532 Dr. C. I. Forsyth Major on the 



ternal cingulum cusps " of Talpa &c., and the inner talon of 

 Chrysochloris — he is not acquainted with tlie Centetid genera 

 in which tliis is likewise distinctly preserved — with the 

 " protocone " of Talpa. He regards the main cusp of the 

 tooth of Gentetes as homologous with the " paracone " of 

 Talpa, wherein he agrees with Winge. He, however, 

 proceeds to say : — '' I do not think that the ontogeny of 

 the trituberculate insectivore molar justifies Mivart's fusion 

 theory, but rather suggests that this tooth corresponds only 

 with the paracone triangle of the Mole's tooth." In other 

 words, Woodward on the whole favours the tritubercular 

 theory ; this merely because the *' paracone " — 5 in his 

 figures on pi. xxvi. — being ontogenetically the first molar-cusp 

 to appear, he infers from this circumstance that it is the 

 primitive cusp in a phylogenetic sense as well. He asserts, 

 but does not demonstrate, " that as regards the primary cone, 

 its ontogeny recapitulates its phylogeny." 



That it is not merely the anterior " trigon " of more com- 

 plicated teeth [Talpa &c.) with which the upper molar of the 

 Centetidai and Chrysochloridaj is homologous, is at once 

 evident from the fact that the sides of the unique trigon in 

 the latter families embrace not only the anterior but also the 

 posterior half of the outer " cingulum-cusps." 



The above-mentioned text-figure in the paper on Megala- 

 dapis was given by me in conjunction with four others, as 

 representatives of four families of Malagasy mammals 

 (belonging to three different Orders) with so-called trituber- 

 cular molars. My object was to show (1) that, as a rule, on 

 close examination the tritubercular teeth are nevertheless 

 found to consist of more components than only three cusps; 

 (2) that the fact that " trituberculy " occurs in different orders 

 is in itself sufficient to point to convergence, and the more so 

 since the genera in question are highly specialized in other 

 respects also ; (3) that the three main cusps of the teeth with 

 which we are dealing have different homologies in each of the 

 three orders (Lemuroidea, Carnivora, Insectivora) —a further 

 support for the assumption that convergence has taken place. 

 This circumstance has escaped Woodward's attention, or he 

 would hardly consider that his interpretation of the molars 

 of Gentetes and Ghrysochloris brings them into complete 

 agreement with the molars of mammals in general. 



So much for the more historical part of our subject. 



That the complicated mammalian molars will ultimately 

 be traceable to a simple form I have never denied, and have 

 even expressly stated as much ; but how far back we shall 

 have to go to find this simple condition the future will show. 



