7, 8 a, b. (7.) Nor a robber our horses, O excellent one; nor the sorceresses our men.
By the most distant roads let the thief, the robber, run; by a distant one let the toothed rope, by a distant one let the malignant hasten (ṛṣ).
As to the division, see under the preceding verse. In the fourth pāda, part of the mss. accent taskarás. In the first, there is discordance among them as to the accent of bhadre. The 'rope with teeth' is of course the snake, as the comm. also has sense to see (rajjuvad āyataḥ sarpādiḥ). Our 8 a, b is identical with iv. 3. 2 c, d; and our 7 c, d resembles a, b of the same verse. ⌊For yātudhānyàs the comm. reads the masculine, -nās.⌋
8 c, d, 9 a, b. (8.) Do thou, O night, make the snake blind, harsh-smoked (?), headless; grind up the two jaws of the wolf; cast (ā-han) the thief into the snare.
⌊Apart from the variation in c, the verse is identical with 50. 1, below: see note thereon.⌋ All the mss. (except, by accident, one of SPP's) at the beginning have ándha; SPP. reads ádha, with one ms. and the comm.; but ádha is plainly out of place, and andham, as emended in our text (it should have been accented andhám; ⌊correct the misprint⌋), a very plausible correction. Our rendering of tṛṣṭadhūma is mechanically accurate; probably the word is corrupt; Ppp. reads the pāda andho rātri tiṣṭadhūmam. The comm. explains as ārtikārī dhūmo viṣajvālādhūmo ni- (ms. vi-) çvāsadhūmo vā yasya; the translators understand -dhūma as "breath" or "odor"; ⌊Griffith renders 'with pungent breath'⌋. In c, d the mss. have jambháyāsthéna táṁ drupadé jahi (but many of them have -bhā́-). SPP. follows them and prints jambháyās téna táṁ dr-: from this our text makes a bold departure* ⌊implying as its pada-reading jambhaya: ā́: stenám: drupadé: jahi⌋; but something had to be done to make sense; any one is invited to do better if he can. The comm. reads with the mss., and forces through a meaningless version. Ppp. has a different and corrupt text: hano vṛkasya jambhayādvāinaṁ nṛpate jahi ⌊cf. end of note to 50. 1⌋.
*⌊The assumption of an ā́ (ā́...jahi) after jambhaya is supported by nírjahyāsténa...jahi at 50. 1 c, d, below, where the collocation is almost unequivocal (see the note); for although jahyās (as given by the pada-mss.) is a good optative of hā, the combination of hā with nis is hardly Vedic, and we must there assume the division nir jahy ā ste-, the locative drupadé fitting well with ā jahi (cf. i. 11. 4; x. 8. 4 c). The rationale of the corruption here is not hard to see: the hiatus between c and d being once covered by the fusion of the final of jambhaya with the ā́ of ā́ stenám, nothing was easier than to see a form jambhayās in the first part of the combination, and then to substitute téna for the vastly less common stenám or for the meaningless tenám (which might be read out of the combination: see Prāt. ii. 40 note); the exigency of the meter occasioned by the blunder with jambhayās then made the insertion of tám easy. With the Berlin solution of the corruption, the meter is in perfect order. The interesting parallel from the Avesta, hām zanva zembayadhwem, Yasht i. 27, adduced by Geldner, KZ. xxx. 514, may here be noted.⌋
