OF THE MESOZOIC MAMMALIA. 
259 
probably occurred in the ancestors of the Dasyuridce} In the Triconodontidce, on the 
other hand, jj i was apparently suppressed. In each case the reduction resulted in the 
typical marsupial number. The mode of premolar milk succession in Triconodon was, 
so far as observed, typically marsupial. 
We thus find in these two families several characters which are shared by one or 
other of the Marsupials, and others which are exclusively Marsupial, leaving no 
reasonable doubt as to their relationship. The evidence in the case of the Tricono- 
dontidcB is, however, much the strongest, and, as Professor Owen pointed out, they 
bear further a family likeness to the Dasyuridce. Triconodon resembles Tliyladnus 
in the shape of the mandible, in the triconodont type of inferior molars, in the very 
late appearance of the fourth true molar, but lacks the internal heel of the superior 
molars. The position of the Peralestidcs is doubtful, turning largely upon our inter- 
pretation of the homologies of the molar cusps. The molars of Peraspalax are very 
similar to those of Didelphys, as seen upon the inner surface, but lack the apparent 
derivation from the tritubercular type. As observed above, in discussing the Multi- 
tuberculata, we must distinguish carefully between real and superficial resemblances. 
I'his obtains also in the following comparison. 
The discovery of the Kurtodon molar pattern apparently adds another mesozoic 
marsupial prototvpo in its likeness to that of Phascolomys (Plate IX, fig. 15.) In 
the mesozoic genus the premolars are rudimentary and separate ; in the recent genus 
all but one have disappeared. In both genera the upper molars are compactly placed 
and present an outward curvature. The Phascolomys molar resembles two Kurtodon 
molars placed side by side, as the columnar crowns present an external groove g, and 
two outward opening V s instead of one ; the enamel e at the sides of the \ is raised, 
and the intervening valley of dentine d is bisected by a faint ridge e' which appar- 
ently is the bottom of a superficial fold of enamel. In both genera the last premolar 
is molariform. The remaining premolars are small in Kurtodon, and have disappeared 
in Phascolomys. An important difference is the large canine ^ in Kurtodon vs. the 
rodent-like median incisor of the recent genus, which seems to show that these genera 
do not belong upon the same line. 
To conclude, the Triccmodmitidai were undoubtedly in the marsupial line ; the 
Amphitheriidce were probably Marsupials and the Kurtodoniidce, were possibly so; 
while there is no means of deciding in regard to the Peralestidcs. No definite 
subordinal character can be assigned, but in view of the retention of several primitive 
1 See Oldfield Thomas, “ On the Homologies and Succession of the Teeth in the Dasyurid®, etc.,” Phil. 
Trans., April 28th, 1887, p. 443. In this valuable memoir, the author reaches the conclusion that pm^, 
was probably the tooth which was suppressed in all the Polyprotodonts, and i 4 of the incisor series. The gen- 
eralized marsupial formula is given as i 5 p 4 m 4. The apparent reduction of p, in the ariconodoMa;, does 
not support the author’s hypothesis thatp^ was uniformly suppressed in the Marsupials. 
» It is possible that this tooth although caniniform may prove to be an incisor. 
