232 ME. W. p. PYCRAFT ON THE MORPHOLOGY AND 



the pre-ilium varies, but I think that this, in spite of variation, combined with other 

 characters may be relied on. 



Apter)jx anstra^s seems to stand alone in the great width of the pre-acetabular 

 ilium, a width due to a highly arched dorsal border and a very considerable lateral 

 expansion of its antero-ventral border. A rather sinuous post-acetabular ventral 

 border and a sudden widening of the post-acetabular ilium caudad are combinations 

 apparently peculiar to this species. In A. haasfii the pre-ilia are sharply truncated 

 forwards, and the pre-ilium is generally broader in jiroportion than in A. oioeni. Tlie 

 post-acetabular ventral border appears to be concave. 



A. australis mantelli appears, superficially, not to be very readily distinguishable 

 from A. oweni. 'Y\\e chief differences appear to lie in the smaller pectineal process and 

 broader ischia of A. australis mantelli. 



The pelves of Dinornithidce and of the jEjJi/or/iithida' very closely resemble one 

 another, and differ from all other flightless members of the Palwognatim in tliat the 

 post-acetabular region of the pelvis is flattened out into a large, pentagonal plate, nearly 

 as broad as long. This is made up partly by the great length of the transverse 

 processes of the synsacral vertebr2e, and partly by the great widening of the dorsal plane 

 of the ilia — a widening only feebly represented among the PaUfoynatha' elsewhere 

 in StrxtMo. 



In the relations of the ischium and pubis the two pelves now under discussion most 

 nearly resemble Apteryx. Apteryx, however, difteis in one respect, in that in this 

 genus the obturator fissure and foramen are confluent. In the Dinornithine pelvis 

 the foramen is shut ofi" from the fissure. The pectineal process is large in Apteryx, 

 very small and wanting in the Dinornithine pelvis. 



In the Dinornithidce the sacral are more or less easily distinguishable from the post- 

 sacral. In the yEpyornithidce this is not the case. 



In the Dinornitliidce the post-sacral neural spines lie in the middle of a deep fossa, 

 the floor of which is formed by the upwardly directed neural spines In the adult this 

 fossa is closed more or less completely by a bony roof formed by tabular lateral 

 e.xpansions from the crest of the neural spines. Caudad, however, this closure is not 

 quite complete : a pair of lateral slits run up on either side of the median neural 

 plate, from behind forwards ; the extent of these slits decreasing with age, but never 

 entirely disappearing. 



In j^pyornis the roofing of the fossa is ample : a double row of foramina only 

 excepted, which run from behind forwards to the crista transversa. 



The pelvis, both in Dinornithidw and JEpyornithidfe, is relatively much shorter in 

 proportion to its width than in the other Paheognathce. 



