V 



1887.] of Animals commonly named Dinosauria. 167 



three sub-orders. The characters used in the classification are drawn 

 from all parts of the skeleton. The chief orders are the Saaropoda, 

 comprising the allies of Cetiosaurus ; the Stegosauria, which includes 

 the allies of Scelidosaurus ; the Ornithopoda, formed for the allies of 

 Iguanodon ; and the Theropoda, which includes genera related to 

 Megalosaurus. The sub-orders grouped under the Theropoda are 

 named from their typical genera Goeluria and Compsognatha. The 

 chief difference of Marsh's system from that of Huxley is that he 

 separated the allies of Cetiosaurus from the Iguanodontidse to form 

 the type of a primary division of the group, as I had suggested (' Greol. 

 Soc. Quart. Journ.,' vol. 30, 1874, p. 690), and named it Sauropoda. 

 Otherwise the Theropoda is identical with the Megalosauridse ; the 

 Ornithopoda is the Iguanodontidse similarly re-named ; while the 

 Stegosauria is the ScelidosauridaB of Huxley, enlarged like the 

 other groups by Professor Marsh's admirable discoveries, and re- 

 named. 



The characters on which these animals should be classified are, I 

 submit, those which pervade the several parts of the skeleton, and 

 exhibit some diversity among the associated animal types. The pelvis 

 is perhaps more typical of these animals than any other part of 

 the skeleton, and should be a prime element in classification. The 

 presence or absence of the pneumatic condition of the vertebras is an 

 important structural difference. Differences in the construction of the 

 base of the skull are indicative of affinities. The presence or absence 

 of armour is less important, since it may show all grades of develop- 

 ment from the perfect shield of Polacanthus to small granules in the 

 skin ; and the condition of the tarsus seems to me likely to be 

 influenced by the habits of life of the animals. Yet the more general 

 of these characters are morphologically preferable to slight differences 

 in dental character, or digitigrade or plantigrade progression, or 

 number of digits, or relative size of limbs. Many of the characters 

 hitherto regarded as ordinal seem to me rather of a nature to 

 distinguish families. 



The ilium at first sight has the aspect of a distinctive character of 

 the whole group, and has been regarded as Avian, because it extends 

 both in front of the acetabulum and behind it. This character is 

 common to birds ; but it is also shared by the Ornithosauria, and to 

 some extent by the Anomodontia. Hence this condition of the ilium 

 does not necessarily imply that the Dinosauria is a homogeneous 

 group. Professor Cope pointed out two distinct types of ilium 

 which he regarded as ordinal. First, there is the ilium which is 

 prolonged forward as a more or less narrow process which is typically 

 seen in Iguanodon and less typically in Scelidosaurus. Secondly, 

 there is the ilium which has its anterior process developed into a 

 vertical plate. The bone varies a little in shape in every genus, but 1 



2 



