^{page 216 note 1} Conington repeats this.

^{page 217 note 1} So E. vi. 39, ix. 48, G. i. 80, 223, ii. 539, iii. 24, 357, A. i. 603, ii. 665.

^{page 217 note 2} So G. i. 370, iii. 133, 428, A. v. 713.

^{page 217 note 3} So E. vii. 51, G. iv. 412. G. iii. 271 continuoque auidis ubi subdita flamma medullis| (uere magis quia uere calor redit ossibus), illae| etc. may be thought only just to escape the rule by virtue of continuoque. This half-exception is in the latest of these books.

^{page 217 note 4} So G. iv. 305, A. i. 341.

^{page 217 note 5} G. iv. 258, 265, 442 are, I think, rhetorically, though not grammatically, parallel.

^{page 217 note 6} Cf. Bentley on Lucan I. 253. Perhaps A. iii. 433 praeterea si qua est Heleno prudentia, uati | si qua fides should be included here.

^{page 217 note 7} So ii. 458, iii. 151, 219, 695, ix. 512, 692.

^{page 217 note 8} So ii. 217, vi. 117, xi. 429, xii. 355.

^{page 218 note 1} The fact that the line belongs to a cancelled passage is probably irrelevant.