THE PROBLEM OF CLASSIFICATION 207 



tilian or amphibian skull, but there were twice or thrice as many 

 bones in the older forms as there are in recent ones. And there has 

 been in general an increase in bodily size in every phylum. The 

 largest animals have always lived at or near the end of their race, 

 and a race of small animals has never been evolved from a race of 

 large animals. Furthermore, horns, spines, protuberances, and ex- 

 crescences occur only in the later history of any race, never at its 

 beginning. 



The two chief factors of evolution have been environment and 

 heredity. There is more or less impulse due to heredity, a sort of 

 vis a iergo that seems to influence evolution along parallel lines in 

 related forms, though we are never sure how much is due to it and 

 how much to similar environmental influences. 



The chief problem, then, in any classification is the relative im- 

 portance of structural characters in the absence of the actual con- 

 necting links insensibly uniting different forms, that is, the determi- 

 nation of the more conservative hereditary characters of the skeleton, 

 those which have been influenced less by environmental conditions. 

 Some parts of the skeleton are very variable even in nearly related 

 forms. The number of vertebrae in the spinal column, we have seen, 

 may vary extraordinarily within an order. Chameleon lizards have 

 only about sixty vertebrae; other lizards may have a hundred and 

 ninety-four, while snakes of the same order may have as many as 

 four hundred and fifty. The number is seldom of more than generic 

 value, and sometimes perhaps not more than specific. It would be 

 absurd, for instance, to unite in the same group a lizard and a turtle 

 because each happens to have eight cervical vertebrae. 



And the teeth of reptiles, unlike those of mammals, have little 

 value as criteria of relationships, so adaptable are they in shape and 

 number to food habits, though their location may be more conserva- 

 tive. The pectoral and pelvic girdles have been influenced less by 

 environmental conditions; the structure of the feet stfll less in adap- 

 tation to life conditions. More conservative is the arrangement and 

 mode of articulation of the ribs. Most conservative of all has been 

 the structure of the cranial region of the skull, that surrounding the 

 brain, and in consequence it furnishes the most reliable characters 

 for the discrimination of the larger groups, the subclasses or super- 

 orders. 



