the moment of their satisfaction however a huger mass than usual thunders down upon them, and they are crushed and broken.

In the section on the Roman MS. “ effrena " is proposed for “ea frena” in Aen. 6. 101. Nothing is said against the reading of the MSS., so that it is not easy to see how to defend it. I can hardly suppose M. Ribbeck not to be aware that "ea frena concutit" is Virgil's way of saying "tam vehementer frena concutit," or that shaking the bridle is a natural expression for making the horse feel the bit: and yet if neither of these supposed difficulties was the motive for altering the text, it is hard to say what can have been.

In the section on the Gudian MS. M. Ribbeck takes occasion to discuss two passages where he changes the received reading. In Aen. 1. 396 Pal. corrected has “captos iam respectare," Gud. originally "captos iam etpectare," words which he supposes to point to what he considers the true reading, "capsos iam respectare." The swans are said to look towards their coops or enclosures, a strange sense, as nothing in the passage leads us to suppose that tame swans are spoken of, and expressed by a strange word, the authority for which in that sense seems to be a single passage in Velleius. He objects to the common reading "in despectando per se spes perfugii nulla:" true: but the poet (as Dr. W. Wagner remarks) is speaking not of escape but of freedom from danger : the eagle is out of the way, just as the storm is over, and the swans are settling on the ground at leisure. The other passage is v. 323 of the same book, where M. Ribbeck thinks Madvig right in objecting to the common reading, or at least to the common punctuation, wrong in connecting "maculosae" (or "maculoso ") "tegmine lyncis" with "cursum," as "tegmine" could not be used for the hide of a living beast. M. Ribbeck himself adopts" tegmina" from Gud., making "tegmina lyncis" co-ordinate with "cursum" as the object of "prementem." I do not know Madvig's remarks, which were published in a Dutch periodical in 1859, so that, as before, I am not sure what I am to answer: I suppose, however, that it is not the huntress wearing a lynx's skin which is objected to, as that might be paralleled, but the apparent awkwardness of “aut” as coupling "succinctam" with "prementem." "Prementem," however, is coupled not with "succinctam" but with "errantem," the two cases supposed being that the huntress is wandering about in search of game and that she is in the full cry of the chase. As to "prementem tegmina lyncis" in the sense of hunting a lynx for its hide, I scarcely think it will be considered to add much beauty to the passage.

In speaking of the Berne MS. which he calls c, he mentions his own conjecture on Aen. 1. 455, "intrans" for "inter se," as partially supported by that copy, which has "intra se." "Artificum manus inter se miratur " is a Virgilian expression for "artificum manus miratur dum aliam cum alia confert :" "intrans," unless I am greatly mistaken, would be exceedingly flat, and not at all like Virgil's manner. “Variis,” which M. Ribbeck extracts from "uaseis," the reading of the aforesaid MS. for "roseis " in Aen. 7. 26, does not seem particularly tempting, though had it been the reading of the MSS. it might have passed without remark. "Croceis," Schrader's and Bentley's conjecture, is far more plausible. But "roseis" is supported by an exact parallel in Ovid, Fasti 4. 713,"Postera cum veniet-Memnonis in roseis lutea mater equis," which I owe to my friend Mr. Nettleship.

In the sixteenth section of the same chapter, in which the inferior MSS. are discussed, M. Ribbeck speaks of six Paris MSS., as containing the "optima emendatio” "Paris" for "creat" in Aen. 10. 705. In his critical note on the passage he had expressed himself more cautiously, "si fides Potterio," Pottier's reliability as a reporter of the readings of his Paris MSS. having been generally questioned by subsequent critics. Some years ago, being anxious to set the question at rest, I requested my friend Mr. Duckworth (of Trinity College, Oxford), who happened to be in Paris, to examine the six MSS. in this passage: and he assured me that none of them contained the

word "Paris" or anything like it, all having "creat" or some slight variety of "creat." After this, I should not be disposed to trust Pottier's report of the transposition of the lines 660-665 in the same book without further evidence. The mistakes which M. Ribbeck has made in this section about the English MSS. of Virgil have been pointed out by Dr. W. Wagner in the postscript to his first paper "On Ribbeck's Virgil" read before the Philological Society.

