I’ll just make one last post (for now) on the Bodmer composite codex and its curious features. This post has to do with one of the leaves of the paschal sermon of Melito, P.Bodmer XIII. As was customary for the early volumes in the Papyrus Bodmer series, only a couple of pages were illustrated with photographic plates. One of these pages was a bit of a puzzle. I reproduce the photographic plate below:

Photographic plate of P.Bodmer XIII, page 36; image source: Michel Testuz, Papyrus Bodmer XIII (Cologny-Geneva: Bibliotheca Bodmeriana, 1960)

The presence of two thin threads woven through the left margin of this page is intriguing. At first glance, one might think of these as portions of the threads used for binding the codex. But there are problems with such an understanding. As the even-numbered page suggests (and the rest of the quire confirms), this page (ⲗϛ = 36) is a codicological verso. That is to say, the inner margin, the one in which we would expect to see the threads used for binding, should be on the right, not the left of this page. And the leaf consisting of pages 35-36 is the first leaf of a quire. What’s more, the pattern of holes through which this thread appears to be woven does not match that of the binding of the rest of the codex. The original editor of P.Bodmer XIII, Michel Testuz, recognized all this already in 1960 and concluded that the presence of the thread was a mystery (“Nous n’en voyons pas la signification”).

The editors of the Kirchenväter-Papyri revisited the question in the 1990s and received a more satisfying answer. They cited a letter from Rodolphe Kasser (19 August 1993), who explained as follows (please pardon my rough translation of the French):

“The folio consisting of pp. 35-36 is indisputably the first of its quaternion, and the thread we see in the picture of p. 36 is not the one that linked the bifolia of this quire (and which has disappeared, but in other contemporary codices at the Fondation Martin Bodmer, in which the thread of the original binding is preserved, we see that it is much thicker than the aforementioned ‘very thin thread’: it is rather a kind of fine string, about 1 mm thick). The pattern of breakage on the edges of this folio shows that when this codex was taken out of the ground by those who found it, this folio was placed with the wrong face up, the top of the page just where it belonged, but with p. 36 facing p. 34, and p. 35 facing p. 37 (the text clearly shows that pagination of pp. 34, 35, 36, and 37 is correct). This folio (by wearing out?) had thus been detached from its quaternion already in antiquity. Someone wanted to repair this accident somehow by securing the folio in its quire to prevent it from being lost. To avoid damaging the text, it was only possible to sew it in its margin (about 7 mm from the edge). The widest margin, which was also the outer margin, was chosen for this purpose. This forced the repairer to place the folio with the wrong face up (see above). We can assume that the beginning and the end of this very thin double string were tied respectively at the top and the bottom of the original thread with which the the quire had been sewn.”

A follow-up query and Kasser’s reply (27 August 1993) clarified what was happening with the conjoint folio, pages 49-50. According to Kasser, the break that detached the leaf containing pages 35-36 “did not occur at the central fold but about 1 cm from the central fold in the inner margin of page 35.” This would explain why we do not see the characteristic pattern of holes near the fold in the image of this leaf.

This explanation makes reasonably good sense of the available evidence. I don’t know of any other early codices that have been repaired in exactly this manner, but I have seen enough variety and creativity in making repairs that Kasser’s explanation does not strike me as at all implausible. It also provides a couple important lessons to take with us when examining more fragmentary remains of codices:

1. The presence of thread in a codex fragment doesn’t always mean you have the original binding.
2. The presence of thread in a codex fragment doesn’t always mean you have an amulet (I’m looking at you, editors of P.Oxy. 64.4406, almost certainly a repaired codex leaf and not an amulet).