The distinction here drawn is so great and obvious that for proof of the German girl's case we need better evidence than Coleridge's rumour of a rumour, cited, as it is, by Hamilton, Maudsley, Carpenter, Du Prel, and the common run of manuals. Not that I deny, _a priori_, the possibility of Coleridge's story. As Mr. Huxley says, 'strictly speaking, I am unaware of anything that has a right to the title of an "impossibility," except a contradiction in terms.'[7] To the horror of some of his admirers, Mr. Huxley would not call the existence of demons and demoniacal possession 'impossible.'[8] Mr. Huxley was no blind follower of Hume. I, too, do not call Coleridge's tale 'impossible,' but, unlike the psychologists, I refuse to accept it on 'Bardolph's security.' And I contrast their conduct, in swallowing Coleridge's legend, with their refusal (if they do refuse) to accept the evidence for the automatic writing of not-consciously-known languages (as of eleventh-century French poetry and prose by Mr. Schiller), or their refusal (if they do refuse) to look at the evidence for telepathic crystal-gazing, or any other supernormal exhibitions of faculty, attested by living and honourable persons. I wish I saw a way for orthodox unimaginative psychology out of its dilemma. After offering to anthropologists and psychologists these considerations, which I purposely reiterate, we examined historically the relations of science to 'the marvellous,' showing for example how Hume, following his _a priori_ theory of the impossible, would have declined to investigate, because they were 'miraculous,' certain occurrences which, to Charcot, were ordinary incidents in medical experience. We next took up and criticised the anthropological theory of religion as expounded by Mr. Tylor. We then collected from his work a series of alleged supernormal phenomena in savage belief, all making for the foundation of animistic religion. Through several chapters we pursued the study of these phenomena, choosing savage instances, and setting beside them civilised testimony to facts of experience. Our conclusion was that such civilised experiences, if they occurred, as they are universally said to do, among savages, would help to originate, and would very strongly support the savage doctrine of souls, the base of religion in the theory of English anthropologists. But apart from the savage doctrine of 'spirits' (whether they exist or not), the evidence points to the existence of human faculties not allowed for in the current systems of materialism.