89 
selection, which, in our present state of knowledge, seems equally 
difficult to understand, and failing this, we must suppose they are due 
to certain laws of heredity of which we are at present ignorant, owing 
to our not knowing what the fundamental laAvs governing heredity 
really are, and it is this last solution I am inclined to favour, first, 
because there is a fair probability of its leading to something by 
which our knowledge will be increased, and secondly because if these 
characters are vestigial it seems difficult to understand how a transverse 
ridge, a longitudinal ridge, a tubercle, and a depression, can in various 
species be relics of the same ancestral character. On the other hand 
it is easy to understand how different characters might arise in 
different species, which need not necessarily be variations of one 
character, but which at the same time are bound by the laws of 
heredity to bear a strong similarity, and consequently the closer the 
relationship of species the closer we should expect to find the secondary 
characters. As a good example I think the characters shown by the 
genus llybius in the Dytiscidae clearly indicate this side of the question. 
The $ characters all run as far as I know on one definite line, and 
that is the last ventral segment of the abdomen is keeled and 
wrinkled in a very peculiar way. In the 2 s the centre of this segment 
is compressed in the middle, and the appearance in both sexes is that 
there is a superfluous amount of chitin, or that the segment grows too 
large for the requirement of the species. In the male the difficulty is 
got over by forming a keel and wrinkles, and in the ? by bulging 
the segment up in the centre. Thus on the one hand the characters 
are so closely allied in the different species as to be almost identical 
to each other, this agreeing with the close similarity between the 
species, but on the other hand the very artificial appearance of the 
segment is opposed to its being a vestigial character. 
In fact it seems a general law, that the closer the relationship 
between members of the same genus, the closer are their secondary 
sexual characters related, but directly we try to tabulate the different 
families on these characters we are at fault, and this is I think 
because these characters are developing divergent peculiarities, rather 
than vestigial convergent ones. Another of the interesting points 
connected with secondary sexual characters, is in reference to those 
characters which suddenly crop up at wide intervals in distant families. 
Take for instance some species of the genus Amara ; we find in some the 
posterior tibias clothed with pubescence on the inner side in the $ , 
and we have no other genus of British ground-beetles showing this 
character, and it is not till we reach the end of the Avater-beetles, that 
we find it in Hydraena gracilis. Again there is a solitary species of 
ground beetle Sphodrus leucoptthalmus, which in the male has large 
hind trochanters, and we must look among the British Staphylinidae 
before we find a similar structure in Homaliumpygmaea. 
Stenus jitno with its pubescent metasternum has the same characters 
as Scaphidium 4^-macalatmn, a clubhorn in this respect. 
If these are instances of similar characters arising de novo in each 
instance, it is a remarkable circumstance well worth noticing, but if it 
is due to any hereditary connection, then the persistence of some of 
these structures is simply amazing, and when connecting links are 
found between ground beetles, water beetles, “ staphs,” and clubhorns, 
the question is when does heredity cease, and what is the cause of 
