94 K. ZIMMERMANN. 



in small particles of food to the mouth by a sweeping movement ; (6) 

 to comb the hairs of antennse and antennules ; and (c) to guard, when 

 folded, the exit channels of the branchial cavity. In Galathea and 

 Munida, the joints of the endopodite are long and thin, and the whole 

 maxillipede approaches in form the type of the primitive Arthropod 

 appendage, save that the endopodite is slightly flattened, and can, when 

 folded, form a fairly plate-like structure. P. platycheles shows greater 

 specialization of the endopodite, the joints being very much flattened 

 and broadened, although the span of the limb when straightened is con- 

 siderable. P. longicornis is in this respect almost intermediate between 

 the tw:o types (Plate 6, A, B). All these species have specialized hairs of 

 two kinds on the ultimate and penultimate segments of the endopodite. 

 These are — (o) sweeping hairs (Plate 4, C), very long and slender, bearing 

 two regular rows of fine branches ; and (6) cotnhing hairs (Plate 4, I, i), 

 not so long as the former, but stouter, and bearing two rows of very stiff 

 points like the teeth of a double comb. 



Comparison with the third maxillipede of Cancer and allied Brachyura 

 is instructive. In Cancer, the broadening and flattening of the proximal 

 portion of the endopodite and the reduction of its distal joints have pro- 

 ceeded far (see Plate 3, D, E, F). The function of the endopodite here is 

 perhaps exclusively to protect the exit channel, and it is converted into 

 an organ primarily plate-like and protective, and destitute of sweeping 

 hairs. In connection with the life high up the shore, antennae and anten- 

 nules are reduced in size and importance (experiments prove the former 

 to be practically insensitive to touch), so that this double reduction 

 renders the combing of antennae and antennules by the hairs of VIII im- 

 practicable. The combing hairs, having lost their function, are obviously 

 degenerate : though they retain their two rows of lateral branches, these 

 branches are so much thickened as to be almost fused in rows, and quite 

 disqualified as combing teeth. The fact that Algal growth has been found 

 attached to the antennae of practically every specimen of Cancer pagurus 

 examined appears to prove conclusively the absence of combing device 

 and sensitivity. Carcinus (a more active Brachyuran than Cancer, and 

 one proved by experiment to have greater power of antennal perception), 

 has the third maxillipede just long enough to clean the antenna tip, and 

 it is noticed that in members of this genus the combing hairs have persisted 

 without degeneration — also Algal growth is less commonly present on the 

 antenna, and never present near the tip. On the other hand, P. ^jk^//- 

 cheles presents, in this matter of the combing hairs, a condition markedly 

 analogous with that of Cancer. This must be connected with the general 



