¹ [All Greek forms in this article have been transliterated by the Editor according to the principles set forth by André Martinet, A project of transliteration of classical Greek, Word 9.152-61 (1953). Most forms will be obvious to those familiar with Greek; note especially that w represents digamma. The letters ē and ō (or ê and ô with circumflex) everywhere represent eta and omega, except that in Laconian āniokhíōn the ō stands for a long omicron.]

² Cf. Deecke, BB 8.149-50 (1884), with references to Ahrens; J. and T. Baunack, Studien 1.16-8 (Leipzig, 1886); R. Meister, Die griechischen Dialekte 2.221, 227, 254, 284-5 (Göttingen, 1889); O. Hoffmann, Die griechischen Dialekte 1.71-2 (Göttingen, 1891); F. Bechtel, Die griechischen Dialekte 1.440-1 (Berlin, 1921); C. D. Buck, Greek dialects ² 97 (Boston, 1928); F. de Saussure, Recueil des publications scientifiques 457 fn. 2 (Genève, 1922); E. Schwyzer, Gr. Gramm. 1.631 fn. 2 (München, 1939).

³ Cf. Walde-Pokorny 1.189-91. See also P. Kretschmer, KZ 31.415-6 (1892); F. Solmsen, KZ 34.450-1 (1897).

⁴ My reconstructions are phonemic, but attempt to give some allophonic interpretation of résonants on Edgerton's principles, however schematized; thus /y/ may appear as y, i, or iy. The laryngeals are treated consistently as consonants. In reconstructions with laryngeals the macron denotes a long-grade or contracted vowel; otherwise a long vowel of any provenience. Next to an a-coloring laryngeal the basic vowel e is written rather than an allophone a (as yet nonphonemic).

⁵ Likewise, if Döhring's old connection of Gk. hyákinthos with Lat. iuuencus needed any further refutation (see Boisacq s.v.), this argument would be helpful.

⁶ For the exceptional lack of metric ‘position’ before the initial Z- of names like Zákunthos and Zéleia, see P. Chantraine, Gramm. hom. 1.110 (Paris, 1942). If we posit *zw-, the reason for similar treatment is still equally compelling (avoidance of a sequence long-short-long). Since the occurrences are all late (Odyssey and Catalog of Ships), a digamma may also have been largely disregarded in this unusual combination.

⁷ [Cf. Werner Winter, On the origin of the saṃprasaraṇa reduplication in Sanskrit, Lg. 26.365-70 (1950).]