customs of the tribe or people where it obtains. Just as in excavations of prehistoric tumuh, or in geologic formations, it is necessary to notice the strata and exact position of the various objects as they come to light, so is it necessary in every excavation into human society to note the strata and exact position of the various phenomena as they are brought into prominence. I do not suggest that such a line of inquiry is needed in order to substantiate conclusions already arrived at. I do not suggest even that before the comparison of custom with custom was undertaken the comparison of tribe with tribe should have been dealt with. I merely wish to put it forward as a proposition which is worth while considering at this stage in the history of the comparative sciences: that some attention should be given to the study of comparative custom based upon the examination of group with group.
Both Mr. Spencer and Dr. Tylor have seen the importance of this aspect of comparative custom. The compilation of the elaborate tables of Descriptive Sociology by Mr. Spencer supplies us with a very good example of the method required for such a study; and Dr. Tylor's recent attempt to elaborate a more scientific method for the study of institutions is the most valuable contribution to comparative custom which has yet been made. By the process so carefully elaborated by Dr. Tylor we are taught to classify the relationship of one custom or belief to another, to pick out what we may call the natural adhesions to any given custom or group of customs. That is to say, given a custom A, we should expect to find associated with it in close relationship customs B, C, and D. But is this all? I venture to think that we may even go a step further and declare that other customs, say, E, F, and G, cannot exist side by side in natural co-relationship with the primary group A, B, C, and D. A very important conclusion follows from this. If in any given country or land two such groups of custom are found to exist side by side the phenomenon must be due to some abnormal
