is rolled back like a cloud leaving the night clear. Commentators discuss in which place it is genuine. Surely, anywhere and everywhere. Such lovely lines, once heard, were a temptation to any rhapsode, and likely to recur wherever a good chance offered. The same explanation applies to the multiplied similes of B, 455 ff. They are not meant to be taken all together; they are alternatives for the reciter to choose from.
And even where there is no flaw in the composition, the formulæ for connection between the incidents—"Thus then did they fight," "Thus then did they pray"—and the openings of new subjects with phrases like "Thus rose Dawn from her bed," and the like, suggest a new rhapsode beginning his lay in the middle of an epic whole, the parts before and after being loosely taken as known to the audience.
Nevertheless, the striking fact about our Homeric poems is not that they show some marks of the rhapsode's treatment, but that they do not show more. They are, as they stand, not suited for the rhapsode. They are too long to recite as wholes, except on some grand and unique occasion like that which the law specially contemplated; too highly organised to split up easily into detachable lengths. It is not likely that the law reduced them to their present state at one blow. All it insisted on was to have the 'true history' in its proper sequence. If it permitted rhapsodes at all, it had to allow them a certain freedom in their choice of ornament. It did not insist on adherence to a fixed wording.
The whole history of the text in the fourth century illustrates this arrangement, and the fact essentially is, that the poems as we have them, organic and indivisible,
