these are shifted one quarter sinistrally; similarly the Brahmanic system in India, or the Javanese system shifted similarly one quarter sinistrally, would be the same as Landa's Mayan symbols, etc.,—the cases of exact coincidence, shift as we will, however, being very few. In some cases a coincidence appears if of two systems one be reversed,—i. e. if we read one dextrally and the other sinistrally, starting with the one point which bears the same color in both systems. Thus, for example, the Apache is the same as the Navaho if the latter is reversed; the Mayan=the Javanese reversed, etc. Or the two methods may be combined when quite a number of new coincidences appear; but the only value of these coincidences as yet seems to be, that they show that there are often several systems in which the same colors appear in the same relative order (or reversed): the tribes whose color-systems thus agree in part seem to have no apparent bond to connect them, however. If there were cases where two systems could be made to coincide by reversing one, and it could be shown that the ceremonial circuit of one tribe was dextral while that of the other was sinistral, then we might be inclined to consider the matter somewhat more carefully, but I have failed to find any such cases as yet.
As a whole, the results of such a comparison as has been made here are to some extent negative ; diversity and not uniformity is the characteristic feature of the symbolism, and no general principle can be laid down as underlying the choice of colors by different peoples. It may be objected that this statistical method of studying such a subject is inadequate, and that the religious motive must be taken into account. That the religious element in the whole matter is of the greatest importance I do not for a moment wish to deny; but by neglecting for the time being the religious motive, which is necessarily somewhat local in its nature, and differs from tribe to tribe, we get a much clearer view of the general factors, which, modified by local influences, have led to the choice in any one instance of the colors associated with the cardinal points.
