of Mr. H. I. Bell and Dr. W. Schubart. Prof. A. S. Hunt, who has been good enough to give an opinion on a photograph of a page of the papyrus, also assigns it to the third century, though he is inclined to place it in the second rather than the first half of the century. I do not think that he, any more than I, would wish to dogmatize on such a point; indeed, no competent papyrologist would on the available evidence care to be positive within a generation or two. A date in the third century may, however, be assigned with some confidence, and the primitive method of quire-formation, described above, and the early type of some of the abbreviations are also arguments in favour of an early date. If this conclusion be accepted, it is obvious that we have in this manuscript the earliest example of a codex including all the four Gospels, as well as theDATE xi Acts. It confirms the evidence, referred to in the General Introduction, of the early use of the codex form of book, in preference to the roll, by the Christian community, and proves that at any rate in the third century it was possible to have the four Gospels in a single volume, instead of being depen- dent on separate rolls for each of them. 5. THE TEXT Only a summary examination of the character of the text can be attempted here, based upon a first collation of the papyrus with some of the principal MSS. The first question of interest is to ascertain, if possible, to which of the main families of text the new MS. belongs. For this purpose each book must be examined separately: for although the codex contains all four Gospels and Acts in a single volume, it must have descended, directly or indirectly, from separate rolls of each of them; and, as is well known, in many MSS. the several Gospels evidently descended from ancestors of different types. The figures given in the tables below cannot be absolutely guaranteed, but they may, it is hoped, serve to give a correct general impression of the character of the text. It is proposed to give, first, a table of the number of agreements and dis- agreements with the principal MSS. in the passages in which differences of reading occur. For this purpose the manuscripts chosen for comparison are the five great uncials SABCD, the later uncial L, the Washington codex (W), the Koridethi codex (©) which represents the Caesarean text, the groups of minuscule MSS. known as fam. 1 and fam. 13, and in Mark also the minus- cules 565 and 700. In this list no account is taken of passages in which the papyrus has a reading singular or nearly singular to itself. A second table shows the number of passages in which the papyrus has a singular reading, and those in which it agrees with one or a small group of other MSS. With regard to all these tables it should be noted that the more marked and peculiar readings of D and its allies are not taken into account. The object is rather to show the relation between the papyrus and what may be taken to be the underlying text of D, apart from those more marked divergences specially characteristic of it and of the African Latin. The real number of differences from D is therefore much greater than appears in these tables; but it is the agreements that are more significant, and these are shown in full, ‘The remains of Matthew arc so scanty that the figures do not signify much. They are, however, given for the sake of completeness.xii INTRODUCTION With Papyrus Against With Papyrus Against 8 4 6 WwW 5 6 A I 7 ° 3 9 B 2 9 Fam. 1 5 6 Cc 4 6 Fam. 13 8 5 D 7 6 = (= Textus L I 9 Receptus) 3 7 Singular readings or agreements in small groups: ‘Singular 3 With D 2 With 2 » W ° » A ° » fam. 1 ° » B ° In Mark the figures are more important, and though the total amount of text preserved is not large, it is sufficient to establish its general character. The total agreements and disagreements are as follows: With — Against With Against x 42 108 ° 65 91 A 54 94 Fam.t 72 80 B 44 106 Fam. 13. 79 73 Cc 31 67 Cod. 565 68 74 D 49 100 Cod. 700 57 87 L 38 97 c 55 94 Ww 107 52 Singular readings and agreements in small groups: Singular 30 With W 25 With = » O 10 » fam.r 15 » fam.13 24 » cod. 565 18 » cod. 700 12 4 Aw NY OA roaOeSTHE TEXT xiii In Luke, of which a much larger portion is preserved, the figures are as follows : With Against With Against x 138 +150 WwW 124 163 A IIo 182 e 96 167 B 158 130 Fam.1 125 160 Cc 85 96 Fam. 13 105 136 D 136 137 ¢ 117 170 L 153 132 Singular readings and agreements in small groups: Singular 116 With L 25 With x 22 » W Ir » A 2 » © 5 » B 27 » fam. 1 7 » C 7 » fam.13° 9 » D 36 In this Gospel © generally agrees with s, but in 76 cases where they differ the papyrus agrees with © in 38 and with ¢ in 38. W also is mainly Byzan- tine in the chapters represented in the papyrus, but in 54 cases where W and ¢ differ, the