On Aquatic Carnivorous Coleoptera or Dytiscidce, 953 



out in Sternopriscus oscillator, where the luiddle tarsi of the male are very elongate 

 and their true fourth joint is developed to such an extent that it actually very 

 nearly equals the third joint. 



It is a further interesting fact that the only two genera, — and they are widely 

 different from one another — in which the true fourth joint of the tarsi is very con- 

 spicuous, are both peculiar to Australia. 



Also it should be noted that some otthe species of Sternopriscus, differing as they 

 do from all other Dytiscidas by the great development of the true fourth joint of 

 the middle tarsi of the males, have no two species in which this part is similarly 

 formed. 1 judge from this that each species has developed the peculiarity 

 independent of the others, and that their all possessing it (in various ways) is no 

 proof of genetic identity. 



There seems reason to believe that the Hydroporides have had ancestors with 

 five-jointed front and middle tarsi ; not only because the joint usually wanting (the 

 true fourth joint) remains quite visible in certain genera, (Necterosoma, Sterno- 

 priscus) ; but also because the posterior or hind tarsi are always five -jointed 

 throughout the tribe. And as in all the carnivorous Coleoptera except these Hydro- 

 porides, the front and hind tarsi have the same number of joints (fivej, and as this 

 is at present the case in a certain number of Hydroporides (where the fourth joint 

 is greatly reduced though still conspicuous), we may reasonably suppose that these 

 Hydroporides which have now only four joints on the front foot, had formerly five 

 joints, and that the true fourth joint has become reduced and lost, in conformity 

 with the plant-frequenting conditions of their existence ; it being the rule that 

 beetles which walk on vegetable substances have only four-jointed tarsi. The pos- 

 terior tarsi remain always five-jointed because they are used not for walking on plants 

 but for sswimming, and for this purpose reduction in area would be very 

 disadvantageous. 



The scutellum displays in this aggregate great diversity, though as it is concealed 

 from view a superficial or misleading conformity among the species is suggested. 



In Hydrovatini it is found in its minimum of development, so that when the 

 prothorax is separated from the after or middle body the scutellum still appears to be 

 absent, indeed the mesonotum in these insects seems to be excessively small and fra- 

 gile, and owing to the perfect way in which the elytra are locked together, I have not 

 succeeded in getting a view of the mesonotum without rupturing and destroying it. 



In Dytiscus ovatus (Hyphidrini) the scutellum exists as a moderately large 

 membranous plate ; the posterior margin is truncate and presents a sharp edge, but 

 this is not produced into a fine triangular plate : the anterior margin of the scutellum 

 is also truncate and is narrower than the posterior one. 



In Dytiscus duodecimpustulatus (Hydroporiiii^ Deronectes,) the anterior pieces of 

 the mesonotum are larger than in Dytiscus ovatus (Hyphidrini) but the scutellum 

 is not larger, and is of a different form, being rhomboidal, and angulate both in 



