"The Gestation of the Codex"
or, "From Scroll and Tablets to Codex and Beyond"

by Robert A. Kraft (spring term 2008, University of Pennsylvania; 14ap08 draft)

The following depends heavily on the previous studies by Colin Roberts ("The Codex" 1954) and Theodore C. Skeat (The Birth of the Codex, 1983), and on Joseph van Haelst "Les Origines du Codex" (1989) with several significant revisions as well as much additional material. 

Desirata Noted in Reviews of Roberts & Skeat:

Review by Eldon Jay Epp, Journal of Biblical Literature 105 (1986) 359-361 -- Mostly a report, with some emphasis on the speculative nature of the relationship to early Christianity.

Review by G. D. Kilpatrick, Vigiliae Christianae 38 (1984) 409-412 --Problem of Pauline authorship of Pastorals; early use of Mark; connection of (Greek) Roman Christianity with Latin Roman evidence; simplistic argument for nomina sacra unity; place of the epistles in canon formation. Kilpatrick himself is inclined to emphasize Christian desire to differentiate from Judaism in both codex and nomina sacra choices.

Review by Rosamond McKitterick [Oxford Journals page] Library (1985) 360-363 -- 
The codex, however, antedates the formation of the Gospel canon and is very difficult to link with it. This part of the discussion would have been enriched had the possible ownership of the surviving books, the market for the new format and the question of who would have been reading these books been considered, for the manuscripts and their proposed context seem to have significant implications for levels and use of literacy among the Christian communities of the ancient world. (363)
[also qustions the actual isolation of Egypt in bookmaking, arguing for possible imports having much effect]

Review by Peter J. Parsons, Classical Review (1987) 82-84
Offers several queries and corrections, noted  in the appropriate places in the electronic file. Also check:  "P. 13 The reference, Asc., In Mil. 29, has dropped out" [presumably was present in the earlier CHR version?].

Note by J. David Thomas in his obituary for T.C.Skeat:
Skeat's other major contirbution to palaeographical and codicological studies was the book he produced with Colin Roberts, The Birth of the Codex ( based in part on a previous article by Roberts in the Proceedings of the British Academy). Although in some respects controversial and by no means the last word on the subject, this book is fundamental for any examination of this extremely complicated problem, and has been reprinted more than once. Also of enormous value is the chapter on Early Christian Book-Production [[1969, reprinted in Elliott]], which he contributed to the Cambridge History of the Bible (an Italian translation appeared as a separate volume in 1976).

Contents:

Tablets and Codexbooks [vocabulary? gk δέλτοι? δελτάριον, γραμματεῖον -- Lat tabulae, pugillares, codices]

For persons in almost every walk of life in the Greco-Roman world, as well as long before and long after, codices composed of rigid tablets of wood and/or other materials were commonplace. Individual tablets were bound together on one edge with some sort of hinging (e.g. leather strips, cord), so as to open either vertically (hinged at the top or bottom) or horizontally (hinged at the side). The hinged tablets could also be arranged accordian (or concertina) style, to emulate more closely the equally common scroll format. [[What constituted "front" or back??]] The tablets themselves often were coated with wax in a framed and slightly hollowed out area and thus could be reused when older markings were smoothed over with a flat ended tool. Sometimes wooden tablets were painted white or a light shade (or even reddish), and writing in ink or chalk was sometimes placed directly on the painted or unpainted wood. [[e.g. Vindolanda]]

