[Footnote 18: RV. iii. 33. 2.]

[Footnote 19: RV. vii. 95. 2. Here the Sarasvat[=i] can be
only the Indus.]

[Footnote 20: Pa[=n]ca-nada, Punjnud, Persian 'Punj[=a]b,'
the five streams, Vitas[=a], Asikn[=i], Ir[=a]vat[=i],
Vip[=a]ç, Çutudr[=i]. The Punjnud point is slowly moving up
stream; Vyse, JRAS. x. 323. The Sarayu may be the
Her[=i]r[=u]d, Geiger, loc. cit. p. 72.]

[Footnote 21: Muir, OST. ii. 351; Zimmer, loc. cit. p. 51
identifies the _K[=i]katas_ of RV. iii. 53. 14 with the
inhabitants of Northern Beh[=a]r. Marusthala is called
simply 'the desert.']

[Footnote 22: The earlier _áyas_, Latin _aes_, means bronze
not iron, as Zimmer has shown, loc. cit. p. 51. Pischel,
_Vedische Studien_, I, shows that elephants are mentioned
more often than was supposed (but rarely in family-books).]

[Footnote 23: Weber, _Indische Studien,_ I. p. 228;
Oldenberg, _Buddha_, pp. 399 ff., 410.]

[Footnote 24: Very lately (1893) Franke has sought to show
that the P[=a]li dialect of India is in part referable to
the western districts (Kandahar), and has made out an
interesting case for his novel theory (ZDMG. xlvii. p.
595).]