344 Mr. C. W. Andrews on the Development of 



and, as was mentioned above, a rough surface marks the exact 

 position on the scapuU\3 of the outer angle of the chivicles. 

 The relative position of the scapulae to one another is thus 

 fixxed, and that of the coracoids is easily determined. 



It may here be remarked that in comparing scapulas of 

 these animals at different ages, one circumstance is of great 

 assistance, namely that the inner edge of the bone forming 

 the outer border of the coraco-scapular foramen undergoes 

 scarcely any growth during life except at the ends, this thin 

 sharp edge in the adult scapula being the thin edge of the 

 young stages almost unchanged. If, therefore, in comparing 

 an older and a younger scapula the latter be superimposed 

 upon the former, so that the corresponding portions of their 

 inner borders are coincident, the area added to the older bone 

 since it was in the condition of the younger one is clearly 

 seen. 



Turning now to the consideration of the conclusions tliat 

 may be drawn from an examination of the series of shoulder- 

 girdles above described, we find that in the condition 

 shown in fig. 4 the pectoral arch, as far as its ossified 

 portions are concerned, is similar in all essential respects to 

 that found among the Nothosauridas and Lariosaurida?, the 

 scapulae consisting of a hinder portion bearing the surfaces 

 for articulation with the coracoids and humerus, a dorsally 

 directed blade, and an anterior ventral portion, against the 

 anterior edge of which the hinder end of the clavicle was 

 fixed, the two bones apparently uniting in a rude suture, 

 as has been already pointed out ; in many members of 

 the families above mentioned the sutural union of scapula 

 and clavicle is very perfect. The clavicles, as in the Notho- 

 sauridas and Lariosauridas, met in median symphysis, but, 

 owing to the antero-posterior expansion of their inner 

 ends and the absence or reduction of the interclavicle, their 

 symphysis is much longer. The close union of the clavicles 

 with one another and with the interclavicle at their inner 

 ends, and with the scapulas at their outer ends, among the 

 Nothosauridte and Lariosaurid^ seems to imply the mechan- 

 ical necessity of great rigidity in the ventral portion of the 

 pectoral arch of these swimming reptiles, and the changes 

 undergone by the shoulder-girdle of Gryptodidus^ in its passage 

 from its immature condition to the adult state, all tend to 

 render this rigidity more perfect. For instance, in the stage 

 shown in fig. o A the ventral plate of the scapula is growing 

 inward towards the middle line behind the posterior edge of 

 the clavicle, and near the outer border it is sending a plate 



