118 J: LIONEL TAVLER [AUGUST 
tends to develop and maintain its own structure, (2) it tends to 
reproduce, under suitable conditions, other organisms more or less 
similar to itself’ We have therefore to consider every living form 
from a somatic and a germinal side. Both somatic and germinal aspects 
exhibit two tendencies which are differently proportioned in different 
organisms, (1) to remain constant in spite of variable external con- 
ditions, (2) to manifest certain changes of structure. According as 
one or other of these tendencies predominate the organism will 
develop and reproduce definitely or indefinitely. In both somatic and 
germinal development natural selection will tend to favour the 
requisite definiteness or indefiniteness of structure. The inheritance 
of somatic characters does not appear to have been established in any 
one of the many alleged examples; the evidence, therefore, that up to 
the present time has been collected, would seem to favour the con- 
clusion that if accommodations are ever inherited it is an event of 
extreme rarity. 
Yet in spite of the lack of evidence in support of the inheritance 
of acquired characters, there seems to be a considerable mass of 
evidence in favour of the contention that germinal variations often 
correspond in their tendencies to somatic accommodations. 
Definite variability corresponding to environmental accommodation 
might however be acquired in the following way. It has already been 
noticed that every organism, both from its somatic and germinal 
aspects, exhibits two tendencies, one towards definiteness, the other 
towards indefiniteness ; somatic indefiniteness appears to be able to be 
modified by environmental influences, therefore those organisms whose 
somatic tendency is predominantly plastic will survive under altered 
conditions of environment where those organisms of a less easily 
modifiable tendency will be eliminated. Now if somatic characters 
rarely or never become germinal, the modifications of the parental 
organisms cannot be transmitted to their offspring, but those offspring 
that happened to be endowed with variations in the same direction as 
the acquired but not transmitted modifications, would start their life 
with a predisposition favourable to their environment, and therefore 
favourable to more complete modification of the somatic side of the 
organism ; this tendency being accumulative under constant conditions, 
coincident variability would arise by the process of selective elimina- 
tion and preservation, without the need for the assumption of use- 
inheritance, which assumption facts appear to negative. 
Coincident variations would thus have a better chance of survival 
simply because they would be present in the surviving organisms, but 
the principle of selection would be the same whether the variations 
were coincident or not. 
It follows from the preceding argument that definite variability 
is a logical necessity, under certain conditions, if the principle of 
natural selection be allowed to be a factor of considerable importance 
