Omar is a poet who can hardly be translated satisfactorily otherwise than in verse. Prose does well enough for narrative or didactic poetry, where the main things to be reproduced are the matter and substance, but it is plainly contra-indicated in the case of poetry like Omar's, where the matter is little else than «the commonplaces of the lyric ode and the tragic chorus,» and where nearly the whole charm consists in the style and the manner, the grace of the expression and the melody of the versification. A literal prose version of such poetry must needs be unsatisfactory, because it studiously ignores the chief points in which the attractiveness of the original consists, and deliberately renounces all attempt to reproduce them. In deciding on the form to be taken by a new translation of Omar, the fact of the existence of a previous verse translation of universally acknowledged merit ought not, of course, to be left out of account. The successor of a translator like Mr. Fitzgerald, who ventures to write verse, and especially verse of the metre which he has handled with such success, cannot help feeling at almost every step that he is provoking comparisons very much to his own disadvantage. But I do not think this consideration ought to deter him from using the vehicle which everything else indicates as the proper one. As regards metre, there is no doubt that the quatrain of ten-syllable lines which has been tried by Hammer, Bicknell, and others, and has been raised by Mr. Fitzgerald almost to the rank of a recognised English metre, is the best representative of the _Ruba'i_. It fairly satisfies Conington's canon, viz., that there ought to be some degree of metrical conformity between the measure of the original and the translation, for though it does not exactly correspond with the _Ruba'i_, it very clearly suggests it. In particular, it copies what is perhaps the most marked feature of the _Ruba'i_,--the interlinking of the four lines by the repetition in the fourth line of the rhyme of the first and second. Mr Swinburne's modification of this metre, in which the rhyme is carried on from one quatrain to the next, is not applicable to poems like Omar's, all of which are isolated in sense from the context. Alexandrines would, of course, correspond more nearly than decasyllables with _Ruba'i_ lines in number of syllables, and they have been extensively used by Bodenstedt and other German translators of the metre but, whatever may be the case in German, they are apt to read very heavily in English, even when constructed by skilful verse-makers, and an inferior workman can hardly hope to manage them with anything like success. The shorter length of the decasyllable line is not altogether a disadvantage to the translator. Owing to the large number of monosyllables in English, it is generally adequate to hold the contents of a Persian line a syllable or two longer; and a line erring, if at