344 O. C. Marsh— American Jurassic Mammals. 
Nearly all the mammals older than the Tertiary, judging 
from their dentition alone, may have lived mainly upon insects, 
with such accessory diet as modern Insectivores affect. The Pla- 
gsaulacide, however, show evidence of marked adaptation to 
some peculiar food, whether animal or vegetable cannot yet be 
determined with certainty. Now that the upper teeth of Cten- 
acodon are known, and trenchant teeth are found opposed to 
the lower cutting premolars, and tubercular molars to those be- 
low, the problem is simplified, but not solved. The evidence 
at present points to an animal, rather than to a vegetable, diet for 
all the Allotheria. 
It is not improbable that there was a gradual change in diet 
in the later forms, until vegetable food predominated. The 
fact that the Tertiary genus WVeoplagiaulax, Lemoine, has 
only a single lower premolar coincides with this view, and if 
Hypsiprymnus is a still later descendant, the additional molars, 
and other herbivorous features, may be the result of this 
gradual change. 
The few Mammals known from the Trias may be placed in 
two families, Dromotheridw, including the American speci- 
mens, and MWicrolestide, those of the old world. They are all 
quite distinct from any of the Jurassic forms, either those found 
in this country or in Europe. Below the Trias, no Mammals 
have hitherto been discovered; and none are known with cer- 
tainty from the Cretaceous. 
Mesozoic Mammals have very generally been referred hith- 
erto to the Marsupialia. An examination of the known re- 
mains of American Mesozoic Mammalia, now representing up- 
wards of two hundred distinct individuals, has convinced the 
writer that they cannot be satisfactorily placed in any of the 
present orders. This appears to be equally true of the European 
forms which the writer has had the opportunity of examining. 
With a few exceptions, the Mesozoic mammals best preserved 
are manifestly low generalized forms, without any distinctive 
Marsupial characters. Many of them show features that point 
more directly to Insectivores, and present evidence, based on 
specimens alone, would transfer them to the latter group, if 
they are to be retained in any modern order. This, however, 
has not yet been systematically attempted, and the known facts 
are against it. 
In view of this uncertainty, it seems more in accordance with 
the present state of science, to recognize the importance of the 
generalized characters of these early mammals, as at least of or- 
dinal value, rather than attempt to measure them by specialized 
