The Limits of Palaeographic Dating of Literary Papyri: Some Observations on the Date and Provenance of P.Bodmer II (P66)

MUSEUM HELVETICUM Herausgegebenvon -tdi1€par - editoacuradi Schweizerische Vereinigung far Al1er1umswissenschaft (SVAW) Redaktion - rtdaction-redazione Thomas Schmidt - Stefan Rebenich - Danielle Van Mal-Maeder - Martin. Guggisberg I FedericoM.Petrucci:TrenocesultestodelMenessenodiPJatone 36 Manuel Sanz Morales: Zwei Interpunktionsprobleme beiCharitonvonAphrodisias(3.3.17,3.10.8) .. ａｮｫ｡ｵＺｾ［ｩ＠ ｊｯ［ｾ＠ ｾ ｾＺｯ［｡､ｵ ｧｺｾ ｾ ｾｬｴＺ･ｮ＠ ｾ ｾＧ［ｩ･ｎａｶＯＺｅｌｔｉｏｂｖｓ＠ MarcDominicy:Notescritiquessur!'fltgie4,lde Prope rce 73 85 O/ivierDevi/lers: Lafindel'anntel4danslesAnnales (:\proposde Tacitc,Annales, 1,54) Archiiologische Berichte Richtlinien filr Autorinnen und AU!oren (Zur Herstellung der Druckvorlage): siehe www.schwabe.ch unter «Verlag Schwabe», dann «Zeitschriften», dann «Museum H elveticum» The Limits of Palaeographic Dating of Literary Papyri: Some Observations on the Date and Provenance of P.Bodmer II (P66) By Brent Nongbri, Sydney Abstract: Palaeographic estimates of the date of P.Bodmer II, the well-preserved Greek papyrus codex of the Gospel of John, have ranged from the early second century to the first half of the third century. There are, however, equally convincing palaeographic parallels among papyri securely dated to as late as the fourth century. This article surveys the palaeographic evidence and argues that the range of possible dates assigned to P.Bodmer II on the basis of palaeography needs to be broadened to include the fourth century. Furthermore, a serious consideration of a date at the later end of that broadened spectrum of palaeographic possibilities helps to explain both the place of P.Bodmer lI in relation to other Bodmer papyri and several aspects of the codicology of P.Bodmer II. Introduction P.Bodmer II, better known to most scholars of the New Testament as P66, is a much-studied papyrus codex of the gospel of John. 1 Both its relative completeness and the early date often associated with this codex have brought P.Bod- MuseumHelveticum11(2014)1-35 BrcntNongbri mer II a good deal of attention in the years since its publication. 2 It probably originally consisted of 156 pages, of which 104 are fully or nearly fully preserved Substantial remains of most of the rest of the pages have also survived. Most palaeographic estimates of the date of the codex have ranged from the early second century to the early third century. In what follows, I argue that the range of possible dates assigned to P.Bodmer II on the basis of palaeography needs to be broadened to include the fourth century. Furthermore, a serious ｣ｯｮｳｩ､ｾ＠ erat ion of a date at the later end of that broadened spectrum of palaeographic possibilities would help to explain 1) the place of P.Bodmer II among the other Bod mer papyri and 2) several aspects of the codicology of P.Bodmer II. The Original Publication and Subsequent Proposals for Dating P.Bodmer 11 The original editor of P.Bodmer II, Victor Martin, assigned the codex to the early third century or, as he is more usua lly cited in English literature, "ca. 200". 3 Martin's statement of the date of P.Bodmer II is as follows: TheLimi1sofPalaeographicDa1ingofLi1eraryPapyri Martin mentioned severa l individual papyri in his paleographical discussion of P.Bodmer II. The papyrus that Martin found most releva nt for the dating of P.Bodmer Il was P.Oxy. 8.1074, a 2.7 x 5.1 cm fragment of a codex of Exodus. To faci litate comparison, a leaf from P.Bodmer II (Figure 1) is reproduced below along with images of the two sides of P.Oxy. 8.1074 (Figures 2 and 3). Figure l: P.Bodmer 11, page 44. By permission of the Fonda1ion Martin Bodmer, Cologny (Geneva) Figure 2: P.O xy. 8.1074 recto. Courtesy of The Spu rlock Museum, University of Ill inois at Urbana-Champaign Figure 3: P.Oxy. 8.1074 verso. Courtesy of The Spu rlock Museum, University of Illinois at Urbana-Champaign BrentNongbri P.Oxy. 8.1074 was assigned by its editor (Arthur S. Hunt) to the third century or possibly earlier. 5 Martin thought that such a judgement "est exactement celui qu'on peut raisonnablement porter sur l'ecriture de notre Evangile".6 Martin was clearly correct to detect similarities between the hands of these two papyri, but since P.Oxy 8.1074 was itself dated strictly on palaeographic grounds it is actually of no help in assigning a date to P.Bodmer JI. Using one pa laeographically dated papyrus to assign a palaeographic date to another undated papyrus results in, to borrow a phrase from Peter Parsons, "only jelly propped up with jelly".7 Thus, P.Oxy. 8.1074 and a number of Martin's other proposed parallels should from the outset be set aside as unhelpful in establi sh ing a date for P.Bodmer Il.8 Fortunately, Martin did gather some parallels for the handwriting of P.Bodmer II that come from more securely dated papyri. These pieces are worth examining more closely. The follow ing papyri that Martin mentioned have features that provide, with various degrees of certainty, a terminus ante quern and thus offer at least the possibility of helpful palaeographic comparisons: P.Oxy. 1.20 (LDAB 1630; Figure 4), twelve fragments from a roll of Homer assigned to the second century since the reverse contains what Grenfell and Hunt describe as "some accounts in a cursive hand of the late second or early third century". While the judgement is sti ll ultimately pa laeographic, examples of securely dated cursive hands are much more plentiful and hence more reliable guides The LimitsofPalaeographicDatingof LiteraryPapyri Figu re 4: P.Oxy. 1.20. By permission of The British Library C British Library Board P.Oxy. 4.661 (LDAB 474; Figure 5), a fragment of a roll of lyric poetry assigned to "the latter half of the second century" because its reverse contains writing in a cursive hand that is, according to Grenfell and Hunt, " not later than the begi nning of the third century". Figure 5: P.Oxy. 4.661. Courtesy of the Centre for the Study of Ancient Documents, the Cairo Museum, the Association Internationale de Papyrologues, and Dr. Adam Bil low-Jacobsen BrentNongbri P.Oxy. 13.1622 (LDAB 4052; Figure 6), a fragment from a roll of Thucydides, assigned by Grenfell and Hunt to the "early second century" because the reverse was reused for a contract with a date of 148 CE (published separately as P.Oxy 14.1710) P.Oxy. 18.2169 (LDAB 490; Figure 7), a fragment of a roll of Callimachus, "attributable to the later part of the second century", reused for an account on the reverse. 