8 THE TANGANYIKA PROBLEM. 



the antennal scale by £ to nearly i its length. There are from 12-15 teeth on its 

 upper edge, three (rarely two) of which are behind the orbit. The teeth become 

 more widely spaced distally, and the last one is generally separated by rather less 

 than half the length of the rostium from the simple, sharply pointed tip. The 

 lower margin of the rostrum bears from 10-20 teeth, which extend quite to the tip. 

 Below the orbit the anterior margin of the carapace is produced into a triangular 

 tooth, but there is no "antennal 1 ' spine such as is present in most species of 

 Caridina, e.g., in C. wyckii (Fig. 3). A little way back on the side of the cara- 

 pace, and below the level of the sub-orbital tooth, there is a well-marked " hepatic " 

 spine. The lower anterior corner of the carapace is evenly rounded, and theie is no 

 pterygostomial spine. 



The peduncle of the antennules (Fig. 4) falls short of the distal tooth on the 

 outer margin of the antennal scale. The first joint is about equal in length to the 

 two succeeding joints together. The basal spine is small and slender, its tip falling 

 short of the distal end of the joint by \ the length of the joint. The short 

 spine on the distal end of the first joint reaches to about \ the length of the 

 succeeding joint. The ocular peduncle is rather shorter than the first joint of the 

 peduncle of the attennule. 



The mandibles (Fig. 5) are somewhat dissimilar on the two sides. The cutting- 

 edge is separated from the molar process by a shallow emargination, within which 

 are set two stout setae (in C. wyckii there is a row of about 10), followed at a little 

 distance by a thick brush of finer setce just in front of the molar process. 



The first maxillae (Fig. 6) differ from those of Caridina, and such allied genera as 

 Atya and Atyaephyra, in the smaller size of the two inner lobes, the inner edges of 

 which are much shorter, while the lobe which in these genera represents the exopod 

 is here absent. 



The second maxilla; (Fig. 7) also depart somewhat from the type characteristic of 

 the Atyidtz. In the other members of the family the middle lobe of the endognath 

 (the proximal division of the lacinia externa in Boas's nomenclature) is very much 

 expanded, overlapping both the other lobes and presenting a very long, straight, 

 inner edge. In the present form this lobe is much smaller, its inner edge being 

 hardly longer than that of the distal lobe, which it does not overlap. The proximal 

 lobe, as in the other Atyidte, is large and is overlapped for a short distance by the 

 middle lobe. The scaphognathite is truncated anteriorly and produced to a point 

 posteriorly, where it bears, as usual in this family, a tuft of very long, slender setae, 

 hooked at the tip but not presenting the curious swelling and tooth near the base 

 which characterise these setae in C. wyckii. 



In the first maxilliped (Fig. 8) the exopod tapers gradually from the base with 

 hardly an indication of the external lobe (marked a by Boas) present in Caridina 

 as in most Enkyphota. The spipod, rudimentary in Caridina, seems to be quite 

 absent. 



The third maxillipeds (Fig. 9) extend forward as far as the end of the first joint 

 of the peduncle of the antennules. There is on the outer surface of the coxal joint a 

 conical curved papilla similar to, but smaller than, the papilla to which the epipod 

 of this appendage, here absent, is attached in C. wyckii. The exopod exceeds in 

 length the joint from which it springs. The terminal joint is shorter than the 

 penultimate joint, and presents a remarkable structure (Fig. ga). About the middle 



