^{page 146 note 1} Thus (rather than New Light) according to S. H. Taqizadeh. The Sogdian word (“New God”) could refer to either.

^{page 146 note 2} ratu = 10 seconds.

^{page 148 note 1} Similar tables existed also in Uyγur Turkish. Two fragments were published by Rachmati, G. R., Türkische Turfan-Texte, vii, Nos. 8Google Scholar and 9 (see also my notes apud Rachmati, , p. 61Google Scholar). The dates are fixed by reference to the Sogdian and the Chinese calendars: this makes the tables somewhat confused. They cannot be understood without the help of the Sogdian MSS.

^{page 149 note 1} Handwriting of the later Manichæan type.—[restored], (doubtful) or (damaged) letters, see BSOAS., xi, 56.

^{page 149 note 2} Written in another hand.

^{page 149 note 3} Restore: [x + bwnmrg 'yg xwrxš]yd ?

^{page 149 note 4} Restore: [x pnjwg o nwgrwc iiii š] mbyd ?

^{page 149 note 5} Possibly (rwc).

^{page 149 note 6} Read and restore: (m)š(y)r(w)c [pncšmbyd] ?

^{page 149 note 7} End of the “bunmarag” (basic number ?) of the moon.

^{page 149 note 8} Name of the second day of each Sogdian month.

^{page 150 note 1} The months (whose names are not preserved in the MS.) are found by calculation.

^{page 150 note 2} Here begins the preamble to the description of the year which followed the year dealt with in lines 1–20. The purport of the number 413 which was written by a different scribe is not clear. It may refer to an era (era of Šād-Ōrmizd ?).

^{page 150 note 3} I thought at first that this number represented the bunmarag of the sun, but it could hardly precede the words bwnmrg 'yg xwrxšyd for which space is available only at the beginning of line 23. One thus has to assume that 388 is the year of an era whose name unfortunately is lost.

^{page 150 note 4} Restore: [name of an era + the basic number of the sun is].

^{page 150 note 5} = bunmarag.

^{page 150 note 6} Restore: [several fifths. Nōgrōč on Wednes]day.

^{page 150 note 7} Restoration doubtful. Miši-rōč (reading very uncertain), the sixteenth day, would be the same weekday as Xumna-rōč. The purpose of giving the weekdays of the second and sixteenth days of the year is not clear. It should have sufficed to give the weekday of the first day (Nōgrōč). But those days may have been feastdays.

^{page 150 note 8} Here began the list of the “New Gods” for the year described in lines 21–4. The possible numbers of ratus are 00, 72, 144, 216, and 288; hence 88 should be restored to 288. The whole line could be restored as follows: [n'wsrδyc iv byynwy o mnspnd rwcyy o iv šmbyd 'xšpy'h o vimyk 'jmny'h cc]lxxxviii rtw= [The “New God” of month Nausarδīč, the first day of which is a Wednesday, on the day of Manspand = 29th day, a Wednesday, in the sixth hour of the night,] after 288 ratus (on the 28th, 11.48 p.m.). From this the number of ratus could easily be restored in the table on the recto page, e.g. in line 8 it should be “fifth hour 72 ratus (passed)”. A slight difficulty is provided by the absence of a number of ratus at the end of line 14. One would expect: 144 ratus. There are two possibilities: either the number was put at the beginning of line 15 (contrary to the scribe's normal procedure), or the year dealt with in lines 21 sqq. is not the year which followed the year described in lines 1–20.

^{page 152 note 1} No other restoration seems to fit the existing traces. I am assuming that the author of these tables made a slight mistake in calculating. He ought to have written: “Thursday and Friday.”

^{page 152 note 2} Here begins the treatment of the year which followed immediately upon the year dealt with in lines 1–9. This is confirmed by the dates given for the lunar phases. For example, in the first year the New Light fell on the 14th or 15th of the 9th month (line 6), while in the second year it fell on the 4th of the ninth month (line 23). This is the correct interval for successive years.

^{page 152 note 3} The number of ratus is preserved only for the 9th and 10th months, but can be found by calculation for the other months. See above, p. 150, n. 8. The lunations are 29 d. 12 h. 144 ratus (= 24 min.) and 29 d. 12 h. in strict alternation. “9th hour 288 ratus passed” means 8.48. “Wednesday night” is the night before Wednesday. Hence “Wednesday night 9th hour 288 ratus passed” means: Wednesday, 2.48 a.m. Note that “eighth hour” (without ratus) means the full hour = 8 o'clock.

^{page 153 note 1} The manuscript has βγyy nwyy “New God” four times, and m'x nwyy “New Moon” five times.

^{page 153 note 2} “δyn', “δyn'ẖ, also spelt (historically) “δyng, is borrowed from Persian (or late Middle Persian) as are the other names of the weekdays. Cf. Chavannes-Pelliot, , Traité Man., 198 [174] sqq., BBB., 85 sqGoogle Scholar. The use by Manichæans of a form corresponding with Persian āδīne does not favour the view that āδîṇe was a purely Muslimic term. More likely it was pre-Muslimic.

^{page 154 note 1} With restorations (in square brackets) for the missing portion of the table.

^{page 154 note 2} The text presumably continued: “on the first and second days, a Sunday and a Monday.” The interval between Full Moon and New Light is seventeen days here (sixteen in Nos. 1 and 2). Actually, the “New God” took place in the night following the 30th day of the 9th month.

^{page 154 note 3} Restore: [nm]'ny ?

^{page 155 note 1} Presumably = Uyγur “čaidan” in the Khwastwanift, according to a suggestion made by me in BBB., p. 9 = Bema. The Bema Monday was probably the Monday which was closest to the day of the Bema, preferably the Monday preceding Bema (Mani died on a Monday).

^{page 155 note 2} Viz. of a Babylonian month, when the moon was full.

^{page 155 note 3} Its table of contents may help in establishing the beneficiaries of the Yimkiprayers. After the Manichaean gospels we find mentioned texts relating to (1) Ohrmizd, (2) Mār Sīsin, (3) Jesus, (4) the community of the Electi. For the first two this agrees with the order of the fasts as established in this article. It is indeed likely that Jesus was included in the number of the great martyrs. If we assume that the “community of the Electi” has replaced the “Three Presbyters”, we would gain the following order of the Yimki fasts: (1) Primus Homo, (2) Mār Sīsin, (3) Jesus, (4) Three Presbyters, (5) Mani (together seven Yimkis and five fasts).

^{page 159 note 1} Was the difference of one day as to the beginning of fasts (bāchāg) between two parties of Bakhshis in the Mongol period attested by Nasīr ad-dīn Tūsī (see my excursus in BSOS., xi, 1, p. 48), a trace of a dispute between two Manichsean parties of whom one, the conservatives, kept to the orthodox Babylonian reckoning as regards the times of the fasts and the other used the Turco-Chinese reckoning ?

^{page 160 note 1} See remark (3) of the above-mentioned excursus (BSOS., xi, 45–6).

^{page 162 note 1} [This difficulty can perhaps be resolved by assuming that the Sogdian calendars were for the use of monks for whom the rules of the Čaχšapat month brought no change.—W. B. H.]