not anomalies of language, but of thought-the mind choosing for the time, for one reason or another, to regard a thing in a less natural, rather than in a more natural relation-e. g. in v. 263 already noticed, the peculiarity lies not in the grammatical force of the ablative, which is perfectly clear, but in its application to the thing spoken of. This consideration is of importance not only in itself, but as furnishing a principle for judging of the legitimacy of alleged constructions. So long as they were supposed to be mere matters of language, grammarians might multiply them arbitrarily and classify them for convenience' sake: but if we look upon them as belonging to thought, we shall see at once that no other latitude is to be allowed them than that which the mind allows to itself in conceiving of things. Thus the old doctrine of hypallage is in many instances the expression of a truth, as it frequently happens that the relations of two objects (indicated by the cases) may be conceived of as reversed without any injury to the general sense; but it will not require much reflection to show that the present passage could not well be explained by such means, since no ordinary twist of thought would lead a man to look upon a countenance as the instrument, and the blush as the object affected, rather than vice versa. At any rate, with Key's explanation before us, we are not driven to so unlikely a hypothesis.

V. 442. The passages cited by Heyne from Aratus and Avienus show that we must understand medioque refugerit orbe, of the sun retiring from the centre of his disk to the circumference, so as to present a hollow appearance; otherwise we might have accepted Trapp's rendering, "with half his orb retires."

V. 459. Another instance of the ambiguity arising from the mixture of poetry and precept. Do the words frustra terrebere nimbis mean "you need not be afraid of rain-clouds, for there will be none, "there may be clouds, but you need not fear them?"

or

The context rather points to the former, the language to the latter. In a prose writer we should have no doubt that as terrebere is the principal word, the terror and its exciting cause were meant to be the principal things in the sentence; in a poet we feel that though put prominently forward for poetical purposes, they may be intended for nothing beyond the merest accessories.

V. 461. Serenas seems at first sight a strange epithet for nubes, and the balance of the rest of the sentence, vesper serus, humidus Auster, might tempt us to think favourably of any correction which would give the epithet to ventus, such as serenus or serenans. In that case, however, as the proposers of these alterations have found, we should have to alter ventus into Eurus, or some name equally definite, as a counterpoise to Auster, and, this done, we only open ourselves to Heyne's objection, that if the name of a particular wind were mentioned the question unde would be unmeaning. We must therefore retain serenas, connecting it closely with agat, as what Donaldson, speaking of the Greek language, calls a tertiary predicate-agat nubes ita ut serena sint-though a prose author even then would rather write ita ut serenum sit cælum. Ventus, then, instead of being balanced with Vesper, Auster, becomes comparatively an unimportant word, the real stress falling on nubes. V. 469. Quamquam serves as a kind of apology for the very decided transition already begun, from the signs of the sun as concerning the rustic to his historical prognostics, thence to other portents occurring at the same juncture of events, and finally to a general survey of past and present politics. "And yet, as for portents, the sun did not stand alone: there were others," &c.

V. 490. It would seem more reasonable, with Heyne, to take iterum with concurrere than with videre" Philippi has seen a second battle, as Pharsalia did the first:" the passages, however, which he

himself quotes, especially those from Manilius and Petronius, are strong in favour of the more exaggerated view, which makes Philippi to have been the scene of two battles, unless we suppose Virgil's imitators to have mistaken his meaning. Something must be allowed to rhetorical necessity and probably the geographical inconsistency, though not unperceived, would startle a Roman of the time of Augustus much less than any educated man of the present day.

V. 506. Non ullus aratro dignus honor aratrum dignum habetur honore.

=

=

nullo

V. 513. The choice seems to lie between addunt [se] in spatia dant se, like adde manus in vincla meas, Ovid. i. Amor. Eleg. vii. 1. and addunt [aliquid] in spatia, like in scelus addendum scelus est, Öv. Met. viii. 484.

ON SOME PASSAGES IN ÆSCHYLUS AND

SOPHOCLES.

