2

restorer of peace and order after civil war and bloodshed', as the upholder of the old Roman customs and religion against the threatened inroad with Antony and his 'Aegyptia coniunx' of Eastern barbarism, and as the visible impersonation of 'Fortuna Urbis,' the deity who watched over the fortunes of Rome. From this point of view the opening lines of G. i and iii and similar passages, however repugnant to modern taste, are neither unnatural nor derogatory to Virgiï's poetic fame. Right or wrong, they express the thoughts not of a courtier, but of a nation; and the poet by whom those thoughts are 'married to immortal verse,' deserves, if ever poet did, the name of 'national.'

II. THE TEXT OF VIRGIL.

A. Manuscripts.

The wide-spread popularity of Virgil's poems during the times of the Roman Empire, and afterwards in the Middle Ages, led to a more rapid multiplication and more careful preservation of MS. copies than in the case of any other ancient writings with the exception of the New Testament: and the evidence for his text is conspicuous not only for quantity, but for quality. It may be assumed as a general principle that the earlier the date of MSS., the better their authority; not that older MSS. necessarily represent a better text, but because the risk of corruption and variation increases with each successive generation of copyists. With the exception of the Egyptian 'papyri' (some of which date from the 1st century of our era3), the 'uncial' or 'capital' MSS. of the 4th and 5th centuries are the oldest extant specimens of writing. Comparatively few Latin authors are represented by even one MS. of so early a date, while for some of the most important (e. g. Horace, Lucretius, Caesar) we have no MS. evidence earlier than the 8th or 9th century: but of Virgil there are no less than four more or less complete MSS., and three sets of fragments, which can be assigned to the 4th and 5th century A. D. A description of these MSS., with a critical estimate of their relative

1 Aen. vi. 795; Hor. Od. iv. 15. 4-11, etc.

2 Aen. viii. 678 sqq.; Hor. Od. iv. 15; cp. iii. 6.

The

3 A fragment from Herculaneum must be older than 79 A.D. accounts on the recto of the papyrus, upon the reverse of which is written the newly discovered 'Athenian Constitution' of Aristotle, are dated in the 11th year of Vespasian, 78-79 A.D.

peculiarities and value, is given in Ribbeck's Prolegomena, Chs. xi-xiii, pp. 218-320; facsimiles of the handwriting employed being appended to the Index. Photographic facsimiles of single leaves may be found in the collection published by the Palaeographical Society, and that of Zangemeister and Wattenbach, published at Heidelberg.

The four great MSS., are

1. 'Vatican' (Schedae Vaticanae, usually cited as F.), in the Vatican Library at Rome: containing portions of G. iii, iv, and Aen. i-viii. It was written probably in the 4th century, though some have considered the miniatures which adorn it to be even older. It belonged successively to Pontanus, Cardinal Bembo, and Fulvius Ursinus (Orsini); the latter of whom gave it to the Vatican Library before his death in 1600.

2. ‘Medicean' (Codex Mediceus-—M.), in the Laurentian Library at Florence: contains Ecl. from vi. 48, Georg., and Aen. A note at the end of Ecl. states that it was read and corrected by one Asterius, 'consul ordinarius': and as Asterius was consul 494 A.D., it must be prior to that date-probably of the 5th century. It once belonged to the Vatican Library, but was purchased by Cosmo de Medici from the heirs of Cardinal Rodolpho Pio, who died in 1564. First collated by Nic. Heinsius in 1676; printed at Florence in 1741.

3. 'Palatine' (Codex Palatinus-P.), in the Vatican Library, brought thither from the Palatine Library at Heidelberg on its capture by the Bavarians in 1622. Contains Ecl., Georg., and Aen. 33 leaves (out of 280) are wanting-viz. E. iii. 71—iv. 52; G. i. 323—ii. 139, iv. 461—Aen. i. 277; Aen. iv. 116–162, vii. 277-645, x. 463-509, xi. 646–692, 737-783, xii. 47-93. Probably of the 4th century. It was the basis of editions (Commelin) published at Heidelberg in 1589, 1599-1603: but the first accurate collation is that made by Ribbeck.

