The Clausula in Animiatius MarcelUmis. 217 



Galatia II. 87. 6 



Lycia II. 87. 10 



Asia II. 125. 15 



Francia II. 208. 22 



Thracia II. 240. 20 



Pamphylia II. 245. 26 



Macedonia II. 245. 27 



Moesia II. 252. 18 



This practice is of course due not only to the long usage of 

 Greek names in -m like Mesopotamia, but to their analogy with 

 names purely Latin like Italia, Graecia. It would have been utterly 

 impossible to preserve any rational distinction. The clausula The- 

 baidem et Libyain (I. 302. 17) shows that Libya (Aii-iv)i) belongs in 

 this category. 



From Eusebia (I. 94. 15; 12G. 11; 240. 7j, ecclesia (II. 160. 3) and 

 philosophia (I. 326. 14) we may perhaps conclude that not only 

 names of countries but also all other Greek loan-words in -ui were 

 accented on the antepenult by Ammianus.^ No such convenient and 

 sensible rule prevailed in later times- -in fact, so great confusion arose 

 that even native Latin words in -ia occasionally had penultimate 

 accent. 2 



There are in Ammianus a few compound words with exceptional 

 accentuation for which the reason is not clear to me : 



theologi I. 127. 16 



philosophus I. 298. 20; II. 169. 19 



lotophagi I. 22. 6 



galactophagi I. 333, 14 



Beyond these exceptions I have observed only the following : 



Helena I. 279. 22 



Homerus I. 296. 32 ; 276. 6 ; 257. 27 ; 334. 15 



Milonem II. 221. 2 



Theodoras II. 163.5; 169. 1 



Nicaea II. 82. 13; 83. 26; 89. 16; 111. 18 



Homerus and Helena may perhaps be accounted for as names known 

 mainly from books and Latinized through long currency in the 

 speech of educated men : they are not quite on a par with names 

 like Alexander, for instance, which on account of its frequency in 

 every-day use could not be Latinized so easily. A Greek slave 

 would probably have answered more readil}^ to Alexander than to 

 Alexander. The pronunciation Milonem as against Platonem may 



> Cf. K. III. 522. 9. « E. ff. I'ltalie, from Italia. 



