98 THE THEORY OF EVOLUTION 



ians, a state of affairs which Professor Steinmann 

 regards as so scandalous to science (see p. 4), but 

 there are good reasons for the uncertainty. Mesozoic 

 mammals are very rare and, though they have been 

 found in three continents, the specimens are all so 

 incomplete that they contribute but little to a solu- 

 tion of the problem. Certain reptiles obtained from 

 the Permian rocks of South Africa decidedly sug- 

 gest that they were closely allied to the long sought 

 ancestors of the mammals, but the gap is still too 

 great for any final decision. 



The number of genealogical or phylogenetic series 

 within narrower limits, which has already been made 

 out, would fill many volumes. These genealogies 

 have been determined in a great many different 

 types of animals, both vertebrate and invertebrate, 

 but I shall confine attention to the vertebrates, as 

 it is easier to make these intelligible, and can select 

 only a few of the most conspicuous and best known 

 instances. So long as it is clearly understood that 

 but a small number of illustrative examples are here 

 set forth, it will serve no good purpose to pile up a 

 great number of more or less similar cases. Some 

 of the best preserved and most complete of the 

 phylogenetic series are among the families of mam- 

 mals, the changes in which may be traced through 

 the successive stages of the Tertiary period. The 

 most complete of all the series so far discovered is 

 that which displays the development of the horses 

 in North America. One cannot but hesitate to tell 



