Buried History 2017 - Volume 53, 35-44 Scott D. Charlesworth 35 A reused roll or a ‘curious Christian codex’? Reconsidering British Library Papyrus 2053 (P.Oxy. 8.1075 + P.Oxy. 8.1079) Scott D. Charlesworth Abstract: Recently, Brent Nongbri has proposed that British Library Papyrus 2053 came from a codex and not a roll. His primary concern is codicology and he pays no attention to scribal tendencies, including the implications of the palaeographical characteristics of the hand. In a careful reassessment that takes into consideration codicology, palaeography, scribal tendencies, and the physical condition of the papyrus itself, Nongbri’s argument is found to be lawed in a number of ways which speak directly to the possible origins of BL Pap. 2053. All indications are that a third-century Christian used the back of a roll containing Exodus to produce a copy of Revelation for ‘private’ use. Introduction In a recent article, Brent Nongbri (2013) argues against Arthur S. Hunt (1911: 5-6, 13-4) that British Library Papyrus 2053 (BL Pap. 2053) 1 is from a codex and not a reused roll. The papyrus preserves the end of Exodus (P.Oxy. 8.1075 [Rahlfs 909]) 2 on its ‘recto’ (ĺ) and the beginning of Revelation (P.Oxy. 8.1079 [ 18 ]) 3 on its ‘verso’ (Ļ) (Figures 1 Τ 2). Because confusion can result from the diferent traditional and papyrological meanings of recto and verso, the symbols ĺ and Ļ respectively are used in what follows as indicators of the direction of the ibres on each side of the papyrus (cf. Turner 1978: 8–25, 54–60, 63–5). Scholars are divided on the merits of Nongbri’s proposal: van Minnen (2013: 245), Blumell and Wayment (2015: 91), and Mugridge (2016: 175-76 [no. 36], 278-79 [no. 255]) have or appear to have rejected it, while Bazzana (2016: 16-7) and Cate (2016: 42 n. 36) are more positive. Nongbri’s argument has several strands. (1) The amount of and format of the text are not inappropriate for a codex leaf. (2) There is now evidence for Christian codices containing an eclectic combination of texts copied by diferent scribes. One such codex, the Bodmer ‘Miscel- laneous’ or ‘Composite’ codex, contains a comparable page because, like BL Pap. 1053, the new text that begins overleaf encroaches on the inner margin of the codex page. (3) When the backs of rolls were reused, the roll was often rotated so that the text on the Ļ was upside down relative to the irst text written on the ĺ, but this is not the case with BL Pap. 2053. This essay challenges these claims, as well as the overarching contention that ‘there is nothing about the physical characteristics of Pap. 2053 that would deinitely oppose its identiication as a leaf of a codex’ (Nongbri 2013: 78). 1. The ‘Codicology’ of the leaf The place to begin is the size of the hypothetical codex leaf (Figures 1 and 2). Reconstruction of the text block is best done by using the text of P.Oxy. 8.1079 (Rev. 1:4-7c) because, after leaving aside possible but unlikely variants, the number of letters missing from the beginning of Revelation (1:1-3) can be quantiied. The reconstructed text block has an average of 23.82 letters per line. Based on the Nestle-Aland Novum Testamentum Graece text (Aland et al. 2012) and allowing for nomina sacra, 4 there are 282 letters missing, which means that there were c. 12 lines (the exact igure is 11.84) on the top part of the papyrus. Thus, there was a total of c. 29 lines, or c. 30 lines if there was a title, on the Ļ. 5 Working from high quality digital images of the papyrus, the height of the text block would thus have been about 23 cm for 29/30 lines. Likewise, based on ll. 5-10 where the text is almost fully preserved, the width of the text block was c. 9.3 cm. 6 The extant bottom margin is 1.6 cm at its longest point. However, because no part of the bottom edge of P.Oxy. 8.1079 is horizontal, it cannot be assumed, as Nongbri does, that 1.6 cm was the full extent of the bottom margin. The location of the title of Exodos at the bottom of P.Oxy. 8.1075, the letters of which sit on a notional line 1 cm below the notional line on which the last line (l. 17) of P.Oxy. 8.1079 sits, supports this inference. It is likely that the margin extended another 1-2 cm below the ǖǟǠǕǦǣ colophon. Using Eric Turner’s (1977: 25) rule of thumb that the proportion of upper to lower margins is gener- ally 2:3 respectively, a 1 cm upper margin should have a 1.5 cm lower margin (total 2.5 cm), and a 1.6 cm upper margin should have a 2.4 cm lower margin (total 4 cm). Therefore, in the absence of physical evidence 2.5 cm, the hypothetical lower limit, and 4 cm, the hypothetical upper limit, should be added to the above estimate of column height. Thus, the page height was about 25.5-27 cm (cf. 26.38-27.16 cm, Nongbri 2013: 80). No side margins are preserved on the Ļ, so Nongbri turns to P.Oxy. 8.1075 to reconstruct the width of the hypo- thetical codex page (2013: 80). His justiication for this approach is that ‘a left margin is preserved’ on that side of the papyrus. But the dimensions of the hypothetical codex page might just as easily be estimated by adding 3 cm (as another general rule of thumb) to the reconstructed size