Aside from the foregoing indications, we have only the evidence of language and style, an uncertain means of determining the date of a poem so short as this, so imitative, and so full of reminiscences of earlier authors. The writer was consciously using Vergil and Ovid as models; his diction frequently echoes that of other classical poets. Thus the language of his own time is overlaid with that of the writers whom he imitates. We may, however, note the following indications of later date:

(1) Metre. The metre is correct according to classical standards, except that in three places' initial h is counted as a consonant in making position. This is a peculiarity of Christian poetry, and first appears in the fourth century. Our author does not follow this rule consistently, for in sixteen places h is treated just as in verse of the classical period. Elision is remarkably rare, occurring only five times."

(2) Syntax. The following variations from classical usage Occur:10

Double negative for emphasis (29).

Fruor with the accusative case (49).

Licet as a conjunction with the pluperfect subjunctive (115) and with the indicative (148-9). The former is first found in the early imperial period, the latter from the time of Apuleius. Perfect infinitive used with the force of the present (115, 128). This is not peculiar to late Latin, but becomes more frequent in the later period, and according to Schmalz" is a favorite use with Christian writers.

798, 119, 132. 8See note on 98.

956, 95, 98, 110, 147.

The infrequency of elision is not of much value in establishing the date; from the time of the Silver Age elision was avoided by some poets, but there is no regular decrease in the frequency of its occurrence; cf. E. H. Sturtevant and R. G. Kent, Transactions of the American Philological Association, 46, pp. 146 ff.

Birt (pp. 61-2) points out a further peculiarity in the very frequent use of two caesuras, dividing the line into three parts. He finds the same peculiarity in the short poem of 20 lines on Phaedra, attributed to Vincentius (Anth. 279). On these grounds he concludes that this epistle also is the work of Vincentius. These metrical peculiarities are not in themselves sufficient reason for attributing the two poems to the same author, in the absence of other marked resemblances; moreover, the poem on Phaedra is much more faulty `metrically, containing several mistakes in quantity within its short limits. 10 For fuller discussion see notes on verses cited. 11Pp. 435-6.

Future participle used more frequently than in the classical period (45, 56, 74, 87).12

Asyndeton (124 and 129) is noted by Teuffel13 as an indication of late date.

(3) Vocabulary. The following words or expressions show certain peculiarities:14

libenter habe (2), incole (3), dictare salutem (6), pendet (144) apex (22), sacramenta (33), convincere (37) and resolvere (9), for the corresponding simple verbs, capit (69), fluctus (75), vota queror (87), durum tuli (103-4), vota cupis (137), sidere (141), meus as vocative (144).

These peculiarities of usage, taken all together, point decidedly to a date not earlier than the fourth century. The choice of subject and the general treatment lead to the same conclusion. In the time of barren imitation which succeeded the great creative period of Roman literature, admiration for the masterpieces of the past so dominated the minds of educated men that it left little room for originality, or rather whatever originality existed spent itself on mere matters of form and language. The old themes were treated again and again, and ornamented with new rhetorical devices. Among the great poets Vergil was revered above all, his works were studied in the schools and used as models for imitation.15 The absurd length to which this veneration was carried is shown in the Vergilian cento, consisting of phrases from that poet fitted together with much ingenuity so as to make a new poem on a totally different subject.16 A regular school exercise was the dictio, an expansion of a line of passage from Vergil." The works of Macrobius and of Martir

12 For statistics of the frequency of occurrence of the future participle in different writers see E. B. Lease, American Journal of Philology 1919, pp. 262 ff. 13 Vol. 3. §398.12.

14 For fuller discussion see notes on verses cited.

15 Dill, pp. 385 ff.

16 Anth. 1.7-18

17 Three of these have been preserved in the Anthology (223, 244, 255) under the heading Locus Vergilianus or Thema Vergilianum. Among the Dictiones of Ennodius is one (in prose) entitled Verba Didonis cum abeuntem videret Aenean (Dict. 28). The grammarian was occupied with Vergil to such an extent that to say he knew Vergil was sufficient designation of his calling, as in the epigram beginning Arma virumque docens atque arma virumque peritus, Baehrens Poetae Latini Minores 5, p. 98.

anus Capella are an indication of the reverence which was paid to him in the fourth century. Of such a period as this our poem is evidently a production.

