^{page 334 note 1} I am deeply indebted to Professors K. J. Dover and G. W. Williams for their help.

^{page 334 note 2} In Sat. 2. 7. 93 f. an adjective standing by itself takes the place of an unqualified noun.Google Scholar

^{page 334 note 3} Notice also -que as second syllable of the fifth foot in Sat. 2. 1. 25,Google Scholar 7. 94 and A.P. 177; this occurs below in, e.g., Virg, . Aen. 6. 765, 9. 338, 2. 636, and is very common in Latin hexameters.Google Scholar

^{page 334 note 4} C.Q. N.S. xvi (1966), 140 ff.Google Scholar

^{page 334 note 5} Ibid., 149f.

^{page 335 note 1} Cf. C.Q. N.S. xvi (1966), 140.Google Scholar

^{page 336 note 1} Cf. C.Q. N.s. xvi (1966), 145.Google Scholar

^{page 336 note 2} Cf. Hor, . Epist. i. 1. 90 (p. 335).Google Scholar

^{page 336 note 3} Winbolt, S. E., Latin Hexameter Verse, 167.Google Scholar Cf. Nillson, N.-O., Metrische Stildifferenzen in den Satiren des Horaz, 12:Google Scholar ‘Daβ zu alien Zeiten (of hexameter verse) die Elision von kurzen Vokalen als die leichteste, die von langen als die schwierigste gegolten hat, geht aus dem statistischen Material unmittelbar hervor.’

^{page 337 note 1} I regard the line as spurious (not because of its word order) but cannot discuss it fully here. Its formal similarity to Aen. 10. 232Google Scholar and 9. 450 (note potiti) suggests that its author modelled it on them. This procedure implies an intimate knowledge of Virgil; cf. Havet, L., Manuel de critique verbale, §§ 1082 ff., who shows that many corruptions in Virgil's capital manuscripts imply such knowledge.Google Scholar

^{page 337 note 2} I follow Klotz, A. (Teubner 1908);Google Scholar the best manuscript P has, according to him, sese trepido, according to Garrod, (Oxford, 1906), trepido sese.Google Scholar

^{page 338 note 1} portitor and Charon enclose the clause; cf. C.Q. N.s. xvi (1966), 140 ff.Google Scholar

^{page 338 note 2} Mus. Helvet. i (1944), 234 ff.Google Scholar Cf. Dover, K. J., Greek Word Order, 52.Google Scholar

^{page 341 note 1} Except in Hor, . Od. 3. 15.Google Scholar 2 and Plaut, . Tri. 642.Google Scholar

^{page 341 note 2} For the bringing out of a contrast by the separation of adjective from noun cf. C.Q. N.s. xvi (1966), 168 ff.Google Scholar

^{page 342 note 1} Cf. Hor, . Epist. i. 1. 90 (p. 335).Google Scholar

^{page 343 note 1} An adjective standing by itself takes the place of an unqualified noun, as in Hor, . Sat. 2. 7. 93 f. (p. 334).Google Scholar

^{page 343 note 2} In the first 200 lines of Books 1 and 9 there are 94 and 101 spondaic first feet respectively.

^{page 343 note 3} This is no doubt a recollection of Virg, . Geo. i. 328 f.:Google Scholar ‘ipse pater media nimborum in nocte corusca / fulmina molitur dextra’… Though we need not therefore take corusca to agree with fulmina; the choicer alternative is; to take it with dextra (Fraenkel, , Horace, 244).Google Scholar

^{page 345 note 1} Though the subject of an need not come first; contrast, e.g., templum in Sil, . Ital. i. 81 ff. (p. 339).Google Scholar

^{page 345 note 2} Cf. Hor, . Epist. i. 1. 90 (p. 335).Google Scholar

^{page 345 note 3} In the first 300 lines of In Eutr. 1 there are 11 o spondaic first feet.

^{page 346 note 1} Cf. the many examples from Silius in the second section.

^{page 346 note 2} Winbolt, S. E., Latin Hexameter Verse, 40 ff. We may here ignore the unusual form cessatum super imperio, with no caesura in the third foot.Google Scholar

^{page 346 note 3} In the first 200 lines of the following books, where there is punctuation at the hephthemimeral caesura in Corp. Poet. Lat., the numbers of forms A and B are: Virg. Aen. 1, 20 (A): 5 (B), Aen. 5, 15: 11, Aen. 7, 14: 6, Aen. 8, 24: 8, Aen. 12, 23: 12, Sil. 11, 15: 16, Sil. 12, 14: 10, Lucan 7, 4: 13, Val. Fl. 2, 4: 21, Stat. Th. 5, 7: 16, Claudian Rapt. Pr. 2 (Loeb), 11: 24. (I give figures from several books of the Aeneid, in order to throw doubt on Winbolt's statement (p. 42) that ‘Virgil in the Aeneid is about equally divided between the two forms’.)

^{page 347 note 1} Cf. on Epist. i. 1. 90 on p. 335.Google Scholar

^{page 348 note 1} For the proper name in the final position cf. on Virg, . Aen. 7. 649 (pp. 338 f.), though velox … Faunus is not a descriptive sentence inserted into a narrative.Google Scholar

^{page 349 note 1} Cf. Hor, . Epist. I. 1. 90 (p. 335).Google Scholar

^{page 349 note 2} Here a noun is separated from a genitive.

^{page 350 note 1} Cf. p. 343, note 2.

^{page 350 note 21} Cf. Petr. 120. 61 tres tulerat Fortuna duces, quos … (not: … tres F. duces tulerat’).

^{page 351 note 1} Here a noun is separated from a genitive.

^{page 351 note 2} I have not counted examples in which the noun might have been separated further from its adjective, but would not have been the last word.

^{page 353 note 1} Landgraf, on Sex. Rose. § 116.Google Scholar

^{page 354 note 1} Blass-Debrunner-Funk § 385.