As part of the EthiCodex project, our team has been revisiting the often fragmentary remains of early codices. Thankfully, many collections have made digital images of their materials widely available. But sometimes it is also necessary to see things in person.

P.Mich. inv. 22 is a fragmentary folium from a papyrus codex that contains remains of Psalm 8:3-9 and 9:7-17 (TM 61984). It was purchased in Egypt in 1920 by Bernard Grenfell and Francis Kelsey (during the same trip that they purchased the famous Rylands fragment of John’s Gospel, P52). P.Mich. inv. 22 was published by Henry Sanders in 1936 as P.Mich. 3 133.

Given that the papyrus contained portions of Psalm 8 and 9, material relatively near the beginning of the Book of Psalms, Sanders offered a fairly confident estimate of the position of this folium within the codex: “Some seven pages, i.e. three and one-half leaves, preceded the beginning of this fragment in the original codex.”

I am not sure how Sanders came to this number, but it seems a little strange based on the surviving evidence.

Because the top lines of each page are preserved, we have a reasonably good idea of the number of words on the “front” page of the papyrus (everything from the words [στόμα]τος νηπίων καὶ in Psalm 8:3 to the words ἀπώλετο τὸ in Psalm 9:7, inclusive). We can then check the number of words and characters in that stretch of text in a printed edition. In this case, we count in the Rahlfs-Hanhart edition 199 words and 1237 characters (including spaces, punctuation, etc.). From the words ἐκ στόμα[τος] in Psalm 8:3 back to the beginning of Psalm 1:1, we have in the Rahlfs-Hanhart edition 1154 words and 6925 characters (again including spaces, punctuation, etc.). From here, it is just a matter of division to determine roughly how many codex pages would be needed to contain the material on the pages preceding our folium (assuming they carried about the same amount of text as the surviving papyrus folium and assuming that we are dealing with a “normal” collection of Psalms):

1154 words ÷ 199 words per page ≈ 5.799 pages

7178 characters ÷ 1237 characters per page ≈ 5.803 pages

So, we would seem to be looking not at seven pages of preceding text but rather at about 5.8 pages worth of text. But this is just if we consider words or characters. If we also take into account some extra spacing for the titles and numbering of the first eight Psalms, then we can say that probably 6 pages (3 folia) preceded the surviving folium. That means that the “front” of P.Mich. inv. 22–that is, the side with vertical fibers (↓) that contains Psalm 8:3-9:7–would be page 7 and the back–the side containing Psalm 9:7-17 written on the horizontal fibers (→)–would be page 8.

And here is where in-person examination is helpful. In fact it turns out that in the surviving part of the upper margin of the back (→) side of our folium, there is a lightly written page number: H̅, that is to say, page 8, just as the rough calculation above suggested. I noticed the faded letter while looking at the upper margin under high magnification:

But once I knew where to look, I could see the number even in the online digital image:

Sanders seems not to have been aware of the page number (in his defense, it is possible that the ink may have become more visible in the decades since he published the papyrus). I haven’t seen it mentioned in subsequent scholarship, either–though I’m happy to be corrected if I’ve overlooked something. In any event, this papyrus offers a good example of why it is worthwhile to revisit these early codex fragments, especially those pieces that were published in the early decades of papyrology.