138 
H. H. BARTLETT 
reasserted by hybridization. This question brings him face to face 
with the philosophical difficulties of Linnaeus. Either he must deny 
evolution, or, with Bateson, confess that the studies carried on by 
Mendelians have thrown no light on the problems of evolution. Bate- 
son suggests that we should seriously consider the possibility that 
evolution has not taken place from the simple to the complex, but 
rather from the complex to the simple; that the original forms of life 
were heterozygous with regard to all the characters which have ever 
appeared in geological history; that for each character there was a 
corresponding inhibitor, and that the characters have come successively 
to light by the segregation of recessives from which the inhibitors have 
fallen away. We cannot believe, from the tentative way in which 
Bateson proposes this fantastic hypothesis, that he really places 
much faith in it. But the mere fact that he should whisper it 
shows the extreme pessimism of the ultra-Mendelian attitude in 
regard, to the problem of evolution. Other experimental workers, 
however, are more optimistic. The Mendelians have been so firmly 
convinced that differences between species were all capable of Mendel- 
ian analysis that they have disregarded facts which did not fit their 
formulations. Professor de Vries' work is set aside with the statement 
that the chief reason why factorial analysis has been declared to be 
inapplicable to the Oenothera mutations is because no one (except 
Heribert-Nilsson) has set about such an analysis in the right way. 
Even those of us who doubt the universality of the Mendelian phe- 
nomena, see no reason to deny that species formation by hybridization 
and subsequent segregation has taken place on a large scale in many 
groups. One may, however, admit the great prevalence of hybridiza- 
tion, without believing that all variation which may take place sub- 
sequently to hybridization is a result of that hybridization. In other 
words, there is no reason why true mutations should not occur in 
hybrids as well as in pure lines. 
In this connection we may examine a little more closely the view 
that recessives always originate by segregation rather than by muta- 
tion, a hypothesis ancillary to the multiple factor hypothesis. In only 
one case among wild plants has it been satisfactorily shown that a 
recessive mutation differs from the parent form in the lack of duplicate 
factors. This is the case of Capsella Heegeri and Capsella Bursa- 
pastoris, which has recently been studied by Shull. He finds that 
there is a difference of two duplicate factors between the derivative 
