On Aquatic Carnivorous Coleoptera or Dytiscidce. 237 



are separated by a vertical wall, or plate ; this wall ascends from the common suture 

 between the two coxte, and at its termination, spreads out above into a kind of 

 winged or bifid process which completely fills up a gap that would otherwise 

 exist between the arched roofs of the two cavities. The articular portions of 

 the coxa project backw^ards, and on each side of the projecting portion there is a 

 hollow or axilla which permits the base of the swimming leg to be rotated forwards 

 till the front edge of the femur can quite attain the longitudinal middle line of the 

 body. In the higher and larger forms of the Dytiscidre which I am now describing, 

 the articular cavities are entirely concealed by the coxal processes, which project 

 from the surface more or less slightly backwards, and are thus very consjiicuous on 

 the under surface of the body, the swimming legs protruding as it were from the 

 hind part of the projection, and being so close together that their inner margins 

 touch one another at the point of articulation. This structure, with modifications 

 in some details such as the size and form of the coxal lobes, prevails throughout 

 the Macro-Dytiscidte, without any exception. But in the smaller Dytiscidse the 

 articulation of the swimming legs is of a different and much less uniform nature. If 

 a Hyphydrus be looked at — and with this object one of the two swimming legs 

 should be carefully disarticulated — it will be seen that the swimming legs do not 

 project from behind and above any prominent coxal processes, and that they 

 are not contiguous, but are separated by a considerable space, and if the 

 leg has been carefully disarticulated, without damage to the articular cavity 

 it will be seen that this is a circular orifice, completely exposed and not concealed 

 by any projecting lobe, and a farther examination renders it evident that the space 

 separating the articulations of the two sides of the body, consists of the coxal 

 processes. Here then we have the articular orifice exposed and placed outside of 

 the coxal process, separated from its fellow and circular in form, instead of being 

 placed above the coxal process, concealed by it, and contiguous with its fellow ; on 

 separating the hind body from the coxa it is farther seen that the upper portion* of 

 the articular box projects further back than the coxal processes, these latter, being 

 adpressed to, and soldered to the former : thus the coxal processes, are not prominent 

 from the level of the base of the abdomen, and there is no axilla formed to permit 

 the flexion of the leg forwards : indeed the articular cavity being circular and exposed 

 permits the leg to be rotated forwards even farther and more freely than in the 

 Macro-Dytiscidae. This is the type of structure which admits of the greatest amount 

 of motion of rotation for the swimming leg and it attains its maximum of develop- 

 ment in the New-World genus Pachydrus, where it is accompanied by a consolidation 

 of the coxa with the abdominal segments. Many of the Hydroporini have the 

 articular cavities constructed in a manner that is intermediate between that just 

 described, and that of the Macro-Dytiscidse ; as a good instance of an intermediate 



* I have in my descriptions occasionally spoken of this upper portion of the articular cavity, when it is 

 visible under the name " pyxal process." 



2 1? 



