and produce the same as proofs of the insidiousness of diffusion. Thus, to take an example from Mr. Hartland's book, if the celebrated test for the fairy changeling—laughter at water boiled in egg-shells—were found in Japan I should not be content to say that the similarity of the Japanese mind had produced a similar story. I should try and trace where the story first arose. That test, one may safely say, was never invented twice.
On the important subject of method—and it is a sign of youth and vigour in a science for its methods to be still undetermined—I venture therefore to disagree with Mr. Hartland. But this by no means causes me to overlook the grasp of material and skill of arrangement shown in his book. His choice of subject, too, argues great judgment. Fairy tales, properly so called, i.e., tales about fairies, are not so much stories as incidents. Hence the Casual Method, as I would venture to call it, can deal with these anecdotes without raising the inconvenient question of diffusion. It would have been impossible, I should fancy, to deal with even any one of the types of story—e.g., Puss-in-Boots—with anything like the same detail as is shown here, without raising the question of diffusion. Mr. Lang's sketches in his Perrault were only sketches after all, and scarcely touched the crucial problems of the subject.
Mr. Hartland dismisses rather cavalierly Mr. MacRitchie's "realistic" theory of the origin of fairies, rather too cavalierly, I think. His chief argument against it is, that where you find stories of fairies, you ought to find traces of Finns. To that there is a twofold answer. First, the stories may have been brought from places where there had been Finns. Secondly, in nearly all places where such stories are told the present inhabitants have been preceded by a shorter race, whom they have exterminated. Tradition about these autochthones might give rise to fairy tales in Mr. Hartland's sense of the word.
I have only touched on Mr. Hartland's main topics; to
