1901.] AFFINITIES OF UDEXODON. 171 



are shown in the little skeleton have their centra ahout half the 

 leugth of those in the thoracic series. The vertebra? are much 

 crushed and fractured, so that it is difficult to be certain of the 

 identification of the fragments. By the side of one of the vertebrae 

 is a well-developed process, which may be a spine, but which I 

 am rather inclined to believe to be an autogenous transverse 

 process. 



Shoulder-girdle and Sternum. 



In the little skeleton of Udenodon gracilis the shoulder-girdle, 

 though present, is almost completely hidden by matrix and could 

 not be displayed without injury to other parts. In the skeleton of 

 (J. baini (spec. 4, PI. XVIII. fig. 10) the scapula (sc), precoracoid 

 {p.co. fig. 11) and coracoid (eo.) are well preserved, and also in the 

 very imperfect skeleton (spec. 5), while in specimen 5 the sternum 

 is also shown. 



The scapula, precoracoid and coracoid agree pretty closely with 

 the corresponding bones in the Dicynodont shoulder-girdle as 

 figured by Owen (2), Seeley (1), and Lydekker (7). The scapula is 

 somewhat stouter than that figured by Lydekker, but on the whole 

 closely agrees with it. The upper part of the scapula is broad and 

 flat and only moderately curved inwards. The anterior border is 

 grooved for the lodgment of the cleithrum. A little below the 

 middle of the scapula there is a well-marked acromion process 

 (ac, PI. XVIII. fig. 10) which passes forwards, upwards, and 

 slightly inwards as a fan-like expansion. The lower end of the 

 scapula has a large glenoid surface (t/l.) which looks downwards 

 and a little outwards, and an anterior flattened continuation which 

 articulates with the precoracoid. 



The coracoid (co.) is comparatively small, but has a large glenoid 

 surface (>jl-) which looks mainly outwards, and which is separated 

 from the outer surface of the bone by a very prominent bony border. 



The precoracoid (jkco.) is a moderately flat bone, but slightly 

 larger than the coracoid. On its upper border is a deep notch (fo.) 

 which closed by the lower border of the scapula becomes a large 

 oval foramen. The precoracoid articulates behind by a straight 

 suture with the coracoid, and above with the scapula. It appears 

 to furni>h a small portion of the glenoid cavity. 



] have failed to identify the interclavicle or cleithrum in any of 

 thr specimens in my possession, but in the skeleton of U. gracilis 

 oik of the clavicles is fairly complete (<•/., PI. XYL. & PI. XVIII. 

 fig. 8). It is curved very much like the human clavicle. It 

 appears to have articulated with its neighbour in the middle line, 

 and to have rested on the anterior part of the interclavicle. The 

 inner hail' of the hunt.' is flattened antero-posteriorly, and the outer 

 vertically. 



The Bternum, or perhaps mure correct!) |>ncsteriiiim (PI. 

 XV1H. fig. 9), ifl i moderately large four-sided mediae plate, with 

 the angles pointing forwards, backwards, a\n\ to the sides. The 



