the facts as seen by Dr Atkinson; or, if not, we may further admit that the general predominance of 'liquid' endings would naturally extend itself by the influence of imitation and analogy. (ii) If a noun or adj. is placed at the end of the second half, it is a natural tendency to place the word agreeing with it at the end of the first. Thus, when we get to the end of the couplet, we are not surprised by an unexpected trisyllable, but have the satisfaction of receiving a word which we have been expecting and for which we are prepared.

I will now sum up the facts as I conceive they should be regarded.

A. Trisyllabic pentameter endings Summary of rediminish steadily in frequency from Ca- sulls.

tullus to Ovid.

B. Most of the examples are nouns. we find adjectives, adverbs, and verbs. latter were avoided by his successors.

In Catullus

But these

Or nouns form 63 per cent. of the whole in Catullus, 78 in Tibullus, and 96 in Propertius. These figures speak for themselves.

Propertius never avoids, or perhaps rather affects the assimilation of the first half of the Symmetrical pentameter to the second, as in

reccidit inque suos mensa supina pedes

construction of pentameter, &c.

v. 8. 44.

The same tendency is seen at work in some

endings that remind us of Greek,

hunc super et Zetes hunc super et Calais

uret et Eoos uret et Hesperios

candida Nesaee caerula Cymothoe

I. 20. 26.

II. 3. 44.

III. 21 (18). 16.

As already said (p. lxxxix.) he does not avoid the concurrence of similar terminations especially when the termination is a (so also Catullus)'. Nor does he observe the artificial rule of not ending a couplet with a nom. part., e. g. egens ends a verse in Iv. 20. 30 and sedens iv. 15 (16). 24. He is not so careful as Ovid not to allow the sense to run on from the hexameter to the pentameter, or from one couplet to another. This latter frequently produces a fine effect; e.g. v. 8. 7

qua penetrat (uirgo, tale iter omne caue) ieiuni serpentis honos2.

In his employment of elisions he stands between < Catullus and Ovid, but nearer to the for

mer.

Elisions.

He has many more than the latter.
He has 23 cases of elision of a long syllable before

1 He does not observe the rule of not putting què &c. after an . Thus we find taleque 111. 6 (5). 26, iungiteque iv. 21. 13.

To an English reader the difference between the Ovidian and Propertian custom will suggest that between the versification of Pope and Morris.

P. P.

k

a short one, chiefly in the first foot; of which four are of a monosyllable. Once this elision occurs in the latter half of a pentameter, III. 6 (5). 18. As L. Müller points out, this elision chiefly occurs in phrases like uidi ego, &c.; or else where the vowel which precedes is ō (afterwards ŏ) as in nemo adeo. For the elisions at the end of the first half of a pentameter see the footnote'.

Propertius has two non-elisions of m,

O me felicem, o nox mihi candida et o tu

Hiatus.

III. 7 (6). 1.

haec eadem ante illam impune et Lesbia fecit

III. 30 (24). 45.

one unelided long vowel, in imitation of Virgil, sed thyio thalamō aut Oricia terebintho

and one with shortening,

IV. 6 (7). 49.

IV. 10 (11). 17.

Omphale in tantum formae processit honorem

He does not object to keeping a short vowel before 8 and a following consonant. He never lengthens it as Tibullus does.

Short vowel before s and consonant.

Thus quoscumque smaragdos 111. 8 (7). 43, bracchiă spectaui Iv. 10 (11). 53, iam beně spondebant v. 1. 41, tu cauě spinosi v. 4. 48, nunc ubi Scipiadae classes IV. 10 (11). 67, Minoa uenumdată Scylla figura Iv. 18 (19). 21; even consuluitque striges v. 5. 17, and probably Iouis cum prolě Scamandro IV. 1. 27 n.

1 He has two instances of the end of the first half of the pentameter being elided before the last; 1. 5. 32 non impune illa rogata uenit, iv. 22. 10 Herculis Antaeique Hesperidumque choros, compare 111. 6 (5). 10 immortalis ero si altera talis erit. The close connexion between the two halves which this produces is a Greek liberty which Tibullus and Ovid abandoned. It is found in Catullus. In the hexameter he affects an elision at the end of the second foot. Some exx. have been quoted above.

The following seem to be all the cer- Lengthenings in

tain instances:

arsis.

II. 8. 8.

v. 1. 17.

Quantity of separate words.

Synizesis.

uinceris aut uincis: haec in amore rotast nulla cura fuit externos quaerere diuos The coalescence of two syllables into one occurs chiefly in Greek words, such as Promethei 11. 1. 69, Enipeo 1. 13. 21, &c. Besides these we find eodem 11. 8. 26 (eadem IV. 5 (6). 36, eosdem v. 7. 7), dehinc 11. 4. 60: i for ii in plur. in Gabi v. 1. 34, probably Deci (adj.) v. 1. 45 and socīs l. c. and i for. ii the pronoun, III. 19 (16). 35, 111. 32 (26). 64 Lauinīs litoribus. In abiegnus v. 1. 42, the vowel becomes a semivowel (abyegnus). So in Suebus Iv. 2 (3). 45. The resolution of one syllable into two is not found in Propertius. The sole exception is the gen. in -ii. Up to his time this gen. from words in -ius and -ium is only found in a few isolated cases; e.g. Catullus 9. 5 o mihi nuntii beati, Virg. Aen. 3. 702 immanisque Gela fluuii cognomine dicta, and in Ennius in the proper name Tarquinius: and the form is the only one in use in the best classical prose.

Diaeresis.

and -ii.

Propertius however found the ii form so convenient for his verse that he used it freely in conjunction with the -i form; and Ovid followed him and extended the use still further. On the one hand we have Mercurii, imperii, gymnasii and on the other Pelusi, Antoni, Mari, Tati.

It only remains to remark that, besides the usual ego', nescio, uolo, we find an isolated ex

ample of o in verbs, i. e. findo IV. 8 (9). 35. Shortening of -0. As is well known, the shortening spread fast after the Augustan period; and in the age of Juvenal final o is indifferently short or long.

1 egō, v. 2. 3, is remarkable.

CHAPTER V.

LITERARY HISTORY.

THE relation of a poet to his predecessors, contemporaries and successors is always a question of literary interest. In the case of Propertius it is also one of critical importance. The happy comparison of some passage of his models or of his imitators may often suggest the right explanation or emendation of a passage classed up till then as corrupt. Much has been done in this way and much remains to be done. But of this upon a future occasion.

Propertius.

Propertius tells us more than once who are his masters in poetry. These are of course the Alexandrine writers of elegy Calli- The models of machus and Philetas'. His ambition is to be the Roman Callimachus' (v. 1. 64); he prays the sacred shades of Callimachus and Philetas to admit him to their sacred grove (Iv. 1. 1).

A. Greek.

Of Philetas we have only some inconsiderable fragments. We have more of Callimachus;

and Philetas.

but not sufficient of his elegiac poems to Callimachus estimate the amount of our poet's debt to

1 The reader will find a long and somewhat barren discussion of Propertius' relations to the Alexandrines in Hertzb. 1. pp. 186 sqq.