The tablet was one of the oldest, if not the oldest, recipient of writing known to the Greeks, who may have borrowed it from their neighbors to the east.\1/  Homer knew of it, for it was on a folded tablet or diptych [["two piecer"]] that Proitos scratched the 'deadly marks' that were intended to send Bellerophon to his death; in Sophocles, Agamemnon orders the muster roll [list] of the Greek princes to be read from a tablet, and it is on a tablet that Zeus, in a fragment of Euripides, records the sins of men.\2/ In later Greece tablets were the familiar recipient of anything of an impermanent or informal nature -- letters, bills, accounts, school exercises, memoranda, writers' drafts. Already by the fifth century bce tablets of two or more  leaves were in use, but the thickness and rigidity of the material limited the number of leaves that could be included, and in fact no specimen surviving from Greco-Roman antiquity has more than ten. [[add rabbinic reference to 12?]] The earliest surviving Greek tablets, seven in number, date from the middle of the third century bce.  Both sides of each leaf were covered with wax, sometimes black, sometimes red; they are hinged horizontally and contain rough accounts of expenses during a journey on the Nile.\3/ In Rome tablet codices were equally familiar from an early date and were employed not only for the casual purposes of everyday life but for legal documents and official certificates [[e.g. debts and birth records]].\4/

\1/ The Mycenaean Greeks used clay tablets and also, possibly, papyrus (cf. clay sealings containing impressions of papyrus fibres, Marinates, Minos, i, p. 40; Maurice Pope, "The Cretulae and the Linear A Accounting System," Annual of the British School at Athens 55 [1960] 201). For other Near Eastern evidence, see C. Wendel, Die griechisch-römische Buchbeschreibung verglichen mit der des Vorderen Orients, 1949, p. 91; L. Koep, Das himmlische Buch in Antike und Christentum,1952, pp. 15-16.  Especially interesting is the set of ivory tablets from Nimrod, dated to about 707-705 bce, which still retained some of their yellow wax coating, and had originally been hinged together on both sides so as to fold up concertina-fashion, whereas the tablets of walnut wood found with them had perforations so that they could have been hinged on one side only by, e.g. leather thongs (Iraq 16 (1954) 65, 97-9; 17 (1955) 3 20. For representations of wooden writing-tablets in Neo-Hittite reliefs of the same period see B. van Regemorter, 'Le codex relié à 1'époque néo-Hittite,' Scriptorium 12  (1958) 177-181 and J. A. Szirmai, "Wooden Writing Tablets and the Birth of the Codex," Gazette du livre me/die/val 17 (1990) 31-31 (see also Tov, Textual Criticism of the Hebrew Bible 209). An Aramaic letter on two folded wood slats was also found among the Bar Kokhba letters from ca 135 ce Palestine, similar in format to some of the Vindolanda wooden tablet letters from slightly earlier in Britain (see Y.Yadin's report in IEJ 11 [1961] 40ff, and the online Vindolanda materials by A.Bowman and D.Thomas).

\2/ Homer, Iliad 6.168 sq. (Proitos sent Bellerophon to Lycia,  /  with a lethal message, coded symbols / inscribed on a folded tablet.  These told / many lies about Bellerophon. The king was angered, but shrank from killing Bellerophon, so he sent him to Lycia bearing baneful signs [sêmata], written inside a folded tablet and containing much ill against the bearer.); Sophocles, fr. (Pearson) 144 (muster roll); Euripides, fr. 506 (Nauck) (Zeus records human failings). On the authority of tablets in ancient Greek thought, see Dziatzko, op. cit., p. 138, quoting a paper by Fr. Marx (not accessible to us); the gods are represented as using δέλτοι, διφθέραι, ὄστρακα, σκυτάλαι (tablets, parchments, ostraca, stick-codes), anything in fact except βίβλοι, written papyrus rolls.

\3/ Published by H. I. Bell and Flinders Petrie, Ancient Egypt 3 (1927) 65-74. For photographs of three of them see Petrie, Objects of Daily Use, pl. lix. One is reproduced here as Plate I.

