^{page 229 note 1} This, however, still remains to be proved. All we know is that the final chapters (xxxi to xxxvi), regarded as a later addition by most scholars, were written in Abbasid times; the date of ch. xxxi, a bowdlerized version of Vd., i (Pahl. tr.), is settled by Baghdād being mentioned (205¹²); confusion of Sūlīk and Sūrīk. On ch. xxix see Christensen, , Kayanides, 51–60Google Scholar.

^{page 229 note 2} For example, Vd. 2²⁵ in 68⁹ and 946^6–7 (see below Note A); Y. 57¹ (or 3²⁰, or par.) in 170^1–6; Tištr Yt. in 63; Pahl. tr. of Nyāyish 3^4–7 (pp. 29–31, ed. Dhabhar, ) in 165Google Scholar; etc. For further details see Christensen, ibid., 47 sq.

^{page 229 note 3} Zeitschriftf. Indologie u. Iran., vol. ii; supported by Reitzenstein, and Schaeder, , Studien, 6 sqq., 209, et passimGoogle Scholar.

^{page 230 note 1} The question whether or not the author of De Hebdomadibus was influenced by Oriental ideas, has no bearing upon the whole problem. “Oriental ideas” and Dāmdād Nask are not synonyms.

^{page 230 note 2} See Note B at the end of this paper.

^{page 230 note 3} See Note E at the end of this paper.

^{page 230 note 4} A = Great Bundahishn, ed. Anklesaria; W = Indian Bd. (Westergaard).

^{page 230 note 5} There is little doubt that brihēnīdan “create” and “predetermine, predestine” derives from brī “to cut”; cf. Av. taŝ-, θwar∂s-, etc., “cut” and “create”. The -h- compares with that in Man.MPers. pryhyn “loving”, whilst the shortening of palatal vowels in front of -h- is exemplified by Parthian frhyft “love” (commonly mistranslated “glory”), or by Persian ǰihān “world” (MPers. gēhān; g/ǰ as in ǰān, ǰā). The late Pazend spelling barhin-, etc., is the correct continuation of brihēn-, cf. Pers. farmāy- from MPers. framāy-; indeed, we know that in later times Parth. frhyft was pronounced farhīft. Nyberg's explanation of the verb as an ideogram (JA., 1929, i, 250 sq.) carries little conviction.

^{page 230 note 6} A recognition of their Babylonian origin might be found in a Dinkard passage (639¹¹ sqq., epitomized 435 ult.) where it is told that Zoroaster explained the proper import of the zodiacal circle to the “Wise of Babylon” (frazānagān-ī Bābelāyīgān).

^{page 230 note 7} whyk = (young) he-goat, cf. Tavadia, Šnš. 129 (ad x, 9). In a Manichæan fragm. (M 235) we have: 'ystynyd ‥‥ 'wd whyg'n 'c xwy = τ δ ρíφια Matthew 25, 33. Cf. Bakht. bīg.

^{page 231 note 1} Read: ušān ham-ba-χšišnīh pad xxvii χwardag āmārišnīg.

^{page 231 note 2} For details see Note E at the end of this paper.

^{page 231 note 3} Read: uš harwisp bundahišnān-ī gētīg māndān aviš kard hēnd(rather than mānišn), equivalent to: uš ō harw. bund.‥‥ māndān k. h.; mānd “house”, etc., as Pahl. Ps. m'nd-y, Man. MPers. mānd.

^{page 231 note 4} Restore: harw aχtar-ē awēšān < xii aχtarān >.

^{page 231 note 5} Thus TD₂ and Ind. Bd. 6,480,000 is 60 × 60 × 60 × 30, i.e. the number of tertiae partes (sixtieths of a second) contained in an arc of thirty degrees (= one aχtar). Hence, the total number of fixed stars was estimated as equal to the number of tertiae partes in a circle, or 77,760,000. The manuscript DH. wrongly has 8,480,000.

^{page 231 note 6} Read: spāhbedān spāhbed-ē abar awēšān spāhbedān gurmārd.

^{page 231 note 7} Read: pad hamzōrīh ud nērōg-dādārīh-ī awēšān aχtarān.

