On Aquatic Carnivorous Coleoptera or Dytiscida. 969 
that the Carabidze in whole or part are modified Dytiscide. Indeed Herr Kolbe’s 
theory and classification are so evidently based on a false premise that it is not, 
I think, necessary to point out in detail their unsatisfactory character. 
The close approximation existing between the Carabidze and Dytiscidze does not, 
in my opinion, at all bear out the theory that it is the result of descent from a 
common ancestor. It is quite certain if the approximation is due to heredity, 
that the most primitive forms of the two families should bear a resemblance to 
one another, not only in one special point but more or less in other points ; for 
it is clear that as we go back in the genealogy and approach the common ancestor 
so must the descendants become alike in more than one point: there must be 
as it were an approximation on all points, not only on some special point. This 
is not, however, what we find to be actually the case : on the contrary, if we take 
the most primitive forms, such as Pelobius, Amphizoa, Methles, Hydrovatus, 
Suphis, and compare them together we are struck with the fact that we have before 
us some of the most dissimilar of the Dytiscide. The same fact is true of those 
Carabide to which the Dytiscidee approximate ; Trachypachini, Pseudomorphini, 
Scaritini, and Cyclosomus are in fact about as dissimilar Carabidee as could be 
selected from the whole of that enormous family. 
If we approach this interesting question from another point of view, that of the 
structure of any one part, we are brought to a similar conclusion: for instance, if the 
highly developed prosternal process of the Dytiscide and of certain Carabidze is an 
indication of genetic community, then it is clear that it is one of the most ancient 
and fixed features of the Dytiscide, and as it is a great advantage to its aquatic 
possessors, we ought to find it absolutely constant amongst them ; but this is not 
the case, the prosternal process remains comparatively rudimentary in several 
widely separated Dytiscidee (Vatellini, Tyndallhydrus, Andex); and it isclear that 
if the Dytiscidee and Cyclosomus possess this structure in common because their 
primitive ancestor possessed it, then Vatellini, &c., must have separated from this 
common Dytiscid-Carabid ancestral stock before this part was developed ; a con- 
clusion which is absurd; or that the Vatellini have lost by degeneration a character 
which they formerly possessed, and whose possession was advantageous to them ; 
a conclusion which there is not the least reason for believing, and of whose truth 
it would require very weighty evidence to convince us: and which is in opposition 
to the law of survival of the fittest. It is in fact quite clear that the highly 
developed prosternal process in Cyclosomus and the Dytiscidze is not the result of 
genetic community, but has been separately acquired. As regards Pelobius which 
is truly a “connecting link” between the Carabidze and Dytiscidee, there is not 
«the least reason too for supposing that it is ancestral to any Dytiscidz or any 
Carabidee, and there do not appear to me to be any grounds for supposing that 
the Carabidze are the ancestors of the Dytiscidee, or the Dytiscidze of the Carabide. 
If however, under the term “‘ Carabidze” we include not only the known existing 
