has been seriously detrimental to the thorough knowledge of Latin compound construction. Geo. i. 415: Haud, equidem credo, quia sit divinitus illis So C. prints, and says, 'it seems better to follow Reiske in pointing "haud, equidem credo" than to keep the common punctuation "haud equidem credo." Equidem credo is thrown in modestly.' This is strange: for, in any case, the reason denied is denied by one who believes the denial just, and that could only be the poet here. Haud equidem occurs four times elsewhere in V., non equidem thrice, and in no place is there any stop between the two adverbs. But he gives no account of the mood insit.' This is left for his reader to discover. Non quod, non quia, it is true, often take a subj. where no verb appears. But why? because the denial of an invalid reason contains in itself a virtual oratio obliqua, after which usually follows an adversative conjunction, as 'verum' here, with the true reason. But here, where the denial is accompanied with the verb expressing the poet's personal belief, for my own part I do not believe (that the rooks caw and rustle in the boughs because &c.: but &c.'), the words of the oratio obliqua may be at once supplied; viz. 'ingeminare &c. et strepitare' &c. See Append. II. to 'P. S. L. Primer.' Aen. v. 621: Cui... fuissent. Here, taught, as he owns, by Jahn and Forbiger, he has a glimpse of the true reason of the mood: 'it makes us think of Beroe as Iris thought of her.' Yes it makes V. say virtually of Iris: voluit se esse eam Beroen, cui sent.' This is a good instance of virtual oratio obliqua. Aen. vi. 199: illae tantum prodire volando Quantum acie possent oculi servare sequentum. fuis Here, again, he says, 'possent is rightly explained by Forbiger, as indicating the object of the doves in flying onwards.' Yes: but if asked how does 'possent' indicate this, could he have answered, 'because its mood shews it to be in dependence on a virtual oratio obliqua, i.e. that the hist. infin. prodire prodire se voluerunt?' No: for he had not a clear rule in mind enabling him to do so. = Aen. viii. 129: Non equidem extimui Danaum quod ductor et Arcas C. says: 'fores seems to be used on the analogy of those cases where quod with subj. gives a reason which the speaker denies to be the true one, though what is denied here is not the reason, but the fact which the reason might have justified.' On the ground so stated the analogy fails altogether. Here is contained a virtual oratio obliqua, non extimescendum mihi putavi.' Aen. viii. 650: Illum indignanti similem similemque minanti C. merely says: 'auderet the subj. expressing Porsenna's feeling.' Too curtly stated. The subjunctives depend on words expressing Porsenna's feeling: indignanti, minanti = indignari se monstranti, minari se monstranti. Aen. ix. 289: Quod nequeam. In my preface to the 'Public School Latin Grammar' 'Ed. 4), I have pointed out the error of Madvig and Conington Lespecting this passage. See also my note upon it showing that 'quod' depends on the virtual oratio obliqua contained in testis (me hanc insalutatam linquere). In nine other passages cited by me in V. Syntax the mood is not touched in Conington's notes. Among these 'possent' in Aen. i. 367 is a very notable omission. See Public School Latin Primer,' p. 160. (3) Indirect will-speech (Petitio Obliqua). More than sixty Virgilian instances of this construction are cited by me (pp. 666-7). C.'s notes give no account of it. Here and there an instance of it is called oratio obliqua, a generic term, which does not distinguish it from indirect statement or indirect question. See 'P. Sch. Lat. Primer,' App. II., and Preface to 'Public School Latin Grammar,' Ed. 4. He nowhere shews that this, or any clause of oblique oration, is sometimes object of a verb (as of sino &c.), sometimes subject (as of licet). His non-appreciation of it is shown by printing in Aen. v. 796, 'quod superest, oro, liceat': and in Aen. iii. 129 hortantur socii: Cretam proavosque petamus,' with this note: 'hortantur seems to mean that they encourage each other . . . Cretam proavosque petamus is doubtless meant to give a notion of sailor language.' Thus he ignores the real construction by which 'petamus' simply depends as object on 