Genetics and Evolution 
127 
The occurrence of domestic species with new dominant characters is no 
argument for the occurrence of positive mutations, for such species may be the 
result of crossing; and even if the species crossed had not the character in 
question there is ample evidence that genes may be transmitted from parent 
to offspring through many generations without contributing to the qualities 
of the individuals. The hybrid between two very similar, almost identical, 
subspecies may have a very striking new dominant character. 
The authors show that either continuous or discontinuous “variation” (in 
the wide sense of “difference”) may be caused by a change in the genotype or 
by differences of environmental factors or by both acting together. So that 
it is impossible to suppose on the one hand, with Darwin, that only continuous 
variation is important in evolution, or on the other, with de Vries, that only 
discontinuous variation has evolutionary significance. 
For the production of a character the co-operation of many genes is 
necessary, and the lack of any one will lead to a failure to develop a character. 
We cannot speak of that one which happens to be absent as the determinant 
of the character, any more than we can speak of one link of a chain from 
which a weight is suspended as holding up the weight. We can only study those 
genes which are in some cases absent. So long as a gene is invariably present 
we cannot investigate its genetic behaviour. And many genes, viz. those which 
are essential for the working mechanism of the organism, must be universally 
present. But it is possible to acquire the faculty of distinguishing differences 
caused by difference of genotype from those caused by environmental condi¬ 
tions by constantly observing, for instance, the differences between pure clones, 
e.g. nursery trees. These differences are often made up of apparently trivial 
points which might be considered as part of the fluctuating variability of the 
plant and attributed to differences of conditions. But the members of each 
clone are in fact often remarkably constant—their variability is very small. 
And selection between these members has no effect on the offspring. Selection 
is of no effect in a population without genovariability, and therefore in such 
a population it cannot accumulate a difference and lead to a modification of 
the race. 
The authors take the view that dominance implies the presence of a gene 
which is absent in the recessive type. They counter the argument that a new 
dominant character must, on this view, be due to the appearance of a new 
gene, by referring to their contention that no gene can influence development 
unless there is present a given combination of other genes and non-genetic 
developmental factors. This is illustrated by the production of red pigment in 
Primula sinensis and a dominant yellow in mice. The case of the well-known 
Emily Henderson sweet peas is another instance. So that if a dominant 
novelty appears we cannot decide offhand that a new gene has appeared; it 
may be that two or more genes, none of which could produce the new dominant 
character unless all were present, have met and have thus produced it. They 
insist that we are forced in any case to look upon the production of every 
character as resulting from the co-operation of numerous genes, and that to 
consider such cases as special “complications” is a wholly mistaken view 
which had its origin in the fact that the first cases investigated depended on 
the simple presence or absence of one gene which when present completed the 
chain necessary for the development of a striking character. 
