128 
Reviews 
With regard to the production of a red-fruited form in Oenothera, recorded 
by Gates, which appears to be a dominant novelty originating independently 
of crossing, the authors argue that since these Oenotheras show a remarkable 
suspension of combination and segregation in hybrids, characters being separately 
transmitted through the pollen and through the ovules, it is quite possible 
to suppose that in a stock originating by crossing in which this suspension had 
been maintained for many generations, a failure of the suspension mechanism 
might give rise, by the combination of genes hitherto kept separate, to a 
dominant novelty. In Drosophila also, the fly which has formed the material 
of so much recent genetic work, there are obviously so many irregul^ities 
in the ordinary segregation processes that it would not be surprising if Morgan’s 
dominant novelties turned out to be produced in a somewhat similar way 
to that suggested for the red-fruited Oenothera. 
Only if we are absolutely certain that the individuals of a given family 
are homozygous can we call the sudden production of unexpected novelties 
mutations in the strict de Vriesian sense. It must be noted that in America 
especially the term mutation is sometimes used in a much looser sense, i.e. as 
the production of any “novelty.” The authors incline to the view that while 
no case of the acquisition of a new gene (positive mutation) has really been 
proved to occur, the loss of a gene (negative mutation) has occurred in real 
“pure lines” (Johannsen, Nilsson Ehle) though in animals it is well-nigh im¬ 
possible to distinguish between this phenomenon and the production of a 
recessive novelty through rearrangement of genes. 
A very important part of the authors’ general theory of evolution is based 
on their view of the relation of heritable variability to groups of individuals 
(clones, families, domestic breeds, or colonies, varieties or species in nature). 
They use the phrase “total potential variability” for the number of genes in 
respect of which such a group is not pure. The potential variability of a group 
of organisms increases if there are taken up into the group individuals which 
either possess a gene or genes not present in any member of the group, or which 
lack genes that are the common property of all the members of the group. 
Potential variability differs from variability because it concerns all genes, 
whether they have been factors in development or not. The authors believe 
that the total potential variability of any group is always tending to decrease 
(in the absence of crossing with other groups) because the offspring of any 
generation are normally produced, both in nature and under domestication, 
from a small fraction only of the individuals of that generation; and this must 
lead to the dropping out from the breeding stock of part of the original total 
potential variability of the group. This will nearly always happen, even where 
the new generation is derived from a random sampling of the old, because the 
chances are that some combinations will be omitted, and at a greater rate of 
course if it is the result of a definite selection from the old. But the reduction 
of potential variability does not depend upon selection, it is automatic, so 
long as some individuals are dropped out of the breeding stock. 
In the case of a self fertilised plant the reduction of variability will proceed 
(averaging 50 per cent, in each generation) even if all the individuals produce 
offspring, because Mendelian segregation will result in the daughter being 
heterozygous for one half the number of genes for which the mother was 
impure. Only in the case of a quite freely crossing group of allogamous 
