10 SCOPE OF COMPARATIVE MORPHOLOGY 
less degree after propagation by seed: and that thus the possibility has 
been demonstrated of origin and establishment of new forms from more 
or less dissimilar parents. This is not the place to discuss the sources of 
variation: whether it arises by a cumulative summation of slight differences, 
or by mutations fer saltum, or both: nor whether the characters acquired 
during the individual life are or are not transmitted to the offspring, thus 
giving a positive direction to variation: nor yet to consider the effect of 
sexuality, and of the subsequent reducing-division of the nucleus in dis- 
tributing the qualities inherited from the parents. It suffices for our 
theoretical position that variations do occur, and that they are liable to 
be transmitted to offspring. The struggle for existence in this greatly 
over-populated world necessarily acts as a sieve upon such variants, and 
though the survival of the fittest is in no sense a positively constructive 
factor in itself, it results in the preservation of what is capable of self-support, 
and the elimination of what is physiologically less efficient. But when 
thus much is granted, it amounts only to this: that living organisms 
demonstrate that such an origin as evolutionary theory contemplates is 
at least possible. It does not necessarily follow that all known forms did 
originate in this way. Still, we are justified in accepting this view as a 
‘‘working hypothesis,” much more probable than any other explanation 
hitherto given of the existence of various living forms. 
But though we may readily adopt an evolutionary view, as a working 
hypothesis applicable to organic forms at large, it is when we apply it in 
detail that the real difficulties begin. We contemplate, for instance, some 
group of plants which have essentially similar form, structure, and develop- 
ment: we find that they differ in certain details and proportions, and 
that it is possible to lay them out in a series extending from one extreme 
form, through minor gradations, to another extreme form. Such a series 
may be strengthened by tracinw parallelism of variations of two or more 
characters. Where this can be done the probability of the series representing 
a real evolutionary line is greatly enhanced. But there are at least three 
ways in which such a series may be read: (1) that the simplest form was 
the most primitive, and the whole series one of progression: (2) that the 
most complex was the most primitive, and the whole series one of reduction : 
(3) that the origin was from some central point, and the development 
atvergent in two or more directions. Any one of these alternatives would 
be compatible with general evolutionary probability. How are we to 
decide which to adopt in any given case? 
The general principle that progress has been from the simpler to the 
more complex gives to the first alternative a prima facie probability. As 
a matter of fact this consideration weighed largely in the phylogenetic 
decisions of a quarter of a century ago, and the opinions on the descent 
of Ferns serve as a good illustration of it. Those Ferns which have 
the smallest sporangia (Polypodiaceae, Hymenophyllaceae) were held to 
be the most primitive, while those with larger and more complex sporangia 
