136 THE MAMMALIA OP THE DEEP RIVEE BEDS. 



to the same individual as a large male skull which was found in the same locality, 

 but on this point I cannot speak with entire confidence. The glenoid cavity is some- 

 what larger and more oval in shape than in the specimen first described, the antero- 

 posterior diameter distinctly exceeding the transverse; the neck is broader, heavier 

 and less contracted, and has well-marked rugose lines for the attachment of muscles. 

 On the other hand, the glenoid border is less elevated and the postscapular fossa is, 

 in consequence, less concave. The coracoid border curves forward much more 

 decidedly than it does in the scapula of Oreodon, making the proximal portion of 

 the blade relatively much wider than in that genus. The suprascapular border is 

 gently arched, but nevertheless forms nearly a right angle with the glenoid border. 

 The spine is almost median in position, giving pre- and postscapular fossae of nearly 

 equal size, but of different shape, owing to the different course taken by the coracoid 

 and glenoid borders. 



The block of matrix which contains the scapula just described, holds also the 

 last four cervical and first four thoracic vertebrae, with their ribs attached, and the 

 proximal half of the humerus. The same block contains also a small bone (Fig. 34, 

 cT) which is removed but a short distance from the coracoid process of the scapula, 

 and runs forward and inward, overlapping the first rib and the transverse process of 

 the seventh cervical vertebra. This bone I regard as a rudimentary clavicle. Nat- 

 urally, such identification will be received with much doubt, and I was at first very 

 skeptical about it myself. It is certainly most unexpected to find this element in an 

 ungulate so far advanced in differentiation and so high in the geological scale as the 

 middle Miocene, while it has not yet been detected in the Condylarthra of the 

 lower Eocene. Nevertheless, in spite of the a priori improbability of the occur- 

 rence of the clavicle in a Miocene artiodactyl, there seems to be but little doubt that 

 such is actually the fact. In the first place, the position taken by the bone in ques- 

 tion is such as a clavicle would occupy if it were present. There is a slight vertical 

 displacement of the whole fore limb and shoulder girdle, but otherwise the bones of 

 all the surrounding parts — vertebras, ribs, scapula and humerus — are in their natural 

 position almost as perfectly as in a living animal. (2) There is no other bone in this 

 skeleton with which this one can be identified ; it is much too slender to be a part 

 even of the smallest rib, and its shape is quite different from that of any of the 

 elements of the hyoid apparatus. Fortunately, we already possess the latter belong- 

 ing to (presumably) the same individual and can definitely state that the bone in 

 question cannot be referred to it. (3) The shape is that which we should expect to 

 find in a rudimentary clavicle; it is slightly arched downward, is very slender and of 

 rounded section, with an inferior keel, which is best marked in the middle and dies 