4. 'Roman' (Codex Romanus-R.), in the Vatican Library, contains Ecl., Georg., and Aen., with 76 leaves out of 309 wanting : viz. Ecl. vii. 1-x. 9; G. ii. 2–215, iv. 37-180; Aen. ii. 73—iii. 684, iv. 217—v. 36, xi. 757-792, xii. 759-830, 939–952. The character (large capitals) resembles that found on the walls of Pompeii, and inscriptions of the 1st and 2nd centuries: but the barbarous errors of the text1 and crudeness of the miniatures preclude its being of such antiquity, and it is referred to the 4th or even the 5th century. This MS., which in the 13th century belonged to the monastery of St. 1 See below, p. xxviii.

Denys in France, was transferred to Rome, and there used by Angelo Poliziano (see below, p. xliii), who notes its authority for the orthography 'Vergilius.'

The following are only fragmentary :

5. 'St. Gall Palimpsest' (Schedae rescriptae Sangallenses-G.), in the Benedictine library of St. Gall in Switzerland: 10 leaves containing portions of G. iv, and Aen. i, iii, iv. Early in the 4th century.

6. 'Verona Palimpsest' (Schedae rescriptae Veronenses—V.), in the Chapter Library at Verona, formerly at Bobbio: 51 leaves from an old MS. akin to Med. and Pal., containing about 1320 lines: probably of the 4th or 5th century.

7. 'Berlin Palimpsest' (Schedae Berolinenses or Puteanae-A.), sometimes known as the 'Augustean' fragment: 7 leaves, partly at Rome, partly at Berlin, containing G. i. 41-280, iii. 181-220: of the 4th century, probably near the end, though Ribbeck considers it older than any extant MS. The inscription CLAVDIVS. PVTEANVS. FVLVIO. VRSINO. D.D. shows that the MS. came from Gaul into Italy before 1595 when Puteanus died, and passed after Ursinus' death in 1600 with his other books to the Vatican library.

Of later (cursive) MSS. the following are quoted as of corroborative value for the text :

8. Codex Gudianus (y) at Berne, 9th century.

9. Codices Bernenses (a, b, and c), of the 9th and 10th centuries. 10. Codex Minoraugiensis (m), perhaps of 12th century.

11. Codex Bodleianus (o), 11th century.

It is seldom, however, that any cursive MS. gives a reading which is not suggested either by an uncial MS. or by the early commentators, and they are therefore of but slight independent value1;

1 Some idea of the quantity and quality of existing cursive MSS. of Virgil may be formed from the fact that the Bodleian Library alone possesses forty-five such MSS., of which Mr. Madan, one of the Sub-Librarians (who has examined them all), estimates three to rank with Codd. Bernenses (a, b, c), nine as respectable, and the remainder worthless.

The complaint of Petrarch against professional copyists (stationarii) shows the depth to which their art had sunk in the 14th century—' Quisquis igitur pingere aliquid in membranis, manuque calamum versare didicerit, scriptor habebitur, doctrinae omnis ignarus, expers ingenii, artis egens . . . nunc confusis exemplaribus et cxemplis, unum scribere polliciti, sic aliud scribunt ut quod ipse dictaveris non agnoscas. etc.'. . . ' De Remediis utriusque Fortunae,' i. 43, p. 42 (quoted in Symonds' 'Renaissance in Italy,' vol. i. p. 130).

the chief interest of the most important of them (Gud. and Bern. a) lying in their close relationship to Pal. and Rom. respectively.