\4/ For the uses to which tablets were put see Schubart, Das Buch...2, pp. 24 sqq., and notes, p. 175; the ninefold wax tablet illustrated on p. 24 must originally have had ten leaves (see Plaumann's article referred to by Schubart, p. 175). P. Fouad 74 of the fourth century C.E. refers to and describes a δελτάριον δεκάπτυχον (ten folded little tablet). [[Jewish rabbinic literature refers to a 12 leafed version -- see Lieberman, etc.]] On "fold" as applied to rigid objects such as tablets, cf. Euripides, I.T. 727, δέλτου μὲν αἵδε πολύθυροι διαπτυχαί (of a tablet, then, many folding pages). Schubart's comment (op. cit., p. 175) that πτυχή (fold) is not strictly applicable to a rigid material such as wood, and that therefore in this passage it implies a previous use of folded leather, papyrus, etc., is misconceived, since πτυχή can be used of the folds of doors. Cf. LSJ and Pollux, Onomast., ed. Bethe, i, p. 207 [= TLG 418, 2nd century CE]: καὶ Ἡρόδοτος (VII 239) μὲν λέγει `δελτίον δίπτυχον,’ οἱ δ’ Ἀττικοὶ ‘γραμματεῖον δίθυρον,’ καὶ θύρας τὰς πτύχας ἄχρι δύο, εἶτα πτύχας, καὶ τρίπτυχον καὶ πολύπτυχον (and Herodotus said "two-fold tablet," but the Attic commentators "two-paged notebook," and pages/doors the folds until two, then folds, even tri-fold and multiple-fold) [[check ET of Herodotous 7.239]].


We have evidence of such tablet codices already in Greek art of the 5th century bce, and on into the Islamic period in Egypt up to modern times. The recommended use of such tablet codices by Roman authors of orations is set out in some detail by Quintilian in the late first century ce (Institutio Oratoria 10.3.6-33), in his  advice to prospective rhetoriticians about preparing their written materials:

[6] In order to do this with the utmost care, we must frequently revise what we have just written. . . ..
[19] The condemnation which I have passed on such carelessness in writing will make it pretty clear what my views are on the luxury of dictation which is now so fashionable. . . . [22] the advantages of privacy are lost when we dictate. . . . [31] There are also certain minor details which deserve our attention, for there is nothing too minute for the student. It is best to write on wax owing to the facility which it offers for erasure, though weak sight may make it desirable to employ parchment by preference. The latter, however, although of assistance to the eye, delays the hand and interrupts the stream of thought owing to the frequency with which the pen has to be supplied with ink [32] But whichever is employed, blank pages (tabellae) must be left in which one is free to make additions at will. For lack of space at times gives rise to a reluctance to make corrections, or, at any rate, is liable to cause confusion when new matter is inserted. The wax tablets should not be overly wide; for I have known a young and over-zealous student write his compositions at undue length, because he measured them by the number of lines, a fault which persisted, in spite of frequent admonition, until his tablets (codicibus) were changed, when it disappeared. [33] Space must also be left for jotting down the thoughts which occur to the writer out of due order, that is to say, which refer to subjects other than those at hand. [4.1] . . . There is good reason for the view that erasure is quite as important a function of the pen as actual writing.

There is no reason to doubt that such tablet codices also were in use in other ancient cultural contexts such as Judaism, although no actual specimen have yet been publicized. The Bar Kokhba letter (above, n. 1) does not seem to have been hinged, but consisted rather of two wooden slats, each of which contained a column of Aramaic writing in ink, with each slat then scored and folded [were the two originally joined?], resulting in four sections (two still connected) as found. According to Yadin, the resulting letter when opened was about 17.5 by 7.5 cm, which suggests that the horizontal folding resembled more a scroll (accordian style?) than a codex. Yadin comments that "the practice of writing on wood was widespread throughout the Orient, and is often mentioned even in rabbinical literature" (41 -- no references provided).

Quintilian's contemporary Martial, also writing in Rome near the end of the first century ce, makes frequent mention of writing materials in his long list of possible gifts for the Saturnalia celebration. Rigid tablets are listed first (14.3-6): of citrus wood (pugillares citrei),  waxed five leaved (quinquiplices [cera]), of ivory (pugillares eborei), and three leaved (triplices); most later references by Martial are to parchment material, as we shall see below.