9 Figure 7: P.Oxy. 18.2169. Image courtesy of the Egypt Exploration Society and Imaging Papyri Project, Oxford The Limits of Palaeographic Dating of Literary Papyri P.Ryl. 1.16 (LDAB 2661; Figure 8), a fragment of a roll of comedy assigned by Arthur S. Hunt to the "latter part of the second century", since its reverse was reused for a letter (published separately as P.Ryl. 2.236) dated to January of either 253 or 256 CE. io Figure 8: P.Ryl. 1.16. Reproduced by courtesy of the University Librarian and Director, The JohnRylandsLibrary,TheUniversityofManchester BrentNongbri PSI 1.8 (LDAB 1443; Figure 9), a fragment of a roll of Homer assigned to the late first or early second century. It was reused for documentary purposes both on the recto in the margi ns and on the reverse. The documentary texts have been assigned to the third century or possibly early fourth century. 11 Figure 9: P.S.I. 1.8. lmage appears courtesy of the Ministero per i Beni e le Attivit3. Culturali, the BibliotccaMediceaLaurenziana,Florence. Reproductionprohibited As the images indicate, the resemblances between these hands and the hand of P.Bodmer II are not especially impressive. T he dated hands Martin has assembled are for the most part of a more forma l and less rounded type t han the hand of P.Bodmer II. Of Marti n's dated comparanda, P.Oxy. 13.1622 wou ld probably be the most legitimate match in terms of overa ll appearance, and even here the resemblances are not overwhelming. Nevertheless, for this second-century piece, the suggestion of a parallel with P.Bodmer U is not wholly unreasonable. TheLimitsofPalaeographicDatingofLiteraryPapyri In 1960, Herbert Hunger critiqued Martin's palaeographic work and argued for a rather earlier date for P.Bodmer I I - the middle or even early second century. 12 Hunger latched on to Martin's passing remark suggesting that P.Bodmer II bore some superficial simil ar ity to P.Ryl. 3.457 (P52; LDAB 2774) a nd the Egerton Gospel (Egerton Papy rus 2; LDAB 4736), ma nusc ripts which at that time were regarded as d ating fro m the first half of the second centu ry.13 Hunger then criticized Martin fo r not assigning P.Bodmer II to the same period. To make his case, he named over twenty-fi ve papyri that he deemed worthy of compa rison with the hand of P.Bodmer IL As was the case with Marti n's pa laeographic assemblage, a number of Hunger's proposed para llels are themselves palaeographically dated and thus are in fact of no independent va lue for assigning a date to P.Bodmer 11. 14 Hunger did, however, offer severa l examples of fir mly dated papyri that he clai med showed good simi larity to the ha nd of P.Bod mer II or were immediate precursors to the style of P.Bodmer II . T he securely dated documentary papyri wr itten in hands that Hunger rega rded as simi lar lo that of P.Bodmer II were: P.Oxy. 1.37 (49 CE), P.Oxy. 2.275 (66 CE), 12 13 14 H. Hunger, "Zur Datierungdes Papyrus Bodmer II (P66)" Anzeigerder Osterreichischen Akademieder Wissenschaf1enphilosophisch-his1orische Klasse91(l961)12 -23. Hungerconcludes his article with the following: " lch halte es daher auf Grund des vorgelegten Vergleichsmatcrials nicht nur fi.ir gerechtfcrtigt, sondern sogar fiir erforderlich, 1*- wenn schon nicht in die l . Hltlfte -so zumindestindieMittedes2. Jah rhundenszusetzen" (23).Beforethispoint, therehadbeen littledebateabout Marti n'sdate.Theonlycriticalopi nionsknownto meare associated with Jean Duplacy who speculated 1ha1 Martin may have dated the codex a bit too ea rty;heoffhandedlydescribedadateof "vcrs200"as"u n peu haute"("Oiienest lacritiquc textuelledunouveautestament? ll ",RecherchesdeScienceReligieuse46[1958]210-313, quotationat294;reprintedinDuplacy,0Uenestlacritiquetex1uelledunouveauresramen1? [Paris 1959J 49). Elsewhere, Duplacy cited an anonymous papyrologist ("trCs competent") as assigningP.Bodmerllto1hefourthcentury("Bulletindecritiquetextuelledunouveautcsta· ment. l n, RecherchesdeScience Religieuse50[1962]242-263,at25 1, n.16) Onproblematicaspectsofthcpalaeographicdatcsassignedtothesemanuscripts(andfor images of the manuscripts), see B. Nongbri, "The Use and Abuse of P52: Papyrological Pitfalls in the Dating of the Fourth Gospel", HTR 98 (2005) 23-48 lnadditiontoP.Ryl.3.457andtheEgcrtongospel,Hunger'snon-securelydatedpapyri included: the Chester Beatty Numbers-Deuteronomy codex (LDAB 3091) assigned by Kenyon to thcearlysecondcenturybutby Hunttothelatesecondorearlythi rdcentury; P.Oxy.17.2080 (LDAB 487), a roll of Callimachus described by Hunt as "li kely to fa ll within the second century"; P.Oxy. 19.2213 (LDAB 491), another portion of Callimachus assigned to the "first partofthesecondcentury";agroupofpapyrusrollsfromOxyrhynchus,P.Oxy.18.2161 -2162 (LDAB 103 and 117) and PS I ll.1208-12IO(LDAB 100, 103, and 102), now all attributed tothesamescribeandassignedadateinthesecondorearlythirdcentury;seeW.A.Johnson, BookrollsandScribesinOxyrhynchw(Toronto2004)18-20and61. Hungeralsocites PSI 11.1211 andP.Oxy.13.1622,whichwerebothnotedbyMartinanddiscussedabove; P.Berol. 9782 (LDAB 3764), a papyrus roll with a commentary on P\a1o's Theaetetus generally assigned to the second century; P.Oxy. 22.2310 (LDAB 320), portions of a roll of Archilochus assigned to "about the middle of the second centu ry"; and P.Oxy. 25.2436 (LDA B 4808), a papyruswithmusicalnolationassignedto tbe"cndofthefirstcenturyo r theearlysecond century" BrentNongbri P.Oxy. 2.286 (82 CE), P.Oxy. 1.94 (83 CE), P.Oxy. 10.1282 (83 C E ), P.Oxy. 2.270 (94 CE), P.Oxy. 4.713 (97 CE), P.Vind. Gr.19812 (= P.Vind.Bosw. l; not long after 87 CE), P.Lond. 2.141 (88 CE), P.Oxy. 12.1434 (1071108 CE), P.Vind. Gr. 12247 (= CPR 1.28; 110 CE), P.Mich. 3.202 (105 CE), and P.Vind. Gr. 2004 (= CPR 1.18; 124 CE). Hunger also noted P.Oxy. 5.841 (LDAB 3713) a collection of fragments of a roll of Pindar that can be dated with a reasonable degree of certainty to the early second century since the text of Pindar was written on the reverse of a doc· umentary roll of the late first century, and the roll contains cursive scholia that Grenfell and Hunt considered "not later than the middle of the second century". Figure 10: P.Oxy. 2.286. By permission of The British Library © British Library Board The Limits of Palaeographic Dating of Literary Papyri Hunger did not offer detailed comparisons with all of these papyr i, and nor will I. Instead, I will focus on what is perhaps Hunger's most comparable secu rely dated papyrus, P.Oxy. 2.286, a piece that Hunger himself singled out for more extensive discussion.ts P.Oxy. 2.286 is a petition from the first year of the reign of Domitian (May of 82 CE). An image is reproduced here as Figure 10. Hunger described the similarities he perceived between P.O xy. 2.286 and P.Bod mer II as follows: P. Oxy. 286 (= P. Lond . 