82 MOSTLY MAMMALS 
would, however, show that while the middle toe is not 
markedly larger than either of the others, the bones of the 
wrist are arranged on the linear plan, while in the ankle 
the upper surface of the huckle-bone is nearly flat, or 
intermediate between that of the elephants and the odd- 
toed ungulates. Omitting mention of certain other minor 
peculiarities in the structure of the limbs, if we now turn 
our attention to the teeth, we shall see that these also 
present features unknown in any living ungulates. We 
find, for instance, in the first place, that the upper jaw is 
furnished with two pairs of permanently growing chisel- 
like teeth, comparable to the single pair of incisors in the 
rodents or gnawing mammals; these being opposed by 
three pairs of nearly similar, although horizontally placed, 
lower teeth. Such permanently growing incisor-teeth are 
paralleled among existing ungulates in the hyrax, but the 
toxodon stands alone in the order from the circumstance 
that the cheek-teeth likewise grow throughout life, instead 
of forming roots. Here, then, we have another point of 
resemblance in the toxodon to the rodent order. When 
we examine the form of the grinding surface of these 
cheek-teeth, there does not appear any marked resemblance 
to those of any existing ungulates. The link is, however, 
furnished by certain allied forms from the older Ter- 
tiary beds of Patagonia, known by the name of Nesodon, 
of which the first fragmentary remains were brought to 
Europe by Darwin, in the Beagle; the toxodon being 
confined to the Pampean deposits and the underlying beds 
of Monte Hermoso. Now, in the nesodons, the structure 
of the cheek-teeth clearly approximates to that character- 
ising the odd-toed ungulates, although belonging to what 
naturalists term a more specialised type. It is further 
noteworthy that in these nesodons, although the cheek- 
