352 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1958 
assessed. By methods such as these, G. G. Simpson has worked out 
that the evolution from Hyracotherium to Equus occupied 60 million 
years. This involved passage through 8 genera, the duration of each 
being on the average 7.5 million years; 30 species of a duration time of 
2 million years each; and 15 million generations each reaching ma- 
turity in 4 years. These data can be compared with those obtained 
from other groups of animals, from which they differ considerably. 
The results show that evolution rate is not correlated with variability, 
nor with generation time, and that it is selection that controls the 
direction and intensity of evolution. 
These results are all the more important because in the past some 
paleontologists, unequipped with knowledge of modern genetics, have 
imagined that, from tracing the course of evolution in the lineage of 
fossils which they established, they were in a position to draw con- 
clusions about the cause of such evolution. Some have thought that 
they had found support for the inheritance of acquired characters, al- 
though they knew nothing about inheritance; others imagined that as 
the lineage of some fossils showed linear progression of certain char- 
acters, they were justified in concluding that evolution involved an 
innate directional component, an expression of “momentum” leading 
to evolution in “straight lines,” which they called orthogenesis. They 
failed to realize that if selection in a particular direction benefits an 
organism, continued selection in the same direction will, up to a certain 
point, benefit it further. Others again have concluded from their ma- 
terials that a distinction in principle could be made out between “big” 
evolution leading to large evolutionary changes, and “small” evolution 
producing trifling results. None of these speculations can stand up 
to the evidence that selection determines the course of evolution, its 
speed or its slowness, the greatness or smallness of the effects produced, 
and its direction, which if constant for any length of time simulates 
orthogenesis. 
It is of interest to consider how far it is possible to extrapolate the 
results of modern genetics into the paleontological past. C. R. Diver 
has shown that in snails the patterns of banding found today were 
already in existence in Pleistocene times. It is necessary, however, 
to beware of concluding that because characters are similar they must 
be controlled by the same genes. Even in one and the same species 
today, the gene complex can undergo permutations which reproduce 
the same structures with different genes. An example is provided 
by the gene “eyeless” in Drosophila which produces flies with very 
small or no eyes; this is, of course, extremely harmful. Eyeless flies 
can, however, be made to breed, and although the mortality is very 
high, progeny can be reared which after a few generations have eyes 
like normal flies. In such a stock the “eyeless” gene is nevertheless 
present unaltered, as can be proved by mating these flies with normal 
