84 
transformation, viz. the reduction and simplification, proceeds from 
the inner side outwards in superior, and from the outer side inwards 
in inferior molars. The now prevalent transverse arrangement of 
cusps or lobes is not the primitive condition, but a specialised 
pattern of the crown.” (pag. 204-205). By a more careful study 
of the dentition of the Sciuridæ we see that the most primitive, 
i. e. most brachyodont forms have got the most polybunous teeth 
(e. g. Sciurus indicus), and that the outer parts of the upper and 
the inner parts of the lower teeth are not only the most constant, 
but also the most complicated tooth-elements. Hence it follows 
that the trituberculism is not a primitive, but a secondarily acquired 
condition, and this is the case in the Ungulata, the Condylarthra, 
the Creodonta and the Lemuroidea as well as in the Sciuromorpha. 
Of the numerous original cusps some few are developed which fit 
in larger grooves or valleys of the opposite tooth, as the so-called 
Sprotocone” in the basin of the "talonid” and the "hypoconid” in 
the "primitive trigone”, The more polybunous — i. e. primitive 
— a tooth is the more distinctly the cusps are arranged in longi- 
tudinal rows. This is not only the case with the rodents, but also 
with the other primitive mammals, among the Carnivora e. g. in 
the Subursi: ,,I will only mention incidentally that, in my eyes, 
amongst recent Carnivora, the Subursi — and, so far as the form 
of the molars is concerned, Ai/urus — approach nearest to the pri- 
mitive carnivorous Mammalia, whilst some of the Arctocyonidæ are 
the most primitive of Creodonta." (pag. 200).  Allotheria which 
are often pushed aside as an aberrant order, are really not more 
aberrant than e. g. Myomorpha are. The most essential difference 
from Mus is that in Allotheria the cusps are arranged in longitudi- 
nal series, in Mus in transverse series, the longitudinal grooves 
being very shallow. Several of the Cretaceous Allotheria are no 
doubt ,,in the direct ancestral line of Eutheria.”" (par. 2791 — 
Therefore it may be supposed that the primitive Eutherian molar 
was polybunous with the cusps arranged in longitudinal series 
(3 above, 2 below); and through a gradual transformation of this 