The rigid codices also make a strong showing in the older Roman legal traditions gathered by Justinian's commission in the 6th century, especially as objects containing records of debts the value of which can be passed on to heirs along with the objects. Ulpian, for example, in the early third century, is reported to debate whether the designation "books" (libri) can include materials other than papyrus and parchment, and concludes that all written objects qualify, with explicit reference to
writing on wood-slabs (in philyra), on ivory (eboreis), and on wax tablets (in ceratis codicillis)..

Probably the most extensive and longest lasting use of the rigid codex was in the "school" context.  The quest for literacy put everyone involved in contact with tablets and tablet books. Raffaela Cribiore has studied this material closely in her Columbia dissertation (1993) and subsequent publications.  In her section on "notebooks" (##379-412), she describes 22 rigid codices dating from the first to the 7th centuries, most of them waxed wood, and the largest with 10 leaves.

Scroll to Codex in Various Greco-Roman Contexts

The transition from scroll to codex in the Greco-Roman world at large followed different patterns with different types of material. On the "high literature side," the statistics provided by M-P\3 (January 2008) show the following for texts of Homer: The total throughout the period covered (500 bce through 800 ce), is 1663 items (1413 Iliad; 250 Odyssey, including some scholia and other paratextual materials and 10 ostraca), of which 161 are codices (112 papyrus and 49 parchment) and 8 are wood tablets (about 10%). For the period up to about 200 ce, of the 1149 total only 11 (01%) codices are listed (9 papyrus, 2 parchment), and from 201-300, of the 559 total, 48 (08.6%) codices (41 papyrus and 7 parchment); after 301 we find 230 total Homer texts, of which 125 (54%) are codices (80 papyrus and 45 parchment). The numbers don't add up correctly, presumably because ambiguous dating shows up twice. But that is only part of the story, since as we have seen, in the school setting in which the producers of these texts learned to write, it was quite normal to see Homeric materials written on tablets and tablet codices. Indeed, under "Homerica" of various sorts, M-P\3 lists 187 items, of which 29 are "codices" of papyrus (27) or parchment (2) and 7 are on wood (2) or waxed wood (5) -- a significantly higher percentage in codex format (almost 20%).  It is, then, somewhat surprising that Homer was so relatively slow in being displayed in the emerging new format. Martial at least could imagine Homer "in hand held parchments ... in many folds of skin" (14.184-186) in the late first century, even if early examples have not survived for us.

Perhaps the situations was significantly different for other types of literature, such as Martial's Epigrams for which we not only have the author's statement that they were available in codex form, but also the name of the Roman bookseller from whom the work could be obtained in that format in the late first century ce. M-P\3 combines "Elegy and Epigrams" as a category, with 36 examples of which only one is a codex. With regard to similar categories that can be explored rapidly:
"Oratorical" -- 95 total, with 9 codices (papyrus; only one earlier than 300 ce)
"Epic poetry, pastoral, didactic, and hymns" -- 151, with 29 codices (7 before 300 ce)
"Unidentified poetry" -- 117, with 14 codices (7 before 300 ce)

More narrative types:
"History and Geography" -- 152 total, with 14 codices (10 on papyrus; 2 papyrus earlier than 300 ce; earliest on parchment ca 100 ce)
"Romances" -- 48 total, of which only one is in codex form (plus two copies of the Jewish/Christian story of Jannes and Jambres, which we will handle separately; all three on papyrus).


