232 Austin Morris Harmon, 



the poets, even by those of Ammianus' own age.' Nor have I l)een 

 able to find any certain indication of its occurrence in tlie rhythmical 

 prose of Sedulius, Cassiodorus or Ennodius.^ In the grammarians, 

 however, there are some traces of the phenomenon. Velius Longus 

 says,^ aquani qnoqiie per q scribcntes nomen ostendiiinis, per c vera 

 verbum ab eo quod est aciio, utinam acuam. Now if men had to be 

 told to write aqiiam when they meant water and acuam when they 

 meant sharpen, we can fairly infer that they made little if any 

 distinction between qu and cu in pronunciation. The remark of 

 another grammarian,* vacua uoii vaqua, vacui nan vaqui, points in 

 the same direction. And there is still more significance in the state- 

 ment of Consentius,^ u qnoquc litteram aliqui pingnius ecferunt, ut 

 cum dicunt ueni putes trisyllabuui iucipcre. To be sure it is the 

 syllabication of initial ii to which his words primarily bear witness, 

 but it stands to reason that any i)erson who made ueni a trisyllable 

 would have treated Quadi, for example, in the same way. In fact, 

 qu appears to have been much more subject to syllabication than 

 initial u, inasmuch as Ammianus, who makes so free with qu, does 

 not allow initial u to count as a syllable. 



There is then some evidence for the existence of a considerable 

 tendency- in Latin toward the syllabication of qu. For aught I see 

 to the contrary, the tendency may have been especially strong in 

 late Latin. Even if a pronunciation as Inroad as that of Ammianus 

 were general in his time, we could not expect to find adequate 

 recognition of the fact either in the poets, who would follow tra- 

 dition, or in the grammarians, whose concern was always with the 

 written rather than with the spoken language. Tliis consideration 

 undeniably lends much weight to the scanty evidence that we ac- 

 tually find, which, although it is far from strong enough to be de- 

 cisive, is yet sufficient to make us hesitate to ascribe the broadness 

 of Ammianus' pronunciation to Greek influence. 



Another fact still remains to be recorded in regard to Ammianus' 



> Yet the syllabication of n in ciii (e. g. Martial I. 104, 22) is parallel. 



' Subsequent investigation may reveal it in the usage of other writers — 

 juay perhaps reveal also that some writers who do not allow qu to ooniit 

 as a syllable have a tendency to avoid using words containing qti in 

 the clausula, which would be hardly less significant than positive 

 testimou}'- in support of our point. 



» K. VII. 75. 10. 



* Probi App. K. IV. 197. 23. 



» K. V. 395. 15. 