^{page 231 note 8} čēgōn gōvēdis the usual formula to introduce a quotation. The implied subject of gōvēd “he says” is the author of the book or tradition quoted (cf. Arab, qāla in Muslim books). When the book cited happens to be the Avesta (as in our Bundahishn passage), the subject of gōvēd is the author of the Avesta, namely Ohrmazd according to Zoroastrian teaching (cf. Dinkard, pp. 9–10). In such cases čēgōn gōvēd (often amplified: č. g. pad dēn) corresponds to the Muslimic formula: qāla (or qauluhu) ta'ālā. The reading guftēt “it has been said” (as if the author of the Avesta were unknown) which Nyberg has proposed for YMLLWNyt (JA., 1929, i, 264; Hilfsb., ii, 84), not only violates the Pahlavi grammar, but is unacceptable also for semasiological reasons.

^{page 231 note 9} See Notes D, F, G at the end of this paper.

^{page 232 note 1} The remaining portion of the second chapter is omitted in the Indian Bd.

^{page 232 note 2} Presumably two of the stars “whose names are known” mentioned before. Pārend = Av. Pārendi-? One can hardly read Pārend-ī mazdadād.

^{page 232 note 3} Read: stārag-ī 'wyd'p'nyk = a-wiyābān-īg. The latter word has survived in Persian as biyābānī, according to Aḥmed b. 'Abd-al-Jalīl Sagzi (cited by Taqizadeh, S. H., Gāh-šumārī, 335, n. 469Google Scholar) = “the fixed stars of the first to the third magnitude and the lunar mansions”. Since verbs derived from wiyābān-mean “to lead astray” (wiyābānēnīdan, etc.), it becomes clear that a-wiyābānig “not subject to being led astray” is a translation of πλανς, inerrans. The Persian astronomers naturally preferred this clear term to the ambiguous aχtar (1) fixed star, (2) constellation, (3) zodiacal sign.

^{page 232 note 4} The copyists evidently did not understand this passage. They left out two, and wrongly divided one word (n-hwstyn). Read: wuzurgīh-ī noχustīn, wuzurgīh-ī <dudīgar, wuzurgīk-ī> sidīgar. The first astronomer to classify the stars according to their “magnitudes” was Hipparchus (second century b.c.); he distinguished six magnitudes.

^{page 232 note 5} See Note B at the end of this paper.

^{page 232 note 6} These seven tethers constitute the “light” counterpart to the seven ties which connect the seven planets with the lower regions, and through which the planets exercise their influence upon terrestrial events. The inventor of this etymology of Haftōreng probably employed the word rag “vein” for these ties (haft rag “seven veins”) for which band “tie, tether” has been substituted here. In unpublished Manichæan texts MPers. rag (also Sogdian r'k) is actually in use for these invisible and indestructible connecting lines (besides words like land, of. e.g. Mir. Man., i, 196). A Sogdian passage (on the “dark” ties, from M 178): 'ty en wyspw δywtyy ky 'ty wy' 'nxrwznyy βstyt xnd wyx r'k 'ty ptβnd w'fnd 'ty ptyw'fnd, “They wove to and fro roots, veins, and connections from all the demons who were imprisoned in the zodiacal circle.” The Kephalaia, chaps. 48 and 49, contain a detailed description of these pipe-lines (Coptio lihme).

^{page 232 note 7} Read: 'YK (DH.).

^{page 233 note 1} Read: uš angārag ud wihēz padiš nēst? wihēz “to leave, or progress (in an upwards direction)” (cf. Nyberg, Mazd. Kal., 60 sq.) is often confused with wīšēb- “to shake, toss”, and with nišēb- “below”, in astrology = “dejection” (Taqizadeh, S. H., l.l., p. 336Google Scholar; “nišast” is a misspelling of nišēb).

^{page 233 note 2} On the two “spheres” see Note C at the end of this paper.

^{page 233 note 3} See Note C at the end of this paper.

^{page 233 note 4} Read: miyān zam <īg ud> spihr.

