THE EVOLUTION OF THE MAMMALIA. 33 
(3) A very large number of similarities between different ani- 
mal groups have been independently acquired (parallel develop- 
ment), and are not due to genetic relationship. This most 
important principle is tacitly or explicitly denied by many zoolo- 
gists, but the palaeontological series demonstrate it in the most 
unequivocal manner, nor can any rearrangement of the series 
produce a different result. Lists of such parallelisms in single 
characters might be multiplied to an indefinite extent, but they 
may go further and involve the entire structure. For example, 
the camels have a great many points of resemblance to the true 
ruminants, and yet it may be conclusively shown that most of these 
resemblances are not due to inheritance from common ancestors. 
Even more remarkable is the case of the horse-like animals which 
Ameghino has discovered in the Miocene of Patagonia, and which 
he believes to be ancestral to the modern horses. In these crea- 
tures may be observed a most wonderful likeness to the true 
ancestral horses of the northern hemisphere, in the structure of 
the teeth, skull, vertebrae, and limbs, and yet a careful study of 
them shows that they can only be most distantly related to the 
horses, and that they belong to a different order of hoofed animals. 
One is tempted to think that nature had at her disposal only a 
limited number of patterns of teeth, skulls, feet, etc., so often 
are these repeated in widely different groups. 
These innumerable cases of parallel evolution greatly increase 
the difficulty of determining the genealogical series, a work which 
would be very' much easier if every obvious resemblance could 
be at once accepted as a proof of relationship. They emphasize 
the utter absurdity of the " single bone method " of palaeontol- 
ogy and the necessity of tracing the series in the most cautious 
manner, step by step. They show, too, how arbitrary and unnat- 
ural every scheme of classification must be that is founded upon 
a single character, and that such schemes must employ the total- 
ity of structure to be even approximately successful. 
In the second place, palaeontology is in a position to make 
valuable contributions to the discussion of the factors or efficient 
causes of evolution and of the facts of heredity, which are insep- 