797) !asst sich mit seinen runden Buchstabenformen , wenn man von dem ande rs gestalteten Xi und kursiven E lementen wie dem Alpha und der Ligatur E psilon-Iota absieht , mit P66 gut vergleichen. So finden wir die Dreiergruppen in z. 11 frn:ioE1v, 12 -coU•rov, 15 µuo:SoUvcu, 23 aU.oU<;; u. [a]. Tau wird in Iuxtaposition manchmal gedrOckt, z. B. 13 -cfi<;, 20 -cfiv,250£-cfiv I have enla rged some of the particular elements Hunger described and placed them next to samples from P.Bodmer II below (see Figure 11): page46,linel ｾ＠ line23 page5, line2 Figure 11: P.Oxy. 2.286 and P.Bodmer IL By perm ission of The British Library C British Library Board and the Fondation Martin Bodmer, Cologny (Geneva) The hands might be described as comparable insofar as both are rounded with occasional ligatures , but the similarities are not at all impressive. P.Bodmer II is genera lly more bilinear (the vertical line extending down into two of the examples from P.Oxy. 2.286 is the tail of a rho, which regul arly descends into the fo llowing line in that piece) and shows fewer tendencies toward cursive styles (note the omega-nu combination in P.Oxy. 2.286). Nevertheless, Hunger's assignment of P.Bodmer II to the early of midd le part of the second century has been taken 15 Acasecouldbemadethat P.Oxy.5.841 isinfactHunger'sclosest relatively securely dated parallel.Platesareavailableforconsultationi ntheoriginalpublication BrentNongbri up by many scholars and has even made its way into a standard palaeographic handbook. 16 Eleven years later, Eric G. Turner described Hunger's proposed dated documentary parallels as "not cogent" and argued instead that P.Bodmer II should be dated in the "earlier iii A.D".17 Turner described the writing of P.Bodmer II both as "Medium-sized, rounded, 'decorated' capital, slowly written. Slightly flattened in appearance, letters being given horizontal extension" and as "a rudimentary form of 'Coptic Uncial"'. 18 The basis for his dating of the codex consisted of several attributes of the hand of P.Bodmer II that Turner regarded as more generally characteristic of the third century: "the broad forms of 9 ... and 8, thee with finial at end of cross-bar, the narrow a in a single sequence are hardly of ii A.D.; apostrophe between double nasals (uy"yeA.oui;) is not normally written in documents till iii A.D". 19 The combination of these various features thus led Turner to a third century date for P.Bodmer Il. 20 Many scholars have found Turner's arguments persuasive. 21 Nevertheless, Hunger's dating is preferred in The Limits of Palaeographic Dating of Literary Papyri some circles, and it is not uncommon to see P.Bodmer II described as dating to "the middle of the second century". 22 A variety of historical conclusions (about second century Christian scribal habits and about the early popularity of the gospel of John, for example) have followed based on that mid-second century date. 23 Palaeographic Assessment and a Proposal Don Barker has recently made a strong case that the ranges of dates assigned on the basis of palaeography to Christian literary papyri have genera lly been too narrow. 24 P.Bodmer II can serve as an especially informative example of this phenomenon. It is not that some of the more securely dated parallels of the second and third centuries do not show some simi larities to the hand of P.Bodmer IL Rather, it is that other potential comparanda (especia lly potential comparanda with later dates) have not been fully explored. In the case of P.Bodmer II, one does not even have to look very far to find a reasonably close palaeographic match with a firm date. Within the Bodmer collection itself, indeed within the same cache of manuscripts purchased with P.Bodmer II, there is a more firmly datable papyrus that bears many striking resemblances to P.Bodmer II, namely P.Bodmer XX (LDAB 220465), a copy of The Apology of Phileas BrentNongbri that describes a trial held before the prefect Culcianus. 25The dates of Culcianus are reasonably secure, and the events narrated in the text cannot have taken place earlier than the first years of the fourth century. 26 Assuming P.Bodmer XX is not an autograph, the papyrus most likely dates to some point in the middle of the fourth century (though a date later in the fourth century cannot entirely be ruled out).27 P.Bodmer II and P.Bodmer XX show a number of what I would suggest are compelling similarities in spacing, letter forms, and overall appearance.28 For convenience, I provide images of a leaf from each codex placed sideby-side (see Figures 12 and 13) and enlarged images of several examples below (see Figure 14). The two hands are noticeably similar in a number of ways. The individual letters are formed in much the same manner, and there appears to be an effort to maintain an equal height and width for the individual letters in both pieces. Groups of letters are in nearly identical positions relative to one another (especially evident in the first enlarged example, the word E:itoiricri;:v). T his likeness in both the horizontal and vertical spacing between the letters is really quite remarkable. T he ligatures are also very similar (note how the middle bar of the epsilon meets iota and rho and the way the horizontal stroke of the tau moves iototheeta). BrentNongbri ::'fetl n!-reJr.l page8,linel0 page44,line6 page\7,line2 page84, line7 Figure 14: P.Bodmer XX and P.Bodmer II. By permission of the Fondation Martin Bodmer, Cologny(Geneva) The chief difference between these two samples is the presence of serifs, or blobs, at the end of certain strokes, in P.Bodmer II. These serifs have been con· sidered a characteristic feature of what Wilhelm Schubart called the "decorated style". Schubart assigned this style a date range from the last century of the Ptolemies to about 100 CE. 29 It is always good to recall , however, that such des· ignations of "styles" are modern scholarly conventions, and we should probably not think of t hese "styles" as fixed classification systems existing in antiquity. 30 The Limits of Palaeographic Dating of Literary Papyri Moreover, in a detailed and compelli ng study, Giovan na Menci has demonstrated that these "decorations" in fact appear on papyri with a much wider range of dates than Schubart had allowed. 31 Her study has brought about a recognition among palaeographers that the "decorated style" should not really be considered as a unified "style" at all, "but a single feature of several styles, spread over a period of at least four centuries from ii B.C". 32 Indeed , it seems writers were capable of simply adding these "decorations" on demand. 33 Thus, the presence of serifs in P.Bodmer Il or their absence in P.Bodmer XX should not distract us from appreciating the overall similarity between these two hands. 