I now come to the seventeenth section, which is chiefly occupied by a defence of the conjectures introduced by M. Ribbeck into the text. He believes that all the extant copies of Virgil can be traced to a single (unknown) archetype, written "parum nitide," and filled with a multitude of conjectures, glosses, and interpolations. The arguments by which he supports his belief are first, the instances of agreement presented by particular MSS. in the transposition of words, secondly, the instances where the true reading has had to be restored conjecturally. So far as I understand the first argument, it is to the effect that instances may be produced where any two of the better MSS. agree against the rest in a particular transposition, which points to the conclusion that there must have been one great repertory of transpositions from which copyists drew ad libitum. I do not profess to be skilled in the art of constructing a conjectural history of the text of an ancient author: but I should have thought that the phaenomena of the text of Virgil were likely to be particularly impatient of being accounted for on any such hypothesis, and it certainly does not seem to me that the instances of transposition which M. Ribbeck adduces are either so numerous or so striking as to require any other explanation than that of accidental coincidence in the case of copies considered to be otherwise independent. But I must hasten to the second argument, or rather to the conjectures which are produced as constituting its strength. In E. 3. 110 "hau temnet dulcis, haut" cannot be called a happy conjecture. The received reading at any rate gives an appropriate sense, "both are worthy of the prize, and so are all who like you can realize the sweet and bitter of love:" the new reading introduces a general maxim for which Palaemon's speech hardly seems the place. "Alte" for "ante" E. 6. 80 is ingenious, but not convincing. The sense it gives is unobjectionable, but it would I think be a little feeble here: and so I would rather take my chance between the two rival interpretations of "ante,” though neither is free from exception. Neither the change of "an" E. 8. 49 into "at," nor the reconstruction of the whole passage which it involves, seems to me Virgilian: nor can I recognize the necessity of exact strophical correspondence in Damon's and Alphesiboeus' songs, though I am glad to introduce an additional burden from Gud. after v. 28.

In G. 3. 402 M. Ribbeck thinks Scaliger's "exportans" necessary. I have endeavoured in my commentary to prove that it is unnecessary, and that the reading of the MSS. is unobjectionable: but as M. Ribbeck adduces no reasons, I cannot defend myself. "Tussos" G. 4. 62 is highly ingenious, though the credit is really due to Reiske: but "iussos" is perfectly Virgilian, being supported by "monstratas aras v. 549 of the same book, "iussos honores" Aen. 3.547. If any one objects to "iussos" of things which are not yet commanded, he may refer to Aen. 8. 629, "pugnataque in ordine bella,” a stronger case. Here the things are said to be already prescribed, because they will have been prescribed before the action enjoined can take place.

The conjectures introduced into the text of the Aeneid are naturally more numerous. Rutgers' "Eurum" for "Hebrum" Aen. 1. 317 would be quite unobjectionable, scarcely going beyond the licence which may be taken in dealing with the MS. spellings of proper names, if "Hebrum" were not placed beyond doubt by the imitation of Silius 2. 73 foll., and by its evident appropriateness to Harpalyce of Thrace. Huet's "auri" for "agri" v. 343 is highly plausible: but it is natural enough that Virgil should be thinking of a wealthy Roman of his own time "dives agris, dives positis in fenore nummis," rather than of the particular kind of wealth which a Phoenician would be

likely to possess, and to talk of land in one place and gold in another is quite in keeping with his love of variety.

In Aen. 2. 76 M. Ribbeck reads "Quive fuat, memores quae sit fiducia capto." I see no objection to "quive fuat," if only Virgil had written it: but the MSS. tell us he wrote "quidve ferat," which is equally unobjectionable. M. Ribbeck says the Trojans could not know that Sinon had anything to tell: but it was natural to presume that he had, as he had purposely thrown himself in their way. But the rest of the emendation, "memores quae sit fiducia capto," which I suppose must mean "recollecting that a prisoner would require to be reassured," does not seem to me in Virgil's manner. In the common reading "quae sit fiducia capto " means, as Dr. Henry well puts it, "why he should not meet the captive's doom."

In the vexed passage Aen. 4. 436 M. Ribbeck gives in his text "Quam mihi cum dederit, cumulatam monte remittam." The last words he apparently understands "I will repay it with a mountain by way of interest." To promise mountains of gold, or to promise seas and mountains, are, as he says, proverbial expressions in Latin for to make extravagant and excessive promises. But there is no reason to believe that such a colloquial hyperbole would be admitted into epic poetry: besides, M. Ribbeck ought to have remembered that nothing is so hazardous as to attempt to manipulate a familiar proverb by varying the expression. Half the stories which are told of blunders made by foreigners in trying to speak an unfamiliar language idiomatically turn upon unadvised experiments of this sort. I fear if Aulus Gellius could come to life again, he would write a chapter on the stranger who wished to introduce "cumulatam monte" into Virgil.