^{page 233 note 5} MSS. pwn ZK-y ms mynwg ZK, to be read: pwn ZK MY'-y mynwgyk? Cf. 37⁶Tištr ‥‥ hām āb stārtēd, mēnōgīhā ō wād abispārēd, etc.

^{page 233 note 6} Read tyštl (TD₂twhšytwl) MY' YNSBWN-yt, w'l'n w'lynyt. Cf. 63¹⁵, 67⁴. 130¹⁰, 136^1–2, 137¹, etc.

^{page 233 note 7} čč-i ktk-ms'd is evidently the same as Av. asma katō.masā Y. 17, 20, cf. Vd. 19, 4. “Stone” fits also Gr.Bd. 19¹⁴ where Nyberg (JA., 1929, i, 222, 291) offers the reading: sēnmurv! In that passage it is related that before the creation of the plants, etc., one-third of the surface of the earth was “hard as čč-d'r” (read čč-s'r = sangsār) = stony or rocky country, another was gard-āgand “filled with sand” (Nyberg: vyarθ aγvand! Cf. 136¹¹ χāk ud gard = 140¹⁴MY'-l'y = 'pl' = χāk ud gard; wād-ī gardag “sand-storm”). Another clear passage is 140¹ where čč' alternates with sng 139¹⁴. Considering that čč is (1) an ideogram, (2) the equivalent of sang “stone”, it can be hardly anything but a strongly corrupt spelling of KYP' (Frah., xvi, 3; Syr. k'p' = kēfā). We have to keep apart the čč' of Frah., viii, 1 (one of the worst lines in that book) on which Nyberg based his explanation.

^{page 234 note 1} Uncertain. The word (deceptively resembling padīχw “thriving”) recurs 44⁹ as “revolving”.

^{page 234 note 2} Hardly čahragān “spinning-wheels”? Possibly to be read č'hrkw'n = čahragwān, cf. Man. MPers. r'stw'n “circuit, circumference” (= Pahl. “r'stk'n” Gr.Bd. 210¹⁰, corrupted “rah-wīrān” Jamaspi, vii, 2, p. 49, ed. Messina? But see Pahl. Riv. Dd. 49¹¹, p. 160, ed. Dhabhar, Google Scholar).

^{page 234 note 3} A similar comparison was contained in a lost Avestic text from which a few words are quoted in the Frahang-i Oim, iv a, p. 15, ed. Reichelt: “And the smallest of those stars are like the head of a medium-sized man.”

^{page 234 note 4} The Avestic original to our passage probably merely said: “The moon is čar∂tu-masah-.” On measures see Note A at the end of this paper.

^{page 234 note 5} The present is spelt wh- in Man. MPers. texts (e.g. in M 819), i.e. weh- from Olr. wid- (= Yaghnobi wid-, Pashto wul-, etc.), cf. the h in Parachi yuh-.

^{page 234 note 6} Since the sun reappears in the same meridian about four minutes later than a star, and the moon 52·7 minutes later than the sun, we should expect the statement that the stars were swifter than the sun, and the sun swifter than the moon. However, according to the Bundahishn the lunar and solar spheres are further from the earth than the sphere of the stars so that, to keep pace with the stars, sun and moon have to travel at a considerably greater velocity to cover the greater distances of their orbits.

^{page 234 note 7} This statement is puzzling. That the “fixed” stars possess “proper motion” is a modern discovery (made by Halley in a.d. 1718), and the stars enumerated here (= lunar mansions 1, 2, 6, 8, 9, and Sirius) are not noteworthy for particularly great proper motion (except for Sirius and Procyon). Possibly the text means that the apparent absolute distance travelled by stars close to the equator is greater than that covered by stars near the poles during the same time (the angular distances measured in right ascension being equal); hence stars in proximity to the equator would appear to move quicker than others. Even so it is difficult to understand the reason for selecting only the stars enumerated here (all of which are fairly close to the equator).

^{page 235 note 1} See Note H at the end of this paper.

^{page 235 note 2} miyān-drang. On drang “period” see Zaehner, , BSOS., ix, 319, 584Google Scholar.

^{page 235 note 3} i.e. a sidereal year.