34 Other papyri with secure dates in the same general time period tend to confirm the appropriateness of this comparison. As an example, we can consider a letter from the archive of Aurelius Isidorus written in 298 CE that bears a strong overa ll resemblance to P.Bodmer II. Below, I place the two pieces side-by-side (see Figures 15 and 16) and offer more detailed comparisons (see Figure 17) While there are differences in the formation of individual letters (note especially the alpha and the mu) between these two pieces, the spacing, ligatures, letter shapes, and relationships among letters all combine to yield a very similar overallimpression. 35 31 32 33 34 35 found in Diocletian's edict on prices of the year 301 CE. See the edition of S. Lauffer, Diokfetians Preisedikt (Berli n 1971) 121 G. Menci, "Scritture grechelibrarie con apici ornamentali (III a.C.-11 d.C.)",Scritlura e Civiltd 3(1979)23-53 The quotation is from P. J. Parsons' discussion of the Greek documents from Nabal l:fever in ｾｩＮｔｯｶＬ＠ Th e Greek Minor Prophets Scroll from Nal,ial /fever(BlfevX/lgr), DJD 8 (Oxford 1990) See, for example, P.Cair.Zen. 59535, a school exercise with hexameter lines in an undecorated literary hand followed by a phrase in a formal hand with ornamental serifs. For a palaeographic discussion of the piece, see C.H. Roberts, Greek Literary Hands 350 B.C. -A .D. 400(0xford 1956)4,item4c Theothernoticeabledifferenceinthesample lhaveprovidedisthemoresharplyangular alpha and upsilon of P.Bodmer XX as opposed to the curved and looping alpha and upsilon of P.Bodmer II . But the variation between an angular looping letter form occurs even within ｾＺｩＧｲｵｳＮ＠ One finds, for example, a looping alpha near the end of line 8 or page 6 of Another documentary papyrus of the fourth century worth comparing to P.Bodmer II is P.Lond. 6.1920, a letter from a Greco-Coptic milieu (the dossier of the monastery of Phathor dating to the mid-330s). While its letter forms are not as rounded as those of P.Bodmer II, itsoverallappearanceisreminiscentofP.Bodmerll.Aplateisavailableforconsultationin G. Cavallo and H. Maehler, Greek Bookhands of the Early Byt.antine Period A.D. 300-800, BICS Supplement 47 (London 1987) plate 8a Figure 15: P.Cair.Isid. 2, declaration of olive trees (d imensions 21 cm w x 25 cm h). Courtesy of the Centre for the S1udy of Ancient Doc- ｾＺ､･ＮＧ＠ ｾＺ｡＼［ｬＭ｢Ａｳｯ｣ｩｴｮ＠ Internationale de Papyrologues, Figure 16: P.Bodmer II , page 9 (dimensions: 14.2 cm w x 16.2 cm h). By permission of the Fondation Marti n Bodmer, Cologny (Geneva) TheLimitsofPalaeographicDatingofLiteraryPapyri P.Cair.Isid.2 P.BodmerII page4,line3 These examples indicate that the range of possible palaeographic dates for P.Bodmer II ought to be extended to include dates in the very late third century and into the fourth century. If one also accepts as compelling some of the proposed parallels securely dated to the second century, then the range of possible palaeographic dates for P.Bodmer II is pushing two centuries. Such a wide span is perfectly reasonable, and thi s point needs to be emphasized. We should not be assigning narrow dates to literary papyri strictly on the basis of palaeography. Four kinds of evidence support this contention1. 2 36 The first type of evidence comes in the form of papyri that demonstrate at least some scribes were capable of writing in multiple different styles generally assigned to d ifferent time periods. P.Oxy. 31.2604 provides an example, in which a scribe puts on a show of skills by copying the same poetic line in different styles, twice in a narrowly spaced hand at home in the third century and once in a spacious uncia l typical of the first century. The second type of evidence is the phenomenon sometimes called "archaism". 36 The classic case is P.Oxy. 50.3529, a papyrus scrap written in a textbook example of a first century Roman hand. The editor of P.Oxy. 50.3529 noted its palaeographic affinities with the hand of P.Oxy. 2.246, a registration of livestock dated to the year 66 CE. P.Oxy. 50.3529 is, however, a copy of the Martyrdom of D ioscorus, so this writing can be no earlier than the year 307 CE. The span for this hand is therefore at least two and a ha lf centuries. Thisdesignationissomewhatunfortunate,sinceitpresumesthatthesecurelyda1ed/a1erexamplesaretheoutliers BrentNongbri 3. 4. T hird , the active working life of a scribe could be remarkably long. Revel Coles has suggested that the same scribe could be responsible for copying parts of P.Oxy. 64.4441 (315 CE) and P.Oxy. 67.4611 (363 CE), which "would result in a working life not less than 49 years". 37 Finally, similarities in hands were passed from teachers to students, so that a given hand could last through multiple generations. 38 A ll of these factors suggest that we should be very wary of assigning palaeographic dates within narrow margins (and we should certainly end t he highly dubious practice of palaeographically dating pieces "circa" a particular year). 39 A reasonable palaeographic date range for P.Bodmer lI would be mid-second to mid-fourth century. 40 Contextualizing P.Bodmer lI among the Other Bodmer Papyri The type of argument I have just made can be quite disconcerting. It is frustrating to point out that we do not know with certainty something that we fo rmerly thought we knew quite well. If palaeography thus leaves us with this wide ra nge of dates, is there any way to establish a date for P.Bodmer II with more precision? I believe that we can. In the case of P.Bodmer II, we are fortunate to have at least a few clues that point toward a somewhat firmer conclusion regarding the date of the codex. The fact that P.Bodmer XX, perhaps the closest piece of securely datable comparative palaeographic evidence for the script of P.Bodmer II, may well come from the same find is a good place to start exploring. The provenance and TheLimits ofPalaeographicDatingofLiteraryPapyri extent of the find that included P.Bodmer II is a matter of dispute . In a series of publications over the last thirty years, James M. Robinson has argued that the vast majority of the Bodmer papyri (along with other papyri now dispersed in the Chester Beatty Library and several other institutions) are to be equated with the Dishnii papers, a group of Greek and Coptic papyrus codices, rolls, and letters allegedly discovered in upper Egypt late in 1952. 41 Further, he has made the case that these Dishnii papers are the remains of the library of the Pachomian monastic order, which was founded in the fourth century. An alternative theory posits a more limited corpus (consisting of many of the Bodmer papyri and some of the papyri in the Chester Beatty collection) that is supposed to have originated in what is variously described as "a Christian school at Panopolis, which also provided elementary training in rhetoric", or "perhaps ... a school [in Panopolis] where both classical and Christian authors were read". 