But if the particular animal which Greek artists selected as a component part of their Centaurs is thus of minor importance, the fact that their Centaurs were always composite in conception, always compounded of the human and the animal, is highly significant. In discussing the various types of Callicantzari in various parts of Greece, we found that, where there exists a belief in their power of metamorphosis, they are stated to appear in single and complete shapes, while, where the belief in their transformation is unknown, they are represented in composite shapes; and having previously concluded that the belief in their metamorphosis was a genuine and original factor in the superstition, we were led to formulate the principle, that a being of some single, normal, and known shape who has originally been believed capable of transforming himself into one or more other single, normal, and known shapes, comes to be represented, when the belief in his power of transformation dies out, as a being of composite, abnormal, and fantastic shape, combining incongruous features of the several single, normal, and known shapes. Now the horse-centaur of Greek Art is a being of composite, abnormal, and fantastic shape, combining incongruous features of man and animal. If then the principle based on facts of modern Greek folk-lore may be applied to the facts of ancient Greek folk-lore, the horse-centaur of Greek Art replaced a completely human Centaur capable of transforming himself into completely animal form. Moreover I am inclined to think that such a development was likely to occur in the representations of art even more readily than in verbal descriptions. For even if the artist belonged to an age which had not yet forgotten that the Centaurs were human beings capable of turning themselves by sorcery into beasts, how was he to distinguish the Centaur in his picture either from an ordinary man, if the Centaur were in his ordinary human shape, or from a real animal, if the Centaur were in his assumed shape? He might of course have drawn an ordinary man and have inscribed the legend, ‘This is a Centaur capable of assuming other forms’; or he might have drawn an ordinary animal with the explanatory note, ‘This is not really an animal but a Centaur in disguise.’ But if such expedients did not satisfy his artistic instincts, what was he to do? Surely his only course was to depict the Centaur in his normal human shape, and by some animal adjunct to indicate his powers of transformation. And that is what he did; for in the earliest art the fore part of the Centaur is a complete human figure, and the hind part is a somewhat disconnected equine appendage[656].