papyrus agrees with W in 31 and with ¢ in 23. The relatively small portion of John shows the following results : With Against With — Against x 40 41 Ww 34 48 A 26 56 ° 42 41 B 40 40 Fam.1 36 44 Cc 9 20 Fam. 13 26 62 D 43 37 ¢ 23 60 L 32 48 Singular readings and agreements in small groups: Singular 25 With L 2 With x 7 » W 5 » A I 5 © 7 » B 4 58 fam.1 3 » C ° 5 fam. 13 1 » D 15xiv INTRODUCTION In this Gospel W belongs to the Neutral group, and © would appear to be of the same character. Consequently the agreements of the papyrus with these MSS. are much more numerous than in Luke. The papyrus agrees 27 times with © against s, and only 7 times with s against ©. With W the agreements are 26 against 14 with s. In Acts there are fewer important MSS. for comparison, and fewer varieties of type of text. A and C join the same family as B and x, and the main point of interest is the distinction between this type and the ‘ Western’ type headed by D. In the following table all agreements with D are shown, but the more extensive and characteristic variants of D are not recorded, nor a considerable number of minor variants in which D stands alone or nearly alone. The figures, so qualified, are as follows : With = Against With — Against x 121 69 Cc 81 46 A 115 77 D 89 75 B 122 72 s 98 89 The number of singular readings is high, amounting to 80. Of agreements in small groups there are only 2 with 8, none with A, 4 with B, 3 with C, and 8 with D. From these figures, inexact as they inevitably are, certain conclusions appear to emerge with some clearness. In the first place it is evident that the papyrus does not range itself exclusively with any of the main types of text generally recognized. It is certainly not of the Antiochian or Byzantine type. It would, indeed, be surprising if it were; and in fact not only do its disagreements with ¢ greatly outweigh its agreements, but A also is by far the least favoured of all the great uncials in the Gospels. But neither does it show a decisive preference for B and x, the main representatives of Hort’s ‘ Neutral’ type, though it has considerable affinities with them, especially in Luke and John. With D and the other authorities of the so-called « Western’ type its relations are interesting and significant. In all the Gospels there are a considerable number of passages in which it supports readings of this type, including a good many which have exclusively ‘Western’ support. But it has none of the more notable variants characteristic of this type, and where D and the Old Latin have peculiar readings, the papyrus is usually against them. Further, it has a considerable number of readings which are either singular or have very slight support elsewhere, though most of these only affect the order of theTHE TEXT xv words or other small details. The most notable new readings are in Lk. ix. 50 (03 yap eorw Kab’ ipav ob8e Snép tuav for bs yap otk éorw Kal’ tucv bmp spar eativ), xi. 15 (€AdAnoray dxupol Aéyovres for elwov), xi. 42 (dvnfov for mfyavov), xii. 4 (ronOire for GonOjre), and xiii. 32 (wovotpar for dmorehG or émureda). On the other hand it is clear that the papyrus has a strong affinity with the group which has of late years been segregated by Canon Streeter and others, and has been identified by them with the text of Caesarea. This group is most noticeable, and has hitherto been almost exclusively studied, in Mark, where it is mainly represented by the uncials W and © and the families of minuscules headed by codd. 1 and 13. Now in the portions of Mark contained in the papyrus the predominance of this type of text is strongly marked. W is the single MS. with which it is most often in agreement ; © and famm. 1 and 13 and cod. 565 follow next in order, with a decisively higher rate of agreement than with either SB or D. There is, indeed, less agreement with SBL than with As, and almost the same amount with Az and D. The agreements with WO famm. 1, 13 in small combinations amount to 74, and those with sABD to 30, of which 13 are with D; while the total agreements with the former group are 323, and with the latter 189. It is clear therefore that in this Gospel the character of the text is definitely Caesarean. It is almost equidistant from ‘Neutral’ and ‘ Western’, with a slight leaning to the latter, and is no further removed from A. The evidence of the Chester Beatty papyrus will be of con- siderable importance in determining the relations between the Caesarean text and Egypt. In this connexion it will be remembered that Prof. Lake and his colleagues at Harvard have so far modified Streeter’s contention that Origen used a ‘ Neutral’ text while he was at Alexandria and a ‘Caesarean’ text after his removal to Caesarea, as to show that he certainly used a Neutral text for a short time after his settlement at Caesarea, though he subsequently abandoned it for the Caesarean type, to which he thenceforward adhered ; while there is at least some probability that he used a Caesarean text in Alexandria.’ The present papyrus seems to show that there would be no impossibility in his finding a MS. of this type in Egypt. With regard to the other Gospels, the remains of Matthew are too slight to enable the character of its text to be determined. So far as the evidence goes it appears to show a slight preponderance of the Caesarean group over the others, with 8 and D next in preference. In Luke the balance of evidence is * The Caesarean Text of the Gospel of Mark, by Kirsopp Lake, Robert P. Blake, and Silva New (Harvard Theological Review, xxi. 4, 1928). pts cxvi INTRODUCTION quite different. The Caesarean text in this Gospel has not been examined, and it is known that W, for the greater part of the book, is substantially Byzantine. The same may apparently be said of ©. We are therefore still very uncertain as to what the Caesarean text of this Gospel was, and it may be that the papyrus is a better representative of it than any witness hitherto known. Its relations to the other principal authorities are, however, markedly different from those which it has in Mark. Its closest affinity is with the Neutral text, notably with B and L rather than 8; D is rather further away, but all these are distinctly nearer than AW© famm. 1,13. The papyrus text can therefore be defined as about equidistant between ‘ Neutral’ and ‘Western’, with a slight leaning to the former, and without the peculiar readings specially associated with the latter. In John the proportions are again somewhat different, so far as can be gathered from the small amount of text preserved. In this Gospel © falls into the ‘Neutral’ group, and shows much the same figures as NB. The papyrus again stands between the Neutral and Western families, but this time is slightly nearer to the latter, LW and fam. 1 are further away, and fam. 13, A,C, and = are decidedly more remote. The character of the papyrus may be further illustrated by showing its relations to other MSS. in readings which do not appear in the Textus Receptus. Omitting Matthew as unimportant, the results are as follows: In Mark: Papyrus with & 24 Papyrus with W 59 ” » A 7 ” » © 36 » » B 28 55 » fam.1 32 » » C 15 » » fam.13 35 i » D 28 55 » cod. 565 40 ” » L 24 ” » cod. 700 25 In Luke: Papyrus with 8 85 Papyrus with L 92 » » A 24 ” » W 28 ” » B 100 ” » © 39 ” » C 34 » » fam. 1 53 ” » DTHE TEXT xvii In John: Papyrus with & 31 Papyrus with L 22 * » A 9 3 5 OW 24 * « 3B 29 * » © 26 = x C 7 iz » fam. 1 21 » » D 32 ” » fam. 13 10 These figures confirm the conclusion already arrived at. ‘They show a strong affinity with the Caesarean group, and especially with W, in Mark ; a less close association in Luke (in which Gospel, however, the Caesarean text has not yet been much studied), and a more frequent alliance, first with B L, and in the next place with D and x (B advances in this table a little, while D falls back) ; and in John an almost equal adherence to NB, D, and ©, but with a noticeable leaning to D and © in particular readings. With A and the distinctively Antiochian group it has decidedly less in common. Before leaving the Gospels it may be of interest to check these results by comparing them with those arrived at by Prof. Lake in his examination of Codex 1. In his edition of that MS." he has given lists in which the variant readings are classified under various headings. Since these lists are compiled only from selected passages, and much of these is not represented in the papyrus, the comparison cannot be exhaustive ; but the results are instructive. In the readings given in list A (readings of cod. 1 which are found in most ancient authorities but not in the Antiochian text) the papyrus has one agree- ment in Mark with cod. 1 and no disagreement. In list B (readings for which the Old Latin version is the chief ancient authority) the papyrus has 4 agree- ments and 2 disagreements in Mark, 1 agreement and 7 disagreements in Luke. In list C (readings for which the Old Syriac version is the chief ancient authority) the papyrus has 2 disagreements in Mark, 1 agreement and 2 disagreements in Luke. In list D (readings which are found in both the Old Latin and Old Syriac versions, but not in NB or in the Antiochian text) the papyrus has 2 agreements and 4 disagreements in Mark, and 4 disagreements in Luke. So far the amount of agreement with cod. 1 is not very great; but in list E (readings for which xB are the chief authorities) the papyrus has 4 agreements and 1 disagreement in Mark, 6 agreements and 7 disagree- ments in Luke; and in list F (readings which are supported by a few other manuscripts but cannot be identified with any