^{page 235 note 4} Read BYRH-xiii (in the place of BYRH-i iii). Sidereal months are meant here. Thirteen sidereal months (355·17 days) are about as much as twelve synodical months (354·36 days), although rather less than a sidereal year.

^{page 235 note 5} The text is corrupt. I have failed to find the correct restitution. One could read YNSBWN-yt in the place of YNSHWN-yt (or dōsānēt), and ŠBKWN-yt in the place of ŠDYTWN-yt.

^{page 235 note 6} Presumably merely different spellings of the same word.

^{page 236 note 1} The height of the average man is eight vitasti acc. to the Indian Bundahishn ch. xxvi (= Gr.Bd. 162¹ wrongly: six v.), or his own vibāzu in Gr.Bd. 189⁷.

^{page 236 note 2} The Frahang-i Oīm chapter on measures (xxvii) opens with the absurd statement that fourteen angust were a paδa. It should have been obvious that 14 is merely a copyist's error for 16. In the same chapter the vitasti is described as of 12 angust, and the frārāst (= 2 vitasti) is defined as 1½ paδa. Hence, 1 paδa= 16 angust (as it should be).

^{page 236 note 3} Possibly the pace of 2½ feet (gradus) was also known, if ZK in Frah. Oīm, xxvii a, line 4: čvaiti aētšaya, čand ZK angust, should be a blunder for the numeral sign for xl (cf. Bartholomae s.v.). At any rate, Av. gāman is always of 3 feet. An alternative expression for gāman is frabāzu = ½ vibāzu, see Bartholomae s.v.

^{page 235 note 4} Sogdian wβ”z (Dhyāna 88) renders Chinese hsün, a measure of eight ch'ih (each of ten ts'un “inches”). In the same passage, Chinese “16 ch'ih” is translated as “16 wyδ't'y”. As Weller, F., Monumenta Serica, ii, p. 394Google Scholar, rightly remarks one must conclude that the wyδ't- is the eighth part of the wδ”z = Av. vibāzu, i.e. a span. This agrees with the meaning of connected words in modern East-Iranian dialects (Shighni wiδēd “span”, etc.), cf. Morgenstierne, , IIFL., ii, 262Google Scholar. The Sogdian translators took Chin, ch'in for “span”, not “foot” (as European translators commonly do).

^{page 235 note 5} Common in Pahlavi texts, cf. e.g. Nyberg, ii, 154, s.v. nāδ (where nāy-ī paimānīg should be read = perches of correct measure, or average length). Occasionally, nāy is used in the place of ǰud-nāy (thus Gr.Bd. 189⁸).

^{page 237 note 1} As far as I know this has not been understood before.

^{page 238 note 1} The longest day is 16 hours long at the latitude of 48° 43' (obliquity of the ecliptic = 23° 42', as in the year ±0), or rather, if “day” = time of visibility of any part of the sun disc, and allowing for refraction, at lat. 47° 20'. Including twilight, the proper latitude would be 37° approx. (allowing 1 h. 24 m. for morning + evening twilight, assumed to begin and end at the sun's zenith distance of 97°); at lat. 36° the longest day + twilight = 15 h. 51 m., and 15 h. 44 m. at lat. 35°. We can perhaps say that the longest day of the Pahlavi texts is based on conditions prevailing in Northern Persia, but that its length was rounded off to be twice that of the shortest night. The shortest day was simply decreed to be of equal length with the shortest night, without having regard to actual conditions. However, the entire scheme may have been borrowed from the Babylonians (cf. e.g. Book of Enoch, chaps. 72 sqq.)

^{page 238 note 2} This explanation seems preferable to taking lypyh (Ipyh, lpy') for the ideogram for rētak (rēdag) “young child” (cf. Bailey, BSOS., vii, 70 sqq.), hence possibly = “small, minutus”; the ideogram in question was originally lpy' (rabyā).

^{page 239 note 1} Another hitherto unrecognized MPers. word of Babylonian origin is Man. MP. šwd'b “companion” (in h'm-šwš'b, BBB.; differently Bailey, , BSOS., ix, 230Google Scholar)= Akkadian šutapu (Syr., etc., šuttāpā, šuttāfā). In unpublished Parthian texts šwšmym “best man (at a wedding)” occurs, also originally Akkadian, cf. Syr. šūšbīnā. In MPers. fragments I noticed m'l'h “sailor” = Akk. malaḫu (Syr. mallāḥā).

