252 SECTIONAL ADDRESSES. 
upon whether the random selection of individuals which originally 
colonised these habitats varied from the genotype of the parent stock 
in characters of survival value in relation to those habitats or not. 
Where two such habitats abut on one another and there is no specific 
adaptation to the two habitats intermediates of hybrid origin are often 
found along the line of contact. 
In plants which are self-fertilised as a rule, but in which crossing 
is not absolutely excluded, numerous species may come to exist in the 
same geographical area and the same habitat, for the changes in geno- 
type brought about by the occasional crossing will be fixed and the 
phenotype purified, i.e. rendered more homogeneous, by the subsequent 
isolation for many generations of the different families. It is in this 
way that the ‘elementary species’ of such a form as Hrophila vulgaris - 
(Draba verna) may be supposed to have originated. The differences 
between these are small but constant, and they must be regarded as true 
species. A variety, on this view, is a relatively transitory form which 
may at any time be reabsorbed by crossing into the general stock of the 
species. 
i We cannot in the present state of knowledge reject altogether the 
possibility of other modes of formation of new species.  Geneticists 
differ as to the occurrence of radical alteration in the nature of a gene, 
or of new genes arising de novo in the genotype, i.e. as to the occurrence 
of ‘ mutations ’ in the narrowest sense, while the interaction of conjugat- 
ing chromosomes by ‘ crossing over’ is well recognised. We cannot, I 
think, exclude the possibility of long-continued action of the enyiron- 
ment actually altering genes or even creating new ones. Thus we have 
only shifted the problem of variation back. We cannot as yet express 
variation in terms of chemistry and physics. We do not know what 
genes are. They may be definite chemical substances, they may be 
physico-chemical complexes, or some may be of one, some of the other 
nature. It is certain that a great number must always be present, and 
that the phenotypic ‘characters’ must depend on their interaction. 
We cannot analyse a race of organisms genetically except in respect of 
those genes that may be present or may be absent. Many genes must be 
present invariably or the working mechanism would break down—the 
organism would be non-viable—and these we cannot separate by breed- 
ing methods. ‘These things being so, we cannot wholly exclude the 
hypotheses of orthogenesis and of epharmosis as causes of evolution, 
much as we may dislike them on account of their vagueness. Modern 
genetic research has been able to demonstrate to a very large extent 
the exact correspondence between changes in phenotype and the drop- 
ping out and new combinations of genes. But it is impossible at 
present to demonstrate exactly how such possible processes as ortho- 
genesis or epharmosis may work. We know nothing of orthogenesis 
except as a phenotypic phenomenon, though we can conceive the possi- 
bility that the genotype tends to undergo continuous progressive change 
in one direction, change which might depend, for instance, on an 
orderly series of dissociations of molecular complexes, and show itself 
by corresponding orderly change of the phenotype in one direction. 
Such a hypothesis would explain certain phyletic phenomena, but we 
