Notes on South African Mollusca. 141 



scrub near the coast, from Algoa Bay and Montagu, in the Cape 

 Province, to the Southern districts of Damaraland. 

 Genotype. T. globulus (Miiller). 



In their general anatomy the species of this genus that have been 

 dissected agree very closely with one another. Only in the repro- 

 ductive organs of T. lucanus do we find any marked divergence from 

 the common type. The radula is very constant throughout the 

 genus, the specific differences being slight ; and the tricuspid condi- 

 tion of the central and lateral teeth forms, perhaps, the most striking 

 character of Trigonephrus. So far as is known, this feature is found 

 in no other genus of the Acavidac ; and Pilsbry has pointed out that 

 the presence of side-cusps in Trigonephrus indicates that it is a 

 relatively primitive member of the family.* 



There can be little doubt that in its tricuspid central teeth 

 Trigonephrus retains a feature which was possessed by the ancestors 

 of the group, but has been lost by nearly all the other members of 

 the Acavidac. It is not so certain, however, that these ancestral 

 forms possessed tricuspid lateral teeth. Pilsbry has justly stated 

 that as a rule "all modifications in the teeth proceed from the 

 median line of the radula outwards towards the edges, the outer 

 marginal teeth being the last to be modified " ; and that " a study 

 of the marginal teeth, therefore, gives a clue in many cases to the 

 ancestral condition of a much modified radula." f Now if we 

 examine the marginal teeth of Trigonephrus, we find that while the 

 ectocone is, from its first appearance on the outer teeth, a separate 

 cusp, the endocone arises by the bifurcation of the mesocone, with 

 which it is united in the marginal teeth (see especially PI. IV., fig. 11). 

 It therefore seems not unreasonable to suppose that the endocones 

 on the lateral teeth of Trigonephrus may have thus arisen from the 

 mesocones in evolution, and that the ancestral Acavidae may have 

 had bicuspid lateral teeth. According to the principles explained 

 when discussing the distribution of the Acavidac, the most primitive 

 members of the family should be found, not in South Africa, but in 

 the more remote regions of South America, which are furthest from 

 the centre of evolution. It is therefore specially significant that in the 

 radula of Strophocheilus rosaceus, King, from Chili (judging from 

 a specimen, found at Coquirnbo, in Professor Gwatkin's magnificent 

 collection), while the central teeth are tricuspid, as in Trigonephrus, 

 both the lateral and the marginal teeth are bicuspid, without 



* Proc. Mai. Soc., 1905, vi. p. 288. 

 f Man. of Conch., 1895, ix. p. xiii. 



