The essence of Rhythm in poetry is the regular recurrence of syllables pronounced with more stress than those intervening. To produce this effect in its perfection, precisely equal times should occur between the recurrences of the stress. But, in the application of rhythm to words, the exactness of these intervals is sacrificed somewhat to the necessary length of the words; and, on the other hand, the words are forced somewhat in their pronunciation, to produce more nearly the proper intervals of time. In different languages these adaptations take place in different degrees; one language disregarding more the intervals of time, another the pronunciation of the words.
The Greek language early developed a very strict rhythmical form of poetry, in which the intervals of time were all-important. The earliest Latin, on the other hand, —as in the Saturnian and Fescennine verse,—was not so restricted. But the purely metrical forms were afterwards adopted from the Greek, and supplanted the native forms of verse. Thus the Latin poetry with which we have to do follows for the most part Greek rules, which require the formal division of words (like music) into measures of equal times, technically called Feet. The strict rhythm was doubtless more closely followed in poetry that was sung than in that which was declaimed or intoned. In neither language, however, is the time perfectly preserved, even in single measures: and there are some cases in which the regularity of the time between the ictuses is disturbed.
The Greeks and Romans distinguished syllables of two kinds in regard to the time required for their pronunciation, a long syllable having twice the metrical value of a short one. But it must not be supposed that all long syllables were of equal length, or even that in a given passage each long had just twice the length of the contiguous shorts. The ratio was only approximate at best, though necessarily more exact in singing than in recitation. Nor are longs and shorts the only forms of syllables that are found. In some cases a long syllable was protracted, so as to have the time of three or even of four shorts, and often one long or two shorts were pronounced in less than their proper time, though they were perhaps distinguishable in time from one
short (see § 608. c, d). Sometimes a syllable naturally short seems to have been slightly prolonged, so as to represent a long, though in most (not all) cases the apparent irregularity can be otherwise explained. In a few cases, also, a pause takes the place of one or more syllables to fill out the required length of the measure. This could, of course, take place only at the end of a word: hence the importance of Cæsura and Diæresis in prosody (§ 611. b, c).