The comparative value of the leading MSS. (1-7) in cases of doubtful reading is thus estimated by Ribbeck (Proleg. p. 320): First he places Pal. (P.), with Vat. (F.) and Verona fragm. (V.) of nearly equal authority; Med. (M.), though akin to P., is inferior to it; while Rom. (R.) is least trustworthy of all. The St. Gall and Berlin palimpsests (G. and A.) are too scanty to have much value. The highest possible authority (p. 309) is the united testimony of F. M. P. V. (Nos. 1, 2, 3, 6). As the result of his investigations he assumes an 'archetype,' or original copy, from which all extant MSS. are derived, 'currenti stilo parum nitide scriptum, oppletumque nube conjecturarum, glossematum atque interpolationum.' This conclusion that our existing uncial MSS. represent a single 'recension' of the text derives some support from the fact that the early commentators occasionally supply readings (some of them intrinsically probable) found in none of the MSS. the inference from which would be that they or their authorities had access to another recension than that represented by our MSS. A few examples of such readings are-G. i. 508 formantur for conflantur (Nonius and Serv. on Aen. xii. 304); G. iii. 177 mulgaria for mulctraria (Nonius and Philarg., citing Valgius Rufus, an Augustan poet); ib. 415 gravi nidore (Non., Serv.) ; Aen. ii. 62 dolo (Non. and Serv.); vii. 773 Phoebigenam for MSS. Poenigenam (due to Probus alone); xii. 605 floros for flavos (Serv.).

In spite, however, of such abundant MS. evidence-partly, perhaps, from its very abundance-the text of Virgil, though not in a bad state, is far from certain. The possibility that all our MSS. represent but one recension detracts considerably from the apparent advantages of number and variety: while the fact that Virgil's writings became a school-book within fifty years of his death, and consequently were copied and recopied over and over again, adds much to the possibility of corruption through the carelessness or ignorance of copyists. And the positive errorsgrammatical, metrical, or orthographical-which are found even in the best extant MSS., indicating as these do a decline of scholarship at the period to which the said MSS. belong, give some idea of the liability to corruption to which the text must have been exposed. As an instructive specimen of such errors we may take G. iii. 181-214, a passage represented in the four great uncial

6

MSS. (F. M. P. R.) and the Augustean' fragment (A.). In about thirty lines we find the following-182 'et qui' for 'èquï'(A.): 183 “bella tantum' for 'bellantum' (F. 1); 184 'audere' for 'audire' (M. 1); 188 'audiat' for 'audeat' (Rom., Med. 2); 192 'quae' for 'que' (Rom.), 'currum' for 'crurum' (Med. 1); 193 'quae' for 'que' (Rom.), ‘auris' (Pal.); 201 'fugas' for 'fuga' (Rom.); 202 'Aelei' for 'Elei' (Rom.); 209 'vela' for 'ulla' (A.); 212‘insula for 'in sōla' (Pal.); ‘religant' for 'relegant' (Rom.); 213 ‘nata ’ for 'lata' (Med.). The proportion and character of the mistakes here attributed to Rom. (R.) quite bear out, it may be observed in passing, the estimate formed by Ribbeck of its comparative

value.

This being the case with regard to the evidence of MSS., the student of Virgilian criticism must enquire further what help is to be got from ancient commentaries. The rapid and continuous popularity of Virgil's poems, especially for educational purposes, made them the favourite text-book of teachers and grammarians throughout the Imperial times, and the basis of an innumerable number of treatises, grammatical, exegetical, a d linguistic. Many of these survive only in scattered references in the works of the later and better preserved among them (e. g. Servius): but it is probable that more of the common material of the scholarship of the 1st century A. D. has been preserved by later grammarians than at first sight appears. Much of the accumulated store of Virgilian learning available for Servius or the writers of the Verona Scholia belonged probably to the time of Trajan or earlier and so far as this earlier learning is imbedded in the notes of later commentators, their testimony may be of greater value than if it only belonged to the actual time at which they wrote. If Servius, for example, embodies for us the opinion of scholars of the Ist century on points of interpretation—of men, that is, who had that living sense of the meaning of words which the student of a dead language cannot have-he cannot safely be set aside as valueless : and a respectful hearing is at least due to testimony which may rest ultimately upon other and better materials than those of our MSS. or existing knowledge. To take a single example, where in the difficult passage Aen. ix. 486, Servius tells us that 'funera' is fem. of a lost adject. 'funerus,' analogous to 'celera' from 'celerus,' and supports it by the authority of Ennius, we cannot but feel that this is at least probable. Or when on G. i. 461, he explains serenas'=' siccas,' (an interpr. also found in Nonius), and we recall Lucretius' use of 'seresco' (i. 306), 'grow dry,'