Pre-animistic Religion. ijq
portion of dried and pounded flesh are potent medicine in themselves, so long as sympathetic niagic is at the stage at which it takes itself for granted. Magical processes, how- ever, as we have seen, specially invite explanation. What more natural then, given an acquaintance with the images of trance and dream, than to attribute the mysterious potency of a dead man's body to that uncanny thing his wraith? Let me quote just one instance to show how easy is the transition from the one idea to the other. A young native of Leper's Island, out of affection for his dead brother, made his bones into arrow-tips. Thereafter he no longer spoke of himself as "I," but as "we two," and was much feared.^ The Melanesian explanation was that he had thus acquired the mana, or supernatural power, of the dead man. Clearly it is but a hair's breadth that divides the mana thus personified from the notion of the attendant ghost, which elsewhere so often meets us.
There remains the difficult question whether Animism is primarily, or only derivatively, connected with the religious Awe felt in the presence of most kinds of disease. I am disposed to say '^ disthigiio." As regards delirium, epilepsy, and kindred forms of seizure, the patient's experience of hallucinatory images, combined with the bystanders' impres- sion that the former is, as we say, " no longer himself," would, I think, wellnigh immediately and directly stamp it as a case of possession by a spirit. Then all convulsive movements, sneezing, yawning, a ringing in the ear, a twitching of the eyelid, and so on, would be explained analogously. On the other hand there is a large and mis- cellaneous number of diseases that primitive man attributes to witchcraft, without at the same time necessarily ascrib- ing them to the visitation of bad spirits. Thus a savage will imagine that he has a crab or a frog, some red ants or a piece of crystal, in his stomach, introduced by magical means, as for instance by burying the crab (perhaps with ' Codrington,/. A. I., xix., 216-7. N 2
