950 On Aquatic Carnivorous Coleoptera or Dytiscide. 
any other genus, and a fortiom, no group the ancestor of any other group ; neither 
can I find it credible that the similarities visible are due to cescent from a common 
unknown ancestor. 
The geographical distribution of the tribe is extensive, but fails to reach the 
cooler temperate regions, or New Zealand and the Pacific islands. 
III. 2.—Tnbe HYDROPORIDES. (Vide p. 319.) 
About five hundred species, arranged in twenty-six genera, and forming five 
secondary aggregates or groups (in addition to two isolated genera) form tls tertiary 
ageregate or tribe. It is therefore by far the most extensive of the tribes of 
Dytiscidee. They are all small insects, the largest size attained being about 6 m.m. 
of length, and the surface of the body is nearly always punctate. 
The tribe is specially defined by the structure of the prosternum, which is 
deflected or bent between the front coxze so as to be very discontinuous in the 
plane of its direction with that of the prosternal process. The anterior border of 
the hind coxa is directed forwards as well as outwards, in such a manner that its 
greatest extension in front is at a point nearer to the epipleura than to the middle 
line (longitudinally) of the body. The metathoracic episternum penetrates to the 
middle coxal cavity. The front and middle tarsi have the three basal joints formed 
so as to show a flat sole clothed beneath with a kind of glandular pubescence: these 
tarsi have usually only four visible joints, the joint between the third and (true) 
fifth joints being reduced to a mere knot ; in the cases where this rudimentary 
joint is quite conspicuous (Necterosoma, Sternopriscus) it usually remains 
small in comparison with the adjacent joints. The scutellum is usually quite con- 
cealed, but in Celina is large and conspicuous. 
Of the above characters, that drawn from the structure of the prosternum is the 
one by which an insect may most certainly and readily be identified as a member 
of the tribe ; but the tarsal structure and other points must be also taken into 
consideration, for if not, Suphis (a primitive form of Noterides in the Dytisci 
fragmentati) might be supposed to belong to the Hydroporides, as the form of its 
prosternum corresponds to the above definition. 
The Hydroporides show no tendency whatever to thickening or strengthening of 
the anterior parts of the prosternum along the middle, as do all the higher and larger 
forms of the Dytiscidee : on the contrary, the anterior parts of the prosternum are 
always feeble and but little developed, and afford but slight protection to the front 
coxee, The anterior transverse band of this part is always small, and has not the least 
tendency to being arched in the transverse direction: the band between the front 
coxx, connecting this transverse part with the prosternal process, is never more 
prominent than the coxe, and is often more or less depressed between them : the 
prosternal process itself is much more variable and may attain a large development 
