The Clausula in Afnnn'ajius Marcellifius. 229 



labication is certainly the rule in adjectives like aequus, and is uni- 

 versal in the nouns equus and aqua. It is universal also in quies, 

 both in the nominative, where it is combined with a shift of accent 

 (qiiies) and in the oblique cases {qiiiete). \\\ some other words, such 

 as Quadi, quattuor and querela, it seems to be universal, but they 

 do not occur often enoui^h to conclude this with certainty. In 

 quaeritabat, in sequcbatur and in propi}iquabaut, among- other words, 

 s3dlabication is optional. 



So unusual is this feature of the pronunciation of Ammianus that 

 at first it seems almost incredible. But it is certain beyond all per- 

 adventure, inasmuch as the strength of the evidence which supports 

 it is in no wise conditioned upon the validity of the h3'pothesis that 

 all of the clausulae in Ammianus were originall}^ ' regular.' Even 

 if we reject that hypothesis, we must still admit that ' irregular ' 

 cadences are very infrequent and exceptional. This admission ampl}' 

 suffices to establish the point in question, for clausulae such as those 

 cited above are too frequent to be classed as ' irregular.' The truth 

 comes out unequivocally if we look at the matter in the light of 

 statistics. In 1811 sentence-endings there are 22 which, taking 

 Gardthausen's text as it stands, unquestionably present irregular 

 cadences, and, as we have just seen, 35 which become irregular 

 unless the syllabication of // after q be admitted. To count these 

 35 clausulae as irregular would not greatl}' alter the relative status 

 of the regular and the irregular cadences. But it would bring it 

 al:)Out on the one hand that irregular cadences, though constituting 

 but 3 per cent of the total number of clausulae, would occur in 

 50 per cent of the clausulae which contain the combination qu, and 

 on the other hand that clausulae containing qu, though amounting 

 only to 4 per cent of the total number of clausulae, would form 

 60 per cent of the irregular cadences. It is plain, therefore, that 

 to avoid creating an inexplicable abnormality we must admit that 

 all or nearly all the apparently irregular clausulae containing the 

 combination qu are in reality regular. And they cannot l^e regular 

 unless qu counts as a syllable in them. 



Not only is the principle certain in itself, but there is ample 

 justification for applying it wherever its application will make a 

 cadence regular. On this point it will take but little reflection to 

 convince the most sceptical. For instance in assuming, in order to 

 explain a single clausula,^ that quin counts as a dissyllable, we are 

 sup])orted by the analogy of the relative pronoun. If quod may so 

 count, why not quin ? And in asserting that quod ma}' so count, 



' I. 159. 8 afuit quin caperer. v. p. 175. 



Trans. Conn. Aoad., Vol. XVI. 15 Oct., 1910. 



