44 Journal of Philology. older writers love to heap together words of the same or similar meaning ; thus iv. 624, I should retain sudantia, and at the end of v. 586, supply ignis with the old editors, not Jlammce, and I, 720, keep undis, and n. 942, omnituentes and quamque tuentur, an intentional repetition; so vi. 1260, I unhesitatingly reject Lachmann's insipid labes, and keep languens ; then in the pre- ceding verse I read ex agris is mceror in urbem, for the is has either been absorbed in agris, or by that frequent error in our MSS. is and mceror have been transposed : n. 387, and vi. 909, I retain ortus, as quite Lucretian. vi. 1281, I fill up this verse by reading " pro re prcesenti mcestus," the prcese having been lost in pro re, the nti in the m of mcestus. Bernays has, I think, not unfrequently been misled by too great a deference for Lachmann ; often however he has rightly restored the readings of older editors, often he has introduced true emendations of his own ; sometimes he has deserted Lach- mann without reason in my opinion, for instance in I. 412, ii. 428 and 381, where the corruption est tali, est ali, est aii is quite obvious; vi. 1175, where mersans appears to me to have much force. On the other hand, not to mention cases in which he only recalls older readings, he is frequently right in new readings of his own, as, to mention a few only, n. 1089; in. 694; iv. 761 and 1282; in. 45, 286, 527 and 1135. He shews good judgment too in rejecting some of Lachmann's transpositions. I am glad to see that he arranges I. 805 807 in the manner I had myself done long ago ; here Lachmann preposterously reads ambusta for arbusta, transposes two verses, and thus turns into an agent of destruction one of the four elements, the combined influence of which is necessary for the argument. Bernays like- wise rightly sees in some cases that there is a hiatus, where Lachmann has failed to observe it, as v. 29 31. Long ago, however, I had arranged these verses in a different order from any other editor, and for the following reasons. I believe that after 28 followed a verse completing the sense of 30, and that it began with Quid, like 28, and thus, by a usual blunder, oame to be omitted; that then, in order to restore the appa- rent grammatical construction, 29 was put before 30. I would read as follows, intending of course by- the verse supplied, only to shew the general meaning of the sentence : Quidve tri pectora, &c. [Quid volucres pennis seratis invia stagna] Tanto-
