Dante: The Divine Comedy, translated by Peter Dale Anvil, 422pp, £25 in UK

THE flood of Dante translations shows no signs of abating. Over a decade in the making, Peter Dale's new version of the Divine Comedy comes little more than a year after versions by Allen Mandelbaum, Robert Pinsky and Steve Ellis.

Of these, only Mandelbaum's went the full distance to the Paradiso and only Pinsky's rhymed; Dale's does both. Pinsky's response to the relative poverty of rhyme words in English compared with Italian was to adopt a flexible system of consonantal rhyme (e.g. "mass/thighs"). Dale rejects this as a compromise; his objective is nothing less than an English equivalent of full-rhymed terza rima.

It is a daunting undertaking. Terza rima is notoriously one of the most difficult verse patterns to reproduce in English, with Shelley's "The Triumph of Life" and the second section of Eliot's "Little Gidding" among the very few successful attempts.

Previous translators of Dante, such as the Victorian Henry Cary, have cut the Gordian knot by deciding that an English Commedia should be in blank verse. But unlike iambic pentameter, the Italian hendecasyllable line has no "feet", merely a syllable count of between ten and thirteen with a single primary stress on the tenth.

Add to this the fact that Dante's rhymes are invariably double or triple-syllabled, where English rhyme is largely monosyllabic, and the scale of Dale's task becomes apparent.

His credentials are impeccable, with excellent translations of Villon and Laforgue to his name. One feature which Dante has in common with Villon is the huge amount of apparatus he tends to attract, with some Italian editions reducing the text to a trickle of one or two lines per page over a torrent of footnotes. Dale dispenses with all this, confining himself to a brief summary before each canto.

The readability of his text vindicates his decision. There is no denying the arcane side of the poem (the passage in the introduction on its numerological symbolism should convince any doubters), but what Dale shows is that it demands to be read as a narrative first and foremost and only afterwards, if at all, as a tract on Florentine politics, Thomist theology or whatever.

In his "Conversation About Dante", Mandelstam suggested that "semantic adequacy is equivalent to the feeling of having fulfilled a command", and Dale proves that musicality need not come at the expense of simple accuracy.

A previous terza rima translation like that of Laurence Binyon in 1943 achieved this only at the cost of numerous archaisms and inversions; Dale's main concessions to translatorese are occasional small oddities of syntax.

Consider, for instance, the placing of "sublime" and "aspire" in Arnaut Daniel's speech to Dante in Purgatory XXVI: "I'm Arnaut that weep and sing at once: in thought/I see my previous madness; I see sublime/Before me too that day of joy I sought./I pray you by the Good that helps you climb/And to the summit of the stair aspire,/Be mindful of my pain in due time."

Similarly, his loyalty to the English monosyllable sometimes gives his line-ends a truncated, unsatisfactory feel, as when Francesca's speech in Hell V, known to every reader of Dante, ends: "listening, faint with pity, it was as though/I would have died; and, as the weakened clay/Falls to its death, I fell down in that throe" - not to mention the lameness of that "throe".

These are minor flaws, though, and do not detract from a hugely impressive achievement. This is the best complete rhymed Dante we are likely to have for a very long time, and one for which all readers of poetry will be profoundly grateful.