( 391 ) 



bnt uot iu A. nmfraUs, have a median ventral canal formed by the united cata- 

 pophysis. In only one species did I find a complete canal to be formed by the 

 catapophysis, that was on ('. T'. 1 1 of .-1. auatralis ; bnt in all the catapophyses of 

 two or three vertebrae came very nearly into contact. This was more marked in 

 A. oweni and A. aiistrulis than in the other species, On the twelfth cervical vertebrae 

 in all species of Apti'njx a median hypapophysis appears, save only in A. haasti 

 where it is the thirteenth. In -1. mantelU, A. aiistralU, and in ^4.. haasti the first 

 hypapophysis is at least occasionally bifid. The hypapophyses in A. oweni continue 

 to the first dorsal vertebra in A. haasti to the third dorsal. In all species of 

 Aptenjx there are sixteen cervical vertebrae, the number of free dorsals, and 

 consequently of sacrals, varies somewhat. In an A. australis I found 6 free dorsals 

 and 6 sacrals in front of the acetabulum. Iu three A. mantelli these numbers were 

 7 and 5 ; in another 8 and 4 ; in an A. haasti 7 and 5 ; in an A. oweni 1 and 5. 



In Apteryx oweni I find nine perfectly free caudal vertebrae. One in front of 

 this series is overlapped by the ilium, but is not completely ankylosed to the vertebra 

 in front. (PI. XVI. f. 3.) 



In A. mantelli there are either seven or eight free caudal vertebrae, not over- 

 lapped by the ilia. In front of these either none at all, or two or three vertebrae 

 which are not ankylosed to each other. 



In A. australis there is a longish pygostyle formed by the partial ankylosis of 

 the last four caudals ; in front of this are three free caudals. (PI. XVI. f. 4.) 



A. haasti has eight free caudals not overlapped by the ilia. In front of this 

 series is one caudal not completely ankylosed. 



Ribs. 



In Apteryx oweni there are two ribs belonging to the cervical series, three 

 ribs which join the sternum below, and four ribs following, of which the first has a 

 sternal division, which does not, however, reach the sternum. Uncinate processes 

 are borne by the last cervical rib, and by the four following ribs ; thus there are five 

 in all. Sometimes four ribs articulate with the sternum or even, according to 

 Mivart, five. 



In A. mantelli there are also nine pairs of ribs ; but only one belongs to the 

 cervical vertebrae. The following four reach the sternum : the rest have not a 

 sternal half. The six first ribs have uncinate processes. 



In A. haasti there are nine pairs of ribs, as in the other species, in addition 

 to a rudiment belonging to the fifteenth cervical. Four ribs reach the sternum, 

 and there are six pairs of uncinate processes. 



In A. australis from Stewart Island there is a rudimentary rib on the fifteenth 

 vertebra ; then follow six pairs with uncinate processes, of which the last five reach 

 the Bternum ; then two largish free ribs without uncinate processes and finally 

 a minute pair. 



Shoulder Giedle. 



Professor T. J. Parker has dealt with the variations of this part of the skeleton. 

 It is chiefly in the proportions between scapula and coracoid, in the size of the angle 

 between the two, and in the bony or membranous condition of the procoracoid, that 

 variations occur. 



This last-named variation, like the others, has no very definite relation to the 



