On Aquatic Carnivorous Coleoptera or Dytiscide. 885 
(i.e., Agabus wasastjernz and its allies), aad by yet other species a great approach is 
made to members of the second group of Agabus; D. chalconotus of this group and 
A. lugens of the second group being in most points of their structure very similar, 
but the difference in the coxal lines essentially distinguishes the two, and is more- 
over confirmed by an important difference in the hind claws of the two species these 
being elongate and simply curved in A. lugens, while in D. chalconotus the curving 
occurs only near the apex and that in a slightly unequal manner, the apex of the 
outer claw being thinner and more curved than that of the inner. 
Group 20. 
Coxal border very wide ; the coxal lines gently, not abruptly, turned outwards 
at the extremity ; prosternal process punctulate, slightly raised along the middle. 
The fact that a large proportion of the area of the coxal processes is formed by 
the coxal border distinguishes these species from all the other groups. In Agabus 
tristis, the border is not quite so large as in Dytiscus bipustulatus. This latter 
species is a most extraordinary one on account of the great variation it shows 
in structural and sexual characters. 
Group 21. 
Anterior portion of prosternum united with prosternal process so as to form a 
sensible angle ; prosternal process compressed ; wings of metasternum rather short ; 
coxal lines not greatly turned outwards at the extremity ; coxal border small; male 
anterior tarsi much developed ; hind legs slender. 
For one of these two species C. J. Thomson proposed the generic name Acatodes, 
but his separation of the insect cannot be adopted at present; he appears to have 
relied chiefly on the elongate basal joint of the hind tarsus, but this is found in most 
of the Agabi with slender and elongate swimming legs. The peculiar bend of the 
prosternum is very marked in Dytiscus fuscipennis, but is comparatively obscure 
in Agabus coxalis ; in this latter species the male tarsi are broader and have larger 
palettes beneath than is the case in D. fuscipennis. 
Group 22. 
Prosternal process extremely compressed; wings of metasternum short, coxal 
lines much approximated, abruptly turned outwards at the apex ; apical joints of 
the antennee dilated in the male. 
C. J. Thomson restricted the generic name Agabus to one of these three species 
(Dytiscus serricornis) ; but the isolation of the group under a separate name can 
searcely be adopted at present, especially as the characteristic structure of the 
prosternum and coxal lines is much diminished in A. clavatus, Lee. ; but of all the 
groups I have adopted for the genus this is the most likely to prove really distinct. 
The excessively compressed prosternal process is very similar to that found in the 
5 X 2 
