1 66 PATAGONIAN EXPEDITIONS! PALEONTOLOGY. 



allel sides, or contract posteriorly ; except in the Planopsidce, it is rugose 

 and strongly convex between the molariform teeth, becoming nearly flat an- 

 teriorly. The premaxillaries are separate from each other and from the 

 maxillaries and are so loosely attached that they are very rarely found in 

 association with the skull and are known only in a few genera of the Meg- 

 alonychidce; in this family each premaxillary consists of an anterior and 

 two posterior branches,, which vary much in length and breadth and in 

 the size and shape of the incisive foramina. There is reason to believe that 

 in the other two families the premaxillaries are of quite a different type. 



FIG. 1 8. 



Occiputs of Santa Cruz Gravigrada, x . a, Hapalops longiceps ; b, Eucholceops fronto ; 

 c, Hapalops elongates ; d, Pelecyodon cristatus. 



In the adult animal the two rami of the mandible are coossified at the 

 symphysis, which is extended into a more or less elongate, edentulous 

 spout or beak, the length and shape of which vary greatly, but in all 

 known examples it contracts anteriorly, ending in a point. The dentary 

 portion of the horizontal ramus is usually very thick and massive and its 

 ventral border may be nearly straight (Mylodontidce), strongly convex 

 (Megalonychidce), or may form a great descending flange (certain of the 

 Planopsidce). The condyle is hemispherical and is set upon a long, 

 slender neck, and the angle, which in the Megalonychidce is a very long, 

 prominent and curved hook, and in the other two families is much shorter 

 and broader, is always more or less inflected, somewhat as in the marsu- 

 pials, but less decidedly and abruptly. The postero-external opening of 

 the dental canal is, with one known exception, large and conspicuous ; its 

 usual position is upon the side of the horizontal ramus at the base of the 

 coronoid process, but in most species of Hapalops it is in the ascending 

 ramus. The exception noted is the very imperfectly known Tretnathermm, 

 in which the foramen is minute and on the point of disappearance. The 

 foramen never has the position seen in the Pleistocene Megatheriidce. 



The hyoid apparatus is but partially known ; so far as it is preserved, it 

 bears considerable resemblance to the corresponding parts of Cholozpus. 



