224 



These cases do not call for extended comment, since the phen- 

 omenon which they exemplify is already well known from the oc- 

 casional practice of the poets, the testimony of the grammarians 

 and the evidence of the Romance languages.^ It is remarkable only 

 that instances of synezesis are relatively so infrequent in Ammianus.^ 

 Examples of dialysis are much more common. 1 submit first a 

 number of instances in which intervocalic /' appears and counts as 

 a syllable. 



Trai'anus II. 203. 25; 248. 12; 249. 8; 260. 1 



Troia 1. 279. 21 -^ 



ai'unt 1. 282. 8 ; 248. 21 



Aquileia I. 247. 20 ; 248. 15 



Pompeius 1. 87. 4 ; 11. 189, 24 



Fonteius I. 75. 10 



plebeius 1. 232. 15; II. 98. 19. 



ei'us I. 258. 31 ; II. 90. 8 



This pronunciation of intervocalic / is occasionally, though rarely, 

 exemplified in the usage of the poets.* In Ammianus, however, not 



> See Miiller de Re Metr. p. 249 ff. : Lindsay, p. 142 ff. : Sommer, 

 pp. 144, 145. 



* It is possible that some cases which I have listed under forms II and 

 IV {adulatioiie Jlagrantintii : decerneret in pla7xitie) should be read with synezesis 

 as I and III. But I liave preferred not to assume it when unnecessary. 



* Priscian K. II. 287. 8 classes Troia with Italia. Phrvgia, Lydia, etc. Cf. 

 Part. K. III. 467. 12. 



* In pure Latin words it is very rare : see L. Miiller, de Re Metr. 

 p. 264. Harper's Lexicon (and others') register Pompejus {trisyl.) or Pompeius 

 [qiiadrisyl.)., plebeiics {-ejus) and Fon tents— on what grounds I know not. 



