130 Annals of the Soutli. African Museum. 



times quite definitely missing, there are other cases when the ring 

 is incomplete or when coalescence with the preceding joint can only 

 be inferred from a line of suture. The eridopod or inner ramus of 

 the uropods may be a single piece or it may be divided into two 

 or three joints of varying relative lengths. The resulting differences 

 are rather easy to observe, and have been, in fact, of much service 

 in classification. But even here perplexities occasionally arise. 

 Among the species of Sympoda earliest described are Bodotria 

 scorpioides (Montagu) and Bodotria arenosus, Goodsir. These are 

 so much alike that their generic separation is hardly to be thought 

 of. Nevertheless the uropod of the former has a two-jointed 

 endopod, while that of the latter is provokingly undivided. 



When the question arises of arranging the families in a natural 

 order, one would probably think precedence appropriate to those 

 which retain the most primitive characters. Among these would be 

 the most complete segmentation of the body and the fullest equip- 

 ment of the segments with their several pairs of appendages. On 

 the first account the families with a distinct telson should stand 

 before those without one. But when the second point is also con- 

 sidered, we find the full complement of five pairs of pleopods 

 combined with entire want of a distinct telson, or in one case with 

 a telson of the smallest type. All other families with the telson 

 distinct have a diminished number of pleopods, varying from three 

 pairs to none. These differences refer only to the male sex, because, 

 so far as at present known, all the females with singular unanimity 

 dispense with pleopods altogether. In some families, however, the 

 males are in this respect like the females. 



The provision of exopods or swimming branches on the perseopods 



in the two sexes has its uses for systematic arrangement. But while 



in the majority of families the adult males have these branches well 



developed on the first four pair of peraeopods, the females are never 



so well provided, having at most exopods well developed on the first 



three pairs and a rudiment on the fourth. In both sexes the 



exopods may be limited to the first pair of peraeopods. For full 



advantage to be taken of these much-varying characteristics it is 



obviously important that both sexes should be observed. But, owing 



probably to the respective habits of these, it not unfrequently happens 



that new species have to be, or at any rate are, founded on specimens 



of a single sex, so that the characters of the other sex have to be 



guessed at or left out of count. 



These are a few of the difficulties which confront the systematist 

 in points the most readily available for his purpose. There are 