^{page 239 note 2} The name Spithridates on which Noeldeke based his opinion, does not prove the existence of an Old Ir. word spiθra- “heaven” (anyway, spihris not “heaven”, but “sphere”, hence also “fate”). For all we know, Spithridates could mean “having white teeth, λευκóδους” (Av. dātā-, Pahl. dāt “tooth”). The first to suggest the identity of spihr with σφαȋρα was Lagarde.

^{page 240 note 1} A similar case is possibly provided by the Parthian spelling of zyncyhr “chains” (Pers. zančīr). The routine etymology (*zaina/i-čiθra-) is proved false by Sogdian zynčry'kh (P 2, 1065), in Man. script jyncry'. Bailey, , BSOS., x, 596Google Scholar, compares Saka tcaṃgalai.

^{page 240 note 2} Cf. e.g. DkM. 130 ^1,2; Antia, Paz.T., 212, apu.

^{page 240 note 3} In Sogdian this deity is even dubbed δynmzt'yzn, βγyy, βēn-mazdayazn βαγi (M 140, unpublished; for the spelling, cf. mzdyzn Cowley, , Aram. Pap., nr. 37, 6, p. 133Google Scholar). The Manichæans, it is well known, unblushingly called their own religion (MPers.) dyn-m'zdys, dēn-māzdēs.

^{page 240 note 4} The Greek word appears in Parthian as zwnws (Mir. Man., iii). The spelling is distressing, but not more startling than that of θρóυoς in Parthian: trnys (ibid., where the translation is wrong).

^{page 240 note 5} Under its common Persian designation, the galaxy is briefly referred to 60⁵. The interpetation of Yasna 9, 26, by Junker, , Aion-Vorst., p. 162Google Scholar, is unacceptable.

^{page 241 note 1} Thus the author of the Bundahishn refers the reader to the passage of the second chapter. Read cygwn <ZK-y> NPŠH npŠt, cf. 135⁶et passim.

^{page 241 note 2} blynnd (DH. blynd) is not clear. It should be the equivalent of paidāgīh (28⁵). Read barēnd?

^{page 242 note 1} i.e. wherever the reading is sufficiently clear to enable one to judge. One name (No. 15) has been left out in Freiman's manuscript, evidently by mistake (owing to the similarity of the following name). There are, however, some small differences in the form of the names, those in Beruni's book showing traits typical of Khwarezmian, those in Freiman's list having a distinctly Sogdian aspect. Several of these variations are due merely to the different age of the two lists. Thus we have: No. 12 Fr. wyšprn = Ber. χšfarn (uχšajarn), No. 9 Fr. my' = Ber. 'my, No. 17 Fr. δ'r'nt = Ber. š'rynd (δārend), etc. Noteworthy is No. 13 Fr. 'strwšk or 'štrwšk = Ber. šwšk, the latter form recurring in late Uyyur lists which otherwise give merely the Skt. names, for Skt. Višakhā (No. 14), see Rachmati, , T.T., vii, p. 55 (on 1, 18)Google Scholar. For No. 24 Freiman's list gives a shortened form, βrwyšt = Beruni Khw. frχšbyθ = Beruni Sogd. frštb'θ (all adapted from Skt. [pūna-]proṣṭhapuda). Beruni uses an abbreviation for No. 25, wbyr, in the place of wbyr-frχšbyθ = Fr. prw-βrwyšt = Beruni Sogd. pr-fršt (also shortened) = Man. Sogd. prw-frwxšpδ; wbyr = ubir is the late Khwarezmian form of the same word as Sogd. prw- (= Skt. uttara-), Av. aparam.

^{page 243 note 1} The question whether the Munich, catalogue agreed with Beruni's Sogdian list, or with his Khwarezmian list and that of Freiman's manuscript, depends solely on the acceptance of this rather doubtful restitution.

