PURE LINES 319 
population as a whole is brought about by the com- 
bination of these two separate factors. 
This statement applies to the case of an organism 
in which self-fertilization is the general rulé, so that in 
this way the separate lines are kept distinct. Where 
cross-fertilization takes place between the members of 
different pure lines the case becomes enormously com- 
plicated, and this is much the most frequent instance 
which we have actually to deal with. It has been 
suggested that the members of.different lines when 
crossed together may display Mendelian phenomena, 
but the existence of so large a proportion of acquired 
variability renders the problem of analyzing the result 
almost insuperable. We have seen, however, that the 
numerical results obtained by the biometricians do 
not appear to be inconsistent with the existence of 
Mendelian inheritance in populations. 
We find, then, that the questions of inheritance of 
acquired characters and of evolution by the aid of 
continuous genetic variations are not yet absolutely 
settled. Recent discoveries by Winkler and Baur 
regarding the nature of so-called ‘ graft-hybrids’ go 
far to prove that acquired characters cannot be in- 
herited in plants. The classical example of a ‘ graft- 
hybrid’ is Cytasus Adamt, which was produced in 1825 
by grafting Cytisus purpureus on the laburnum. This 
plant, which is more or less intermediate between the 
two parent species, has been reproduced by further 
grafting, but its seeds always give rise to plants which 
areindistinguishable from thelaburnum. Adventitious 
