HEREDITY AND MUTATION AS CELL PHENOMENA 52 1 
Considering now certain features of mutations as we know them, 
one of the most interesting is their variety. Many attempts have been 
made to explain all mutations in terms of one idea. These universal 
explanations have involved (i) redistribution of Mendelian char- 
acters, (2) loss of factors, (3) reduplication of gametes, and many 
other hypotheses. They have nearly all, however, involved the idea 
of the mere loss of qualities or the recombination of those already 
existing. Such views have been brilliantly advocated, particularly 
by Mendelian writers. But they consider the phenomena of muta- 
tion largely from the outside. The cytological and anatomical 
combined with the experimental investigation of particular mutations 
already reveals that mutations belong in various categories and are the 
result of different types of change.^ 
Germinal changes are not due merely to plus or minus variations, 
in the loss or addition of MendeHan factors; but on the other hand 
each change results from a morphological or physiological alteration 
affecting one element of the germ plasm. This being the case, there 
is plenty of m.aterial at hand for progressive and divergent evolution, 
and it is unnecessary to imagine that the endless variety of organic 
structure has resulted from successive germinal simplifications by 
means of loss. 
In a brief analysis of the various known types of germinal change, 
they may be classified into (i) those which are fundamentally morpho- 
logical and (2) those which are primarily chemical. It may be ex- 
pected in general that these two classes of changes will frequently be 
inherited in different ways. This brings us to the question of the 
relation between heredity and particular variations in another aspect. 
We have already observed that variation has to do with the origin of 
differences between organisms, and heredity with the perpetuation of 
those differences. It may here be pointed out that in the inheritance 
of any character-difference such as we are considering, there are two 
features to be taken into account; (i) the nature of the character it- 
self, and (2) the mechanism of its inheritance. It is true that in 
Mendelian inheritance the mechanism by which the germinal deter- 
miners of the various character-differences are distributed in the germ 
cells during meiosis is the chief feature to be considered. But since 
it is now clear that in certain cases these characters are diluted and 
1 For a discussion of this subject see Gates, The Mutation Factor in Evolution, 
Chap. IX. MacMillans: London. 1915. 
