266 Evolution and Adaptation 
Thus the mode for beech trees in general is sixteen; but, as 
shown in the other table, this mode does not correspond with 
either of the two individual modes here ascertained. The 
illustration shows that the racial mode may differ from the 
individual mode. There are also cases known in which 
the mode of a group of individuals living in one locality is 
different from that of another group living in another locality. 
This difference may be a constant one from year to year, 
although so slight, that unless actual measurements are made, 
the difference cannot be detected, because of the overlapping 
of the individuals from different localities. If evolution took 
place by slow changes of this sort, it might be possible to de- 
tect its action, even when very slow, by means of measure- 
ments made on a large number of individuals. At least this 
has been suggested by those who believe new species may 
result from changes of this sort. 
There is some evidence showing that by selecting particular 
individuals of a series, and breeding from them, the mode 
may be changed in the direction of selection. Thus it has 
been stated by Davenport that the descendants of twelve- 
and thirteen-rayed daisies give a polygon with a skewness of 
+1.92; while the descendants of twenty-one-rayed plants 
give a polygon with a skewness of —.13. 
Pearson has described very concisely the possibilities in- 
volved in the selective action of the environment. He states 
that if we examine the frequency distribution of a set of 
organisms that have just become mature, and later make a 
similar examination on the same number of individuals (but 
not the same individuals) during the period of reproduction, 
we shall probably find that a change has taken place which 
may have been due to selection of some sort. The same 
thing might be found in the next generation, and, if it did, 
this would indicate that “selection does not necessarily mean 
a permanent or a progressive change.” The selection in this 
imaginary case would be purely periodic and suffice only to 
