202 Austin Morris Harmon, 



We have seen that there is no essential distinction in Ammianus 

 between the clausulae which end sentences and those which end 

 interior kola. In both cases the same forms and types occur. "We 

 fomid, however, that in the interior clausulae he displayed somewhat 

 more laxity in his observation of positional quantity than in the 

 final clausulae because it is the final clausulae that dwell longest 

 in the ear. And we noted in passing another difference which was 

 reserved for fuller consideration here, in the point that neither in 

 the relative frequency of the forms with reference to one another, 

 nor in the relative frequency of the various types of each form is 

 there close agreement between the statistics furnished from Book 

 XXI and those from the collection of sentence-endings. As to the fre- 

 quency of the forms the statistics are here expressed in percentages. 



Following the hint given b}' this comparison I have examined 

 the facts in Book XXI ; they are 



Interior Final 



I 45.6 45.7 



II 30 11.1 



III 20.9 40.4 



IV 3.3 2.6 



The relative proportions of the forms are obviously the same in 

 the final clausulae in this book as in the other nine books, though 

 I is slightly less frequent, IV slightly more frequent than usual. But 

 the relative proportions in the interior clausulae are quite different. 

 Form II shows a great increase, form III a great decrease in fre- 

 quency, while I and IV remain aliout the same. We find, then, 

 that II is preferred over III before minor pauses and that III is 

 much more favored than II as fiual clausula. If we pursue the in- 

 vestigation into the typology of these cadences, it becomes clear 

 that thr falling off in III as an interior clausula is all in type 6 

 icollibiis abdiderunt) ; the falling off in II as a final clausula is marked 

 in all t3i)es, but greatest in () (opifcx callidus)} Furthermore, though 



* The fact that II ('lausulae with f caesura (v/vus exiistus est) are more 

 comniou at the end of the sentence than in the interior has uo signif- 

 icance; it is simply because verbs are more frequent at the end of the 

 sentence. 



