If any one finds it difficult to admit that the savage can so easily be deceived, I would direct his attention to the well-known instances of children's self-deceptions. Most of them behave, at a certain age, as if their dolls were alive and, to all appearances, there are some moments when they think so. What they think at other moments is another matter. We need not suppose that the savage cannot take, at times, a critical attitude and perhaps undeceive himself. It is sufficient that at other moments, when under the pressure of needs or in the excitement accompanying ceremonies of considerable social significance or of much personal importance, he should be able to assume the attitude of the believer. The behaviour of certain mentally deranged persons throws some light on this point. Such a person may believe that his hands are always dirty and be constantly washing them. If reasoned with, he may perhaps be convinced that they cannot be dirty. Yet a few seconds later he will exclaim, 'But I feel they are dirty,' and return to the wash-basin. The savage is under the control of his impulses and feelings to a degree approaching that of the person instanced. In this connection, the effect of repetition, and of the tribal sanction obtained by magical customs, should not be overlooked. They tend to make doubt and criticism next to impossible. What need is there in cases of this kind to introduce a middle term between the actions of the magician and their expected effect? None whatsoever. The thought of an efficient agent or power passing out of the magician or of his instrument to work upon the victim is no necessary part of this type of Magic. (_d_) The belief at the root of a great variety of magical practices, that 'like' produces 'like,' may have arisen in still other ways than the one just indicated. Nothing is more common than the invisible passage of things, be they heat, cold, light, thunderbolt, odours, diseases, etc., from one person or object to another, either by contact or through space. The frequent instances of diseases spreading by infection among men, animals, and vegetables, seem in themselves sufficient to suggest the belief that 'like' produces 'like.' The idea of contagion must have appeared very early indeed. Now, as the savage is quite unable to distinguish between the different agencies involved in the variety of experiences of this sort, he cannot draw the line between the 'likes' that really produce 'likes' and those that do not; hence his very strange expectations. This class of Magic also is independent of the conception of an agent effecting the connection between the objects related as cause and effect.