syllables by a quaver on the ω. Similarly σπείους is an attempt to make the Attic σπέους fill the place of the uncontracted σπέεος, and εὐχετάασθαι is an elongated εὐχετᾶσθαι. Spelling, of course, followed pronunciation; the scribe wrote what the reciter chanted.
The historical process which these forms imply, can only have taken place when Athens looked nowhere outside herself for literary information, when there were no Ionic-speaking bards to correct the Attic bookseller. Some of them, indeed, can only have ceased to be absurd when the Koinê, the common literary language, had begun to blur the characters of the real dialects and to derive everything from the Attic standard. That is, they would date from late in the fourth century.
But to eliminate the Attic forms takes us a very little way; there is another non-Ionic element in 'Homer's' language which has been always recognised, though variously estimated, from antiquity onwards, and which seems to belong to the group of dialects spoken in Thessaly, Lesbos, and the Æolian coast of Asia including the Troad. Forms like Ἀτρείδαιο, Μουσάων, κέν for ἄν πίσυρες for τέσσαρες, intensitives in ἐρι-, adjectives in -εννος, and masses of verbal flexions are proved to be Æolic, as well as many particular words like πολυπάμμονος, Θερσίτης, ἄμυδις.
There is also another earlier set of 'false forms' neither Æolic nor Ionic, but explicable only as a mixture of the two. κεκληγῶτες is no form; it is an original Æolic κεκλήγοντες twisted as close as metre will allow it to the Ionic κεκληγότες; ἤπυτα τέττιξ, for 'singing cicada,' is the Æolic ἄπυτα brought as near as metre permits to the Ionic ἠπύτης. Most significant of all is the case of the Digamma or Vau, a W-sound, which disappeared in
