A PERSISTENTLY TROUBLESOME QUESTION IN LATIN PHONOLOGY has been the accentuation of enclitic composites.' One consequence of this problem is the uncertainty it may cause in the pronunciation of classical dactylic hexameter, where the separable enclitics (-que, -ve, and -ne) are to be found in abundance. The present study attempts to review and clarify the areas of phonetic controversy through a metrical analysis of all enclitic rhythms in the poetry of Vergil. Particular attention will be given to the putative regression of accent that may occur when an enclitic is followed by elision. Since late antiquity, there has always been some theoretical confusion about the precise nature of the Latin accent shift caused by the addition of these common enclitic particles. Most modern scholars readily agree that, in the classical period at least, all final heavy syllables must have received a syntactical accent before the enclitic, regardless of the isolate accentuation of the base-word: multtumque, multosque, multaeque, imperiumque, imperi6que, etc.2 This principle could be inferred from the metrical evidence of Augustan Latin literature even if it were not confirmed by Roman grammatical theorists.3 The elements of uncertainty occur in three main areas: (1) the accentuation of enclitic composites in which the base-word ends in a light syllable (multaque, limindque, armentdque, Saturn'dque, etc.); (2) the potential tendency for elision to cause a regression of accent within the enclitic composite; and (3) the possible dynamic effect of the verse-ictus upon the accent of the composite. The first problem of the trochaic and dactylic base-words has arisen because of conflict between the testimony of some imperial Roman grammarians, who appear to have advocated the accentuation multdque/