[946] Ehrenreich, loc. cit. p. 45 ff.; von den Steinen, loc. cit. p. 153 ff.
[947] It should be stated that a like conclusion was reached by Lucien Adam from the vocabularies brought by Crevaux from the Upper Japura tribes—Witotos, Corequajes, Kariginas and others—all of Carib speech.
[948] A. C. Haddon, The Wanderings of Peoples, Cambridge, 1911, p. 109.
[949] Described by E. F. im Thurn, Among the Indians of Guiana, London, 1883.
[950] A. C. Haddon, The Wanderings of Peoples, pp. 110-11.
[951] V. d. Steinen, Unter den Naturvölkern Zentral-Brasiliens, p. 157. "D'après Gonçalves Dias les tribus brésiliennes descendraient de deux races absolument distinctes: la race conquérante des Tupi ... et la race vaincue, pourchassée, des Tapuya...."; V. de Saint-Martin, p. 517, Nouveau Dictionnaire de Géographie Universelle, 1879, A—C.
[952] Novos Estudios Craniologicos sobre os Botocudos, Rio Janeiro, 1882, passim.
[953] Possibly so called from the Portuguese botoque, a barrel plug, from the wooden plug or disc formerly worn by all the tribes both as a lip ornament and an ear-plug, distending the lobes like great leathern bat's-wings down to the shoulders. But this embellishment is called tembeitera by the Brazilians, and Botocudo may perhaps be connected with betó-apoc, the native name of the ear-plug.
[954] They are the Cambebas of the Tupi, a term also meaning Flatheads, and they are so called because "apertão aos recemnacidos as cabeças entre duas taboas afim de achatál-as, costume que actualmente han perdido" (Milliet, II. p. 174).
[955] Such "identities" as Tic. drejà = Aym. chacha (man); etai = utax (house) etc., are not convincing, especially in the absence of any scientific study of the laws of Lautverschiebung, if any exist between the Aymara-Ticuna phonetic systems. And then the question of loan words has to be settled before any safe conclusions can be drawn from such assumed resemblances. The point is important in the present connection, because current statements regarding the supposed reduction of the number of stock languages in South America are largely based on the unscientific comparison of lists of words, which may have nothing in common except perhaps a letter or two like the m in Macedon and Monmouth. Two languages (cf. Turkish and Arabic) may have hundreds or thousands of words in common, and yet belong to fundamentally different linguistic families.