960 Qn A<iuatLC Carnivorous Coleoj^jtera or Dytiscidce. 



coxal lines are not greatly turned outwards towards their termination then the 

 supra-articular border is very broad. 



There are no stigmatic rugse on the first ventral segment, and tlie terminal two 

 stigmata are small or of moderate size. The swimming legs are highly developed, 

 but vary uiueh in this respect, being in certain forms rather elongate and slender, 

 in others short and thick ; their tarsi are always terminated by two very distinct 

 nearly straight claws, which may both be elongate, although the upper or outer 

 one is usually much shorter than the other ; the hind margins of the four basal 

 joints of these hind tai'si are furnished with beautiful cilise or elongate scales which 

 are set close to one another, and overlap the following joint, to the surface of which 

 thev are so extremely closely pressed that they scarcely interfere with the perfect 

 smoothness of the face of the tarsus ; their scales vary greatly in the tribe, but 

 they ai'e always conspicuously present, and occur in no other water beetles out of 

 the tribe. In Eretes they are short and broad, and form a single series, and as 

 they are easily removed by friction with a hard surface, they may escape detection 

 in it, but even when quite removed by ill usage there remains a series of punctures close 

 to and parallel with the hind margin, marking their jDoints of insertion and so indi- 

 cating their existence in an incontrovertible manner. In the species where they 

 are most highly developed as in D. vittatus (No. 1049), they are very conspicuous, 

 and arranged so as to form an arch, and are so extensive so to cover quite one-half 

 of the area of the outer face of the tarsus. 



In many other respects there occurs a considerable amount of variation in the 

 structure of the Hydaticides, and the greater part of these exceptional features 

 occur in Eretes ; some have been noticed in the description of that genus, but 

 others still require mention. In all the Thermonectini and Hydaticini the anterior 

 and middle legs are separated by a considerable space, but in Eretes they are much 

 more contiguous, and this difference is accompanied by some important correlative 

 discrepancies in the structure of the adjacent j^aits. The front of the prosternum 

 in Eretes shows a kind of longitudinal face in the middle, placed nearly at right 

 angles to the plane of the prosternal process (and so simulating to a superficial 

 observation the prothoracic structure of the Uydroporides); whereas in Acilius and 

 Hydaticus there is no trace of this longitudinal thickening of the middle of the 

 front of the jirosternum. In (Ethioiiectes howevei', the prosternum is greatly 

 incrassate along the middle and conq)ressed, and a similar btructure, though not 

 developed to so great an extent, is approached by some other Thermonectini. '1 he 

 prosternal process is acuminate, narrow, and elongate in Eretes, whereas in all the 

 other genera it is short and broad and obtuse or rounded : the inter-coxal process 

 of the metasternum is marked with an elongate narrow groove in Eretes, and onl}' 

 connects in an imperfect manner with the mesosternal fork, whereas in the other 

 members of the aggregate the connection with tlie mesosternal fork is intimate and 

 perfect, and the depression for the prosternal process is peculiarly short, broad and 



