system would utterly mar the effect of a tetralogy in which the tragedies were linked by story. Let us imagine, for example, a performance in which the Agamemnon was followed by two plays wholly unrelated to it and to each other; then the Choephoroe; next, two more plays on other subjects; and then the Eumenides. It is manifest that the impressiveness of the Oresteia as a whole would be destroyed; the unity of the poet's large design would be broken up; his work, as now presented, would be no longer the work which he had planned. But we have seen that tetralogies of the Oresteia type continued to be exhibited, at least occasionally, down to the end of the fifth century. Take, for instance, the year 429 B.C., when Philocles brought out his tetralogy, the Pandionis, and suppose that on the same occasion two other poets offered tetralogies not linked by story. Are we to suppose that the sandwich-system, as it might be called, was applied to all three? If so, then manifestly Philocles was placed at a serious disadvantage as compared with his two competitors. Or are we rather to assume that Philocles was allowed to have his Pandionis performed as a whole, and that only the other two tetralogies were interfused? In that case, we have two different systems in operation at the same festival, to the detriment of its symmetry. There would have been small inducement to institute the new plan, when, besides being complicated and troublesome in itself, it was one which could not be uniformly enforced; or, if uniformly, then