42 For the purposes of my own investigation, I do not need to resolve this dispute. In what follows, I will refer chiefly to the texts that (as far as I can tell) everyone agrees were part of the find that included P.Bodmer II. Even this pared down corpus is a somewhat motley assemblage. 43 Included are Greek and Coptic papyri and parchments BrentNongbri along with some bili ngua l specimens. There is classica l and Christian material (on occasion bound in the same codex). The majority of the pieces are codices, although there is some material in roll form. 44 The bulk of the pieces have been assigned dates in the fourt h and fift h centuries, although there are outliers ranging in date from the second century CE to perhaps the sixth century. 45 We ca n begi n to contextualize P.Bodmer II within this collection by first limiting the fiel d of view to the Greek Christian materials. There are seven codices with identifiably Christian material in G reek. The chart below outlines the codices and the dates assigned to them in Rodolphe Kasser's inventor y: Designation(s) P.Bodmerll P.Bodmer V, X, XI, VII, XIII, XII, XX, IX, VIII (the Bodmer "Composite" or "Miscellaneous" Codex) P.Bodmer XIV- XV P.Bodmer XXIV P.Bodmer XLV, XLVI, XLVII,XXVII P.Bodmer XXIX- XXXVIII ("Codex Visionem") P.Monts.Roca. inv. 126 178, 292,338(theBarcelona / MontserratGreekLatin"Miscellaneous" Codex) Contents Gospel of John Genesis of Mary, Corr. of Paul and the Corinthians, 11th Ode of Solomon, Jude, Melito On the Passover, liturgical hymn, Apology of Phileas, Psalms 33- 34, 1- 2 Peter Gospels of Luke and John Psalms 17- 118 Visions of Hermas, Vision of Dorotheos, hexameters Date 2nd- 3rd cent 3rd- 4th cent. 46 3rdcent.41 3rd- 4th cent. 3rd- 4th cent. 4th- 5th cent 4th cent TheLimitsofPalaeographicDatingofLiteraryPapyri If we take into account the Coptic codices in the Bodmer publications, the picture does not appreciably change, as nearly all the Coptic literary materials thus far published have been assigned dates in the fourth and fifth centuries: Designation(s) P.Bodmerlll P.BodmerVI Contents Gospel of John and Genesis Proverbs P.Bodmer XVI P.BodmerXVIII P.Bodmer XIX Exodus Deuteronomy Gospel of Matt hew and Romans P.Bodmer XXI (= P.ChesterBeattyac Joshua and Tobit Sthcent. 50 Lamentations, Epistle of Jeremiah, Baruch 4th cent 4th cent 4th- 5thcent.48 5th(- 6th)cent.' 9 4th cent 1389) P.Bodmer XXII (=Mississippi Coptic Codex II) P.Bodmer XXIII P.Bodmer XL P.Bodmer XLI ｃｲｯｳ｢ｹﾷｓ｣ｨｾ･ ｮ＠ Codex (=Mississippi Coptic Codexl) s1 49 50 Isaiah Song of Songs Acts of Paul Melito On the Passover, 2 Mace. 5:27- 7:41,1 Peter, Jonah, liturgical exhortation 5th cent. 4thcent. 52 BrentNongbri The only codex among the Bodmer papyri composed of entirely non-Christian literature, the Bodmer Menander (P.Bodmer XXV, IV, and XXVI), has been assigned by Turner to the early fourth century. 53 If one takes into consideration Robinson's expanded corpus, we could add perhaps an additional three codices with Greek Christian materials, all dated to the fourth century. 54 Robinson's expanded corpus would also add at least two Coptic codices with a similar date range. 55 What emerges from this quick survey is that most of the codices are dated to the third to the fifth centuries with a clustering in the fourth century. Thus, P.Bodmer II, as traditionally dated, would sit at the early end of the range of dates assigned to the Christian manuscripts (it would stand out all the more if one were to assume Hunger's mid - or early-second century date). In fact, of all the Bodmer papyri in any language or format, only P.Bodmer XXVIII, fragments of a papyrus roll of a Greek satyr play, has been assigned definitively to a date earlier than the third century, and Turner has quite plausibly suggested (based on physical features of the fragments) that these were torn, folded, and used as material to stiffen the covers of one of the codices. 56 At "ca. 200", then, P.Bodmer II would be the earliest item in the Bodmer hoard that was actually a part of the collection proper. Robinson, however, has argued that P.Bodmer II was not a product of the library in the way that most of the other codices were. Rather, he has claimed that P.Bodmer II and a few of the other papyri must have "entered the library as gifts from outside". Indeed, he wrote that this sort of explanation must be posited for such "early Greek New Testament texts as P.Bodmer II (P66, the Gospel of John ... ), and P.Bodmer XIV-XV (P75 , the Gospels of Luke and John ... ), where one might even think of Athanasius living in hiding with the Order while in exile as the source of such gifts". 57 Robinson claimed that these "early" Greek New Testament texts functioned as "venerated relics" at the monastery. He illustrated this claim by reference to the physical properties of P.Bodmer XIV-XV: T his "valuable old codex was rebound in late antiquity, by pasting fragmentary leaves of the quire together as cartonnage to thicken the leather cover, and by TheLimitsorPalaeographic DatingofLi1erary Papyri sewing the binding thongs through the inner margin of the quire so near the writing that the codex cou ld not be opened wide enough to be actually read. One is inclined to think that the codex had become a relic". ss P.Bodmer II has also been repa ired in a way that interferes with reading parts of its text (bands of reinforcing papyrus cover parts of the inner margins of some pages). Yet, to describe these codices as "relics" already in the fourth century on this basis seems misleading, because it is unclear at what point these repairs and rebindings took place. Jt is particularly curious that Robinson should make this argument, since if one were to assume that Robinson's claims about the ex tent of the library are correct, the latest items in the find {Chester Beatty ac. 1494 and 1495, small papyrus rolls containing copies of Horsiesios' letters 3 and 4 in Sahidic dated to the seventh century) would indicate a deposition date for the hoard in the seventh century at earliest. 59 Even if one considered the trimmed down inventory without Robinson's proposed additions, the latest material would suggest a deposition date no earlier than the late fifth or early sixth century. A codex produced in the early fourth century would have period of at least two centuries for wear and tear to take place and repair and rebinding to occur. At what point between the copying of these codices and their deposition in or after the sixth century were these pieces rebound? How often might they have been rebou nd? Clearly some attention to the construction of P.Bodmer IJ is in order. The Codicology of P.Bodmer II In his original description of the codex, Martin noted a number of feat ures of the make-up of the codex and recognized that P.Bodmer II had been rebound in antiquity. The dimensions of the codex are 14.2 cm (width) x 16.2 cm {height) lt falls in Turner's Group 9. In the set of plates published in 1962, a stay, or strip 58 59 Robinson, The Story of the Bodmer Papyri 156; see also "The Pachomian Monastic Library" 5-6. On the fragments in the cover, see M.