Having dealt with the Homeric derivations of Hellas[509], let us now ascend to the word, from which it is itself derived; Hellas being evidently, in the Greek tongue, the country which had been occupied by the Helli. Of the people who are so termed, either under the form beginning with the aspirate, or else under that of Σελλοὶ, we find obvious Homeric vestiges in the Hellespont, Ἑλλήσποντος; in various rivers termed Σελληείς; and in the invocation of Achilles to Jupiter, which places the Selli in the north of Thessaly, about wintry Dodona, and seems to stamp them as then still remaining a people of the rudest habits in their mountain home[510]; Δωδώνης μεδέων δυσχειμέρου· ἀμφὶ δὲ Σελλοὶ σοὶ ναίουσ’ ὑποφῆται ἀνιπτόποδες, χαμαιεῦναι. _The Σελλοὶ of Homer._ The word Ἕλλοι would appear to be not the most probable reading of the text of this invocation; for it presumes an inconvenient loading of the sentence with the double pronoun σε and σοι. But there can be no doubt whatever as to its identity with Σελλοί. Independently of philological argument, there is the strongest presumption that in this place Achilles intends to name his own national ancestry, as being the ministers of the god; who give him, as it were, the right to invoke the aid of the Pelasgic indeed, but therefore genuine and original, Jupiter of Dodona. But no circumstance seems to be better established by philological research, than that in many cases of Greek words, which now begin with the aspirate, there was one (or more than one) initial letter, and that frequently that letter was the _sigma_. Much obscurity has hung about this subject, from the fact that discovery has proceeded piecemeal, and that for a length of time the word _digamma_ was used to signify what had originally filled the void now existing in so many places of the Homeric versification. What this _digamma_ might have been was disputed; but it was, almost insensibly perhaps, assumed to be some one letter or sound only. But as inquiry has made further advances, many forms of a lost letter or letters have been discovered: and it has also been made clear that the gaps ought to be filled up variously, and not by any one uniform expedient. To take very simple examples, there can be no doubt about the identity of ἓξ, ἕπτα, ὓς, with _sex_, _septem_, _sus_: nor any doubt about the essential identity of ὕδωρ and _sudor_, ἡδὺς and _suavis_, ἑκυρὸς and _socer_: none therefore that the σ ought to be supplied, and not _f_, _w_, or _v_, in the passage φίλε ἑκυρέ[511]. While indeed a presumption arises[512] from the German words _schwieger_ and _schwäger_, that a double or even treble loss may have occurred, and that the passage may have run φίλε σϝεκυρέ. Under these circumstances, in the case before us, where we have both forms represented, there can be no hesitation as to the identity of Ἑλλοὶ and Σελλοί: the first represented in Ἕλλας, Ἕλληνες, Ἑλλήσποντος, and the Ἑλλοπία of Hesiod: the other and older