^{page 212 note 1} In v-onja v- is prosthetic, like the unwritten ω- in Eng. one: ο comes either from ο or α.

^{page 213 note 1} Germ, welle and its cognates, gathered by Boisacq s.v. είλúω, are best ascribed to the root [S]WEL. The s well of the sea is living English to-day. With salum, to which σ⋯λος probably is cognate, the I.E. word for ‘salt’ is probably to be connected. The Greek, Latin, and Celtic forms have al from ll between vowels. In the other tongues al is from OL, and the cognates are graded differently because they are differently suffixed (ςṓl-d-). It has not been observed before that Skr. κςār⋯-ς ‘salty’ is cognate with Lat. sāl, so that the root has to be written KSWEL, and the choice is open whether to define ‘salt’ from ‘sea’ (salum), or to define as ‘burning, caustic,’ as the lexica define κςαr⋯ς (deriuationis causa). As regards the phonetics of the initial group KSW, which alternated with SKW (see examples in § 16), the derived tongues exhibit any two (but w/u may not precede), or any one (see Walde, s.vv. uibro dissipo; Persson, P., Beitr. p. 997Google Scholar , in medio; Fay, JAPA. XLIV. 108 § 3; and, with particular bearing on the present article, Zupitza.in BB. XXV. 93). Under these conditions the sept of κόλπος (see Boisacq): Germ, wölbung will come from (S)KWEL-P. O.Pruss. klup-stis ‘knee’ will first have meant ‘swelling,’ and Baltic cognates meaning ‘kneel, stumble, trot’ (κ⋯λπη) will be denominative thereto (pace Solmsen ap. Boisacq l.c.). Here also will belong σκ⋯λος ‘thigh, dick-bein’ (in Homer ‘ham’), named from its swelling, and Lat. scelus will have meant ‘insolence’ (cf. Walde s.v. insolesco) to start with, as in fact Plautine scelestus means ‘insolent’ Also κ⋯λлις ‘urna’ and its cognates will have been named from an original bellying shape.

^{page 214 note 1} Prellwitz properly derives κvτος ‘anything that contains or covers’ (Liddell and Scott) from the root κν-κν ‘tumere,’ in κν⋯ω and Boisacq properly correlates κvτος and έγ–ντí ‘usque ad cutem’ with Lat. cutio. On the κ-κ alternation see Hirt in BB. XXIV. 218 sq.; Weidemann, Lit. Gram. § 43; Bartholomae in Gr. Iran. Phil. I. § 54; also Boisacq's article on σχáζω.

^{page 214 note 2} Thanks to the misfortune of having read é instead of ē in Lith. ςνéςtας ‘butter’ I denied its connection with Av. χśνíd ‘milk,’ l.c. § 5.

^{page 215 note 1} This semantic observation leads to a simple etymology for Skr. κѕ⋯rαti ‘fundit, fluit’ and we finally ‘euanescit,’ a secondary sense development found also in the sept of χέω (cf. Lettic cognates in Prellwitz, s.v.). Thus κѕ⋯rαti may but come from the root KS[W]EL ‘tumere,’ beside which stood a root (K)SᾹU;tumere’ (see § 16), whence Skr. κѕó–dαs– ‘flood, swell’: Av. χšαο–dαh-. By a sort of blending of Av. χšαν ‘tumere’ with zav ‘fundere, fluere’ (in zao-tor- ‘pouring-priest’) we might expect a something like yžαν (yž=χšs sonantized). Now in the Iranian tongues neither χšαν nor zav nor yžαν are found in verb forms, but in place of χѕαγ Skr κsárαti we do find yžαγ ‘fundere, fluere’ As I shall show elsewhere the explanation of κs⋯rαti as cognate with φθε£ρω is mere bravura of algebraic phonetics.

^{page 216 note 1} In Av. χšᾱ-ν vαnt- ‘splendidus’ the prius is an instrumental from KSƏ-T- ‘splendor,’ reduced— perhaps in composition, say with hu- ‘bene’ or duš– ‘male’—to the absolute null grade.