240 On Aquatic Carnivorous Coleoj^tera or Dytiscidce. 



ruent and are at their maximum in Dytiscus, where all are large, and those placed 

 on the two apical segments are, contrary to the usual rule, larger than the others, and 

 are of a transversely elongate, elliptical form. In Cybister and Megadytes the 

 abdominal stigmata are all small, and the quite small apical one, passes through the 

 segment in the form of a tunnel, instead of opening directly through it as it does 

 in others of the family ; moreover in the two genera just named, the terminal 

 segment appears to be much more retractile and mobile than it is in other forms. 

 The side pieces of the hind body consist of a hard membrane, becoming externally 

 harder and corneous, and attached to the side of the turned up edge of the ventral 

 plates ; the side piece of the basal segment is marked in the Colymbetini and 

 Dytiscini with transverse rugse, which do not occur in any other Dytiscidse, and are 

 probably of some assistance in the process of respiration. 



The ventral plates or segments, are six in number, and are hard and corneous ; 

 they are transversely arched and each one has its outer margin turned upwards, 

 and even somewhat inwards, and this part is marked off from the rest of the surface 

 by a raised margin, so that the upjser edges of the ventral plates look at first sight 

 as if they did not belong to this part of the body, but were rather the abdominal 

 side pieces ; the margin giving them this appearance is however nothing but a 

 raised carina for the accommodation of the edges of the elytra, a similar margin 

 being developed, on the outside of the hind coxa and even along the side pieces of 

 the metasternum. The basal segment is very much modified in form to accurately 

 adapt it to the hind coxa, and for this purpose its middle appears to be completely 

 cut away, so that in an undivided insect, the first visible ventral segment appears 

 to be separated into two pieces placed one on each side and separated in the middle 

 by a considerable interval ; but on dissecting off the hind body it is seen that the 

 two pieces are connected together by a slender isthmus, but that this middle piece 

 or isthmus is concealed by being turned upwards at right-angles to the rest of the 

 plate. The basal segment carries on its upper face (and therefore as it were in the 

 interior of the body) a transverse corneous partition, extending all across the hind 

 body, and adapted to the reflexed postei'ior portion of the hind coxa ; this concealed 

 transverse partition is the true first ventral plate, the basal plate just spoken of 

 being really a portion of the second segment : it results from this arrangement that 

 the first dorsal plate is attached to what is truly the second ventral plate, the 

 second dorsal plate to the third ventral, and so on ; but this nomenclature is not in 

 use and the first visible ventral segment or true second plate is called the first plate. 

 The second ventral plate is more or less slightly emarginate in the middle in front, 

 by being adapted to the projecting internal laminte of the coxfe ; in some cases this 

 emargination is but slight (Cybister, Dytiscus) but in others Dytiscus duodecim-pus- 

 tulatus No. 462 e.g.) it extends nearly to the hind margin of the segment, so that 

 in such a case if the ventral plates are counted along the middle line they appear 

 to be only four in number. The sutures between the first and second, and 



