928 On Aquatic Carnivorous Coleoijtera or DytiscidoB. 



points as above mentioned, while the lower forms have become nearly or quite 

 extinct. Hyph3'drus is the highest form by far, and it is rich in species, while the 

 remaining fragments indicate that the ancestors were probably more numerous in 

 species. I should expect that we maj' find that Dytiscidse with similar posterior 

 articular cavities, but with smaller coxse and less developed swimming legs, have 

 existed at a past geological epoch in South Africa. 



It seems clear that South Africa is the metropolis of the Hyphydrini, all the 

 fragmentary forms are found there, and there only ; and although the highest form, 

 Hyphydrus, is widely distributed, yet the distribution is such that South Africa 

 may be said to be its centre. No species is known from the western hemisphere. 



II. 9. — Group Hydroporini. {Vide p. 389.) 



This extensive secondary aggregate or group comprises ten genera and no less 

 than three hundred species. They are small insects, the largest size attained being 

 about 6 m.m. of length : the surface is alwaj's more or less punctured, and frequently 

 bears a delicate pubescence. 



The prosternal process is attenuate and acuminate towards the apex, and never 

 has a truncate hind margin; its apex always attains the intercoxal process of the 

 metasternum. The mesosternum is always placed at a considerable angle to the 

 metasternum, so that it is but little visible between the prothorax and metathorax. 

 The hind coxal cavities are either contiguous or approximate, never widely separate; 

 the coxal process always shows an outer angle, (or more or less rudimentaiy lobe) 

 projecting over the coxal cavity, so that this latter is never placed completely 

 external to the coxal process. The hind margin of the posterior coxa is free, and 

 not soldered to the ventral segments. The epimeron of the mesosternum is never 

 reduced to a linear band. The scutellum is invisible although a minute portion of 

 its apex is occasionally exposed. 



The large number of insects connected by these characters show a great deal 

 of variation in different points of their structure ; indeed there is scarcely any part 

 of the external skeleton that remains unvaried throughout the aofSfresfate. Never- 

 theless by its positive and negative characters it is a perfectly distinct group in the 

 Hydroporides. The coxal processes never show the two widely separated coxal 

 notches with a largely developed extra rimal lobe, such as is seen in the Hydrova- 

 tini ; and the ventral segments are not soldered with the coxa5 as in Bidessini. 

 While from the Hyphydrini the group is distinguished by the coxal cavities never 

 being quite exposed and much separated. Sternoi^riscus again has the mesosternum 

 very invisible, and with a very narrow, sublinear epimeron. 



The genera, with one exception, have the fork of the mesosternum unconnected 

 with the intercoxal process of the metasternum, and thus form an exception (but 

 not an unique one) in the Dytiscidae, the vast mass of species of the family having 

 these two parts connected. 



