OFTIIEHUMANFAMILY. 7 



and overcome the difficulties of another, and, perhaps, radically different form. 

 With these considerations in mind it is a much greater cause for surprise that so 

 many schedules were completely executed than that a considerable number should 

 have failed to be so. 



The schedule is necessarily self-corrective as to a portion of the persons described, 

 since the position of Ego and his or her correlative person is reversed in different 

 questions. It was also made self-confirmatory in other ways, so that a careful 

 examination would determine the question of its correctness or non-correctness in 

 essential particulars. This was especially true with respect to the classificatory 

 system. Notwithstanding all the efforts made to insure correctness, it is not sup- 

 posable that the tables are free from errors ; on the contrary, it is very probable 

 that a critical examination will bring to light a large number. I believe, however, 

 that they will be found to be substantially correct. 



It was a matter of some difficulty to determine the proper order of arrangement 

 of the materials thus brought together. The natural order of the subject has been 

 followed as closely as possible. All the forms of consanguinity exhibited in the 

 tables resolve themselves into two, the descriptive and the classificatory. Of these 

 the former is the most simple in its structure, and for this reason shoiUd be first 

 considered. It embraces the systems of the Aryan, Semitic, and Uralian families, 

 Avhich are identical in their radical characteristics. The classificatory system has 

 one principal form, the Indo-American, and two subordinate forms, the Malayan 

 and the Eskimo. Of these, the Malayan is the most simple, and probably under- 

 lying form, and, as such, wovdd come first ; after this in its natural order would be 

 either the Turanian or the American Indian, at convenience, since each stands in 

 the same relation to the Malayan ; and after these the Eskimo, which stands discon- 

 nected from the systems of either of the families named. But it was found advisable 

 to reverse this order, as to the classificatory form, on account of the preponderating 

 amount of materials, and to consider, first, the American Indian, then the 

 Turanian, and after all these the Malayan and Eskimo. 



In Part I., after discussing the elements of a system of relationship considered 

 in the abstract, the Eoman form of consanguinity and affinity is taken up and 

 explained with fulness and particularity, as typical of the system of the Aryan 

 family. This is follov/cd by a brief exposition of the forms which prevail in other 

 branches of the family for the purpose of indicating the differences between them 

 and the typical form; and also to ascertain the general characteristics of the 

 system. The systems of the Semitic and Uralian families are then treated in the 

 same manner, and compared with the Aryan form. By this means, also, the 

 limits of the spread of the descriptive system of relationship are determined. 



In Part II., after discussing certain preliminary facts, the Seneca-Iroquois 

 form is first explained with minuteness of detail, as typical of the system of the 

 American Indian family. After this the several forms in the remaining branches 

 of this family are presented ; confining the discussion, so far as could properly be 

 done, to the points of difference between them and the typical system. 



In Part III., the Tamilian form is first presented and explained as typical of 

 the system of the Turanian family ; after which the forms that prevail among tho 



