88 
at one time was useful to the progenitor of the several species display¬ 
ing these characters, and which has gradually become obsolete from 
disuse. This, of course, puts the closure on the whole question at 
once, and if correct all well and good, but if not we are settling the 
matter in a way that is likely to stop all further investigation. 
There is another point which renders this explanation of these 
characters in Pterostichus a particularly difficult one, and that is that in 
the genus under consideration there is one species, Pterostichus niger, 
which shows the same character in the $ as in the S • That the 
character has in this case originated in the S , and is a genuine 
secondary sexual character, there can, I think, be no doubt, as in the 
other species of the genus which have similar characters in the S 
there is no sign of a corresponding structure in the $ . If then these 
are vestigial characters how is it that P. niger has them not only well 
marked in the S but in the $ as well, and shows no sign of losing 
them. I admit that there are of course great difficulties in the way of 
classing them as useful, and consequently developing, structures, but it 
is better to take this side of the case, which will prompt closer investi¬ 
gation, than to refer them to the limbo of ancestral characters on the 
decline. 
What we really want to know in these matters are the laws which 
govern heredity, and until we have some broad generalisation on 
which to base our arguments we cannot but wander in the dark. That 
there are simple laws which would cover our present knowledge of 
facts I have no doubt whatever, but what those laws are, and how long 
we shall have to wait for an exponent of them, time only will show. 
There are many facts which show how close and important a part 
heredity plays in these obscure secondary characters. No matter 
what genus of beetles we take up, if they display secondary characters, 
then we find that within the limits of the genus the characters run on 
similar lines, and this holds good in nearly all cases, and to such an 
extent that it becomes a most obvious truth to anyone studying the 
subject. It is doubtless this fact which has caused the bulk of these 
structures to be considered relics of some characters displayed in the 
common ancestor of the genus. To my mind this question opens up 
an interesting field of study as to how far natural selection can effect 
its ends by utilising small variations from the type. I know it is 
generally argued, and generally admitted, that variations can occur in 
all directions, and that it is by this means the innumerable adaptations 
in nature are effected through natural selection; but on the other 
hand we do not find the members of a genus diverging indiscriminately 
in all directions, but, as a rule, we find them varying in one general 
direction, or along certain lines, no matter whether we take secondary 
sexual characters or others, and it seems superficially as if natural 
selection did not have it all its own way, but that the hereditary 
character, history or tendencies, were so firmly ingrained into a genus, 
that natural selection was powerless to utilise variations except in the 
direction which was governed by the hereditary tendencies of the 
genus in question. Once more reverting to the secondary characters 
in the genus Pterostichus, I have tried to show the difficulties and 
objections of referring these to vestigial characters, and it seems 
equally obvious that sexual selection can have nothing to do with the 
matter. If then they are developing characters they are due to natural 
