5384 Dr. W. D. Matthew—Relationship of the Sparassodonta. 
a parallel case. The facial vacuities of the Artiodactyla are certainly 
secondary developments and are independently developed in different 
families. The hypothesis that they are retained from reptilian ancestry 
is certainly not supported by any adequate evidence. Nevertheless, 
the palatal vacuities of Marsupials may date back to the earliest 
differentiation of the order, and their loss be an independent speciali- 
zation in each group where it occurs; the uniform presence of the 
character in all the more generalized members of the order seems to 
point to this conclusion. 
The principal marsupial characters of the Sparassodonta thus appear 
to be of much greater weight than anything that can be adduced in 
favour of their placental relationship. They are definite, constant, 
and exclusive characters of the marsupial skeleton, and no material 
approximation to any of them occurs among the Creodonts.1_ There is. 
but one character of real importance which appears to connect the 
Sparassodonts with the Creodonts and modern Carnivora, viz. the 
histological structure of the enamel. The investigations of Mr. Tomes 
on the enamel structure in Marsupials, Creodonts, and Carnivora have 
yielded some very significant results, and it may be as well to 
summarize them briefly and point out what their real bearing 1 is, He 
shows that— 
1. The marsupial enamel is Astindashed by two principal 
characters: (a) it is penetrated by tubes from the dentine, and is thus. 
less completely differentiated; (4) its component prisms are parallel, 
lacking the interlacing arrangement of Carnivora. 
2 Tn the, Inadaptive Creodonts (Sinopa and Hye@nodon, family 
Hyznodontide ; Oxyena, family Oxyenide ; Pachyena and Mesonyz, 
family Mesonychide) the enamel structure is identical with that of 
- modern Carnivora. 
3. In the Adaptive Creodont Didymictis and the primitive Cynoid 
Cynodictis the enamel is not penetrated by tubes from the dentine, but. 
its component prisms are parallel. 
4. In Borhyena among the Sparassodonta the eel structure 
is that of Inadaptive Creodonts and modern Carnivora. 
5. Traces of the more primitive enamel structure are seen in the 
lower orders of Placentals (Insectivora and Rodents); and in Hyraz, 
the most primitive of hoofed Placentals, the enamel is like that of 
Marsupials. 
6. Among the Marsupials, the more specialized carnivorous types 
exhibit the nearest approximation to the higher development of the 
enamel characteristic of Placentals. Phascolom ys alone among the 
Diprotodonta resembles the Placentals in its enamel structure. 
From these facts we may deduce that— 
1., The Marsupials are not derivable from the Creodonta, since their 
enamel structure is much more primitive. 
2. The Creodonta are closely related to the Fissipede Carnivora, 
and, like them, show a high grade of evolution in enamel structure. 
1 There is a slight inflection of the jaw in the later Mesonychide, but it is not at. 
all like the marsupial inflection. Certain other alleged approximations are equally 
slight, superficial, or unimportant. aan 5S Ca 