^{page 243 note 2} Of Rachmati's Uyγur fragments, some start with Kṛitikāḥ, some with Aśvinī; the Arabs began with aš-šaraṭāni = Ašvinī.

^{page 243 note 3} But Skt. No. 23 Śatabhiṣaj = Sogd. No. 22 sdmyi (? Sachau šδmšyr) = Fr. No. 22 stmyš = Khw. No. 22 sdmysyj.

^{page 243 note 4} “Kaht” = name of a constellation. The transcribed form suggests Cetus (κτoṭ and qyṭ'), but this works out only moderately well.

^{page 244 note 1} Nyberg has not seen that the passage refers to the lunar mansions.

^{page 244 note 2} Presumably misspelling.

^{page 244 note 3} = Pahl. *lhwt. It is difficult to say which form (Ihwt or lhyt?) is right.

^{page 244 note 4} = Pahl. *spwl; stlprobably incorrect.

^{page 244 note 5} Pahl. gāw = ideogr. TWR' DkM. has twy twr' (Nyberg's *dit-gāv), but twy should be cancelled (the scribe miswrote twy in the place of TWR' and corrected himself without striking out the wrong form).

^{page 245 note 1} = “goat's head”? However, the reading of 'pysr = “crown” is equally possible.

^{page 245 note 2} Pazend gōī. In Pahlavi script, gōy “ball” and yōγ “yoke” are indistinguishable. I read Yōγ because that is the name of the equivalent (21st) Sogdian and Khwarezmian lun. man. (corresponding to Skt. No. 21 Sravana = Altair). The longitude of the 22nd Pahlavi l.m. is 280°–293° 20', that of the 21st Sogdian mansion should be 282° 53'–295° 45'.

^{page 245 note 3} Based on the values for right ascension and declination (interval of 100 years) in Neugebauer's Tables (Chr. i).

^{page 245 note 4} See Taqizadeh, S. H., BSOS., ix, 133 sqqGoogle Scholar.

^{page 245 note 5} See Nallino's paper in A volume of Oriental Studies presented to E. G. Browne, and cf. Taqizadeh, , Gāh-šumārī, 316–322Google Scholar; BSOS., ix, 136 sq.

^{page 245 note 6} This conclusion, however, is necessarily based on the assumption that the Persian astronomers were able to find the point of the vernal equinox and to measure the longitudes fairly accurately, and that they exercised some care in fixing the lunar mansions.

^{page 246 note 1} The ascendant is given as Cancer 19°, the time being noon of the day of the vernal equinox. At that moment Sirius was rising. Should these data reflect actual conditions (correctly observed), it should be possible to determine the latitude of the observer, and the date of the observation. So far as I have been able to calculate the latitude would be 40° 33', the time BC. ± 330 (a point λ = 109°, β = 0° is rising, or a = 110° 37', δ = + 22° 22'; hence φ = 40° 33'; Sirius' hour angle at rising = its right ascension; position of Sirius in BC. 330: α = 75°·6, δ = – 16°·24). The result (Northern Sogdiana at the time of Alexander's invasion) is rather unsatisfactory, probably because the data are unreliable.

^{page 246 note 2} Presumably misreading of drafšag (1) “banner,” (2) “hem,” but most suitably (3) “awl” = NPers. dirafš, durōš “awl”. The names of the preceding lunar mansions are (in Pazend): No. 16 srōb, srōi, srub, No. 17 nur, nōr, No. 18 gelu, gaelu. I should like to suggest the following restorations: No. 16 srūy “claws”, No. 17 var “breast, front”, No. 18 dil “heart”, namely of Scorpio. Note that the Arabic name of No. 18 is also “heart” (qalb). [No.16, Persian du surūyi gazdum, beruni, Pers. Tafhīm, p. III; cf. Greek χηλαì τoȗ Σκορπíου, Ptolemy, , Tetrabiblos, i, 9, p. 24 (p. 50 ed. Bobbins, , 1940)Google Scholar; ibid., “grafša” = κντρον= nīš-i gazdum.]

^{page 247 note 1} [Persian pas-ravande-i Parvīn, Beruni, Pers. Tafhīm, p. 108.]