The Clausula in AiuDiianus Marcellinus. 209 



whereas in others it has an accented penult (I. 105. 14 exinde 

 acciderat; 248. 13 exinde lovinum; 293. 5 exinde transferri\ 332. 8 

 exi'nde perductf). 



The statement of Priscian that aliquando was pronounced with 

 antepenultimate accent is not l:)orne out hy the usage of Ammianus.i 

 Our author accents the word on the penult: so I. 46. 16 aliquando 

 nee lacessitus; 260. 31 aliquando est visus; 263. 16 aliquando audivit; 

 II. 234. 5 aliquando contingit; 242. 16 aliquando Roinanuni.'^ 



We learn from ancient authorities that when the enclitic -que is 

 attached to a word, the final syllable of that word is accented 

 even if it be short.-' For this I have noted the following evidence 

 in Ammianus,— I. 196. 3 pkraque correxit (cf. 296. 19 and II. 147. 27); 

 11.1 felicitateque correxit; 101. indeque cindoriim; 168. 29 crudeli- 

 taieqne terrebat; 179. 15 telaqiie dispositis; II. 176. 28 sperandaque 

 similia. 



In patefacit and venumdata we have two verbs with accented 

 short penult. The one occurs only in the clausula patefacit ingentes 

 (I. 315. 13), where, however, it is certain, for the sense of the 

 passage excludes the possibility of altering to patefecit. It is sup- 

 ported not only by the testimony of Priscian, but by the fact that 

 in this and kindred words there is no such vowel-weakening as in 

 afficif, for example.* For the accent of venumdata I have found 

 no direct evidence from the grammarians ; it occurs in the clausula 

 four times to my knowledge (II. 129. 12 gentibus venumdata; 213. 6; 

 227.21; 247.19), and is to be contrasted with pessihndata (11.93.12 

 pessiimdatae visae sunt). 



A .similar peculiarity is to be observed in the penultimate ac- 

 centuation of qui I i bet and its diiferent case-forms. Thus I. 107. 30 

 quilibet tunc praesens; II. 147. 1 interftcere cuiuslibet; 136. 21 quahbet 

 inventa; 237. 8 quolibet occiso ; 236. 16 transenntes quolibet. 



Parallel to quolibet is quovfs, which is to ])e found in the sentence- 

 ending II. 91. 10 internecivo bello quovis graviorem. I have not 

 found any other instances of this pronoun in clausula. 



To the best of my knowledge nihilominus does not occur in 

 Ammianus; on the contrary, m'hilo minus serves as a clausula in 

 I. 250. 10. 



> Inst. XV. 29; cf. XV. 10. 



* Hence II. 105. 21 aliquando Romanorum is to be emended into ali- 

 qiia7ido Romanum, comparing the instance last cited. 



» E. g. Priscian, Inst. V. 63 and Part. 26; cf. Sommer, p. 101. 



* Inst. VIII. 35 ; Part. 127 ; cf. Sommer, pp. 102—103. 