·L Lakmann, " Papyrus Bodmer XIV- XV (P75) Neue Fragmente", MH64 (2007) 22-41 and J.M. Robinson, "Fragments from the Cartonnage of P75", H TR 101 (2008) 231- 252. On the issue of the alleged re-stitching of P.Bodmer XIVXV through the inner margin, I bel ieve Robinson may have misread the editors' introduction to P.Bodmer XXV (part of the Menander Codex), which pointed out that , li ke the Menander codex, P.Bodmer XIV- XV showed evidence of rebinding(the fra gments in the cover), no! that P.BodmerXIV- XVwas treated in exactly the same manner as the Menander codex, which was in fact rebound using the "stabbing" method through the inner margi n. See R. Kasser and C. Austin (ed.), Papyrus Bodmer XXV Minandre: La Samienne (Cologny·Geneva 1969) 16- 17. Indeed, this is the conclusion Robinson himself reaches (The Story of the Bodmer Papyri 151) Some caution is in order fo r at least two reasons. First, 1he dates of Chester Beatty ac. 1494 and 1495 are themselves based on palaeography. See T. Orlandi , "Due Rotoli Copti Papiracei da Dubl ino (Lettere di Horsiesi)", in Proceedings of the Sixteenth International Congress of Papyrology (ed. R.S. Bagnall et al., Chico, Ca. 1981) 499- 508. Second, the case that these Pachom ian materials were part of the same find as the Bodmer materials is not air-tight (for the argument, see Robinson's account in The Story of 1he Bodmer Papyri 130- 150) BrentNongbri of protective parchment, is visible along the fo ld of the central sheet in the first two quires (between pages 8- 9 and 26- 27). Apparently, some of these parchment stays have been removed, since Martin reported that such strips were present "dans tous nos cahiers". 60 Martin concluded these parchment stays, punctured with two holes for the binding strings to pass through (at B and C in the image below), were part of an ancient rebinding of the codex (see Figure 18) Figurel8:P.Bodmer ll ,page49,showingsecondarystayandpositionofholesalongthecentralfold By permission of the Fondation Martin Bodmer, Cologny (Geneva) The Li mils of Palaeographic Dalingof Literary Papyri for a discussion of the origina l binding technique, Martin relied on the expertise of Berthe van Regemorter, whose report on the codex he quoted at length. She observed that while the most recent binding of the codex used holes at Band C, the original binding actually employed two pairs of holes along the central fold in the sheets- an upper set, labeled A and Bin Figure 18 and a lower set, C a nd D.61 She noted more wear between the pair A and B and the pair C and D, but not so much between B and C. Th is pattern of holes along the fold, two on the top half of the sheet and two on the bottom half, with little wear along the fold between the upper and the lower pairs (that is, between B and C), indicates that the original binding likely consisted of a link-stitch or chain-stitch at two independent pairs of sewing stations.62 T he technique is illustrated in the drawing below (see Figure 19). Figure19:Quiresjoinedbyalink-slitchaltwoindependentpairsofsewingstations;drawingadapted from J. Vezi n, "La realisation mattrielle des manuscrils la1ins pendant le haul Moyen Age'', Codicofogica 2: Eitments pour une codicofogie comparte (ed. A. Gruys and J.P. Gumben, Leiden 1978)15- 51at37. 61 62 lnthemost recentlypublishedse1ofplates,theupper holesaremosteasilyseenonp.24, the lowerholesonp.51.Page88shows 1racesofbothsetsofholes The clearest discussion of ancient bookbindi ng, with ample illustration, is J.A. Szirmai, The Archaeology of Medieval Bookbinding (Aldershot 1999). Also helpful, though now dated, are the discreet studies gathered in B. van Regemorter, Binding Structures in the Middle Ages: A Selection of Studies (!rans. J. G reenfield, Brussels 1992). There is a growing bibliography on bookproduction in Egypt.SeetheliteraturecitedinC. Kotsifou,"Bookbi ndingandManuscript Illumination in Late Antique and Early Medieval Monastic Circles in Egypt", in Eastern Christians and their Wri11en Heritage: Manuscripts, Scribes and Context (ed. J. P. Monfer· rcr-Salaet al.,Leuven2012)213- 244 BrentNongbri According to Regemorter, this binding technique was without doubt "la couture primitive des manuscrits les plus anciens" and "probablement celle a laquelle on etait habitue au Ille siecle". 63 But such an assessment raises questions: Primitive relative to what? And why typical of the third century? By way of comparison, Regemorter mentioned only two items: In regard to the spacing of the holes through the pages for the binding, Regemorter referred to a waxed wooden tablet (British Library Add. Ms. 33797) that has been described as dating to "perhaps [the] third century A.D.", not exactly a confidence-inspiring designation for comparative dating purposes.64 In regard to the manner in which the quires were most likely attached to their cover, she referred to the manuscripts now known as Nag Hammadi codices, which are generally assigned to the middle of the fourth century at earliest. 65 Thus, the assignment of the original binding to the third century is open to some question. In fact, the binding technique of using a link-stitch at two independent pairs of sewing stations is employed in a number of the Coptic multi-quire codices from the Bodmer find. 66 The chart below outlines the data: 67 68 ｬｦｷｯ｣ＢＧｩ､ｒ｢ｐｾｴｨ＼ｂﾰﾷｮ＾Ｐ＠ The LimilS of Palaeographic Dating of Literary Papyri Name Contents P.Bodmerlll Gospel of John and Genesis Proverbs Exodus P.BodmerVI P.Bodmer XVI P.Bodmer XIX Gospel of Matthew and Romans P.Bodmer XXI Joshua and Tobit (=P.ChesterBeatty Page dimensions Date (widthxheight) 16.5cmx23.25cm 4th cent 12cmxl4.5cm 13.5cmxl6cm 12.5cmx15.5cm 12.5cmx18.5cm 4th-5th cent Sth(- 6th) cent. 67 4th- 5th cent 5th cent. ac.1389) P.Bodmer XXII (=Mississippi CopticCodexII) 68 Jeremia h, Lamentations, 12 cm x 14 cm Epistle of Jeremiah , Baruch 4th cent In addition to the shared binding technique, the similarity of format with P.Bodmer II is also noteworthy. The pages of P.Bodmer II are, like a number of these Coptic codices, relatively small and square-ish at 14.2 cm x 16.2 cm. The closest parallel to P.Bodmer II in terms of both size and format is, however, a papyrus codex of the fourth century, the Bod mer "m iscellaneous" or "Composite" codex, which contains P.Bodrner XX along with several other texts. 