( 394 ) 



separating its two horns. The shajie of the vomer in this region is in fact not unlike 

 that of certain Limicolons birds, especially perhaps Haetnatopus. In other skulls of 

 Apteryx I did not find this peculiarity nearly so well marked. 



The basis cranii is a little different in the various skulls which I have examined. 

 In A. mantelli the basi-sphenoid bar widens out in a fusiform fashion, to contract 

 again just before it is hidden beneath the vomers. This does not cliaracterise the 

 other skulls. But A. oweni and .-1. haasti agree to diflfer from A. mantelli in that 

 the basi-sphenoidal rostrum is ridged on either side where it widens out into the 

 basi-temporals. In A. austrnlis these two lateral ridges exist as well as a median 

 one lying between them, equnlly marked bat not quite so long. 



Pelvis. (PI. XVI.) 



A comparison of the pelves of the different species of Apter>/x shows a great 

 uniformity in the characters of this part of the skeleton. The principal point of 

 difference which I could detect concerned the length of the prepubic process. The 

 following measurements show how short this process is in A. mantrlli, and how long 

 it is in A. oweni : — 



The two ilia diverge from each other anteriorly, and here is the widest part of 

 the pelvis. The two bones, instead of being almost parallel with the vertebral column, 

 as they are posteriorly, are flattened out laterally, which is of conrse the cause of 

 the greater width of the pelvis here. Posteriorly the pelvis is widest just behind 

 the acetabnla, across the antacetabular processes. In rare cases the width of the 

 pelvis at these two points is eijnal or snbequal. These examples are so far nearer 

 to the other genera of struthious birds, in none of which are ilia wider in front than 

 across the antacetabular processes. The following are a few measurements : — 



