indeed of these short and fragmentary pieces is of more philological than literary interest, and belongs rather to the head of French language. They are, however, evidence of the progress which, continuing for at least four centuries, built up a literary instrument out of the decomposed and reconstructed Latin of the Roman conquerors, blended with a certain limited amount of contributions from the Celtic and Iberian dialects of the original inhabitants, the Teutonic speech of the Franks, and the Oriental tongue of the Moors who pressed upwards from Spain. But all these foreign elements bear a very small proportion to the element of Latin; and as Latin furnished the greater part of the vocabulary and the grammar, so did it also furnish the principal models and helps to literary composition. The earliest French versification is evidently inherited from that of the Latin hymns of the church, and for a certain time Latin originals were followed in the choice of literary forms. But by the 11th century it is tolerably certain that dramatic attempts were already being made in the vernacular, that lyric poetry was largely cultivated, that laws, charters, and such-like documents were written, and that commentators and translators busied themselves with religious subjects and texts. The most important of the extant documents, outside of the epics presently to be noticed, has of late been held to be the _Life of Saint Alexis_, a poem of 625 decasyllabic lines, arranged in five-line stanzas, each of one assonance or vowel-rhyme, which may be as early as 1050. But the most important development of the 11th century, and the one of which we are most certain, is that of which we have evidence remaining in the famous _Chanson de Roland_, discovered in a manuscript at Oxford and first published in 1837. This poem represents the first and greatest development of French literature, the chansons de geste (this form is now preferred to that with the plural _gestes_). The origin of these poems has been hotly debated, and it is only recently that the importance which they really possess has been accorded to them,--a fact the less remarkable in that, until about 1820, the epics of ancient France were unknown, or known only through late and disfigured prose versions. Whether they originated in the north or the south is a question on which there have been more than one or two revolutions of opinion, and will probably be others still, but which need not be dealt with here. We possess in round numbers a hundred of these chansons. Three only of them are in Provencal. Two of these, _Ferabras_ and _Betonnet d'Hanstonne_, are obviously adaptations of French originals. The third, _Girartz de Rossilho_ (Gerard de Roussillon), is undoubtedly Provencal, and is a work of great merit and originality, but its dialect is strongly tinged with the characteristics of the Langue d'Oil, and its author seems to have been a native of the debatable land between the two districts.