The Peinciples of Palaeontology. 159 



undergo the modifications already mentioned,' but furthermore, 

 the three bones of the pelvis are closely joined, and the caudal 

 vertebraB are united in one bone (coccyx). In the Archa^opteryx, 

 on the contrary, the bones of the pelvis are separated, and the 

 tail is formed of 21 vertebrae ; the transition manifests itself 

 in young birds, where the bones of the pelvis are very weakly 

 united, and where the vertebrae of the tail are clearly distinct, 

 especially in the Ostrich. 



This is a clear example of variations in correlation. The 

 principle of correlation was enunciated by Cuvier, and applied by 

 him with a success which has remained a lasting triumph. 

 According to Cuvier, '' The parts of a living organism are so 

 closely related, one with the others, that no one can be changed 

 without necessitating change in the others." Hence, given the 

 form of one organ of an animal, it ;s possible to deduce the form 

 of all the others. This is a consequence of another principle, 

 that of the conditions of existence^ according to which an animal 

 especially created to live in certain conditions, must have all its 

 organs adapted to this end. 



Cuvier, a strenuous partisan of the creation and immutability 

 of species, did not seek to investigate the cause of the principle 

 he enunciated, but confined himself to illustrating it by examples. 

 He showed, for instance, how among tiie carnivorous mammals 

 the teeth are incisive, and the jaw so articulated that only 

 vertical motion is possible ; among the herbivora, on the con- 

 trary, the teeth are cuspidate, adapted to grinding, and the 

 articulation of the condyle of the jaw is so elongated as to favor 

 a lateral motion. These characteristics with others are always 

 associated in the same individual.' 



We shall presently see what estimate we should make of the 

 general application of this law. 



The question here concerns, as is understood, different organs 

 adapted to the accomplishment of the same function. 

 ^ But it frequently happens also, that variations appear in correla- 

 tion without the preceding condition being realized. Let us 

 consider, for example, the series of fossils of which the horse is 

 the last term. 



The study of the dentition shows a series of progressive dif- 

 ferentiations from the molars with omnivorous tubercles in 

 Eohippiis to the molars with herbivorous lamellae in the living 

 horse. There exists a correlation between these variations and 



