August 26, 1886] 
natural selection as sexual selection, correlated variability, the 
law of homology, or any other of the Darwinian factors. The 
expression used by Mr. Romanes for his new factor—the 
“« Segregation of the Fit’’—seems to imply fitness for something, 
presumably for the conditions of life, and if the survival of the 
“* fit” race is determined by natural selection, then I venture to 
think that natural selection must still be regarded as the theory 
of the origin of species and as something more than a theory of 
the origin of adaptations. 
To the foregoing Mr. Romanes will probably say that physio- 
logical selection is a mecessary adjunct of natural selection, and 
that no new species can arise without the co-operation of his 
factor. If this be the case, then, bearing in mind the views 
which I have expressed with reference to the subordination of 
physiological to natural selection, it seems to me that the most 
likely course to pursue is to appeal to the latter for an explana- 
tion of the ‘‘ primary ” specific character, viz. sterility. It is 
true that Darwin and many of his followers have attempted in 
vain to account for this primary specific distinction by natural 
selection, but I still venture to think that the solution lies in 
this direction. Indeed, the elements of the solution appear to 
me to be furnished by the original theory of Darwin and Wal- 
lace, as I will now briefly attempt to show, hoping to elaborate 
the idea on some future occasion, or, still better, leaving it 
for development or extermination in the hands of professed 
biologists. 
The struggle for life being the most severe between the most 
closely allied forms, diversity is in itself an advantage, because 
the individuals which depart from the parent type may (but not 
necessarily zws/) ‘‘ seize on many and widely diversified places 
in the economy of nature, and so be enabled to increase in 
numbers.” Hence Darwin’s principle of ‘‘ divergence of cha- 
racter,” so well restated by Mr. Romanes in his paper. Now, 
if diversity is an advantage, natural selection can deal with it 
like any other advantageous character, and would seize upon 
any means afforded to secure its perpetuation, provided always 
that the divergence was in the direction of some unoccupied 
“place in the economy of nature.” This last condition amounts 
to nothing more than that the divergence is to be of an advan- 
tageous character. Among the most effectual means of securing 
permanent diversity must be sterility with the parent form: 
hence, given a variability in the reproductive capacities of 
different pairs of individuals (which I have already conceded), 
the question is whether natural selection could not develop out 
of this more or less imperfect fertility a more or less complete 
sterility. If it is to the advantage of some particular variety not 
to resemble its parent form, out of the immense number of 
divergences which are always taking place (by ordinary varia- 
tion) those varieties which showed the greatest infertility with 
the parent form would in the long run survive, for the very 
reason that their progeny, tending to inherit the characters of 
their parents, would possess the advantageous characters of the 
latter, which led to their survival in the first instance, and, 
among these characters, that diminished fertility with the 
parent form which lessens the chance of their extermination by 
imtercrossing. 
As the foregoing argument is necessarily expressed in general 
terms, it will be of use to specialise our ideas by an appeal to a 
hypothetical case. Suppose, for instance, that a dominant 
species gives rise to the twelve varieties A, B, C,. ... L, out 
of which four, B, D, K, L, possess some slight advantage over 
the parent form, adapting them to some new conditions in their 
environment. The four varieties thus stand a chance of surviv- 
ing, while the eight others would be the ‘‘unfit.” Now these 
four varieties in their incipient stage (and in the absence of iso- 
lation) would be subject to extermination by intercrossing in the 
next generation with the parent form. But the chances against 
these four varieties being eguadly fertile with the parent form 
must be exceedingly great: let us suppose, therefore, that B 
and Kare less fertile with the parent form than D and L. Under 
these circumstances the latter would be wiped out by inter- 
crossing, while the former would tend to retain their peculiarities 
and thus survive. The peculiarities both of B, K, and D, L, 
were originally advantageous, but those of B, K, are alone 
allowed to survive. The parent species has, as it were, 
attempted to give rise to four derived species, and has only 
succeeded in producing two. It is a case of selecting out of a 
number of advantageous modifications those particular varieties 
which are the least fertile with the parent form. From the 
slight sterility thus produced in the initial stages, natural selec- 
NATURE 
\ 
385 
tion, by acting in the same direction, might evolve the more or 
less perfect sterility which we now behold, because every de- 
parture on the part of the derived form in the direction of 
fertility with the parent form would be a retrograde step tend- 
ing to obliterate those advantageous characters which led to the 
first survival of the new form, and the descendants of such 
partially fertile departures would constantly be weeded out 
owing to the dilution of their peculiarities with the less advan- 
tageous characters of the parent form. To put the case in 
another way, it may be said that natural selection is constantly 
endeavouring to develop the most advantageous modifications 
of every species, but succeeds the better the less the degree of 
fertility of the advantageous modification with the parent form, 
and succeeds only perfectly by producing perfect sterility with 
the parent form. Fertile advantageous modifications, on the 
other hand, would be swamped by absorption into the parent 
form. 
