ANIMAL RELATIONSHIPS. 251 
iations occur, and that they may be inherited, is cer- 
tain. 
If a florist wishes to obtain a new variety of a certain 
plant, he watches for variations; and by carefully 
selecting those plants which vary in the desired manner, 
he is able, after several generations, to produce what 
he wishes. 
As we have previously pointed out, variations which 
help an animal to maintain its place against enemies, 
and in the face of obstacles, are the ones likely to be 
transmitted to offspring, and so perpetuated. Natural 
selection, which depends on both heredity and varia- 
tion, must be considered in determining the relation- 
ship of animals. 
A Genealogical Tree. To show relationships, resort 
is often had to devices known as genealogical trees. 
These are supposed to show the ancestry of animals. 
As so many facts must be known to determine ances- 
try, such schemes must always be considered as only 
attempts to show relationships in a very general 
manner. Greater knowledge may at any time compel 
changes to be made. 
Classification. A scheme of classification should 
tell the same story of kinship as the genealogical tree. 
It must also be subject to change, for the same reasons 
as the tree. At present there is no classification of 
animals acceptable to all the persons qualified to judge 
of such matters. What follows is a somewhat pro- 
visional extension of the outline presented in Chapter I. 
