[814] Cf. A. M. Tozzer, Comparative Study of the Mayas and the Lacandones (of Yucatan), and the literature given in articles "America, South" and "Brazil" in Hastings, Encyclopædia of Religion and Ethics.
[815] J. W. Fewkes is of opinion that the great Snake dance (an economic function) was formerly conducted by the Snake clan (Sixteenth Annual Report of the Bureau of American Ethnology, p. 304).
[816] The choice of the object is determined by local conditions that are not known to us. Sometimes, probably, the object is the one most important for the welfare of the community; sometimes it may have come from accident. See below, § 554 ff.
[817] The artificial objects that are regarded, in a few cases, as totems are probably of late origin, the product of reflection, and thus differing from the old totems, which arise in an unreflective time. However, the artificial totems are doubtless sometimes looked on as powerful; in some cases they may be little more than badges.
[818] This is Frazer's definition (in his Totemism p. 1), supplemented by the words "not worshiped." Cf., on the whole subject, Tylor, in Journal of the Anthropological Institute, xxviii, 144; F. Boas, in American Journal of Psychology, xxi; A. A. Goldenweiser, "Totemism," in Journal of American Folklore, xxiii (1910).
[819] For a preciser definition of totemism see below, § 520.
[820] The details are given in Frazer's Totemism and Exogamy.
[821] Certain Arunta traditions appear to point to a time when the totem was freely eaten. The bird-mates of the clans may be regarded as secondary totems—perhaps a survival from a time when a clan might have more than one totem.
[822] Spencer and Gillen, Northern Tribes of Central Australia, pp. 173, 318.
[823] The clan-names may formerly have been totemic, but data for the decision of this point are lacking.