^{page 638 note 1} For a thorough study of the very intricate questions relating to old-Iranian computs, see Lewy, Hildegarde, “Le Calendrier perse”, in Orientalia, 1941, pp. 1–64.Google Scholar

^{page 638 note 2} The etymology *átr-yâiya is rejected by Henning, J.R.A.S., 1944, no. 2, on the ground that this “could appear in O. Persian only as ât(a)ryâdiya: tr becomes ss in Iranian only where a vowel follows”. Now, must the group ocolusive + ry always vocalize its r, never its y? It is true that Old-Iranian, in contrast with Sanskrit, shows a general tendency to do so, as can be judged from several passive forms in −ya: av. kiryeti (= k yati): skr. kriyáte; av. miryete (= *m yatai): skr. mriyáte. Still, there is at least one exception: in the passive of star “to strew ”, av. avâstryata, uslryamnô (*-striya-), the same treatment of r and y is exhibited as in skr. slriyate. Our Persian word can readily be explained by exactly the same treatment (r is consonant, y develops a i): *âtr-yâdiya > *âtriyâdiya. O.-P. amariyata is ambiguous, as has been seen by Bartholomæ, since it can represent either *amryala or—with the same guna-degree as in skr. smaryáte — *amaryata.

^{page 639 note 1} When the beginning of the year was shifted, Aramati fell at the end. Nyberg has rightly observed that the intercalary days, devoted to the dead, inserting themselves at the end of the year, came immediately after Aramati; and the dead give their name to the month which opens the year. Thus, Aramati seems to have its Armenian value of “abode of the dead”: sandarametk’ “hell, inferi”. But, in the primitive numeration, one can only say that Aramati ranges in the “first” function and the artâvan are placed between two terms to which they are etymologically akin: aramati and arta.

^{page 640 note 1} In writing this note, on the basis of my paper to the XXIst Congress of Orientalists in Paris, I have used suggestions or references by Messrs. Bailey, Barr, Dumézil, and Nyberg, for which I am grateful.