² Met. 6. 536Google Scholarconiugialia iura, 7. 715Google Scholariura iugalia, Am. 3. 6. 82Google Scholarsocii iura … tori, 3. 11. 45Google Scholarlecti socialia iura, Her. 6. 41Google Scholarconubialia iura, 9. 159Google Scholariura sacerrima lecti, 16. 286Google Scholarlegitimi … iura tori, 21. 140Google Scholarpolliciti debita iura tori.

³ Note especially Her. 2. 94Google ScholarPubMedoscula iungere, 4. 147Google Scholarfoedera iungere, Trist. 2. 536Google Scholarfoedere iunctus amor, 3. 4. 44Google Scholaramicitias iungere, Ars 1. 492Google Scholarsodas tu quoque iunge moras.

¹ Cf. Housman, C.R. xi (1897), 104.Google Scholar

² Madvig, sese avetGoogle Scholar, Koch, se iubetGoogle Scholar, Birt, se pavet (!)Google Scholar, Sedlmayer, se vetat (!), conjectures fit only for oblivion.Google Scholar

¹ Something similar has probably happened at 12. 65 orat opem Minyis: alter petit alter habebit: so P; petit altera et altera habebit P2Gς. Housman, (loc. cit., p. 286) suggested that habebit was ‘tacked on at the end’ to fill up a metrically defective line.Google Scholar

¹ Bentley appears to have made a similar conjecture, but apparently wanted to write Quid? quasi nescires …, which by disjoining quid from the pentameter stultifies the correction.

¹ I had thought that we might have here yet another corruption of numquid: (at 16. 366 non stands in the MSS. and the correct form num is preserved only in the margin of P; at 16. 367 numqnid has produced non quid in two MSS. and numquam in several others; and the same has probably happened at 7. 45, where Shackleton Bailey conjectured numquid for quid (quod) non of the MSS.; at 21. 177 the Parmensis has gone half-way with nun quid). Certainly corruption of the rare numquid would help to explain the proliferation of variants; but in Ovid numquid never means more than num, and quid has not the pronominal force which would be demanded of it here.

² Compare Am. 3. 5. 33 f.Google Scholarnoctumae sic dixit imaginis augur, / expendens animo singula dicta sua, followed by a catalogue; Tr. 3. 5. 13Google ScholarPubMedet lacrimas cemens in singula verba cadentes, ‘tears falling at each individual word’; also Ex P. 1. 5. 20, 4. 6. 34.Google Scholar

¹ It is difficult to retain tuum (‘your wound = the wound you caused me’): I know no parallel in elegy. Once vulnus became vultus the change to tuos was inevitable. For similar corruptions at the end of the line cf. 7. 124 suis PEG²ς tuis D meis E²Gς, 20. 76 sui P suis G meis w tuis ς, 21. 52 tuis or meis.

² In the text I give Housman's improvement of the MSS. reading prima fuit vultus nuntiafana tui. Vulnus belongs to Palmer.

³ I thank the editors of the Thesaurus Linguae Latinae for so promptly answering my request for information from their unpublished article on multus. Having thus supplemented my collection with two further examples I think that the list I give is virtually complete.

¹ Doubtless the rule applies in some degree to other accusative adverbial forms. I simply cite two examples found in Wölfflin: Val, . Max, . 4. 11 quantum domo inferior tantum gloria superiorGoogle Scholar; Veil, . Pat. 2. 11 quantum hello optimus tantum pace pessimus.Google Scholar But at Livy, 10. 35. 2senserunt quantum in sua parte plus volnerum ac caedis fuisset, quantumGoogle Scholar is probaly to be explained by the fact that plus is used as a noun. One might add Val, . Max, . 5. 9. 3: in filo aliquantum taetriore.Google Scholar