IN Æsch. Prom. 711, ἑκηβόλοις τόξοισιν ἐξηρτημένοι, Enprvμévo has been substituted from one MS. by Blomfield, Dindorf, and Paley, with the approbation of Wordsworth, in Philolog. Museum, i. 243, who remarks, that such instances as πώγωνας ἐξηρτημένας (= suspensi loculos) cannot justify the construction with the dative in a similar sense. The question is whether τόξοισιν ἐξηρτημένοι can stand for τὰ τόξα ἐξηρτημένα ἔχοντες, or whether the dative is a bar to such a construction. A precisely similar instance is found in Soph. Ed. T. 3, ἱκτηρίοις κλάδοισιν ἐξεστεμμένοι, where the numerous parallel passages quoted in Wunder's Excursus may convince us, as they might have convinced him, that Matthiæ is right in considering the dative equivalent to the accusative, as the sense must be “ having boughs wreathed [with wool]."

In Prom. 814, τὴν μακρὰν ἀποικίαν, μακρὰν has, I believe, been taken by all the commentators as an adjective agreeing with amouкíav. Hence they have been anxious to establish that μakpòs means distant, Blomfield quoting Laudaturque domus longos quæ prospicit agros from Hor. Ep. i. 10. 23, where longos agros means not distance, but simply a length of country, while Liddell and Scott endeavour to persuade themselves that such expressions as μακρὰ κέλευθος, μακρὰ βιβὰς, μακρὸν αὐτεῖν are in point. The simple truth is that μaкρàv is an adverb, and that ifÆschylus had expressed himself in the nominative, he would have written not ἡ μακρά, but ἡ μακρὰν ἀποικία. The same mistake has been committed in a precisely similar

instance, ǹ μaкрàν пóλш, Sept. 613, where the editors have been still more perplexed, Blomfield taking T μακρὰν with πομπὴν, Paley most unaccountably connecting it with both Toμm and Toλìv, and Dindorf rejecting the whole line as spurious.

In passing I may remark that in Prom. 693. ἀμφάκει κέντρῳ ψύχειν ψυχάν, where Dindorf now reads ψήξειν after Meineke's ψήχειν, the old reading is amply confirmed by Eum. 157. μεσολαβεῖ κέντρῳ

....

πάρεστι

μαστίκτορος δαΐου δαμίου βαρὺ τὸ περίβαρυ κρύος ἔχειν. Soph. Cd. Τ. 31. θεοῖσι μέν νυν οὐκ ἰσούμενον σ ̓ ἔγω οὐδ ̓ οἵδε παῖδες ἑζόμεσθ ̓ ἐφέστιοι. All the commentators seem agreed in governing σe by Kρívovτes, v. 34. "illud sponte suâ quisque intelliget," says Wunder. It seems at once more natural and more idiomatic to make it the accusative after ἑζόμεσθ' ἐφέστιοι = λισσόμεθα. looúμevov will then resume its ordinary sense, that of equalled or reckoned equal, instead of being taken to mean no more than ἴσον.

174. ἰηΐων καμάτων ἀνέχουσι γυναίκες. We may account for this construction by supposing that as avexew is used here in the sense of λήγειν Or παύεσθαι, it takes the case which they would naturally govern-a principle just exemplified in the foregoing note, and of very wide application to the Greek language, which determines the regimen of cases by the sense of a verb at least as much as by its form. See an article on the Metaphysics of Greek Syntax, by W. J. Hickie, in No. XXV. of the Classical Museum.

325. ὡς οὖν μηδ' ἐγὼ ταὐτὸν πάθω. Those of the editors who notice these words at all acquiesce in the explanation of the Schol. who supplies on, Wunder throwing out an additional suggestion that the speech may be considered as abruptly broken off. It is hardly credible that no one should have seen that it is an elliptical expression with a quasi imperative force, such as would usually be filled up by a word like σKETTÉоV. The attempt, however, made by Dindorf and Wunder to alter a similar passage, Ant. 215.