As a religious practice it is a special characteristic of Phrygia, a land always fascinating to the student on account of its strange freaks of religious psychology; and from Phrygia the practice spread into some adjacent communities, such as Bambyke. One may be allowed {257} to pause a little to consider the original motive that prompted it. At first sight one is tempted to explain it as due to a morbid exaggeration of the craving for purity. But elsewhere, where this impulse was most powerful, for instance, in the later Orphic and Isis cults, and in early and mediaeval Christianity, it produced many mental aberrations but not this particular one. Nor, again, have we any reason for supposing this craving to have been strong in the devotees of Phrygia; the Galloi of Bambyke, according to Lucian, were possessed by strong though impotent sexual desires and were allowed full license with women. The form of communion most ardently sought with the Phrygian goddess and with the later Sabazios was a marriage symbolised by a sexual act; and Greek and Latin writers, both pagan and Christian, agree in reprobating the obscenities of Kybele-Attis worship; we may note also that Phrygian sacred mythology is somewhat grosser than the Hellenic. We are compelled to seek another explanation, and I can suggest none other than that which I have hazarded elsewhere; namely, that Phrygian religious emasculation was an act performed in a frenzy of exaltation by the priest or _mystes_ desirous of assimilating himself as far as possible to the female nature of the great divinity.[257.1] {258} The worship was under male ministration for the highest part; but for the full exercise of divine power the male priest must become quasi-female and wear a female dress, the latter part of the rôle being common enough in primitive “theurgy.” The priest is himself at times the incarnation of the young god, and is called Attis. Therefore Attis was himself supposed to have performed the same act, even at the cost of his life. How early was this institution of a eunuch priesthood in Phrygia we have no direct evidence to prove. It may be a “Hittite” tradition; for figures that Perrot reasonably interprets as eunuch-priests are seen on the reliefs of Boghaz-Keui. Returning to the topic of the death of the divinity, we may assume that in Phrygia this was a very ancient tradition, enacted yearly by a ceremonious laying out of the vegetation puppet on a bier, or the suspension of it on a pine-tree. We have no direct or otherwise trusty evidence for the immolation of the priest who incarnates the god; doubtless the stories about the death of Marsyas and the harvest-sacrifice of Lityerses point to a ritual of human-sacrifice; whom Marsyas stands for is doubtful, but in the Lityerses legend it is merely the passing stranger who is slain, and neither of these traditions is explicitly linked with Attis-cult.