I thus venture to think that the theory of natural selection as 
sketched out by Darwin and Wallace is still a theory of the 
origin of species. The production of the sterility of species by 
this agency is, according to the present views, to be referred to 
the same causes as all the other modifications produced thereby, 
viz. the natural selection of a ‘* spontaneously ” occurring varia- 
tion in the function of one particular organ. In the case of 
domesticated races no such selection with reference to the func- 
tions of the reproductive system has been effected, but the 
varieties have only been kept from interbreeding by what 
amounts to isolation. It is not surprising, therefore, that such 
artificial forms, which have been selected only with reference to 
external characters, should be fertile z#tex se, while natural 
species, in which fertility z7z/er se has been rigorously suppressed 
by natural selection through long series of generations, should 
exhibit a greater or less degree of sterility.! In other words, 
“physiological” appears as one particular phase of natural 
selection, and as far as we can see there is no reason why there 
should not be other modes of variation leading to the same 
result by acting indirectly upon the reproductive system. But 
all such modes of variation would still be subject to develop- 
ment or suppression by natural selection. R. MELDOLA 
Greenock, August 21 
THE Duke of Argyll’s letter about organic evolution, pub- 
lished in your last week’s issue (p. 335), calls for a few remarks, 
as it is very misleading, and bespeaks some misconceptions on 
the part of the writer. He has evidently read his own views 
into the two articles on organic evolution contributed by Spencer 
to the April and May numbers of the Wineteenth Century. In 
those articles Spencer makes no ‘‘admission”’ ; what he says 
there with respect to natural selection has been held by him for 
the last twenty-six years. He does not deny that the natural 
operations denoted by natural selection do constitute an operat- 
ing cause in the evolution of species. Only, he goes deeper: he, 
with his characteristic truly philosophic insight, sees in natural 
selection a froximate cause; sees behind it the primordial 
operations of forces of nature which rendered natural selection 
possible, and supplied it with a point-d’appui. Then he assigns 
use and disuse as another cause in the origination of species. 
Now all this is not a ‘‘declaration against’? what your corre- 
spondent pleases to call ‘* Mechanical Philosophy,” but is a part 
and parcel of it. It is rather a ‘‘ declaration against ” all sorts 
of teleological philosophy. Let him remember also that Spen- 
cer’s philosophy is the acknewledged philosophy of evolution ; 
and he may rest assured that, even if the theory of natural selec- 
tion as a cause in the genesis of species be proved untrue, that 
philosophy will still stand opposed to any philosophy that will 
attempt to bring back ‘‘ Mind” as one of the cazses of organic 
evolution. 
Your correspondent is a little too hasty in his rejoicings over 
Mr. Romanes’ paper on ‘‘ Physiological Selection.” He will 
see from the second part of the paper that even Mr. Romanes is 
unable to deny that in some cases at least natural selection is 
quite competent to originate species. 
Then your correspondent thinks the theory of natural selection 
“‘not only essentially faulty and incomplete, but fundamentally 
erroneous,” ‘‘in so far as it assumes variations to arise by acci- 
dent.” Now by ‘‘accident” or ‘‘ chance” in this connection, 
evolutionists (including Darwin) have simply meant the action 
* From the above it follows that local races or species produced by isola- 
tion should be more or less fertile with the parent form. This is a point 
, which might well be tested experimentally. 
n 
