OF THE MESOZOIC MAMMALIA. 
197 
and differs from it in the greater distinctness of the anterior basal cusp; the cingulum 
is well marked upon both the inner and outer surfaces (see also Mes. Mamm., PI. Ill, 
hg. 18). In two specimens the anterior basal cusp of is wanting; in the third 
(No. 24,778) it is faintly developed on the outer surface; this tooth is larger than 
pm^, and has a pronounced and much indented cingulum. The molars resemble 
those of the mandible inverted ; the central cusp is slightly the most prominent ; the 
cingulum is bold and deeply indented. M^, in (No. 47,779), is very instructive, as 
showing the manner of wear; the posterior surfaces of the cusps, a and h, are worn 
slightly concave, probably by the anterior faces of the cusps, h and c, on the lower 
molars. The external cingulum is also indented, and descends low upon the outer 
face of the crowns. 
Varieties of Triconodon. — It is probable that the numerous specimens of Trico- 
nodon, as Professor Owen has suggested, represent several species. Besides the varia- 
tions in the mandible which have already been described, we observe many differ- 
ences in the forms of the teeth. In T. occisor the canine is nearly straight. The pre- 
molars vary in the development of the anterior and posterior basal cusps. The molar 
variations are still more marked ; the molar cusps are sometimes erect, subequal in 
size and only partially confluent (Mes. Mamm., PI. Ill, fig. 9) ; in other specimens 
they are retroverted (fig. 19) ; sometimes the middle cusp is much the highest (PI. IV, 
fig. 2) ; the cingulum is in some cases smooth, in others it does not embrace the 
anterior and posterior faces of the molars; this seems in a measure to be due to age. 
Of still greater interest are the variations in the number and succession of the 
molars, because they serve to connect the genus Triacanthodon with Triconodon, and 
to show that in Triconodon we have, in all probability, the same relation of milk and 
permanent dentition which we find in the modern marsupials.^ The evidence for this 
hypothesis, which was suggested to the writer by Mr. Lydekker, is found in a com- 
parison of a series of specimens beginning with Triacanthodon serrula. When these 
are placed in line, as in the accompanying cut, we at once observe a constant increase 
in size of the jaw taking place in direct proportion to the succession of the third and 
fourth molars. In the smallest of the series, a, T serrula (Mes. Mamm., PL IV, figs. 
7 and 8), behind three teeth of the premolar pattern are five teeth shaped like molars. 
The foremost of these is somewhat undersized, the hindmost 9??4 is still in its forma- 
tive capsule, and it is only owing to the fracture of the jaw that it is visible ; the pen- 
ultimate molar, m^, is still partially covered by bone. The second of the series, h, is the 
specimen described as T. occisor by Professor Owen (PI. IV, fig. 2) ; it is slightly 
larger ; is only partially protruded as its third cusp is still covered by bone ; the tip 
’ In describing the dentition of Triacanthodon (Mes. Mamm., p. 731, Professor Owen pointed out that the 
fourth tooth, counting from the canine, may belong to the deciduous series. In discussing with Mr. Lyddekker 
the question of uniting the above genera he suggested that the fully adult mandibular formula of Tricmodon 
may be pm 4, m 4. 
