without accurate collations of several of the most important; in the case indeed of one of them, the Palatine, we seem to have been without a collation at all. We now possess collations of all the uncial MSS., fragmentary and entire, and of four or five of the most important of the cursives, which for minute and painstaking accuracy apparently leave but little to desire: and great care has been taken not only in collecting the testimonies of the different grammarians who quote passages from Virgil, but in noting the readings of the various MSS. of each witness. In the absence of Prolegomena, we are still left in some doubt about the comparative importance to be attached to these various authorities and about other details connected with them: in particular I would mention that we do not seem to have the means of distinguishing those different classes of readings, which in the case of the Medicean MS. Wagner discriminates respectively as 'prior' and 'posterior lectio' and 'lectio a manu prima' or 'secunda.' But there can be no doubt that we already possess critical materials surpassing most of those with which we have had to content ourselves till now, not only in degree but in kind, and that their use is likely to effect a considerable change even in that text of Virgil which, since the time of Nicholas Heinsius, has been generally accepted as the best. That text indeed has now but little to fear from the competition of the text or texts which it superseded: the authority which they were supposed to derive from the Palatine has disappeared for ever now that that copy has been actually examined, and their real support is apparently to be found in most cases partly in copies of no name or weight, partly perhaps in the arbitrary conjectures of early editors. But the testimony of the Medicean, on which Heinsius chiefly rested, has been considerably weakened by the results of the new collations: in very many instances the other uncial MSS. are seen to be arrayed against it, while its readings may not unfrequently be accounted for by the

parallelism of other passages in Virgil, which the transcriber apparently remembered. Probably however it is premature as yet to decide on the whole question: we shall learn the real value of our newly collated MSS. better as we become used to them, and there may be a danger of accepting novelties of reading simply as novelties-a danger which I seem to see exemplified in Ribbeck's text, and which my readers will perhaps find to be exemplified in mine. The general result certainly confirms what I ventured to assert in my former Preface, both as to the existence of many varieties of reading which can hardly be accounted for on palaeographical or other external grounds, and which must often be estimated by the somewhat wavering measurement of individual preference, and as to the sufficiency of a text made up from one or other of the MSS. or early authorities without critical conjecture. In the more important of the two instances in the Eclogues where, following others, I had ventured to depart from the MSS., I have now learnt from Lachmann and Madvig that no change was necessary: and if there are any places in the present volume where a word has been introduced from the dictum of a critic without some ancient authority, it will be found, I think, to be in a case which, to a transcriber, was really a case of spelling, such as Cyclopius' for 'Cyclopeus,' or 'deripere' for 'diripere.' Here I am sorry to say Ribbeck is still less to be commended than in the choice of MS. readings. In several places he has introduced emendations into the text, generally conjectures of his own, which are in every case, in my judgment, worse than needless : nor is he in general more happy in his attempts to point out interpolations or to indicate lacunae. Hitherto the text of Virgil has enjoyed a singular immunity from arbitrary criticism. In the last century, while Horace was being transformed alternately by the splendid audacity of Bentley and the more formal and pedantic dogmatism of Cunningham, Virgil remained nearly in the state in

which Heinsius had left him. Cunningham indeed proceeded from Horace to Virgil, whose text he reformed in obedience to certain canons which he supposed himself to have drawn from a scrutiny of the best MSS.; but his edition, though curious and interesting, seems to have produced no effect, whether as being a posthumous publication, or from the absence of the eclat which attended a controversy with an adversary like Bentley, even when that adversary declined to reply, or perhaps because the labours of Heinsius rested on a basis too firm to be easily disturbed. Gilbert Wakefield, towards the end of the century, edited both Horace and Virgil: but his attempts at innovation were too desultory seriously to affect either. Probably the greatest amount of misapplied ingenuity that has been bestowed on Virgil, till we come to Peerlkamp in the present century, is to be found in the conjectures of Schrader, which I know only as reported by Heyne and Ribbeck. They are always, or almost always ingenious, showing that degree of insight which is required to perceive an anomaly of expression, and that degree of tact which hits on a word that might possibly have been used instead; but there their praise must cease. Such ingenuity is, I believe, almost wholly inapplicable to an author like Virgil, whose text, supported as it is by an ample variety of testimony, requires not emendation but illustration. If he has hitherto escaped the fate of Sophocles, whose peculiarities of expression, so curiously analogous to his, have too often been changed by critical licence, the gain is his and that of Latin literature. Whether it would be desirable that our knowledge of MS. materials should be still further extended by an equally accurate collation of the cursives not examined by Ribbeck, I do not presume to say. There can be no doubt that an apparatus criticus like Ribbeck's is far preferable to one like Heyne and Wagner's: as little doubt can there be that to collate the remaining copies satisfactorily would be an almost endless task. In the Bodleian

