Variation as an Organic Function 177 
It is probable that we are really dealing here with three distinct 
classes of phenomena: 
(a) Sporadic individual mutations affecting the less important 
characters. These are not apparently related in any way to the 
efficiency of organic functions, and must be regarded as more or less 
autonomous hereditary changes in subordinate physiological pro¬ 
cesses that are not important factors in somatic equilibrium. They 
probably account for the majority of the differences between 
domesticated varieties. 
(b) Truly specific mutations. These also affect only the physio¬ 
logically less important characters, but they are determinate and 
occur sooner or later in all members of the species, if conditions 
remain unchanged. They are often co-ordinated and may affect 
several characters simultaneously, but they may be individu¬ 
ally not distinguishable from the first class. Variations of this 
type are mainly responsible for the observed morphological differ¬ 
entiation of natural species, and their origin will be discussed 
later. 
(c) Sporadic individual variations affecting the physiologically 
important characters and which are always harmful. These are 
sudden changes in stable structures and functions that have a long 
phylogenetic history, such as the mechanism of vision, or the func¬ 
tions of the blood. They are essentially abnormal phenomena equi¬ 
valent to foetal monstrosities in somatic development. 
Now so far as the individual mutations are concerned, they will 
be subject to the operation of natural selection in the Darwinian 
sense. It is probable that many of the superficial specific adaptations 
to special environmental conditions, such as protective coloration 
and complicated floral mechanisms in relation to insect pollination, 
have arisen in this way. 
On the other hand, the second class of variations which are truly 
specific, and not fortuitous, will be incapable of individual selection, 
and will presumably occur sooner or later in all members of the 
species if environmental conditions remain unchanged. But as we 
have seen that observed mutations of this type do not generally 
affect the physiologically important characters, we have still to 
determine the type of variation which is responsible for the pro¬ 
gressive increase in the efficiency of organic functions in the course 
of evolution. The probability is, that this is due to similar specific 
mutations, each so minute as to be non-recognisable individually, 
but cumulative and mutually compensative. There is no evidence 
Phyt. XXII. 4. 
12 
