414 PEOF. T. JEFFERY PAEKEK ON THE CRANIAL OSTEOLOGY, 



importance to distinguish a genus, and he therefore returned to his earlier practice of 

 placing all the species in the single genus Dinornis. 



Reichenbach (26) was the first, in 1850, systematically to divide the family into 

 genera, proceeding upon the simple plan of erecting into a genus each of the seven 

 species known to him. His material must necessarily have been very imperfect, and 

 two of his genera [Moa and Movia) are undoubtedly synonyms of a third (Dinornis). 



Von Haast (6) was the next, in 1873, to attack the problem. He divided the Moas 

 into two femilies, each containing two genera ; but the characters upon which the 

 definition of the families was based were shown by Hutton (8) to be quite unreliable ; 

 many of the generic characters are incorrect and others inconstant ; and in at least three 

 instances, in the Canterbury Museum, the skull of one species was assigned to the 

 skeleton of another. Moreover, the brief account of the proposed classification, given 

 in a Presidential address, was never followed up by a detailed statement, and was 

 unsupported by measui'ements or figures. Under these circumstances the wide 

 acceptance of Von Haast's views is rather remarkable : they are adopted without 

 remark by Wallace (28) in 1876, and with a qualifying note by Newton (14) in 1885 ; 

 and Lydekker (15) in 1889 discusses the question and comes to the conclusion that the 

 distinction between the two families is a valid one, — a decision which this author's later 

 enquiries ( I ■i) have led him to reverse. Fiirbringer (3) has a long discussion in his 

 usual judicial manner, and concludes that there is no evidence for the establishment of 

 more than one family. 



Last year two classifications were propounded, imfortunately independently : one in 

 England by Lydekker (12), the other by Hutton in New Zealand (9). Both agree in 

 recognizing only a single family, which Lydekker divides into five genera — or four if 

 Megalapteryx be excluded — and nineteen species, Hutton into seven genera and twenty- 

 six species. In both schemes the definitions of certain of the genera are wanting in 

 exactness, especially as regards the skull, which Prof. Hutton rightly considers the most 

 important part of the skeleton for generic distinctions. Mr. Lydekker supplies valuable 

 corrections of many of Owen's determinations, but he has only examined the British 

 Museum collection, which is evidently deficient in many important respects. Prof. 

 Hutton, on the other hand, besides examining the large public and private collections in 

 New Zealand, has himself collected Moa-bones in various parts of the colony, and was the 

 first to recognize the important bearing of the geographical distribution of the species. 



The table on p. 415 gives a comparison of the arrangement of the Dinornithidae by 

 the four authors referred to with that adopted in the present paper. 



It will be seen that Haast, Lydekker, and Hutton are all agreed as to the limits of 

 Dinornis, and that Reichenbach's Anomalojderyx corresponds with the similarly-named 

 genus of Hutton and with Haast's Meionornis. Lydekker 's jlnomalopteryx includes 

 four of Hutton's genera and part of a fifth. Hutton's Enrydpteryw includes species 

 from two of Haast's and from two of Lydekker's genera, and so on. 