Library at Oxford alone there are about twenty MSS. of Virgil, hardly any of which seem to have been collated (I except of course the Canonician MS. which Mr. Butler has examined so thoroughly); the College Libraries too contain a few, the readings of one of which, a copy in Balliol College Library, No. 140, referred by Mr. Coxe to the fifteenth century, have been noted with scrupulous care by my friend Mr. E. Palmer, and placed at my disposal. I myself examined ten or eleven of the Bodleian MSS. to discover the authority for the readings 'litus arenosum Libyae' in Aeneid 4. 257 (see Additions and Corrections at the end of the volume) and 'Trinacriis' in Aeneid 5. 573 (see note there), doubts having arisen about the existence of each; but almost the only other passages I turned to were Aeneid 1. 668, where all agreed with the Medicean in giving 'iniquae,' and the celebrated lines about Helen in the Second Book, which they were unanimous in omitting in the text, one of them adding the passage in the margin. On the whole it would seem that while it may be advisable to apply to an inferior MS. in a case like that which I have mentioned, to ascertain a reading not otherwise certified, it would be waste of time to perform partially a work which, to have any value, should be performed entirely. It is one thing to find that a particular reading which seems necessary to the sense has probably some better support than mere conjecture: it is another to collect all the readings of a copy without knowing what place it holds among the members of one or other of the various families of MSS. through which the text of a popular classical author has been transmitted to us, or indeed before it has been distinctly ascertained what those families are, and what their history has been. A critic of the New Testament may be laudably employed in establishing a theory of recensions inductively by the examination of cursive no less than uncial MSS.; but in the present state of classical studies we shall probably have to wait long before any

one will think it worth while to qualify himself for writing a detailed history of the text of Virgil.

The commentaries which I have used have been in general the same as those employed for the Eclogues and Georgics. I have lost the companionship of Mr. Keightley, and have gained that of Gossrau and Dr. Henry. Gossrau's commentary is neat and compendious, more convenient than Forbiger's, though not so full, and with more traces of independent judgment. He has studied Servius with care, and quotes him at times very appositely and he has paid considerable attention to his author's peculiarities of language and metre, to the latter of which subjects he has devoted an elaborate appendix. His fault is an occasional tendency to see insuperable difficulties and suspect interpolations: but it is kept within bounds, and may perhaps only operate on the student as awakening a wholesome spirit of inquiry. Dr. Henry's work is rather a collection of copious observations on numerous detached passages (Notes of a Twelve Years' Voyage of Discovery,' as he somewhat quaintly calls it) than a regular commentary: but I have found it of the greatest use, as my frequent references to it will show. The form is perhaps a little cumbrous, and the endeavours after precision not always successful: but there is freshness and originality in every page: a large number of the views are at once novel and sound: and the illustrations from other authors are good and apposite, though we may sometimes feel that the more obvious sources have been neglected for the less obvious. I am only sorry that he has not explored' as yet beyond the Sixth Book.

For the notices I have given from time to time of varieties in the Trojan legend and the story of Aeneas' migration unknown to Virgil, or recognized only in the way of distant allusion, I have been indebted almost entirely to Heyne's Excursuses, which seem to me to present a rare union of learning, sagacity, and sobriety.