69 The dimensions of the different sections vary, but the most widely represented page sizes are 14.2 cm x 15.5 cm and 14.2 cm x 16 cm.70 Thus, in terms of both its size and its original binding, P.Bodmer II fits comfortably in a fourth-century context among other Bodmer papyri. The original arrangement o[ the quires of P.Bodmer II is not ent irely clear.71 As I mentioned above, it seems fair ly certain that P.Bodmer II was rebound in antiquity at least once.n The presence of the parchment stays in the central 69 70 71 72 11vangelium Sai'disch: Text der Handschr1ft PPalau Rib. Jnv.·Nr. 183 mit den Varian/ender Handschriften 813 und 814 der Chester Beatty Library und der Handschrift M 569 (Rome 1984) Sincethiscodexseemstohaveconsistedofsmallerindividualcollectionsgatheredtogether, itisnotsurprisingthatthepagesshowevidenceofmultipledifferentbindingtechniques.Autopsyinspec1ionofthecodex'spartsinGenevaandtheVaticanisneededbeforemakingany firmpronouncementsonitsbinding. SeethediscussioninTurner,TheTypologyoftheEarlyCodex79- 80 For a comparison Manin's reconstruction of the arrangement of the qui res with that of Kasser, seeTurner,TheTypologyoftheEarlyCodex70 I am unsure why Turner expresses doubt about this point in The Typology of the Early Codex "Moreover,thiscodexmayhavebeenreboundinantiquity"(60). ln1974,Turnerhadwritten with more confidence, "Clearly the codex was rebound in antiquity" ("Some Questions about the Typology of the Codex'', Akten des XJJJ. lnternationalen Papyrologenkongresses [1974] 427- 438,quotationatp.436) BrentNongbri quires (as reported by Kasser) provide some help in reconstructing the final binding of the codex. Furthermore, the partial remains of quire signatures on pages 17 and 77 provide evidence of a stage prior to the last bindi ng of the book.73 I reproduce the very scanty remains below (see Figure 20) Figure 20: P.Bodmer II , upper right corners of pages 17 and 77 showing remains of quire signatures By permission of the Fondation Martin Bodmer, Cologny (Geneva) Turner quite plausibly reconstructs t he first signature as > B. < on the basis of the quire signatures found in the Tourah codex of Origen's Dialektos and Peri pascha (Publ.Sorb.Pap. 1.683- 684; LDAB 3509), which has been dated on fairly secure grounds to the latter part of the sixth century (see Figure 21 below).74 The Limits of Pa laeographic Dating of Literary Papyri If Turner is correct, then page 17 would be the beginning of the second quire. The reconstruction of the signature on page 77 is unclear. It cou ld mark the beginning of either the fourth (delta) or the fifth (epsilon) qui re. In any event , it seems certain that the quire signatures do not correspond to the latest binding of the codex. As I mentioned ea rlier, one set of repairs to the codex in antiquity involved attaching strips of papyrus the centers of severa l sheets to strengthen the area of the central folds. Kasser has drawn attention to the fact that one of these papyrus patches joined page 59-60 to page 77- 78 (this observation can be confirmed by the corresponding horizontal fibers and on the papyrus patch near line 7 of page 60 and near line 6 on page 77). Thus , in this stage of the codex's existence, page 77 was no longer the first page of a quire. If the quire signatures, then, do not correspond to this later repair job, the question then arises: Are they to be associated with the or iginal construction of the codex, or were they added during an intermediate rebinding? Turner believed that " the ink and ductus [of the quire signatures] seem compatible with an allocation to the original scribe, who in this case also added the page numbers". Thus, he associated the quire signatures with the original construction of the codex.7s Because of the difficulty (presumably) involved in writing in a quire or codex that was a lready bound, Turner believed the usual scriba l practice, and the practice of the scribe of P. Bodmer II , was to in scribe the pages "when the sheets were still detached (before the volume was stitched)".76 Jn such a scenario, keeping the pages and quires in proper sequence for binding would be a priority.77 Thus the presence of page numbers and quire signatures.78 If Turner is indeed correct that this was the usual practice (he does make excep- 75 76 77 78 thoughtcouldbeassigned 101hesixth century.Sincesomeof1hccursivewritinghaslet1er formsnearlyiden1icaltothemoreformaluncialof lhemanuscriptitself, Schererat1ributed thiscursivehandtothecopyistof1hemanuscript Turner, The Typology of the Early Codex 77. Thal the quire signatures appear to have been intentionally !rimmed off is noteworthy, but it is impossible to tell at what s1age th is trimming took place Turne r,TheTypologyoftheEarlyCodex74 l amunable 1o followTurn er'slogici nth isinstance.Noton ly doesthisproposedprocessof copyingunboundsheetsseemcu mbersome,italsocomplicatesTurner'solherobservation about the copying of P.Bodmer II , namely, the growlh in the size of the writing evident in the laterpagesofthecodex:"thescribebeginstorealizehehastoomuchroomandwrites larger", according to Turner (The Typology ofthe Early Codex 74). But if the scribe was copying unbound pages and had this realization , why not then simply main1ain 1he writingsize and adjustthenumberofpagesbyrcmovingsheelsfromtbefinal(asyetunbound)qui re? Therearecounter-examplesinwhichitseemsfairlyclearthattheinscribingtookplaceafter the binding of the codex. See, fo r instance, P.Chester Bcauy XIII . The editor of that manuscript notes that "one bifolio (folios 4and5) still has a hemp binding string in its top holes withpartofanoverliningstrokeonit, indicatingthatthe qui rewasboundbeforethetext was inscribed"; see A. Pietersma, 7\vo Manuscripts of the Greek Psalter in the Chester Beatty LibraryDublin(Romel978)1 BrentNongbri tion for single quire codices), it is surprising that he does not speculate on why some codices lack quire signatures (and even page numbers) As it stands, Turner's proposal raises a number of questions. Given that the quire signatures are so fragmentary, can they reasonably be assigned with any degree of confidence to the scribe of the codex (a nd here I do note Turner's appropriate caution in the phrase "seem compatible")? If they are indeed the work of the original scribe, the quire signatures on P.Bodmer II would be a remarkable feature on the traditional dating of the codex. Even if the codex were assigned Turner's preferred date (early third century), P.Bodmer II wou ld still predate the next earliest exa mple of a papyrus codex with quire signatures by a century or more.79 Two other possible explanations of the quire signatures are therefore worth considering. First, if Turner is correct that the signatures are the work of the scribe of the codex, then a date for the production of the codex in fourth century, when we begin to have evidence for the use of quire signatures, would be more sensible. If, however, the quire signatures are to be associated with a later rebinding of the codex, then Turner's comparandum for the format of the quire signatures is suggestive. The Toura h codex likely dates from the sixth century. Could the quire signatures in P.Bodmer lI have been added to facilitate a rebinding in the sixth century? Such a scenario seems unlikely, given that another substantial repair job (the reinforcing papyrus strips) would have taken place after the insertion of the quire signatures but before the deposition of P.Bodmer IL A repair, however, in the late fourt h century or in the fifth century (or both) seems more plausible. 