
1891.] A Reply to Professor Marsh. 781 
constitutes at most an order included in the Prototheria or Meta- 
theria. We may consider it fortunate, therefore, that the term 
was not given a distinctive ordinal definition, but one which, 
upon the author’s own statement,” failed to separate it from the 
Marsupialia,—viz: 1. Teeth much below normal number; 2. canine 
teeth wanting; 3. Premolar and molar teeth specialized; 4. 
Angle of lower jaw distinctly inflected; 5. Mylohyoid groove 
wanting. The best criterion of the definition and of the inutility 
of this term Allotheria is the fact that it has been rejected 
by every subsequent writer. 
The third and fourth points advanced by the author in reply 
are extremely comprehensive, and, if they can be substantiated, will 
to some extent invalidate my criticism. It will be observed, 
however, that both points are advanced very cautiously. 
3. He states that “no true Plagiaulacide are known with 
three rows of tubercles upon the upper molars.” This i$ equiva- 
lent to saying that we have no positive evidence that the upper 
molars have three rows; it is not- stated by the author that any 
upper Plagiaulax molar is zown with. two rows. I may recall 
the fact that in my review I called attention to this lack of posi- 
tive evidence, and enumerated the strong cases of indirect evidence 
to the contrary which we find, first, in the association of loose 
molars of three rows of tubercles with the lower jaws of Nco- 
plagiaulax, not to speak of Cimolomys; second, in the three 
rows of the maxillary molars of Polymastodon, a genus very 
closely related to Plagiaulax ; third, in the analogy of Tritylodon. 
I consider this indirect evidence so strong that it is very unlikely 
to be refuted by subsequent discovery, and believe that the author 
will never be able to substantiate this first statement. 
4. Finally, he states that “no Allotheria (Multituberculata) 
are known with certainty to have three rows of tubercles in the 
lower molars.” In support of this cautious assertion, the author 
first refers to the type of Stereognathus as probably a portion of 
a maxilla; this probability we have already shown rests upon no 
stated evidence whatever. He observes, secondly, that “ there is 
now conclusive evidence that the Cretaceous molar teeth with 
17 Amer. Jour. of Science, 1880, p. 239. a 
