About the appearance of the god we need say little. We have seen above that an epiphany of some Divine Being or a Resurrection of some dead Hero seems to have been an integral part of the old ritual and thus has its natural place in tragedy. His special duty is to bring the action to a quiet close and to ordain the ritual on which the tragedy is based--thus making the performance itself a fulfilment of the god's command (see above p. 66). The actual history of this epiphany is curious. As far as our defective evidence allows us to draw conclusions we can make out that Aeschylus habitually used a divine epiphany, but that he generally kept it for the last play of a trilogy; that he often had a whole galaxy of gods, and that, with some exceptions, his gods walked the floor of earth with the other actors. (The evidence for this is given in Miss Harrison's _Themis_, pp. 347 ff.) Sophocles, moving towards a more "natural" and less ritual tragedy, used the divine epiphany comparatively little. Euripides, somewhat curiously for one so hostile to the current mythology, intensified this ritual element in drama as he did all the others. And he used it more and more as he grew older. He evidently liked it for its own sake. There is one view about the _Deus ex Mâchinâ_ which needs a word of correction. It is widely entertained and comes chiefly from Horace's _Ars Poetica_. It takes the _Deus_ as a device--and a very unskilful one--for somehow finishing a story that has got into a hopeless tangle. The poet is supposed to have piled up ingenious complications and troubles until he cannot see any way out and has to cut the knot by the intervention of something miraculous--in this case, of a machine-made god. Now devices of this sort--the sudden appearance of rich uncles, the discovery of new wills, or of infants changed at birth and the like--are more or less common weaknesses in romantic literature. Hence it was natural that Horace's view about Euripides's god should be uncritically accepted. But as a matter of fact it is a mere mistake. It never in any single case holds good--not even in the _Orestes_. And there are some plays, like the _Iphigenîa in Tauris_, in which, so far from the god coming to clear up a tangled plot, the plot has to be diverted at the last moment so as to provide an excuse for the god's arrival. Euripides evidently liked a supernatural ending, and when he had to do without a real god--as in the _Medea_ and the _Hecuba_--he was apt to end with winged chariots and prophecies. Can we in the least understand what he gained by it?