80 When th is possibility is considered, Robinson's proposal that P.Bodmer was already a "relic" in the fourth century loses its force. Based on the evidence currently avai lable, then, I wou ld posit the construction and initial copying of P.Bodmer II took place at some point in the fourth century, and at least one subsequent rebinding (though more likely two) took place, perhaps as late as the fifth century. 81 79 80 81 Turner reports that the next oldest examples of codices with quire signatures arc BM Ms. Or. 7594(LDAB107763,aCopticcodexgenerallyassignedtothefirsthalfofthefourthcentu ry on thebasisofcu rsivepapyrireused initscoversandcursivewritingononeofitspages) and theChesterBeattyManichaeancodices(generallyassignedtothefourthorfifthcentury)(The Typology of the Early Codex 77). I am not aware of any new evidence that has come to light thatwouldchange thisassessment l nregardtomedieva\bindings, Szirmaihasobservedthat"even thesturdiestbindingwhen used extensively will inevitably wear out in 25 or 50 years" (The Archaeology of Medieval Bookbinding l 37) l havenotmentionedherethecorrectionstothecodex,whichalsohavethepotentialtoadd tothediscussionoftheusageofthecodex(especial ly thecorrectionatJohn1 3:19onpage99), butbecauseofthedifficultyofassessingthecorrectionsinthepublishedplates, l hesitateto comment on them without autopsy inspection or consultation of high-quality digital images For an informed discussion of the corrections, see Royse, Scribal Habi1s 409- 544 The Limits of Palaeographic Dating of Lite rary Papyri Before bringing this section to a close, I will briefly mention one other feature of the codex that may also be suggestive of a fourth century date. At severa l points in the fragm entary final pages of P.Bodmer II , forms of the terms o-mupOc; and 01cmpOoo are abbreviated in a manner that involves combin ing the letters tau and rho to form a monogram , f , generally referred to as a staurogram. 82 I provide an example below (see Figure 22): Figure 22: Detail of P.Bodmer II , page 137, line 3: abbreviated form of otaupco&f\. By permission of the Fonda1ion Martin Bodmer, Cologny (Geneva) Larry Hurtado and others have plausibly argued that these examples of the staurogram should be interpreted as visua l representation of the crucifixion of Jesus. 83 If this understanding is correct, then this fact wou ld point to a date for the production of this codex in the fourth century, when Christian use of the imagery of crucifixion begins to become more common.11-4 Hurtado, assuming a date of "ca. 82 83 84 Thistau-rholigature isfoundinavarietyofmediathatpredateourearliestsurvivingChristian manuscripts. It occurs in papyri such as LDAB 3551, a roll of Philodemus from Herculaneum, in which, according to Hermann Diels, it abbreviates form s of ｴｰＶＱｾＮ＠ It is found on coins, such as certain issues of Herod the Great, in which the meaning of the monogram is disputed; see D.T. Ariel and J.-P. Fontanille, The Coins of Herod: A Modern Analysis and Die Classification (Leiden201 2) 124- 126. Italso occurs ininscriptions,suchaslGXll.l.4,a listof namesfrom the Flavian period from Rhodes, in which it abbreviatestpu:111:fu;. A free-standing staurogram became a more common feature in Christian documentary papyri in the fourth and firth centuries. See M. Choat, Belief and Cult in Fourth-Century Papyri (Turnhout 2006) 11 6-1 18 Hurtado'sfullestdiscussionofthephenomenon(withamplcbibliography)istobe foundin his essay "The Staurogram in Early Christian Manuscripts: The Earliest Visual Reference to the Crucified Jesus?", in New Testament Manuscripts: Their Texts and their World (ed T. l Kraus and T. Nicklas, Leiden 2006), 207-226. While one standard handbook of early Christianiconographyseemstotakeitrorgrantedthatthestau rograminthesepapyriconstitutes "a kind of pictogram, the image ora man's head upon a cross" (RM. Jensen, Understanding Early Christian Art [New York 2000] 138), art historians are by no means in agreement ｾｴｨＺｳｮＮ＠ I am grateful to Felicity Harley-McGowan for alerting me to the complexities See G. F. Snyder, Ante Pacem: Archaeological Evidence of Church Life Before Constan tine (2nd ed., Macon, Ga. 2003) 58-64. For the possibility that a gem engraved with an image of the crucifixion may pre-date the fourth century, see J. Spier, Late Antique and Early Christian Gems (Wiesbaden 2007) 73-75 and F. Harley, "The Crucifixion" in Picturing the Bible: Th e BrentNongbri 200 CE" for P.Bodmer II, argued that the appearance of the staurogram in this manuscript, and in the Chester Beatty Gospels-Acts codex (P45) and P.Bodmer XIV- XV (P75), provided proof of Christian use of the imagery of the crucifixion in the form of the tau-rho monogram "at least as early as the final decades of the second century, and quite plausibly somewhat earlier". 85 Yet, in light of the evidence laid out above, it would seem more prudent to interpret this feature as further support for a fourth century date for P.Bodmer 11.86 At the very least, such experimentation with cruciform imagery would appear less out of place in the fourth century than in the late second or early third century. Conclusion Palaeography of literary papyri can be an extremely frustrating process for people uncomfortable with ambiguity. In the case of P.Bodmer II, reasonable palaeographic parallels with secure dates can be found from the second century into the fourth century. When, however, one considers that a very close parallel (P.Bodmer XX) can be dated securely on the basis of its contents to the fourth century, and that this piece was very likely part of the same find as P.Bodmer II, a fourth century date for the latter becomes more plausible. When one further takes into account the codicological features of P.Bodmer II, a fourth century The Limits of Palaeographic Dating of Literary Papyri date for the codex becomes even more probable. 87 Certainty in these matters wi ll likely always be elusive, but the combined weight of these considerations poi nts to a date for the production of P.Bodmer II in the early or midd le part of thefourth century. 88 Correspondence Bren1Nongbri W6A 336, Department of Ancient History Macquarie University NSW2109,Australia