ment is a rime subsisting between the syllables preceding the

two caesuras; e.g. I, 1, 47:

aut, gelidas
hibernus aquas

cum fuderit auster.

In still other cases there is a similar sound, but not a perfect rime.

Tibullus employs an even smaller proportion of feminine caesuras, 19 in all, but has also 19 bucolic diaereses, which looks as if he did not regard these as blemishes.

(c) Lygdamus is so orthodox as to be positively dull, having but 10 of his 145 hexameters that are not of the penthemimeral type. Of these, 7 are perfect trithemimeral-hephthemimeral cases, I is a feminine, and 3 are bucolics.

II. PENTAMETERS

All the elegists show in these rather more care than in the hexameters.

(1) Monosyllabic endings. Catullus has one instance; Tibullus, Lygdamus, and Ovid, none; Propertius, with characteristic independence, 4, all being of the same form, viz. sat est.

(2) Verse endings longer than a dissyllable. Catullus has 83 trisyllabic endings, Tibullus but 22 out of twice as many verses, Lygdamus but 3. Of polysyllabic endings Catullus has 92 (18) pentasyllabic, and 1 heptasyllabic), Tibullus 23, Lygdamus but 7: Indeed, Lygdamus in such matters of formal comparison usually more than holds his own. In Ovid the law of a uniformly dissyllabic ending is thoroughly established.

(3) Endings of first half of pentameter. The tendency toward the dissyllable here is not so completely followed. Catullus has 36 monosyllabic endings, Tibullus 7. Almost as many trisyllables as dissyllables appear in Tibullus; but Ovid holds closely to the dissyllable.

(4) The separation of the two halves of the pentameter becomes increasingly careful. In Catullus there are 18 cases where they are run together by elision; e.g. 67, 44:—

speraret nec linguẩm esse nec auriculam.

(5) The preference for dactyls or spondees in the first half varies. Catullus seems slightly to prefer verses of the form, dactyl, spondee, long syllable; but the form, spondee, spondee, long syllable (ie. 5 successive long syllables) is a close second, which can hardly be true of any of his successors. Next comes the form, spondee, dactyl, long syllable; last, dactyl, dactyl, long syllable.

In Tibullus, however, there is an overwhelming preference for opening the verse with a dactyl.

(6) Middle rime. 22 per cent of the pentameters of Catullus exhibit this, and 17 per cent have similar endings. In the later writers the proportion frequently far exceeds this. Often, too, this rime is combined with the same phenomenon in adjacent hexameters, to a noteworthy extent. In Propertius, 2, 34 (a poem of 94 vv.), there are 38 instances of the middle rime, and the 6 consecutive vv., 85-90, have it throughout.

(7) End rime. There are over 200 cases in Catullus, Tibullus, and Lygdamus, fewest of all in Lygdamus. Propertius has I in every 14 verses. Sometimes they occur in triplets. Propertius has one quadruplet rime.

43. The studies of Haupt, a half century ago, showed that Catullus was relatively careless in allowing the elision of a long syllable before a short one.1 The recent elaborate studies of Siedow 2 show that Ovid was more careful than either of the other three elegists in avoiding elision, as well as in avoiding a plurality of elisions in a single verse and the elision of long syllables or diphthongs; that Catullus is most free of them all in

1 Cf. Haupt, Opuscula, Vol. 1, pp. 88 sqq.

2 De elisionis aphaeresis hiatus usu in hexametris Latinis ab Ennii usque ad Ovidii tempora, 1911, with valuable bibliography and tabular statements.

eliding monosyllables; and that Lygdamus leads in avoiding hiatus, not exhibiting a single instance of that phenomenon.1 Similarly interesting studies can be made with reference to the arrangement of words in the verse as a whole, or in different parts of the verse.2 Diaeresis is particularly common in solvo and its compounds.

In other matters, e.g. prosody, progress will be noted after Catullus. Lengthening the final short syllable in the thesis occurs rarely in Tibullus. Shortening final -o in verbs is the opposite phenomenon. In the treatment of quantities before a mute and liquid Tibullus is quite orthodox.

1Cf., on the hiatus in Catullus, Friedrich's note on Cat. 3, 16. Cartault thinks Tibullus shows greater looseness in elision in Book 2 than in Book I as well as in other metrical matters; but Hörschelmann undertakes to show a distinct advance in these respects in Book 2 (cf. elision tables in Hosius, p. 180).

2 Cf. Braum, De Monosyllabis ante caesuras hexametri Latini collocatis, Marburg, 1906; Isidor Hilberg, Die Gesetze der Wortstellung im Pentameter des Ovid; Hornstein, Die Wortstellung im Pentameter des Tibull und Ps.-Tibull, Czernovitch, 1909; Petrus Rasi, De Elegiae Latinae Compositione et Forma, Padua, 1894; Smith, pp. 103 sqq.

3 For the instances of the same in Propertius cf. Hosius, p. 184.

4 Cf. Rasi, de positione debili, etc.; Brenner, Die prosodischen Funktionen inlautender muta cum liquida im Hexameter und Pentameter des Catull, Tibull, und Properz; Winbolt, Latin Hexameter Verse, 1903.

CATVLLI CARMINA

65

Etsi me adsiduo confectum cura dolore sevocat a doctis, Ortale, virginibus, nec potis est dulcis musarum expromere fetus mens animi: tantis fluctuat ipsa malis:

65. 1. confectum G defectu 0.