conditions which need explanation and investigation. The stress I am inclined to lay upon the phenomenon of inconsistency in custom and belief, as opposed to natural association of custom and belief, has never, so far as I am aware, been confirmed by other writers, but so far as my own researches have tested it in some limited spheres, it presents to the inquirer a set of facts which need to be taken into account somewhere.
What is here stated of comparative research generally is pre-eminently applicable to the case of studies on marriage institutions. After the epoch-making work of Mr. McLennan, and the laborious tabular results of Mr. Lewis Morgan, no work of such importance has been issued as that of Mr. Westermarck's. And yet Mr. Westermarck seems to approach his study of human marriage with less than usual emphasis on the adjective which he for the first time introduces, and also with less attention to the institution of marriage as a part only of the social system of humanity. He insists upon the non-gregariousness of early man, and turns for proof of this to the wretched outcasts of savage society, such as the Veddahs, Bushmen, etc., who have no trace of tribal organisation. But is the absence of tribal organisation a necessary proof of nongregariousness and of family interdependence? The use of the word family to describe the associations of the sexes among the rudest specimens of modern man seems peculiarly unfortunate, and it leaves out of consideration the local organisation which is at the bottom of these associations of human beings. Mr. Darwin has taught us the influence of locality in the development of species, so that on the biological evidence, upon which Mr. Westermarck properly lays so much stress, it is not the small separate groups of human beings, wrongly termed families, but the whole local group which must be considered as the starting point. It is the local group of Bushmen, of Veddahs, of Victorian savages, of ancient Finns, etc., which first present themselves for observation and for