Dr. W. Wagner thinks M. Ribbeck's conjecture of "funibus " for "finibus " Aen. 5. 139 deserving of much praise. I cannot myself think it plausible. The new expression seems to me slightly harsh: the old is quite unobjectionable. As for the assertion that "fines" is never used for the station of ships in a harbour, what is to be said of it? "Fines" is a word of general and almost universal application for the limits within which ́anything is contained: and no English poet would scruple to use "limit" on a similar occasion.

In Aen. 6. 520 the MSS. make Deiphobus talk of himself as "confectum curis " on the night when he slept his last sleep. M. Ribbeck objects that we have just been told that the Trojans spent their last night in revelry. True: but it was a short revelry succeeding a long agony of care: and in taking, as they thought, their first rest after the departure of the Greeks, they must have had heavy arrears of weariness to get rid of. There is exactly the same feeling in "mortalibus aegris" Aen. 2. 268, where, though the expression is general, the poet evidently intends to excite our special sympathy for the Trojans. Schrader's "choreis" is ingenious: but it is precisely one of those changes which critics make from taking a too contracted view of a subject.

Peerlkamp's "arva " for "arma" Aen. 7. 430 is not at all impossible: but I think it more likely that "arma" after "armari" is a carelessness of Virgil's. On this however there may well be two opinions, and probably my own judgment is biassed by my general belief in the integrity of the MSS. In v. 667 I do not think M. Ribbeck's insertion of "os" after "inpexum" is required by the sense, while it certainly does not improve the rhythm.

In Aen. 8. 211 there is no occasion to couple "raptos" with "tractos" and construct it with "versis viarum indiciis." Those words belong to "tractos:" "raptos" goes closely with "occultabat." At the same time I do not mean to deny the plausibility of Wakefield's "raptor," if we were dealing with an author whose text was more liable to suspicion.

M. Ribbeck complains of Lucian Müller's invective against his emendation “qua vi clausos" in Aen. 9. 67, saying that if he had not been aware that the rhythm intro

duced was an unusual one, he should not have apologized for it. It is really a question of ear and there are doubtless many ears to which the new line will seem hardly Virgilian, in spite of G. 3. 276 and Aen. 7. 634. Via" in the received reading is synonymous with "ratio," as in Aen. 12. 405. "Sic" for "sed" in v. 146 of Aen. 9 depends on a transposition which we have already seen reason to reject. In v. 226 "et," though not found in the MSS., is said to be necessary before "delecta." I do not know what is the objection to taking "delecta iuventus" in apposition with "ductores," but I suppose it is either that the leaders would be too old to be designated as "iuventus," or that the word naturally implies the rank and file, as distinguished from the chiefs. To the first I reply that "iuventus" means little more than fighting men, and that Aeneas and Achates are addressed as "iuvenes" Aen. 1. 321; to the second that Catillus and Coras, who are unquestionably leaders, are called "Argiva iuventus" Aen. 7. 672. V. 403 is critically difficult, as the MSS. vary, and the best supported reading is not the most likely intrinsically; but that seems no reason for introducing a conjecture. V. 676 "freti armis " is unobjectionable, as the opposition is not between arms and personal strength, but between the protection afforded by walls, and that which a warrior can give himself by his use of his weapons. It is conceivable, however, that as in Aen. 4. 11, Aen. 11. 641, and possibly other unsuspected places, "armis " may be from "armi." At any rate we do not need to read "animis."

As to "transiit" Aen. 10. 785, I must refer to the Excursus on G. 2. 81 in the second edition of my first volume. Peerlkamp's "quamvis dolor alto volnere tardet " for "quamquam vis alto volnere tardat " (or "tardet ") is really ingenious; far more so than Hoffmann's "vis alti volneris ardet." The received reading is difficult: "vis,” in Virgil at any rate, is generally used for offensive force, and the intransitive use of "tardo" is rare, though we might give it its active meaning, and say that his physical strength keeps him back by reason of the wound. On the whole I am not sure that the 'perversa ratio" of Servius (as M. Ribbeck calls it) is not right, and that “vis" is not the violence of the wound, as the use of the instrumental ablative instead of the possessive genitive is quite in keeping with Virgil's other manipulations of language.

