178 



as in one species of the latter genus they are shaped entirely as in Libystes. Indeed I see no 

 reason, judging from the evidence available, to maintain the genus Catoptrus (which is of a 

 later date), but yet I shall keep the genera separated, because I had no opportunity of examining 

 a species of Libystes. 



Key to the species: 



1. Margins of carapace entire, unarmed. Last two joints of 



posterior pair of ambulatory legs not paddle-like . . 2 

 Margins of carapace toothed anteriorly. Last two joints of 

 posterior pair of ambulatory legs paddle-like, apparently 



adapted for swimming L. edwardsi x-\lcock i) 



2. Carapace subquadrilateral ■ L.. alphonsi Alcock ') 



Carapace elliptical ^- nitidus A. Milne-Edwards ">). 



Catoptrus A. Milne-Edwards. 



1870. Catoptrus A. Milne-Edwards. Ann. So. Nat. (5), t. 13, p. 82. 

 1888. Goniocaphyra de Man. Arch. Naturgesch., Jahrg. 53, i., p. 339. 



This genus does not belong to the Catometopous Crabs, as has been lately proved by 

 BoRRAD.'ViLE *), who rightly ranged it among the Portunidae, in the vicinity of Carupa, and 

 instituted a new subfamily {G onto cap hyrinae), apparently then in ignorance of de Man's statement ') 

 that Goniocaphyra -truiicatifrons is identical with Catoptrus nitidus; already de Man regarded 

 his genus as belonging to the Portunidae. It is true, that the most characteristic feature of the 

 Portunidae, the paddle-like transformation of the last two joints of the posterior legs, is absent 

 in Catoptrus^), but on the other hand the form of the carapace, the toothing of the antero- 

 lateral margins, the elongated chelipeds, the shape of the abdomen of the cf (broadly-triangular, 

 with all but the last two segments fused) and last, but not least, the peculiar lobes at 

 the first maxillipeds, so characteristic of the Portunidae, are likewise 

 present in Catoptrus, as has been already shown by Borradaile '). These lobes, as is shown 

 in the figure, form two rather thick membranous expansions at the inner side of the endopodite 

 of the first maxilliped (PI. 9, fig. 3^, a). 



Though thus the systematic place of the genus is doubtless established among the true 

 swimming-crabs, I have in the present paper retained Catoptrus among the Goneplacidae, 

 where it is closely related to Carcinoplax. 



1) Journ. As. Soc. Bengal, v. 69, prt 2, 1900, p. 306; 111. Zool. "Investigator", Crust., prt 10, 1903. pi. 6:, f. I. Hab. Persian 

 Gulf and Andamans. 



2) Journ. As. Soc. Bengal, v. 69, prt 2, 1900, p. 306: 111. Zool. "Investigator", Crust., prt 10, 1903, pi. 61, f. 2. Hab. Andamans. 

 This species, though only known by a single young specimen, is probably identical with L. nitidus. 



3) Nouv. Arch. Mus. Paris, t. 4, 1868, p. 83, pi. 20, f. 5 — 7: Kobili, Ann. Se. Nat. (9), t. 4, 1906, p. 297. Hab. Zanzibar 

 and Djibouti. 



4) Proc. Zool. Soc. London, 1900, p. 577. 



5) Notes Leiden Mus., v. 12, 1890, p. 67. 



6) Libystes., which is certainly very closely related to Catoptrus.^ presents such swimming paddles in the species L. edwardsi Alcock. 



7) Fauna and Geography ilaldive and Laccadive Arch., v. i, 1903, p. 425, textfig. no. 



30 



