irus felly, as A 486; cf. English withe.

oikos house, as a 232; cf. Latin vicus, English War-wick, Berwick, etc.

olvos wine, as T 300; cf. Latin vinum and the English word.

b. It is probable that ἦρα (ἐπὶ ἦρα φέρων Α 572), Ἴλιος, and 'Ipis also were pronounced with initial F.

ε. ἁνδάνω, ὅ, ἕκυρος, ἕξ and others seem to have begun originally with two consonants, of.

d. The verse alone affords no sufficient criterion for the former existence of F in any word; it only indicates the loss of some consonant. This is not conclusive evidence for F, since and j (y) were also lost. Which consonant originally was present has to be learned in each case from inscriptions of other Greek dialects, from a few notes of ancient grammarians, and from other cognate languages (cf. épyov work, olvos wine).

Rem. The Alexandrian scholars did not know of the existence of F in the Homeric language, and consequently they did not use it to explain peculiarities in the Homeric text. The great English scholar Richard Bentley (1662-1742) was the first to discover that its restoration removed many difficulties of Homeric prosody.

e. The sound of Ƒ evidently was going out of use in the Homeric period; it is not infrequently neglected in our texts and sometimes this neglect seems to be due to the poet himself, but F can be restored in many passages by minor changes: kedvà idvîa (Fidvîa) has been restored for the Ms. reading κέδν' εἰδυΐα α 428. For υἱὸν ἑκηβόλον Α 21 it is possible to read υἷα ξεκηβόλον, for ἡβήσῃ τε καὶ ἧς κτλ. α 41 it is easy to read ἡβήσῃ καὶ τῆς, and πᾶσιν δὲ [ανάσσειν for πάντεσσι δ ̓ ἀνάσσειν Α 288. For ἑπτάετες δ ̓ ἤνασσε γ 304, ἑπτάετες δ ̓ ἐάνασσε may be read (§ 25 ). μένος θυμόν τε Γεκάστου may have been the original form of μένος καὶ θυμὸν ἑκάστου Ε 470. αὐτοὺς δὲ ἑλώρια Α 4 became αὐτοὺς δ ̓ ἑλλώ dè A ρια in many Mss. οἳ μὲν οἶνον α 110 is now read where the a

Mss. have oi pèv ap' oivov. As the Alexandrian grammarians and the copyists had no knowledge of this lost letter in Homer, they were solicitous to fill each hiatus by a v movable, ¿' (pá), T' (té), y' (yé), or whatever other addition or change suggested itself. οὐ σύ γ' ἄγεις ρ 576 is probably for οὐ σύ F' (Fe) ἄγεις, and ὃς ἄξει Ω 154 for ὅς F ̓ ἄξει (§ 41 q).

f. F was less constantly preserved in derivative than in primitive words: oivos generally retains its F, but Oivóμaos E 706 has lost it; apvós preserves its F four times, but apveiós shows no trace of it.

g. It has been computed that F in Homer assists in making position 359 times (only in the accented part of the foot or else before the third personal pronoun- before ev once, before the enclitic of 39 times), but prevents hiatus 2995 times (2324 of which are after a short syllable, in the unaccented part of the foot). The force of F is neglected about 600 times in ordinary texts (about half of which passages can be readily changed to restore F).

h. In later poets, e.g. in the elegiac poets and Pindar, Ƒ seems never to make position but often prevents hiatus, poetic precedent allowing hiatus before a word which formerly began with F. The consciousness of the consonantal initial sound of of was retained longest and most clearly. Before that word, no attempt was made to fill a hiatus by v movable or by oux for où, and before it a short vowel was often long by position (§ 41 m).

i. That the sound of F was still alive in the Homeric age is shown by the accuracy of the poet in its use where comparative philology shows that it once existed.

j. F sometimes leaves a trace of its existence in its cognate vowel v: εὔαδεν Ξ 340 for ἔξαδεν, αὐέρυσαν Α 459 for ἀμέρυσαν (§ 11 e), αὐίαχοι Ν 41 for ἀξίαχοι (α privative and fιαχή), ταλαύρινον Ε 289 for ταλά-Γρινον. So doubtless ἀπούρας A 356 for aro-Fpás. Cf. the Pindaric avárav, Pyth. II 28,

for ἀράταν, and εὔιδον Sappho I 7 for ἔριδον. This latter evidov may have been pronounced often where our Homeric texts have εἴσιδον, as ἔσιδεν seems not infrequently to have been substituted for ἔριδεν.

Some irregularities of quantity may be explained by this vocalization of F. Thus ȧжōЄπóv T 35 may have been ȧπoΓειπών pronounced nearly as ἀπονειπών. αὐίαχοι finds its analogy in γένετο ἰαχή Δ 456 (γένετο ιαχή).

k. A neighboring vowel is sometimes lengthened to compensate for the loss of F (§ 41 d).

1. An e was sometimes prefixed to a digammated word and remained after the F was lost, as ἐέλδωρ, ἐείκοσι, ἐέργει, ἑἑ.

m. Sometimes the rough breathing represents the last remnant of a lost consonant (especially in the words which once began with of, as ȧvdávw ктλ., cf. c above), as éκóv, ἕσπερος. Often the same root varies in breathing, as ανδάνω and ἡδύς, but ήδος, — ἕννυμι, but ἐσθής.

n. For the augment and reduplication of digammated verbs, see § 25 h.

DECLENSION.

