Mani's Šābuhragān

represent the true state of the manuscript, as the ends of lines 3–6

do not fit the preceding words by any stretch, however imaginative, of the Middle Persian language. It did not take long to note that they do, however, fit very well into the gap in the immediately preceding sheet, M 482 I,9 thus (based on Salemann's version):

This discovery sent me for the first time to the photographs of the fragments fortunately preserved in the collection of F. C. Andreas in the Göttingen State and University Library, which were readily available to me. Although these sepia prints are all sixty years old, they are still admirably clear,10 if a little lacking in contrast. My suspicion that the puzzling section of text was indeed from a piece of M 482 I—still adhering to and concealing a part of M 472 I (and naturally its own verso)—was soon strengthened. I imparted this notion to Dr. W. Sundermann of the Academy of Sciences of the GDR, Berlin, in the form of a provisional new transliteration of the whole text of the published fragments. He very kindly and with great speed not only confirmed my theory, but also provided me with further details and readings from the original manuscripts. I quote from his letter of January 3, 1978: "This discovery immediately made sense to me. At the institute, I was then able to find it confirmed in the original text. The place where a piece of M 482 was stuck to M 472 was thicker than the rest of the sheet. When one held the fragment against the light, one could see it clearly. I have now detached the piece of M 482 as best I could and put it in its correct place. As I see now, there are still text fragments stuck together in several other cases, the separation of which sometimes allows for additions. For example, a tiny piece of M 475 was stuck to M 477." And later, "the last seven lines of M 477 I V are covered up on your photo by text from the end of M 482."

Accordingly, Sundermann was able to provide me with complete new readings of the first lines of M 475 I V, M 477 I R, M 482 I R and V, and M 472 I R, and of the end of M 477 I V and M 482 I R (in the complete text below, that is, lines 73–6, 97–8, 138–44, 145–51, 162–8, 169–74, and 193–200). These I was in small part able to confirm and to a very slight extent "improve" on the basis of the Göttingen photographs, in the sense that some letters of lines 146–7, marked by Sundermann as doubtful or illegible in the manuscripts, appear more clearly in the photographs.11

9. W. B. Henning had attempted a reconstruction of lines 5–9 in "A farewell to the Khagan of the Aq-Aqatärān," BSOAS, xiv, 3, 1952, 516.

10. Boyce, Catalogue, p. XXIV.

11. Boyce, Reader, 76: "The ms. from which the fragments come is of very soft paper, and the surface is in many places rubbed, with the loss of letters or words." In this connection it