198 Austin Morris Hariuo)i, 



clausulae are to l:)e thus interpreted, for beside the three pure forms 

 the only clausulae that he uses with any frequency are 

 ^ -^ ^ — : — ^=^ itineris errore 



^ ^ v.. ^ : -1- :=i ipse teneatur 



j^^ — : -:!- w ^ rursus imposuit 



These are obviously all modihcations of form 1, each produced 

 by the resolution of a single long. All together, the three simple 

 forms and the three moditications of the tirst form are found in 

 about 92 per cent of Cyprian's clausulae, and are therefore become 

 practically universal in him.^ 



The connecting link between these quantitative clausulae and the 

 accentual clausulae in Ammianus is to be found in the coincidence 

 of accent and ictus in the former. Just as through coincidence of 

 accent and ictus the hexameter-close in Latin usually presents the 

 accentual scheme f<j r>u r>^ i<j r\j, so the vast majority of unresolved 

 metrical clausulae of form 1 present the accentual scheme I, those 

 of form 2 the accentual scheme II and so on. In both cases the 

 coincidence is purely natural in origin : it is a consequence of the 

 penultimate law. A little reflection will suffice to convince the 

 reader of this fact, if he stand in need of conviction. Nearly all 

 the possible ways of putting into words either the hexameter-tag 

 or the metrical clausulae must inevitably give a coincidence of 

 accent and ictus. Combinations of words that do not give it are 

 of such a sort that they would not suggest themselves frequently 

 as a means of reproducing the desired quantities (as miles gravis 

 annis). How far the natural preponderance of combinations giving 

 a coincidence of accent and ictus was increased in either case by 

 conscious elimination of combinations which do not give such a coin- 

 cidence is a questionable point, and one that does not concern us here. 

 Let us look at the result of this coincidence of accent and ictus 

 in a concrete example. Cyprian's clausulae are metrical, as we have 

 said. But the accentual cadences which they exhibit are as follow^s.- 



I 30 per cent 



II 18 per cent 



III 26 per cent 



A 14 per cent 



88 per cent 



1 The statistics in this paragraph are taken from Zielinski {Philol. 

 Supplemcntbd. 9 and 10). 



^ The figures are based on the clausulae in the first four tractates, as 

 given in de Jonge's list {Les clansules metriques dans Saint Cyprien^ Louvain. 

 Paris, 1905. 



