66 PEOF. W. K. PAEKER ON THE 



The instance of the tooth-bearing Ichfliyornis, and also that of the existing 

 Tinamous, show that the keelless sternum is a comparatively recent secondary modifi- 

 cation of this part of the skeleton. The dying-out of the keel has manifestly taken 

 place as part of the same withering of unused structures as that which is seen in the 

 wing of absolutely terrestrial birds. These two groups of Pre-Ratitse — the Opistho- 

 comidse and the Tinamidge — probably co-existed with strong, if short-winged, ancestors 

 of the existing struthious birds. It is very remarkable that in the first of these the keel 

 is at its lowest development, apparently in a primary condition; whilst in the Tinamous 

 the sternum is one of the longest and most carinate of any kind of bird. The shoulder- 

 girdle of Oj)ist]iocomus is still more remarkable than its sternum. In the Vertebrata 

 generally, this part is of great morphological interest, for in it, as in the skull, the 

 cartilaginous tracts, ossified or unossified, are supplemented by parostoses or superficial 

 membrane-bones. It is an awkward necessity of this branch of science that its termino- 

 logy is dominated by the terms of Human Anatomy, which, when applied to parts of 

 lower and simpler forms, are often either incorrect or absurd. Thus the term " coracoid " 

 for the lower part of the shoulder-plate is absurd — the only coracoid which is like a 

 Crow's beak is that of Man and a few Mammals ; whilst the term " clavicle " is 

 incorrect as applied to a lateral parostosis of the shoulder-girdle in most of the cold- 

 blooded types, for they have a simple membrane-bone, whereas the clavicle in Man 

 and his nearest relatives is a compound structure, in which the parostosis is soon 

 blended with endoskeletal elements. This is due to the hot condition of the blood in 

 Mammals, and the same thing appears in most birds — that is, in those that are furthest 

 removed from the Amphibia and Reptilia. Parts of the endoskeleton that are con- 

 tinuous or unsegmented in the cold-blooded types are variously segmented and 

 abortively developed in the nobler Vertebrata; and thus we meet with vestiges or 

 remnants of archaic structures that are used up in many ways in the metamorphosis of 

 the' skeletal elements. In the existing RatitcB, and in this ancient Carinate type, the 

 parts of the shoulder-girdle are in a very primitive condition ; in Ojyistliocomus the 

 transformation that takes place is mainly due to arrested growth and to the blending 

 of parts originally separate. At first sight the structure of these parts in the adult 

 does not seem to be different from that of ordinary Carinate birds ; the scapula forms 

 a single and complete bone, and so does the coracoid, and they are bent upon each 

 other at an angle less than a right angle, as in flying birds generally. 



In the Eatitse the axes of these two bones are coincident ; that, however, is a relapse 

 into a degraded condition ; that this was not the case in their ancestors I feel quite 

 certain. In this typically bent and typically narrow shoulder-plate there is, in the 

 half-ripe embryo, a character as unsuspected as it is instructive : the supra-scapula is 

 segmented from the scapula, as the latter is from the coracoid. Even in Man the 

 " posterior edge " of the scapula is feebly ossified, for a long while at least, thus making 

 a suprascapula ; but in all known adult birds, extinct or existing, the scapula is ossified 



