The Clausula in Animiamis Marcellimis. 205 



141,4; 142,2,20; 143,3,32; 145,32; 146,17,21; 148,17; 149, 8; 

 150,34; 153,6, 11; 155, 8, 11; 156, 8; 157, 21; 161, 7, 9, 18; 162, 16; 

 163, 6, 13; 164, 34; 165, 1, 8, 11, 22. 



Instead of occurring regularly once in every page, the cadences 

 are mostly in groups ranging from 2 to 7 clausulae. But there 

 are also numerous more or less isolated occurrences. The two facts 

 point at once to the predominance of responsion in Ammianus, and 

 to the lack of S3'stem in it. 



Responsion of course brings it about that in short passages of 

 Ammianus' text the relative proportions of the clausulae vary con- 

 siderably. But any passage of three or four pages will be found 

 to reproduce quite closely the ratios which occur in Book XXI, and 

 which may be mentioned again here as the general norm of use in 



Ammianus : 



I 46 



II 27 



III 24 



IV 3 



In proof of this consistency in practice I cite the relative per- 

 centages of the cadences in the sentence-endings of the nine books 

 that I have examined. Being linal clausulae, they naturally do not 

 correspond to the ratios in continuous composition, but the uni- 

 formity which exists in them testifies to a like uniformity in general 

 st3le. 



XIV XV XVI XVII XVII XIX XXIX XXX XXXI Average 



I 48.8 51.2 56 46.7 44 42 49 53.9 52.1 48.8 



II 9.5 12.1 9 14.6 11.2 8.9 11.7 11.8 11.7 11.3 



III 41.1 35 34.3 37.5 41.6 46.8 36.2 34.2 33.9 38 



IV 0.4 1.5 0.5 1.1 3.2 2 2.9 2.1 1.8 



There are discrepancies in these ratios, to be sure, but they are 

 really small — never large enough to indicate any abnormal difference 

 in composition. 



The incessant repetition of so limited a number of cadences gives 

 rise to two phenomena which deserve brief comment : the produc- 

 tion of unintentional clausulae and of prolonged clausulae. In m}' 

 text of Book XXI I have signalled the existence of numerous ca- 

 dences which upon the whole do not seem to me to be intentional 

 because the pauses which they precede are very slight. ^ But there 



' Naturally I do not maintain that I have drawn the line bet^veeu 

 intentional and unintentional clausulae with any gieat accuracy ; the 

 definition is, and must needs l.e, subjective in large measure. 



