Professor Johannsen's Experiments in Heredity. 239 
continued selection for many generations may lead to an alteration 
of type, although selection for one generation only gives no 
sensible effect. “ The burden of proof lies however with those who 
assume such an effect of selection.” In the second place one must 
consider “ Mutations, or the possibility of discontinuous variations 
of the type.” “That they do occur” says Professor Johannsen, 
“ appears to me beyond doubt.” The latter statement seems to 
indicate that the writer does not consider the existence of mutations 
would invalidate his theory. 
But, surely, the truth or otherwise of the hypothesis, that 
continuous selection within the pure race will ultimately affect the 
type, is quite independent of the nature of variation ? If variations— 
germinal variations—in the required direction arise in any way, that 
is all that is needful. If mutations occur they will be picked out by 
the selector in the first instance, and the proportion of mutants to 
somatic variants will increase as the selection is repeated, because 
all (or a large proportion) of the offspring of the former will be 
retained and only a small (or smaller) proportion of the progeny of 
the latter. Is not the denial of the possibility of effecting a change 
of type, by selection within the pure race, a denial of the possibility 
of evolution itself? The existence or non-existence of the effect is 
not merely a criterion—nor a criterion at all—as to the nature of 
variation. 
It seems difficult then, on very wide grounds, to admit that the 
effect of selection within a “ pure line ” or the intensity of heredity 
within such a line can be rigidly zero, i.e. the “ burden of proof” lies 
with those who hold such a conception, which is inconsistent with 
the conception of evolution itself. All that can be proved by such 
experiments as Professor Johannsen’s is that the effect is small 
(compared with the probable error of the result)—an interesting 
result, but a very different matter ; for if the effect in given cases be 
not zero but only small it may in other cases be sensible. If 
Professor Johannsen believes in the occurrence of mutations he 
ought to believe in the effect of continued selection within the race, 
whether accepting the hypothesis of continuous variation or not. 
It is unkind to “ ask for more ” where so much is already given, 
but one may point out that it would be of the highest value for 
comparative purposes to have data of similar form for a character 
more strongly inherited. Would it be possible, for instance, 
without making additional experiments, to regroup the material 
so as to deal not with the weight of the single parental seed, but 