In regard to the place of composition we have no indication except the fact that the Salmasian Anthology was compiled at Carthage. Apparently all the contemporary poets included in the collection are African,18 and it is natural to suppose that many of the anonymous works also have the same origin. We know that Carthage was a flourishing centre of culture and literary activity during the early centuries of the Christian era.19 Many minor works must have been produced there by mediocre writers whose names have not survived, and it is much more probable that compositions of slight merit found their way into an anthology put together at the place where they originated, than that they came from other parts of the Empire. The subject of Dido and her sorrows is one that might naturally be expected to appeal particularly to a Carthaginian poet.

The epistle itself shows that the author was some one who was familiar with the early poets, especially with Vergil, Ovid, Horace, and Lucretius, for it abounds in verbal reminiscences of their works. The correctness of the versification, too, proves that he was well trained in the schools. On the other hand it shows little originality, and there is an entire absence of poetic inspiration or deep feeling. Such a writer might naturally be found among those whose profession was the study and teaching of literature, and it seems reasonable to conclude that the author may have been a Carthaginian rhetorician or grammarian of the latter part of the fourth or the beginning of the fifth century. The poem is valuable in showing what were the interests and literary activities of educated men of this period, and in indicating the strength of the influence which was still exercised by Vergil and Ovid.

18 Riese, p. XXIX.

19 Boissier, pp. 238 ff.; Monceaux, pp. 459 ff.

STYLE AND LITERARY CHARACTERISTICS

By the majority of those who mention it, the poem is briefly dismissed as a mere rhetorical exercise. Palmer20 is the only one of its critics who allows it any merit. The choice of a theme already treated by Vergil and Ovid inevitably challenges comparison with the works of those poets, and such comparison is not to the advantage of our poem. Yet if we accept the author's own estimate of himself as modicus poeta,21 and keep in mind the narrow limits within which he confines his work, it is not altogether without charm and interest. While in a sense it is nothing more than a rhetorical exercise, a comparison with the dictiones referred to above shows that it is of much greater length than they, and much more original in its treatment of the subject chosen.

The poet doubtless had Ovid's epistle of Dido22 before him as his model. The setting in the story is the same as Ovid's, namely the time when Dido asks her sister Anna to carry her appeals to Aeneas.23 The conception of Dido's character also is Ovid's rather than Vergil's-she is gentle and forgiving, resigned to death. She reproaches Aeneas, it is true, after the manner of Vergil's heroine, but her remonstrances are only a mild echo of Vergil's passionate words, and her brief anger is soon suceeded by gentler emotion. Her last wish is for Aeneas' safety. The language, as well as the general treatment, is more suggestive of Ovid than of Vergil, as might be expected; for the former could be imitated much more easily by a writer of mediocre ability. Again and again phrases occur which echo those of Ovid,24 showing that the author must have been sufficiently familiar with his works to reproduce his language unconsciously.

Yet at the same time he altogether lacks the smooth, easy flow of Ovid's verse. The thought is frequently difficult to follow, partly because the language is condensed to the point of obscurity. A noticeable characteristic is a certain poverty of vocabulary, which shows itself in the repetition of the same words and phrases, e. g. vota nocentis (39, 99), honesta pericula (97, 136), peritura (74, 87), perfide (35, 119, 124). Nocens in

20p. XX, footnote 1. 215 22 Epist. 7. 23 Verg. Aen. 4.437–8; cf. Palmer p. 339. 24See notes passim.

different forms occurs six times, and rota eleven times. Quite striking is a tendency to repeat a word three or four lines after its first occurrence, e. g. ipse dolor (11), dolor ipse (16), cf. ipse dolor (89); ligavit (17), ligat (20); conscia (29 and 32); pendet (14), pependet (18); alitura (56), alit (59); reparare (69), reparant (73); miseram (90 and 92); manus (92), manu (94); nefas (97 and 99).

The most original and the best part of the poem is the long passage25 contrasting the constant changes in nature with the enduring grief of unhappy love. The idea is by no means new, but it is developed here to an unusual length and with considerable skill. As we should expect, the details are of literary origin, not drawn directly from nature. The passage is divided into stanzas of equal length by the refrain Sua taedia solus fallere nescit amor. Less effective is the second refrain,26 Cui digna rependes, si mihi dura paras? It does not mark off a distinct division of the poem, as does the first one, the stanzas are not so clearly separated in thought, and one of them is longer by a line than the others.

A marked feature of the poem is the use that is made of alliteration. The following lines show striking instances of this: