( 390 ) 



It appears, therefore, that in all the species examined the claws may be longer 

 or shorter, and more or less curved. I do not see sufficient grounds for emphasising 

 any specific distinctions. It may be generally remarked, however, that ^1. australis 

 and .4. haasti agree in having as a rule longer claws than .1. oweni (which is. after 

 all, a smaller species) ; and, as a rule, A. mantelli is distinguished from its allies by 

 its smaller claws. 



2. SOME NOTES UPON THE OSTEOLOGY. 



The osteology of Aptenjx has been so fully described in the adult and in 

 the embryo, especially by Owen, Mivart, and T. J. Parker, that it seems to be 

 unnecessary to enter into an elaborate description of the skeleton here. 



In certain features, however, the skeleton varies rather more than is perhaps 

 generally thought. A few of these variations coincide with differences in external 

 characters and in other points of structure, and are therefore possibly of specific 

 value. These will be found referred to iu the course of the following pages and 

 at the end of this paper where it is attempted to lay down the specific characters of 

 the different species of Apteryx. But the bulk of osteological variation has not 

 apparently any such meaning. 



Vertebkal Column. 



The vertebral column of Apteryx has been elaborately described by Mivart for 

 A. oweni and A. australis. 



I find that the centrum of the atlas is very deeply notched in ..4. oweni and 

 A. mantelli* ; indeed, in the latter the superior boundaries of the circular notch 

 very nearly meet and convert it into a foramen. The same is the case with A. haasti. 

 A peculiarity which A. mantelli has, and does not share with the other species 

 mentioned, except occasionally in A. oweni, is the incompleteness of the neural 

 arch of the atlas ; there is a median division perfectly complete, though the two 

 halves come into actual contact. The hypapophyses of the atlas of A. oweni are not 

 quite as figured by Mivart. In one specimen I find no hypapophysis at all ; in 

 another, instead of the squarish median process figured by him as extending along 

 the base of the centrum, it was divided into two very strongly marked and more 

 laterally situated processes. As, however, in A. haasti there is a median tubercle as 

 well as these two, it seems probable that the former is the real hypapophysis, and 

 that the latter correspond to the lateral hypapophysial processes described by Mivart 

 in the ostrich. In A. mantelli the lateral processes alone are present. The hypor- 

 apophyses are well developed; iu all species, in A. oweni they are long and rather 

 narrow ; in A. haasti medium, and in A. mantelli very deep. 



As to the axis, I observed but slight differences in the different species. The 

 spine is massive, particularly in A. mantelli : the hyperapophyses are long, and 

 there is no hypapyophyses. The rib-like elements exist in all species, being stontest 

 in A. oweni. 



Dr. Mivart remarks that the tenth and eleventh cervical vertebrae in A. oweni, 



' In one of three specimens of ^ . mantelli the " body " of the atlas, instead of being completely 

 perforated by the odontoid process, was only deeply excavated above a narrow vertically oval foramen, 

 even turning the excavation forwanl through the body. In this individual the arch of the atlas wa« 

 complete above. 