There is not much force in M. Ribbeck's objection to "acceperit ultro," Aen. 11. 471,"qui accipit sequitur voluntatem alterius, ergo nihil ultra id facit quod voluit alter." A person may be compelled to accept a thing, or he may accept it voluntarily; and it is the latter of these situations in which Latinus would gladly have been. "Asciverit urbi" is better than “acceperit urbi :" the one implies that Aeneas would have been the "gener" of the state (comp. Aen. 11. 105): the latter could only refer to Aeneas' admission within the walls, a much poorer thought. In v. 728 I cannot agree that "iniicit iras" is weak, though Heinsius' "incutit," if Virgil could only be shown to have written it, would be an exceedingly good word. "Iniicio" is a strong word in itself: the only question is whether it can be used idiomatically with "iras,” and that the dictionaries, with their "iniicere metum," "formidinem," &c., set at rest. Last of all is a passage in Aen. 12. 55, where it is said of Amata, "ardentem generum moritura tenebat." M. Ribbeck objects that "moritura" would mean that she was actually going to die, and substitutes "monitura." Is it possible? Virgil, in the rapidity of his passion, says that the queen clung to her son-in-law with the tenacious grasp of one with death before her: the critic says she held him in order to advise or reprove him. Utri creditis, Quirites?

As I said in my former paper, I have no wish to derogate from the undoubted merits of M. Ribbeck's work: but I cannot but think that such criticisms as many of those which I have been noticing are a serious drawback to its value. English scholarship has not a few deficiencies: is it not preserved from some errors by the practice of Latin verse composition ?

JOHN CONINGTON.

ON MADVIG'S EMENDATIONS IN VIRGIL.

In the second volume of his Adversaria Critica (Copenhagen, 1873) Madvig has an important chapter in which he proposes a number of emendations in the text of Virgil. It need hardly be said that the remarks of so great a Latin scholar are in the great majority of cases powerful and suggestive, worth recording and examining even by those who find themselves unable to agree with them in detail: and though it cannot, I think, be considered that Madvig's touch is so sure when he is dealing with poets as when he is handling the text of prose authors, it must at the same time be allowed that any scholar who ventures to differ from him is bound to show good reasons for his difference, and can, in any case, only derive benefit from the necessary process of examination and discussion.

Madvig's general conclusions on the right method of dealing with Virgil's text tend in the conservative direction. "Nam in poeta," he says (p. 30), "ab Alexandrinorum exemplo pendenti nec ipso ad facilem et simplicem orationis formam, ut videtur, facto et exercitato, et, quod ad Aeneidem attinet, in poemate ultimam manum non experto multa ferenda sunt et paulo artificiosius et obscurius cogitata et tumidius dicta verbisque adumbrata et nove durcque posita quaedam." I should feel disposed to develope a little more fully, and more in favour of the poet, this true and terse exposition of the difficulties of Virgil's style. Virgil's language was to a great extent the new creation of a plastic genius, consciously accepted as such by the literary sense of his time and of the century succeeding him: a consideration making, so far as it goes, in favour of the MS. tradition where it does not support absolute nonsense. All poets are to a certain extent innovators in language, the extent of their innovations being determined by their instinctive sense of what the spirit of their language requires : and Virgil's mastery of the Latin language, whether displayed in conservation or in invention, was, I venture to think, greater than that of any other Latin poet. The more need, therefore, of caution before attempting alteration in his text: the less the chance that any given alteration proposed (granting some alteration to be necessary) will certainly correspond with what Virgil himself would have written.

But it is time to attempt some justification of these remarks in detail. I will begin with Madvig's proposal on the last three lines of the third Eclogue—

"Non nostrum inter vos tantas componere lites:

Et vitula tu dignus et hic, et quisquis amores
Aut metuet dulcis aut experietur amaros."

"Non quisquis," he says, " in amore infelix est aut, ne sit, metuit, dignus illo praemio dici posse videtur, sed quisquis eam infelicitatem certo modo fert solaturque, velut cantando lenit, eiusque rei significatio in eo vocabulo, quod dulcis scribitur, inesse debet ineritque, si id nominativo casu acceperimus, altera syllaba in arsi producta . . . Ac languentis genitivo casu, producta in arsi ultima, Vergilius posuit Aen. xi. 69: nominativo autem is produxit in pulvis i. 478, sed in forma nihil ambigui habente.” To make dulcis the nominative, when an antithesis is clearly required to amaros, seems very forced nor do either of the analogies to which Madvig appeals support the lengthening of the final syllable of dulcis. Pulvis has a distinct justification in

: