116 BIOMETRY 
gradual return towards the mean character of the 
original race.* 
It must be remembered that in the calculation which 
led to this result perfect normal variability was assumed, 
and the contribution of every ancestor of the same 
degree to the hereditary endowment of the offspring 
was supposed to be exactly equal. Since both these 
assumptions are very unlikely to be realized in any 
actual case, the statement here given must only be 
regarded as an approximate indication of what is 
likely to take place. 
Some remarkable observations have recently been 
published by Professor Johannsen, of Copenhagen, and 
from them are drawn conclusions which seem likely to 
lead to a distinct advance in our understanding of 
the process of so-called continuous variation, and of 
the way in which such variations are transmitted. 
Johannsen’s conclusions have recently received remark- 
able confirmation from the work of Jennings on pro- 
tozoa, so that we are now justified as accepting the 
theory of the pure line as a well-established working 
hypothesis. We shall confine our present account to 
Johannsen’s now classical observations. 
The experiments in question were made upon plants 
which could be self-fertilized for a series of generations. 
In this way many complications were avoided which 
are inevitably introduced in the case of biparental 
* From this it seems necessarily to follow that it is impos- 
sible to establish a permanent breed simply by a process of 
selection. Professor Pearson, however, avoids this conclusion, 