authority generally recognized © Codex 1 of the Gospels and its Allies (Cambridge Texts and Studies, vii. 3, 1902).xviii INTRODUCTION as primary) there are 8 agreements and 5 disagreements in Mark, 3 agree- ments and 10 disagreements in Luke. Finally in list G (readings which appear to be peculiar to fam. 1) there is only 1 agreement and 8 disagree- ments in Mark, and 5 disagreements in Luke. The agreements in list F, in readings for which hitherto no early authority has been known, are especially significant. In Acts the textual problem is rather different. It is a question of a choice between what Ropes calls the Old Uncials (sABC, with which may be ranked the minuscule 81) on the one side, and D and the Old Latin on the other. Here there can be no doubt that the papyrus ranges itself with the former group. It has none of the major variants characteristic of the text of D in Acts, and in respect of minor variants it has a lower proportion of agreements with D than with the other uncials. On the other hand it must be observed that it does contain a certain number of the minor variants for which the evidence has hitherto been ‘ Western’ in character. The papyrus therefore is evidence for the existence in Egypt of readings subsequently absorbed into the Western text but rejected by the others, though it gives no support for the more marked Western variants. As between x, A, B, C, the proportion of agreements and disagreements with the papyrus is almost the same. The general result to which this evidence points would appear to be that this MS. is a witness to the existence in Egypt in the first half of the third century of a type of text distinct from that found predominantly in B, and with a strong infusion of readings found in the early authorities which are grouped together as ‘ Western’, though with none of the larger divergences which are found in some of these authorities. Further, it is closely akin (at any rate in Mark) to the type which found a home in Caesarea. The fact that this type of text is found in a MS. written in Egypt during the lifetime of Origen (or not much later) is some evidence that the type did not originate in any editorial revision undertaken at Caesarea, but existed already in Egypt, whence it was taken (it may be by Origen himself) to Caesarea, where it was adopted by the great scholars who used the library of Pamphilus. Its * Western’ features do not imply any connexion with either Rome or the Syriac Church. It only confirms the conclusion as to the misleading character of the term ‘Western’. The wider divergences, which are characteristic of D and the Old Latin version on the one hand, or the Old Syriac version on the other, do not occur in it. The readings which do occur in it are not geographically Western or Syrian, but are early readings which did not find a place in B, butTHE TEXT xix which, in varying degrees, are preserved in Western, Syrian, or Caesarean authorities. What further conclusions may be drawn as to the textual history of the New Testament are matters of speculation, which must be left to the mature consideration of scholars. The papyrus was not, as the Codex Vaticanus evidently was, the product of a great centre, which had the command of the best scribal facilities and presumably also the best critical materials. The Vaticanus, when once its Egyptian origin is admitted, can hardly have been produced elsewhere than in Alexandria, where many manuscripts would have been available for comparison. It is possible to believe that its text represents the result of critical revision, and yet that it gives on the whole the best representation of the primitive text. By the time that it was produced, in the first half of the fourth century, the primitive text probably did not exist any- where in a fully accurate form. All the evidence that we have goes to show that in the third century great divergences of text in small details existed and were spread universally over the Christian world. A scholar working on the text at that time would have had to make his choice between a variety of readings. But in Alexandria he would have had a larger choice than in most other places of early authorities, and he would have been brought up in an atmosphere of critical scholarship. It would not be surprising, therefore, if he was able to produce there a purer text than was to be found in at any rate most of the manuscripts which had been produced under less favourable conditions throughout the Christian world during the previous century. The question of the general predominance of the Vaticanus therefore still remains open. On the other hand, it is impossible to claim verbal inspiration for such a scholar; and the Chester Beatty papyrus unquestionably reinforces the view that readings found in other early