These forms were rather typical than stringently binding as Dante himself notes (De Vulg. El., ii, 11); many variations were possible. The first seems to have been the most popular type. The poem might also conclude with a half stanza or tornada, (French envoi). Here, as in the last couplet of the Arabic gazul, were placed the personal allusions, and when these were unintelligible to the audience the joglar usually explained the poem before singing it; the explanations, which in some cases remain prefixed to the poem, were known as the razos.

Troubadour poems were composed for singing, not for recitation, and the music of a poem was an element of no less importance than the words. Troubadours are described as composing "good" tunes and "poor" words, or vice versa; the tune was a piece of literary property, and, as we have said, if a troubadour borrowed a tune he was expected to acknowledge its origin. Consequently music and words were regarded as forming a unity, and the structure of the one should be a guide to the structure of the other. Troubadour music is a subject still beset with difficulties[13]: we possess 244 tunes written in Gregorian notation, and as in certain cases the same poem appears in different MSS. with the tune in substantial agreement in each one, we may reasonably assume that we have an authentic record, as far as this could be expressed in Gregorian notation. The chief difference between Troubadour and Gregorian music lies in the fact that the former was syllabic in character; in other words, one note was not held over several syllables, though several notes might be sung upon one syllable. The system of musical time in the age of the troubadours was based upon the so-called "modes," rhythmical formulae combining short and long notes in various sequences. Three of these concern us here. The first mode consists of a long followed by a short note, the long being equivalent to two short, or in

time

. The second mode is the reverse of the first

. The third mode in modern