PHYLOGENY OF THE PALJEOGNATILE AND NEOGNATHJE- 231 



of the free end of the pubis with the ischium, both in Rhea and Struthio, and only in 

 these. 



The pelvic girdle of Apteryx resembles that of Dromceus and Casuarius in some 

 respects. This is what we should expect, since Dromceus probably represents the 

 ground form from which the other Palceognathce have been derived. 



In the extraordinary relative length of the pre-acetabular ilium Apteryx stands 

 alone. It extends forward so as to overlap some four thoracic ribs. The post- 

 acetabular ilium is as short relatively as the pre-acetabular is long ; it never exceeds 

 half the length of the pre-ilium. As in Struthio, the post-iiium and ischium are widely 

 separated, but the separation is more marked since the backward rotation of the 

 ischium is somewhat less than in Struthio. 



The innominates meet in the mid-dorsal line as in Rhea, but with this difference : 

 whereas in Rhea although the median borders meet, yet a dorsal plane is present as a 

 widening out and flattening of that portion of the dorsal aspect of the pelvis which lies 

 between the supra-trochanteric processes, in Apteryx these dorso-median edges meet 

 along their whole length, save in the region of the sacrum proper, where a narrow 

 chink reveals the presence of some three neural spines. There is no supra-trochanteric 

 process, but the ilia meet in the middle line to form a long knife-like ridge. The 

 fovea iliaca anterior extends backwards into the post-acetabular ilium, terminating 

 caudad of the anti-trochanter. Behind the acetabulum the vertebral column is bent 

 sharply downwards, and has undergone great lateral compression, so that the centra 

 of the vertebrae have become vertically compressed plates but little thicker than their 

 neural spines. Furthermore, owing to the downward flexure of the posterior synsacral 

 vertebras the post-acetabular ilia have come to embrace the neural spines of the 

 vertebras instead of the centra. The synsacrum extends, moreover, by a single vertebra 

 beyond the level of the free (caudad) end of the post-acetabular ilium. 



The ischium is a relatively long flattened blade, forming a wider angle with the 

 ilium than in any other of the Ratitce. 



The pubis is slender, with a more or less well-marked ventral curve. Caudad, its 

 free end may touch the postero-ventral angle of the ischium, but it never fuses 

 therewith. 



The obturator foramen and fissure are confluent. 



The pectineal process is large and pointed, and formed in part by the descending 

 process of the ilium and in part by the pubis. 



Specific differences in the pelves of Apteryx are, as Beddard has already shown, very 

 slight. According to him the pectineal process is shorter in A. oweni than in 

 A. australis. In our specimens if there is any real difference the reverse would appear 

 to be the case. I have managed to find sufficient differences to construct a key for the 

 species, though with difficulty. Mr. Beddard has already shown that the width across 



2 e2 



