1910.] Piirotherium. 381 



details from the proboscidean types (including Ma'rithcrium): for instance, 

 the mode of wear of the cheek teeth is entirely different, the premolars are 

 different, the milk teeth are different, the section of the procmnbent lower 

 tusks shows no suggestion of the peculiar proboscidean "engine turning." 

 The palate is very narrow, the orbit is placed above the fourth premolar. 

 The atlas and axis differed widely from the proboscidean type, the atlas 

 not being pierced by the vertebral artery and having a prominent median 

 hypapophysis, the odontoid of the axis being very large and short and sup- 

 ported on the enormous anterior border, while the neural tunnel in both 

 bones is very circumscribed. The head was probably pointed downward 

 and M. Gaudry ventures the hypothesis that the beast had the proi)ortions 

 of a gigantic cavy with bent fore limbs, but post- like hind limbs. A cervical 

 vertebra is flattened, as in Arsinoitheriuvl and the Proboscidea, but the 

 lumbar vertebrse differ from the latter type. The most striking contrast with 

 the Proboscidea lies in the forearm. The scapula has the spine turned 

 forward instead of backward, the coracoid process is very long and promi- 

 nent, the glenoid greatly extended. The massive humerus is extremely 

 broad with very stout ento- and ecto-condylar and deltoid crests, large 

 tuberosities and an enormous head. The radius and ulna are very stout but 

 absurdly short. M. Gaudry concludes from a study of the muscular attach- 

 ments that the very powerful forearm may have been used in digging. The 

 lunar and pyramidal (cuneiform) resemble those of Elephas, but are nar- 

 rower. The acetabuhan, as in the elephants, faced downward rather than 

 outward. The straight hind limb had a long femur held almost in line with 

 the tibia; the astragalus was greatly flattened and the navicular facet was 

 directly below the tibial facet, and this indicates that the foot was strictly 

 rectigrade, i. c, with the digits in line with the tibia. To compensate for the 

 relati\e immobility of the pes the knee joint could double up at a very sharp 

 angle. 



Upon this material M. Gaudry bases the important conclusion that 

 Pyrotheriuvi is not an ancestral proboscidean, and that all its resemblances 

 to members of that order result from the assumption of bilophodont cheek 

 teeth and post-like, rectigrade hind limbs; that these resemblances are 

 accompanied by more numerous and fundamental differences, and that, in 

 brief, Pijrotherium is not closely related to any other of the great "])achy- 

 derms" of different orders, such as Asirapotherium, Dinoceros, Arsinoi- 

 therium, Brontotherium etc., and does not fit into any known order. 



The supposed relationship of Pyrotherium with the Proboscidea being 

 thus at least temporarily disposed of, it seems not impossible that the most 

 likely. point of origin lies somewhere near the Homalodotheria, or rather 

 in the protunguiate ancestry of that group. To judge by analogy with 



