The Clausula in Ammianus Marcellinus. 177 



Turning- now to the instances of Meyer's clausula among the sen- 

 tence endings, we find still less evidence than in Book XXI. Four 

 out of the six cases (1, 2, 4, 5) are not founded on the manuscript, 

 which gives in each instance a regular clausula. Another (3) is 

 certainly to be held corrupt on internal evidence, and the sixth may 

 be explained on the basis of the peculiar but not otherwise unsub- 

 stantiated pronunciation qiiibus} None of these cases is supported 

 by responding clausulae in the surrounding text. 



If the testimony for this clausula is weak, it is weaker still for 

 the other types of A, and for the cadences B, C and D, for the 

 admission of which there is no a priori argument. Yet it differs 

 only in degree and not in kind, since there are examples of each 

 of these clausulae that are superficially sound. More than that, we 

 have among the sentence-endings (17, 18) an example of responsion 

 in C quite as specious as the case in Book XXI just discussed, and I 

 can cite one in D (II. 215. 15 proferrc gestiens = aequitatis tramitem) 

 which can only be disposed of by transposition in both clausulae. 



For this reason it would not be logical to admit Meyer's clausula 

 and exclude the others. Nor would it be logical to hold that cases 

 of responsion in any cadence are valid, and to shut out the isolated 

 cases, for in the first place the responsions are entirely too few in 

 proportion to the isolated cases, and in the second place they are 

 no more difficult to dispose of than the latter. Therefore we must 

 assume either that Ammianus made an illogical and extremely in- 

 frequent use of all these irregular cadences, or that he did not 

 employ any of them at all. 



Between these alternatives the decision must be made, it seems 

 to me, in favor of the latter, when the general condition of 

 the text of Ammianus is taken into consideration. Were his text- 

 tradition perfect, we should of course have a perfect reproduction 

 of his clausulae; but, as we have already seen, it is so far from 



mstittita iavt revertamtir I. 201. 13 ; ordines r. II. 100. 24 ; ordinem r. II. 

 315. 25; fleximus r. II. 137. 27; dedinavimus r. II. 214. 18; tandem ad 

 coepfa I. 75. 23 ; referamus ad coepta II. 103. 21 ; ad ordinem remeabo coep- 

 torian I. 306. 32 ; (cf. II. 134. 18) redejindum ad texttim I. 23. 6 ; pergamns ad 

 reliqua I. 284, 7; II. 66. 5; ad residua narranda pergamns II. 63. 15 ; rfde- 

 amtis ad cetera II. 151. 22 ; hinc ad exorsa I. 128. 3 ; repetetur ordo gestorum 



I. 30. 29 ; redeam unde diverti I. 87. 14 ; II. 246. 10 ; regrediar institiitum 



II. 209. 28. The list is, I think, complete; transitiones^ marking the 

 introduction of a new subject, are of course omitted. 



» See p. 228, 230. 