One subject area in which the codex format was adopted relatively early was for calendric matters, with nearly 30 such codex fragments (all on papyrus, plus two wooden tablets) listed in M-P\3 (January 2008) as earler than ca 300 ce, and 13 more (all papyrus) in the 4th century. In the same periods, 191 papyri rolls and one parchment are listed up to 300 ce, and only 14 papyrus and no parchment in the 4th century. [detail: up to -101, 5 rolls; -100 to 001, 3 (includes1 wooden tablet; the search function finds 54 total but this includes some 1st ce and a great many undated!); between 001 and 100, 22 (no codices!!); 101-200, 94 (6 codices); 201-300, 96 (21 codices, one wooden tablet); 301-400, 28 (13 codices); after 401 only 8 (3 codices)]
M-P2021.71 Almanach, pour 136-144 (?) = P.Oxy. 61.4189 Jones. Oxyrhynchus    after 144 ?
M-P2021.83 Almanach mensuel, Jupiter, pour 14-6a = P.Oxy. 61.4199 Jones. Oxyrhynchus    II
M-P2040.02 Traité astrologique : distinction des planètes en bénéfiques et maléfiques = P.Med. inv. 160. Prov? II    avec indication de la pagination
M-P2014 Tables astronomiques ("Tables faciles") = P.Lond. 3.1278 (Brit.Libr. inv. 1278). Prov?. c. 200 (éd.; III Baccani)
M-P2021.38 Tables pour le soleil = P.Oxy. 61.4162 Jones. Oxyrhynchus    II/III
M-P2043.16 Sur les qualités des signes du zodiaque = P.Oxy. 65.4476. Oxyrhynchus    IIex./IIIin.
M-P2113 Traité de palmomancie = PSI 6.728. IIex./III (P. Degni, dans Mostra2 ; IV éd.)

M-P2021.75 POxy 61.4193 almanac 195-203 (after 203)
M-P2021.69 POxy 61.4188a almanac 201-208 (after 208)
M-P2021.87 POxy 61.4203 almanac Saturn for 215/216 (after 216)
M-P2021.78 POxy 61.4196 almanac 218-220 (after 220)
M-P2021.72 POxy 61.4190 almanac 241-243 (after 243)
M-P2021.73 POxy 61.4191 almanac 236-245 (after 245)
M-P2021.91 POxy 61.4205 five day almanac 257/258 (after 258)
M-P2021.92 POxy 61.4205a five day almanac [see 2021.91 "same hand?"]
M-P2021.39 POxy 61.4163 tables (after 259)
M-P2021.93 POxy 61.4206 five day almanac 272-274 (after 274)
M-P2021.74 POxy 61.4192 almanac 276-280 (after 280)
M-P2021.76 POxy 61.4194 almanac 272-286 (after 286)
M-P2021.94 POxy 61.4207 five day almanac 289-291 (after 291)
M-P2021.81 POxy 61.4197 perpetual almanac (3rd c)
M-P2044 POxy 3.470 horoscope (3rd c)
M-P2053 PSI 3.158 on planets (3rd c)
M-P2066.4 PSI 15.1495 (unedited) fragment (3rd c)
M-P2033.1 POxy 46.3299 planetary tables (3rd c?)

M-P2024 POxy 31.2551 kings of Egypt/astrological text (3/4th c)
M-P2039 PErl 14 astrological treatise (3/4th c)
M-P2021.77 POxy 61.4195 almanac 300-304 (after 304)
M-P2021.98 POxy 61.4211 five day almanac 306/307 (after 307)
M-P2021.12 POxy 61.4143 Ptolemaic tables (after 329)
M-P2010 PHeid 34 (after 349)
M-P2037 (4/5th) PVindob 19370+ for year 348 or 424
M-P2017.1 POtago frg (later 4th c)
M-P2021.14 POxy 61.4145 calculations (4th c)
M-P2021.47 POxy 61.4173 tables (4th c)
M-P2021.48 POxy 61.4173a tables (4th c)
M-P2021.84 POxy 61.4200 almanac Jupiter (4th c)
M-P2067 PGM 4.835-849 astrological section (4th c)

