est' could not have been accurately written before B.C. 19, when Agrippa completed the conquest of Spain, and the word auspicio' seems to hint that Augustus was not present in person. We must therefore suppose, what à priori is probable, that the history was published in parts. There are however two passages which have been quoted by Niebuhr from the first decad to show that it was written much later.

(1) 9. 36, 'Silva erat Ciminia magis tum invia atque horrenda quam nuper fuere Germanici saltus, nulli ad eam diem ne mercatorum quidem adita.' Here the allusion is supposed by Niebuhr to be to the German conquests of Drusus from 12 to 9 B.C. F. Lachmann, however (p. 47), thinks that the passage may refer to Caesar's campaigns, which had at least made the German forests somewhat better known. The wars in Germany about B.C. 25, mentioned by Dio Cassius 51. 20, 21 and 53. 29, may have contributed to the same result.

(2) In 4. 20 he calls Augustus 'templorum omnium conditorem ac restitutorem;' Niebuhr thinks this title could not have been given him so early. But the temple of Juppiter Feretrius was restored during the triumvirate at the advice of Atticus (Corn. Nep. Att. 20). We read in Suet. Oct. 29, 'Aedem Marti bello Philippensi pro ultione paterna suscepto voverat,' from which we may presume that it was built before this time. The temple of Julius was built soon after the battle of Actium. It was in B. C. 28 that the great temple of Apollo on the Palatine was dedicated. Lastly, Virgil connects the building of 300 shrines with the triple triumph of Augustus (Aen. 8. 714). And it was at this time that he held the 'censoria potestas' to which the repair of temples properly belonged.

On the whole then the first decad appears to have been written between 27 and 20. The Aeneid was being written at the same time, and thus the greatest poet and the greatest prose writer of the age were occupied at the same time in calling the attention of the Romans to their origin. We are not here concerned with the later decads, and I do not undertake to discuss the work as a whole.

THE TEXT. It appears that near the end of the fourth century after Christ a recension of the first decad was made by a certain Victorianus, and that the sixth, seventh, and eighth books were emended by Nicomachus Flavianus; the third, fourth, and fifth by his son Nicomachus Dexter. This is proved by notes which appear at the end of some of our MSS., of which the following are specimens:—

'Nicomachus Flavianus, v. c. 3. praefect. urbis emendavi apud Hennam.'

'Nicomachus Dexter v. c. emendavi ad exemplum parentis mei Clementiani.'

'Victorianus v. c. emendabam dominis Symmachis.'

This is illustrated by a passage in Symmachus (Ep. 9. 13), 'munus totius Liviani operis quod spopondi etiam nunc diligentia emendationis moratur.'

Now all our existing MSS., with one exception, seem to be founded on this recension. Criticism therefore has two distinct tasks before itthe first to recover from the existing MSS. the text of this recension, which is their archetype; the second and more difficult to recover from the text of the archetype the original text of Livy.

Of the MSS. the two most important are the Medicean and Parisian, called by editors M and P. P is considered somewhat inferior to M. A complete knowledge of these two fundamental MSS. we owe to Alschefski, whose edition of Livy appeared in 1841.

Closely resembling M was another MS. now lost, but many readings of which (beginning at 1. 20, and ending at 6. 28,) are preserved to us by Beatus Rhenanus. This is called by Madvig R, by Hertz V (Codex Vormacensis).

Next to these Hertz places Codex Bambergensis (B), of which Heerwagen collated the first book, and Codex Einsiedlensis (E) collated by Haupt.

After these come the first Leyden (L) and the 1st Harleian (H1). Closely agreeing with L, is a MS. in St. Mark's Library at Florence, called by Hertz F.

The other MSS. spring, according to Madvig, from some MS. cognate with L. H1 and P are inferior, full of errors, corrections and interpolations.

To determine the reading of the archetype, the consent of M and P is the most decisive evidence.

But Madvig in his Emendationes Livianae,' and more recently Madvig and Ussing in their edition of Livy, have gone beyond other modern editors in venturing to correct the archetype itself. Madvig endeavours to show that certain particular kinds of error were common in it, e. g. the omission and insertion of m at the ends of words; the altering of words so as to force them into agreement in case, number, or gender with those next to them; the doubling of letters and syllables; the insertion into the text itself of marginal annotations, etc. Starting

1 These Nicomachi are not mere names. A long inscription has been found which refers to them. A thorough investigation of it by G. B. Rossi will be found in the twenty-first volume of the Annali dell' Instituto di Corrispondenza Archeologica,' (Rome).

with this distinct conception of the archetype and applying his admirable knowledge of Latin and his great ingenuity, Madvig undertakes to reproduce Livy himself. Our own eminent Latinist, Mr. Munro, pronounces that he has done more for the text of Livy than was done for Greek tragedy by the whole Porsonian School. Mommsen's opinion may be gathered from the following sentence:- Pertractare autem ejusmodi quaestionem et quantum nostrae aetati datum est absolvere cum unus homo possit ex iis qui hodie sunt Madvigius, hoc optamus ut telam a nobis inchoatam et retexat, ubi opus est, et detexat.'

Another MS., containing parts of the first decad of Livy, had long been known to exist at Verona, where it was discovered by Mai, but scarcely anything was known of it, and no editor hitherto has had its readings before him. After Blum, Detlefsen, and A. W. Zumpt had given specimens of it, it was completely copied by Mommsen in 1867, and published in the Commentationes of the Berlin Academy in the following year. This MS. contains the greater part of the third, fourth, and fifth Books, and the early chapters of the sixth. But the publication of it is an important incident in the history of the text of Livy, because in the opinion both of Zumpt and Mommsen it differs from all the other MSS. in not being founded on the Nicomachean recension We thus obtain a means of deciding differences between M and P, and Mommsen finds that the character of M (and R) is raised by the new MS., which he calls V. But he also finds readings manifestly wrong, which, notwithstanding, are common to V and the Nicomacheans, so that we come in sight of a new archetype, itself considerably removed from Livy's original, and already a good deal corrupted. For the rest, he finds V to be considerably inferior to the Nicomachean archetype, yet still in some instances to have alone preserved the true reading. His opinion is in the main approved by Madvig in the second edition of his Livy.

The editor of a classical author undertakes three things: to give a true text, to explain difficulties in the language, and to explain difficulties or add necessary information in the subject-matter. It sometimes happens that in one of these three departments there is little to be done. This is not so with Livy. An editor might find full occupation in settling his text or in explaining his Latin. But when the text is settled and the Latin explained, the labours of an editor of Livy are hardly begun. It is true he might decline to handle the subject-matter and leave it to historians of Rome. Historians of Rome, however, as Mommsen, begin to omit altogether or treat very shortly the matter of Livy's early books; I have, therefore, been led, not only by my own taste, but by a consideration of the interests of classical students, to

include the subject-matter in my province. The reader will see that the present edition owes more to Niebuhr, Schwegler, Becker and Marquardt, Newman, Lewis, Mommsen, Ihne, and other authors of the same class, than to Madvig, Weissenborn, or Hertz. It is difficult to be equally thorough in three departments so distinct, and each so large, as Roman antiquities, Latin philology and grammar, and textual criticism. My chief attention has been given to the first, but I shall be disappointed if this edition is not judged to deal conscientiously and thoroughly with the difficulties of idiom and construction which Livy presents. In textual criticism all I have done is to exert an independent judgment upon the materials furnished by Weissenborn, Hertz, Alschefski, Madvig, etc. I have made it a rule, even when I did not accept Madvig's reading, to put the reader in possession of it.