§ 15. SPECIAL CASE ENDINGS. a. The suffix -pɩ(v), a remnant of an old instrumental case, added to the stem forms a genitive and dative in both singular and plural. It is generally used as an instrumental, ablative, or locative case. The suffix is most frequent in set expressions and in the last two feet of the verse.

1st Declension, always singular: ¿§ evvîþɩv from the couch, ἧφι βίηφι with his own might, θύρηφι at the door.

2d Declension: ἐκ ποντόφιν out of the sea, δακρυόφι with tears, èπ' èoxapópiv on the hearth. The final o of the stem always receives the acute accent.

3d Declension, only with a stems except vaûdi and KотUAndovód (which has gone over into the 2d declension), and

always plural except κράτεσφι Κ 156: ἐκ στήθεσφιν from the breasts, opeσpiv on the mountains.

b. Many of these forms are found only where they are convenient for the verse ; e.g. ναύφι always stands for νηών, not for νηυσί which has the same metrical form as ναύφι.

c. This ending is not used with designations of persons, except αὐτόφι Τ 255, Υ 140, θεόφιν as Η 366.

d. The suffix -0 is added to the stem to denote place where: θύρηθι at the door, οἴκοθι at home, κηρόθι at heart, πόθι where, keîlɩ there (èkeî is not Homeric), weɩ in the morning.

e. The suffix -Oev is added to the stem to denote place whence : "Ιδηθεν from Ida, οὐρανόθεν from heaven. It forms a genitive with the pronominal stems έμε, σε, έ, αὐτο: ἐμέθεν, σέθεν, ὅθεν, αὐτόθεν. Sometimes a preposition is used with it, as ἐξ ἁλόθεν Φ 335, ἀπὸ Τροίηθεν ι 38, ὅθεν εἵνεκα Γ 128, πρὸ ἕθεν πρò éОev E 96. When affixed to adverbial stems, it may lose its final v: ὄπισθε, ἄνευθε, πάροιθε, ἔνερθε.

f. This ending -Oev has lost its original force in certain adverbs; ἐγγύς, ἐγγύθι, ἐγγύθεν do not differ essentially in meaning; cf. ἔνδον, ἔνδοθι, ἔνδοθεν, — πρόσθεν, ὄπισθεν, κτλ. The Aeolic form of the ending appears in urаila Þ 493 from under.

g. The enclitic -de is added to the accusative to denote more distinctly the limit of motion: oikóvde homeward (also otkade, especially of the return of the Achaeans to their homes) ὅνδε δόμονδε to his own house, ἅλαδε seaward, κλισίηνδε to the tent, Τροίηνδε to Troy. So also 'Αϊδόσδε to the abode of Hades; cf. ἡμετερόνδε (sc. δώμα) to our house, εἰς Αίδαο. With Πηλειωνάδε Ω 338, cf. εἰς 'Αγαμέμνονα Η 312. Cf. -δις in ἄλλυδις, ἄμυδις, χαμάδις, -ζε in ἔραζε to the earth, θύραζε, χαμάζε. φύγαδε to fight (for φύγηνδε which is not used) is formed as from a noun of the third declension.

§ 16. FIRST DECLENSION. a. n is found for final a of the stem with the exceptions mentioned in § 5 a f.

b. The nominative singular of some masculines ends in -τα for -της : αἰχμητά spearman, μητίετα counsellor. Cf. the Latin poetă, naută. evρúoπa far-sounding (perhaps a petrified nominative) is used also as accusative, e.g. A 498.

All of these words are adjectival (titular) except Ovéσтa B 107.

c. The genitive singular of masculines ends in -ão or (by transfer of quantity, § 5 d), -εw. After a vowel this ending may be contracted to -ω: ἐυμμελίω Δ 47, Αἰνείω Ε 534, Βορέω y 692. The ending -ew is always pronounced as one syllable by synizesis (§ 7).

し

d. The genitive plural ends in -awv or -εων: θεάων, βουλέων. After o this ending may be contracted, as παρειών Ω 794. After a long syllable (i.e. everywhere except in πvλéwv H 1, M 340, and Ovpéwv 191) synizesis of -ewv occurs, as ναυτέων.

e. The dative plural ends in -noɩ(v) or rarely in -ns, as θύρῃσιν, πέτρῃς, - in -ais only in three words: θεαῖς € 119 (cf. θεά § 5 α), ἀκταῖς Μ 284, πάσαις χ 471.

e

f. The short form of this dative ending is rarely used before a consonant; when it stands before a vowel, it may be said that the final has been elided.

し

§ 17. SECOND DECLENSION. a. The genitive singular has preserved the old ending to which affixed to the stem vowel makes -olo.

b. According to tradition this ending does not suffer elision; but elision is metrically possible, e.g. dvooμévov “Tπepíovos a 24 might be δυσομένοι” Ὑπερίονος. See § 7 a.

c. The termination -oo is indicated by the metre in certain places where all the Mss. give a corrupt form: TéλeσTOV ὅο κλέος οὔ ποτ ̓ ὀλεῖται Β 325, Πολύφημον ὅο κράτος ἐστὶ μέσ γιστον α 70, ἀδελφείο φρένας ἥρως Η 120, Ιλίου προπάροιθε X 6, Aióλoo μeyaλýtopos κ 36. It is to be recognized also in Πετεώο Β 552 for Πετεάοο, from Πετεως for Πετεάος. It