¹ Parts of the subject of the present article have recently been dealt with in three articles: Jules Bloch, Le traitement du groupe sanskrit sifflante + m, MSL 23.261-70 (1929) ; Helmer Smith, Quatre notes à propos de l'article précédent [i.e. J. Bloch's article], likewise MSL 23.270-3 (1929); and R. L. Turner, Indo-arica I: Sanskrit , Bull.SOS 9.41-2 (1937).

² Turner, Dict., sub bāph 434, and Bull.SOS 9.41⁵.

³ Wackernagel mentions bāṣpa-, without explanation, §§162 and 208b β.

⁴ Turner, Bull.SOS 9.41⁵, sets forth the three types without attempt at a reconciliation.— J. Bloch's interpretation of the doublets (Mar. 153), on the other hand, was on the right track.

⁵ Cf. Wackernagel §208b α.

⁶ J. Bloch, MSL 23.267; Turner, Bull.SOS 9.42.

⁷ I had independently the same idea.—The n-stem, *-puṣman-, is posited by Smith, because he attributes the change to dissimilation with following n.

⁸ Immediately before Smith, Bloch, MSL 23.206, had explained the relation of in a different way. Taking as the Old Indic base form, he assumed a change of by dissimilation with the initial. The same phenomenon he saw in Pkt. (Pischel §212), the ss-form living on in Kāsh. bras-wār ‘Thursday’ ().

But this theory does not explain the second p of . As to bahassaï, I doubt whether this is really an inherited form. It may be an adapted loan from Sanskrit. Sp was unpronounceable Middle Indic ; it had to be replaced by a geminate ; if the component perceived as dominant was s, then sp could be replaced by ss. In my opinion, therefore, a Middle Indic change or sp > ss is not yet proven.

⁹ Beng., . ch may be from ś, but is perhaps only a graph for s (Grierson, Indo-Aryan vernaculars 161).

^9a One might think that in *puṣma-, with initial labial, ṣm perhaps developed differently than in *tuṣm°, with initial dental, *tuṣm° becoming *tuṣv°, *tuhv°, but *puṣm°, by dissimilation with the initial, *puṣv°, *puṣ̣° (Kāsh.pōś) ; but Kāsh. bāha ‘steam’ (§7) shows that initial labial did not hinder the development ṣm > h.

¹⁰ By some mistake, B-R left this passage without reference.

¹¹ Cf. Bloch, L'indo-aryen 192.

¹² Kāsh. bāha, therefore, has probably nothing to do with Pkt. bāha- (§1).—Turner, Bull.SOS 9.41, who regards bāṣpa- as the Old Indic base form, still derives Panj. bhās by Bloch's dissimilation theory (cf. fn. 8), with ṣp > ṣṣ after initial labial. The later different explanation of púṣpa- by Helmer Smith he does not accept. For Kāsh. bāha, he tentatively assumes (without support) a further development of ṣṣ after long vowel into h.

¹³ Pischel 1.c. saw in -mhus- a variant with initial s, *smṛṣ.

¹⁴ Turner, Dict., sub bādai 433b, and Bull.SOS 3.313 (1921).

¹⁵ Pischel §209 erroneously separates bhibbhala-; not only could aspirated v° dialectically become bh°, but bhibbhala- is probably not from *vhibbhala-, but rather directly from ribbhala- by assimilation.

¹⁶ This word, against J. Bloch, Mar. 300 (Turner, Dict., does not discuss it), is doubtless ūṣmán- (so already Platts). The original pph still appears in the Gujarātī form; bb(h) is from pph by ‘sonorization of geminate’ as in AMāg. tubbhe ‘you’ = Asokan tuphe, cf. below §17. Deś. uvvā is probably inverse writing for *ubbā.

^16a J. Bloch, Studia Indo-Iranica 17-20 (Ehrengabe für Wilh. Geiger; 1931), thinks that between the neuter vartma and the feminine vaṭṭā there was a masculine *vartmā or *vaṭṭā, due to the influence of the synonym ádhvā.

¹⁷ From Shaw, Ghalchah Languages (Journ. As. Soc. Beng. 1876) 80.

^17a Differently Sturtevant 1.c.

¹⁸ Asoka Jau. Sep. 1.2 has a nom. phe corresponding to Dhau. Sep. 1.4 tuphe. But I should not attach much weight to this isolated form (otherwise Jau. also has tuphe), as it may easily be an error (so Woolner).

¹⁹ That AMāg. bhe is from *ṣme, not *sme (which would at first seem possible), appears from the different development of amhe (cf. §17).

²⁰ A Proto-Osco-Umbrian nom. *tē (Sturtevant, Hitt. Gram. 191) seems to me not assured by the evidence.

²¹ LSI 9¹.630.

²² E. O'Brien and J. Wilson, Glossary of the Multani Lang., Gramm. 20; just so the Lahndā of Shahpur.