M-P2021.49 (undated) POxy 4174 Table des mouvements moyens de la lune
M-P2021.79 (undated) POxy 4196a Almanach
M-P2021.82 (undated) POxy 4198 Almanach perpetual
M-P2021.92 (undated; but see 2021.91 "same hand?") POxy 4205a Almanach for 5 days
M-P2111.02 (undated, provisional) P.Prag. inv. G IV 71 + 156 Traité de palmomancie (inédit)
M-P2762 (undated) Sept signes isolés du "commentaire" (tétrades 72-73, 78, 83-84, 88-89  P.Monts.Roca 1) = P.Ant. inv. 2. Antinoé ?
M-P2021.61 (undated) POxy 4182 Ephéméride
M-P2021.62 (undated) POxy 4183 Ephéméride
M-P2020 (5th c)
M-P2021.58 (5th c)

Another subject area in which codicess made a relatively strong  showing overall is "law."  Indeed, Elizabeth Meyer has argued that "legal tabulae ... seem the most likely model that a Christian codex could be imitating" (316-317), basing her case especially on similar dimensions. In her third appendix (334-338) she lists 37 "Roman Legal Documents on Wooden and Wooden-and-Wax Tabulae" of which 16 or 17 are birth certificates, most of which have their own special formatting for security purposes. Many of the other tablets are similarly formatted; all are dated before the end of the 3rd century ce. None of these tablets are found in the "law" section of the online M-P. Most of them are in Latin (23), or Latin and Greek (13); only one is in Greek alone.

On the other hand, M-P does list 17 papyri codices under the category "law." None are dated earlier than 3/4th CE, and most in 5th or 6th CE. Of 10 listed codices on parchment, several are dated 4th or 4/5th, only one possibly 3/4th CE. Here are the earliest listed examples of parchment (CM) and papyri (CP) codices under the heading "law":
M-P2985 = P.Berol. inv. 6757 Fragment juridique 1e moitié IV (Seider; III/IV éd.; IV/V Turner, Typ.)    CM
M-P2993.2 = CLA 10.1527 (P.Vindob. inv. L 59 + 92) Fragment légal Fayoum ?    IV (Seider; IV/V éd.)    CM
M-P2989 = P.Berol. inv. 11323 Fragment juridique (inédit) Hermopolis ?    IV/V    CM
M-P2990 = P.Berol. inv. 11324 Fragment juridique (inédit) Hermopolis ?    IV/V    CM
M-P2979.31 = BKT 9.201 (P.Berol. inv. 21595) Texte juridique ? Hermopolis ?    IV/V    CM
M-P2972 = P.Grenf. 2.107 (Bodl.Libr. inv. Lat.cl.g.1(P)) Fragment juridique : Societas IV/V (éd.; IV Seider)    CM
M-P2979 = P.Ant. 1.22 Fragment juridique Antinoé    IV    ® (¯ texte non identifié)    CM

M-P2978 = P.Amh. 2.28 (P.Pierpont Morgan inv. Pap. G 28) Fragment juridique IV (Seider; IV/V éd.)    CP
M-P2282 = P.Ryl. 3.476 Registre de constitutions impériales, en grec et en latin IV/V    CP
M-P2991 = P.Coll. Arangio-Ruiz s.n. + P.Haun. 3.45 (inv. L 1 + G 169 c-e + 172 b-c) Manuel juridique IV/V (P.Haun.; IV P.Coll.Arangio-Ruiz    CP 
M-P2281 = P.Berol. inv. 16976 + 16977 Fragment sur Longi Temporis Praescriptio (?) ou sur "prescriptiones temporis" (Schönbauer) (en grec, avec quelques mots latins) IV/V (Lowe; V ? éd.; c. 400 Seider)    CP
M-P2984 = P.Vindob. inv. L 110 Rubriques d'un ouvrage sur le droit criminel Fayoum ?    c. 400 (Seider; VI Wessely)    CP
M-P2988 = PSI 13.1346 (P.Cairo inv. SR 3796) Fragment juridique (?) Antinoé    V (Lowe et Turner, Typ.; IV ? éd.)    CP

