It will be well to mention certain principles which guided my work in Melanesia, and have also been used in ethnological analysis elsewhere, especially in the work of W. J. Perry in Indonesia [The Megalithic Culture of Indonesia, Manchester, 1918]. The first of these principles is that of common distribution. When certain elements of culture are found in association with one another in several localities, we regard this as a ground for assigning the associated customs, institutions, and material objects to one culture, and if the associated elements have no necessary connection with one another, as, for instance, is the case with megalithic architecture and sun cult, we assume that this association, which is meaningless in its present area of distribution, came into existence elsewhere, and reached its present home by transmission.

A second principle is that of organic connection. When two elements of culture are found to be so closely associated with one another that they form constituent parts of one organization, it is assumed that they belong to the same culture. Thus, if megalithic monuments and sun cult are found to occur together as elements in the ritual of a great society, this is regarded as evidence that they belong to one culture, and if the formulas of the ritual of the society are in a language different from that of ordinary life, we have a case in which the principle of organic connection points, not only to transmission, but also to the original home of the language as the region from which the transmission has taken place.

A third principle is only a special case of the second, but it is so important that it deserves special mention. I have called this principle that of "class association." In many parts of the world there is reason to believe that certain social classes or sections of the community represent, and are descended from, settlers from outside. In Polynesia, Melanesia, and Indonesia there is reason to believe that the ruling classes are the descendants of immigrants, while the general mass of the population represent the inhabitants of the country before these settlers came. If an element of culture is found to be especially associated with one or other class, it is, according to the principle I am now considering, assigned to the people whose culture is represented by the class in question. Thus, when I find that the chiefs of Polynesia practice desiccation or other form of preservation of the dead, while the commoners inter their dead in the sitting position, I infer that these forms of disposal of the dead belong to two different peoples. In this case I infer that the desiccation of the chiefs is the later, and interment in the sitting position the earlier, practice. Mr. Perry has found this principle also to hold good in Indonesia, where the association of the cultural use of stone and the sun cult with the chiefs has been held greatly to strengthen the argument based on common distribution that these two elements of culture were introduced by one and the same people.

Each of these three principles standing alone may be subject to exceptions, but when all three lead in the same direction, it is possible to assume, with a high degree of confidence, that associated elements of culture were introduced by one and the same people.¹

¹Rivers, W.H. R., n/an/an/an/a , 11–13.