252 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1949 
It is conceivable that plant species behave differently. Small 
has analyzed several groups of plants from the chronological point of 
view, notably the diatoms. He came to the conclusion that the mean 
duration of species is measured in millions of years. Furthermore, 
he distinguished in the diatoms long-lived species (up to 100 million 
years) and short-lived species (one million years or less). In other 
groups he encountered shorter rates than this. These time rates are 
of much the same order as in the animal kingdom. But there is one 
interesting difference: while in animals the evidence suggests gradual 
change of specific characters, Small regards the changes in the diatoms 
(and perhaps other plants) as sudden. 
Summarizing, we may say that half a million years is a short life- 
time for a species and, so far as our admittedly incomplete evidence 
goes, something ike a minimum. The lifetime of genera appears to 
be rarely less than 5 million years. Short rates are observed in ter- 
restrial groups like most Mammalia, while long rates are frequent in 
mainly marine groups like the bivalve shells and the diatoms. There 
appears to be a certain amount of correlation between the rates of 
species evolution and the degree of changeability of the environment, 
slow rates being frequent in environments like the sea in which living 
conditions are exceedingly stable. 
Hitherto we have been discussing changes in single lines of descent. 
Let us now consider the phenomenon of phyletic splitting. Branch- 
ing is a frequent event in phylogenesis. An ancestral species may 
evolve into two divergent descendant species by the disappearance of 
intermediates, or the ancestral species may remain unaltered and a 
new type emerge as aside branch. In fact, few groups would survive 
without splitting, since owing to the action of internal (genetic) and 
external (environmental) factors, many species become extinct. 
If the rate of splitting equals the rate of extinction, the number of 
species in a systematic category, like a genus or a family, remains 
constant. But if the rate of splitting is greater than the rate of ex- 
tinction, the number of species in the group under consideration will 
rise increasingly steeply along an exponential curve. 
There are many different ways in which the rate of splitting can be 
plotted in relation to time. I have selected two simple methods of 
plotting which can be applied readily to any group from which suffi- 
cient fossil material has been adequately studied by a taxonomist. 
One is to plot the number of species existing in any period or sub- 
period, or smaller stratigraphical division, on the absolute time scale. 
The other consists of doing the same for the newly appearing species 
only. In the first, the surviving species modify the picture. The 
second is a truer presentation of the rate of splitting, but the material 
is often not complete enough for its application. 