'hortantur': 'our comrades exhort that we make for Crete and our ancestors.' At other times he seems to have something like a right idea of the relation between verb and object. Thus in Aen. iii. 457 he supports poscas ipsa canat; in vi. 76, ipsa canas oro, and in xi. 442 he prints very properly et vocet oro. No clearheaded grammarian can venture to deny that a terminal distinction ought to be made between the three species of oratio obliqua, statement, question, and will-speech, which by historians are so frequently exhibited as occurring in the same Oblique Narration. See 'Primer,' p. 158. Take the shorter example of this zeugma, Aen. iii. 234 : Edico, et dira bellum cum gente gerendum. So C. prints; but the commas in the former line should be removed. His note is merely: 'Comp. x. 258.' Turning to x. 258, what do we read? Principio sociis edicit, signa sequantur, Atque animos aptent armis, pugnaeque parent se. (where commas are better away). Here we find only indirect will-speech, while C.'s note refers back to iii. 234. So that in iii. he takes no notice of the zeugma, by which the indirect statement bellum gerendum (esse)' is bracketed with the indirect will-speech 'arma capessant' under the government of the single verb 'edico.' The feebleness of aesthesis in higher syntax could not be more clearly shown. His treatment of the three following places deserves remark. praemittit Achaten, Ascanio ferat haec, ipsumque ad moenia ducat. His note is 'ferat-ducat are apparently to be explained not as depending on praemittit, as in "volo facias," but as an oratio obliqua, Ascanio fer ipsumque duc.' (See below.) Aen. viii. 506: mandatque insignia Tarchon, Succedam pugnae Tyrrhenaque regna capessam. The note is: 'succedam-capessam, an oratio obliqua, expressing the words of the charges.' Aen. xi. 513: Arma praemisit, quaterent campos. Note: praemisit implies an order, and so is followed by an oratio obliqua: comp. i. 645, where the distinction attempted in the note is nugatory, the two constructions being really the same.' This censure of his own note on i. 645 proves that between B. i. and B. xi. he had made some progress in Latin compound construction: but the defective teaching of Madvig still kept his knowledge of this subject on too low a level. (4) In his first volume C. has several times called the Thought-mood Conjunctive: but in his Index (writing 'Conjunctive: see Subjunctive') he seems to have succumbed to the mischievous practice of those writers, who use only the latter name. If we look at such passages as: Quid faceret? quo se rapta bis coniuge ferret ? Geo. iv. a and hundreds like these, we cannot see what justification can be found for calling verbs subjunctive where they are neither subjoined nor subjoining in any sense whatever. To do so is merely to set at defiance all propriety in the use of terms. Such a solecism was not committed by English writers half a century ago, when they divided the Thoughtmood into two uses, Potential and Subjunctive. It is not committed by Madvig or any of the German writers now, who agree in using the term Conjunctive alone. They, however, err, and so fall into confusion, from not using the term Subjunctive in its proper place, where the verb is subjoined, i.e. in dependence. I have avoided both errors, by calling the Mood Conjunctive or pure Conjunctive in such places as those last cited, but, when dependent, Subjunctive. This ungrateful topic I leave with an earnest suggestion. The defects of grammar in Prof. Conington's notes on Virgil speak for themselves. They are undeniable: but they are not irremediable. His literary representatives ought to find a remedy; and they might easily find it. If there is any special difficulty in varying the notes, an Appendix of Grammar might be added, and reference to this made in every place where a grammatical note is required. Nobody would have more pleasure in knowing that such a measure had been adopted, or was in progress, than he who has felt so much pain in drawing attention to the requirement. As to Prof. Madvig, his learning and textual skill deserve the honour they have won: but his teaching of Latin mood is best described by Virgil's lines, Aen. vi. 270: Quale per incertam lunam sub luce maligna |