Otherwise, the earliest of the 15 papyrus "Law" rolls are:
M-P2986 = P.Mich. 7.431 (inv. 513) from 1st CE;
M-P2987 = P.Mich. 7.456 (inv. 5604 br) + P.CtYBR inv. 1158r from perhaps 1/2nd; and
M-P2993.6 = ChLA XII 544 (P.Monac. inv. L 2r) from perhaps around 100 CE
M-P2286.1 = P.Brux. inv. 7172 is 1/2nd CE; 2983 = P.Aberd. 130 (inv. 2 c) from about 100 CE or later;
M-P2279.1 = P.Oxy. 46.3285 is dated 2nd half of 2nd CE;
Several of these 15 pieces are opisthographs or are written on reused materials. The 6 "Law" parchments rolls are dated 4th to 6th CE.

It is difficult to know what to do with this evidence. If Meyer's argument based on tablet dimensions relative to early codex dimensions is cogent (ca 1:1.33 width to height), we might expect the earliest papyri and parchment law codex fragments to fit that pattern. Unfortunately the evidence is spotty and, as has been noted, late, making any evaluation of that sort difficult. Where measurements are given in Turner's list, the law codices tend to be in the ratio 1:1.5; but more importantly, many other types of material have similar size ratios in all attested periods from the 3rd century onward. Dimensions and content do not seem to be closely enough related to be determinative.

Scroll and Codex in Jewish and Christian Contexts

Many of the earliest preserved codices contain works that later came to be canonized as Jewish and Christian scriptures as well as some related texts. This is remarkable, given the small percentage of identifiable Jewish and Christian papyri in relation to other such remains from Egypt in the first three centuries of the common era. This has caused some scholars to look to Christianity as a major factor in the acceptance of the codex format in the Greco-Roman world and in any event, it calls for explanation. Almost noone has explored the possibility that the codex also gained significant popularity in some Jewish circles apart from Christianity, despite the clear evidence that most Christian ideas and practices developed in continuity with the Jewish origins of the Jesus movements. In what follows, an attempt will be made to examine the remnants of Jewish and Christian writings without identifying as "Christian" materials that do not contain any  identifiable Christian indicators.

The survival of horizontal scroll formats among Christians also deserves note. Around the year 400, Jerome comments on the use of scrolls for collections of his letters [R&S 24 n.68], and the "Deeds of Zenophilus" [c 395] regarding Diocletian's attempt to collect and destroy Christian books in the early 4th century includes references such as "when they came to the house of Proiectus [in Cirta, Numedia] he brought out five big and two little books. Victor the schoolmaster brought out two books, and four books of five volumes each." It is likely that at least the mention of multi volumed "books" referred to scrolls. [William Stearns Davis, ed., Readings in Ancient History: Illustrative Extracts from the Sources, 2 Vols. (Boston: Allyn and Bacon, 1912-13), Vol. II: Rome and the West,  289-290; online Paul Halsall's site]. Fragments of Jewish and Christian scrolls from the end of the first century  and later (as well as earlier Jewish ones -- of pre-Christian date) also have been preserved: \n/
POxy4443 of Esther E + 8-9 (1st/2nd ce, papyrus roll; paragraphing, ekthesis, spacing),
PFouad 203 prayer/amulet? (1st/2nd ce, papyrus roll) [no image yet] [vh911]
P. Oxy. 3.405. Irenaeus, Adversus Haereses. Van Haelst 671; papyrus roll (7 fragments), ca 200 ce. [LDAB 2459]
P. Oxy. 10.1228v. John 15-16. Van Haelst 459. papyrus roll (|, -- blank) 3rd c. [LDAB 2779]
P. Oxy. 9.1166. Genesis 16. Van Haelst 14. papyrus roll (--, | blank) 3rd c. [LDAB 3114]
P. Oxy. 8.1075 & 1079.  Exodus 40.26-32 (--; nom sac KS), 3rd ce; other side (|) has Apocalypse 1.4-7 in a different hand .Van Haelst 44, 559; 3/4th ce. [LDAB 3477 & 2786]
Stud. Pal. 11.114. Psalms 68/69, 80/81 (Symmachus ?). Van Haelst 167. parchment roll (--, | blank) 3/4 c. [LDAB 3492; nom sac tetragrammaton]
++ Leiden, Private collection Scherling 126 + Cairo, IFAO Copte 379. Coptic Ascension of Isaiah. 3/4th c. [LDAB 107888]
P. Lit. Lond. 207. Psalms 11-14, papyrus roll (--, 3/4 ce); other side (|) has Isocrates [M-P 1245]. Van Haelst 109. 3/4th c. [LDAB 3473]
P. Alex. Inv. 203. Isaiah 48. Van Haelst 300. papyrus roll (--, | blank) 3/4th c. [LDAB 3127; nom sac KS]
P. Lit. Lond. 211. Daniel 1 (Theodotion) [reused in cover of a Sahidic codex]. Van Haelst 319. early 4th c. parchment roll (other side blank)  [LDAB 3493]
P. Oxy. 10.1225. Leviticus 16. Van Haelst 48. papyrus roll (--, | blank) 4th c. [LDAB 3185]
P. Harr. 31. Psalms 43. Van Haelst 148.papyrus roll (--, | blank) 4/5th c.  [LDAB 3198]
Stud. Pal. 15.234. Psalms 9. [roll or sheet?] Van Haelst 104. papyrus (--, | blank) 5/6th c. or later [LDAB 3295]


Van Haelst catalog evidence ala R&S 42ff: "Apocrypha" -- both manuscripts of the Gospel of Thomas are rolls; so too is the so-called Fayum Gospel and an Oxyrhynchus fragment dated to about C.E. 200 and now plausibly assigned to the [[subsequently??]] banned Gospel of Peter, while a roll is also the format of the only surviving manuscript of the Greek Diatessaron.
Shepherd of Hermas roll and oposthograph; of early patristic texts 6, including 3 manuscripts of Irenaeus, are rolls
In view of the persistent use of the roll in the liturgy of the Eastern Church (see below, p. 51, n. 6[[??]]) it is not surprising that 6 of the 11 texts in the Liturgical section are on rolls.
In the miscellaneous section we have 16 rolls and 21 codices; the rolls (if we ignore 2 the nature of which is quite uncertain) are all treatises or homilies (only one is opisthograph), and their significant proportion testifies to the maintenance of the literary tradition.
[Philo (4) and Josephus (1) are codices,  no scrolls]
\n/ The following are reused scrolls on which the Jewish or Christian material is secondary:
P. Mich. 130. Hermas, Shepherd.  Van Haelst 657. (written on verso of a 2nd c. land register). [LDAB 1096]
P.S.I. 8.921v. Psalms 77 [on other side of register of Arsinoite diagrafai from 143-144 = PSI 8.921r]. Van Haelst 174. 2nd/3rd c. [LDAB 3088]
++ Cairo, IFAO P. 237 b. Revelation 1.13-20. Back of a used roll. 2/3rd c. [LDAB 2776]
P. Oxy. 4.657 + P.S.I. 12.1292. [P\13] Hebrews 2-5, 10-12; on back of P.Oxy. 668 Livy epitome (--, 3rd c) = LDAB 2574. Van Haelst 537. papyrus roll (|, 3/4th c.) [LDAB 3018]
P. Lond. Inv. 2584 [10825]. Hosea-Amos Greek-Coptic glossary (|) on other side of a papyrus scroll with a land register (--, c 200 ce]. Van Haelst 286. 3/4th c. [LDAB 3141]
P. Lips. Inv. 39. Psalms 30-55 (almost complete in 35 cols., | of roll; on other side (--) of accounts from up to 338 ce = "P. Lips. 1.97"]. Van Haelst 133. paprus roll, 4th c. [LDAB 3168