194 Austin Morris Harmon, 



79 per cent of the cases, long by nature without making position 

 in 17 per cent and short in 4 per cent. In II ;' it is long by pos- 

 ition in 76 per cent of the cases, long by nature without position 

 in 16 per cent, and short in 8 per cent. To ascertain how far these 

 proportions vary from the natural, I took as the basis for an estim- 

 ate the penultimate trisyllaliles and antepenultimate tetrasyllables 

 that begin the clausula in various forms and types (as venire prae- 

 cepit, conjluxerit Romam). Here I found that words like venire had 

 the first syllable long by position in 31 per cent of their occur- 

 rences, long by nature without position in 34 per cent, and short in 

 85 per cent, while words like conjluxerit showed 52 per cent long 

 by position, 20 per cent long by nature only, and 28 per cent short. 

 These ratios may not be very accurate, for they were not based 

 on large counts ; but they are accurate enough for our purpose. 

 They show clearly that there is a marked tendency toward po- 

 sitional lengthening of the syllable which we are considering. Though 

 the actual results attained are similar in both forms, the tendency 

 is stronger in I than in II, for in the latter form the natural ratio 

 of syllables long by position is much higher than in form I.^ 



We come now to the d and ;' cV types, in which the second un- 

 accented syllable precedes a caesura (as passiZ'//^ citis, vilis et parva). 

 The two tables which follow will show that position is sought in 

 I d and ;' 6 and in II ;' cV, and that it is neglected in II (5 and in all 

 types of III and IV with 6 caesura. 



aj Quantity observed 

 makes 

 position does not make position total 



cons. -j- cons, vowel -^2 final -nt, -ns vowel -|- cons., hiatus 

 Form consonants -|~ "^'owel cons. -|- vowel 



I (J 201 7 1 14 223 



yd 110 1 10 112 



II 7 d 46 1 47 



357 7 3 15 382" 



* In I y there seems to be a slight inclination toward conscious 

 lengthening- of tlie syllable befoi'e the caesura when the S3^11able follow- 

 ing it is short, as in remanerejit magistros. It is not at all consistently 

 carried out, for it is only illustrated in 7 sentence-endings, while thei-e 

 are on the one hand 15 cases in which the short syllable is not preceded 

 by a lengthened one (as in ruentes aperta^ ferieiida dederimt)^ and on the 

 other hand 10 in which the lengthened syllable is not followed by a 

 short one (as in tandem perveni). Yet in view of the fact that each of 



