"Cor | néliús Lucíus || Scípió Barbátus,"
and this structure prevailed, so that in the fragments of Livius and Naevius by far the greater number exhibit it.
When Greek patterns of versification were introduced, the Saturnian rhythm seems to have received a different explanation. It was considered as a compound of the iambic and trochaic systems. It might be described as an _iambic hepthemimer_ followed by a _trochaic dimeter brachycatalectic_. The latter portion was preserved with something like regularity, but the former admitted many variations. The best example of this _Graecised_ metre is the celebrated line--
If, however, we look into the existing fragments of Naevius and Livius, and compare them with the Scipionic epitaphs, we shall find that there is no appreciable difference in the rhythm; that whatever theory grammarians might adopt to explain it, the measure of these poets is the genuine trochaic beat, so natural to a primitive people, [22] and only so far elaborated as to have in most cases a pause after the first half of the line. The idea that the metre had prosodiacal laws, which, nevertheless, its greatest masters habitually violated, [23] is one that would never have been maintained had not the desire to systematise all Latin prosody on a Greek basis prevailed almost universally. The true theory of early Latin scansion is established beyond a doubt by the labours of Ritschl in regard to Plautus. This great scholar shows that, whereas after Ennius classic poetry was based on quantity alone, before him accent had at least as important a place; and, indeed, that in the determination of quantity, the main results in many cases were produced by the influence of accent.
Accent (Gr. _prosodia_) implied that the pronunciation of the accented syllable was on a higher or lower note than the rest of the word. It was therefore a musical, not a quantitative symbol. The rules for its position are briefly as follows. No words but monosyllables or contracted forms have the accent on the last; dissyllables are therefore always accented on the first, and polysyllables on the first or second, according as the penultimate is short or long, _Lúcius, cecídi_. At the same time, old Latin was burdened with a vast number of suffixes with a long final vowel. The result of the non-accentuation of the last syllable was a continual tendency to slur over and so shorten these suffixes. And this tendency was carried in later times to such an extent as to make the quantity of all final vowels after a short syllable bearing the accent indifferent. There were therefore two opposing considerations which met the poet in his capacity of versifier. There was the desire to retain the accent of every- day life, and so make his language easy and natural, and the desire to conform to the true quantity, and so make it strictly correct. In the early poets this struggle of opposing principles is clearly seen. Many apparent anomalies in versification are due to the influence of accent over-riding quantity, and many again to the preservation of the original quantity in spite of the accent. Ennius harmonised with great skill the claims of both, doing little more violence to the natural accent in his elaborate system of quantity than was done by the Saturnian and comic poets with their fluctuating usage. [24]