386. In a few cases the use of a vowel as long appears to be merely due to the necessities of the meter. Such are:
α in ἀθάνατος, ἀκάματος, ἀπονέεσθαι, ἀποδίωμαι, ἀγοράασθε
ε in ἐπίτονος (Od. 12.423, ζεφυρίη (Od. 7.119)
ι in Πριαμίδης, διά (in διὰ μὲν ἀσπίδος κτλ. Il. 3.357, etc.)
υ in θυγατέρες (Il. 2.492, etc.), δυναμένοιο (Od. 1.276, etc.)
In these cases there is every reason to believe that the vowel was naturally short, and the lengthening must therefore be regarded as a necessary license, to be compared with the neglect of pοsitiοn before Σκάμανδρος, etc. (§ 370), or the synizesis of Ἀίγυπτίη and Ἱστίαια (§ 378 fin.). The diphthong οf εἰαρινός (ἔαρ), εἰρεσίη, οὐλόμενος, οὔνομα, Οὐλύμποιο, is of the same nature. The ου of πουλύς perhaps began in compounds in which it was required by the meter, as πουλυβότειρα, etc., and was extended to the simple word. It is apparently a poetical form only (but see H. W. Smyth, Vowel System, p. 98).
Similarly a short vowel between two long syllables is sometimes treated as long: as in ἠγάασθε (Od. 5.122), Ἡρακληείη (properly -κλεειη), Ὀϊκλείης (Od. 15.244). So τετράκυκλος is scanned ˉ ˉ ˉ ˘ in Od. 9.242, but ˘ ˘ ˉ ˘ in Il. 24.324.
387. The short final syllable of the vocative appears in several places as a metrically long syllable.
Il. 4.155 φίλε κασίγνητε, θάνατον κτλ.
and so 5.359. Also
Il. 19.400 Ξάνθε τε καὶ Βάλιε
Il. 21.474 νηπύτιε
Od. 3.230 Τηλέμακε
Il. 4.338 ὦ υἱὲ Πετεῶο κτλ.
Il. 18.385 ὄρσο Θέτι τανύπεπλε
Od. 24. 192 Λαέρταο πάϊ
Il. 14.357 Ποσείδαον ἐπάμυνε
So Il. 24.569, Od. 8.408, etc.
Il. 23.493 Αἶαν Ἰδομενεῦ τε
The reason may be found (as Hartel thinks[fn]Homerische Studien, i. p. 54.[/fn]) in the nature of the vocative as an interruption of the natural flow of a sentence. It is very possible, however, that the nominative ought to be read in these places : see § 164.