298 HISTORICAL PALZONTOLOGY. 
of both jaws are iurnished with tooth-like denticulations, which 
differ from true teeth in being actually portions of the bony 
— 
ee / PRO en BES 
Fig. 227.—Skull of Odontopteryx toliapicus, restored. (After Owen.) 
substance of the jaw itself, with which they are continuous, 
and which were probably encased by extensions of the horny 
sheath of the bill. These tooth-like processes are of two 
sizes, the larger ones being comparable to canines ; and they 
are all directed forwards, and have a triangular or compressed 
conical form. From a careful consideration of all the dis- 
covered remains of this bird, Professor Owen concludes that 
“* Odontopteryx was a warm-blooded feathered biped, with 
wings; and further, that it was web-footed and a fish-eater, 
and that in the catching of its slippery prey it was assisted by 
this Pterosauroid armature of its jaws.” Upon the whole, 
Odontopteryx would appear to be most nearly related to the 
family of the Geese (Azserine) or Ducks (Anatide); but the 
extension of the bony substance of the jaws into tooth-like 
processes is an entirely unique character, in which it stands 
quite alone. 
The known JZammatls of the Mesozoic period, as we have 
seen, are all of small size; and with one not unequivocal 
exception, they appear to be referable to the order of the 
Pouched Quadrupeds (JZarsufials), almost the lowest group 
of the whole class of the Wammata. In the Eocene rocks, 
on the other hand, numerous remains of Quadrupeds have 
been brought to light, representing most of the great Mam- 
malian orders now in existence upon the earth, and in many 
cases indicating animals of very considerable dimensions. We 
are, in fact, in a position to assert that the majority of the 
great groups of Quadrupeds with which we are familiar at the 
present day were already in existence in the Eocene period, 
and that their ancient root-stocks were even in this early time 
separated by most of the fundamental differences of structure 