authorities have a full claim for considera- tion on their merits. One of its special claims to importance is that it gives us some insight into the condition of the Bible text a century before the period of the great vellum uncials; and while on the one hand it seems to reveal a state of things in which there was a considerable diversity in respect of minor details, and when the distinct types of text which we find later were in process of formation, on the other hand it carries back for about a century the proof of the substantial integrity of our textual tradition. It shows no marked divergences of its own from the main tradition, and it countenances none of the marked divergences found in certain of our early witnesses. It is true that it is very imperfect; but it covers such a substantial portion of theXX INTRODUCTION Gospels that it is legitimate to draw general conclusions from it, and these show us in the early part of the third century a text of the Gospels and Acts identical in all essentials with that which we have hitherto known on the evidence of later authorities. It must be repeated that the conclusions which have been indicated above rest on a first and somewhat superficial examination of the text of the papyrus. In particular it would have been more satisfactory if it had been possible to include the results of a comparison with the versions, especially the Coptic, Latin, Syriac, and Georgian, and with the quotations in Origen from Luke and John. This, however, would have caused considerable delay, since the materials for such an examination would have had to be collected from several sources, and the present editor has not the linguistic equipment for such a task ; and it did not appear right to withhold the publication longer than was absolutely necessary from those far more competent scholars who will be able to complete the study of it and to determine with greater authority the character of the text and the conclusions to be derived from it. 6. THE PRESENT EDITION In the present edition the transcript of the manuscript has been made as exact as possible. It can be checked by the photographic facsimile which is published separately. But it is to be remembered that in a papyrus with ragged edges, such as this, many letters are represented by very slight remains, although these remains may be sufficient to establish the reading beyond doubt. Letters im- perfectly preserved are indicated by a dot beneath them, but only an examina- tion of the photograph will show to what extent the preservation is imperfect. Even the photograph, however, is not decisive, for (unlike a vellum MS., in which photography will sometimes show the writing more clearly than the un- assisted eye can see it) the shadows of fibres of papyrus sometimes give an appearance of ink which does not exist, or again conceal indications of writing that exist. In doubtful cases only an examination of the original can decide. The collation of the text of the Gospels rests mainly on Tischendorf, and is therefore subject to the imperfections of that apparatus. I cannot claim that it is exhaustive, nor hope that it is free from mistakes or omissions. In the case of Mark, however, I have had the advantage of a revision by the Rey. S. C. E. Legg, who has used the materials collected by him for the new Oxford edition of the Greek New Testament. I have also used the collationsTHE TEXT xxi by Prof. H. A. Sanders and Prof. E. J. Goodspeed of the Freer manuscript (W),' and the editions of the Koridethi codex (©) by Beerman and Gregory,* of Codex 1 by Lake,’ and of the Ferrar Group (fam. 13) by Ferrar and Abbott.* For the collation of Acts much use has been made of the apparatus given by Ropes.* I have to thank the Trustees of Dr. Williams’s Library for the loan of several volumes employed in the work. It is not claimed that the apparatus is exhaustive or final. In particular, variations in spelling (especially itacisms, the use of final », and the like) are generally ignored. The apparatus is only intended to make possible a first estimate of the character and value of the new MS. As already mentioned in the General Introduction, the official number assigned to the MS. by Prof. von Dobschutz (who keeps the generally accepted register of New ‘Testament MSS. in succession to Gregory) is P*. * Sanders, The Washington Manuscript of the Four > Codex x of the Gospels and its Allies (Texts and Gospels (The New Testament Manuscripts inthe Freer Studies, vii. 3, Cambridge, 1902). Collection, Part 1, New York, 1912); Goodspeed, ¢ Collation of Vour Important Manuscripés of the The Freer Gospels (University of Chicago Press, Gospels, Dublin, 1877. Chicago, 1914). 5 The Beginnings of Christianity, vol. iii: The > Die Koridethi Evangelien, Leipzig, 1913. Text of Acts (London, 1926).