As to some other points of difference (branchial grooves on carapace, shape of tooth 

 at inner angle of wrist, absence or presence of longitudinal median ridge on palm of cheliped), 

 I cannot tïnd them really constant. 



Each of the two species were split up by H. Milne-Edwards l ) into a number of 

 "speek ;", that seem to have been founded on merely individual variations and have been 

 withdrawn by subsequent authors. Yet each of both species seems to offer constant varieties. 



As to Grapsus maculatus, that is cosmopolitically distributed throughout the warmer 

 ions, Miss. Rathbun remarks 2 ) : "The common rock crab of the tropics, Grapsus grapsus, 

 is separable into two forms, one in which the lobe on the wrist is very broad and terminates 

 in a short point [G. grapsus typical), and one in which the same lobe is narrovv and terminates 

 -in a long narrow spine [tcuuicristatus Herbst). The former inhabits the coasts of America, 

 including the outlying islands, such as the Galapagos, and also the eastern shores and islands 

 of the Atlantic Ocean; the latter is restricted to the oriental region". I can confirm this state- 

 ment, though it must be owned, that the difference in shape of the inner angle of the wrist 

 is scarcely perceptible in some cases, where specimens of either Atlantic or oriental origin 

 are compared. 



It seems to be of more importance, that both Gr. maculatus and Gr. strigosus possess 

 a slender-legged form, named resp. Gr. gracilipes H. Milne-Edwards and Gr. longitarsis Dana, 

 that have been given the rank of separate species. 



In order to elucidate the difference between Gr. maculatus and its subspecies gracilipes, 

 the penultimate leg of both is figured on PI. 4, Fig. 2 and 3. Both are represented natural 

 size; the breadth of the carapace in the specimen of Gr. maculatus is 46 mm., that of 

 Gr. gracilipes*) is much less, $7 mm. Nevertheless we may remark, that the length of the 

 carpo- and propodite together is nearly the same in both specimens, and that in Gr. gracilipes 

 the breadth of the propodite is one-fifth its length measured in the median line, in typical 

 Gr. maculatus more than one-fourth its length. 



As to Gr. longitarsis, Miss Rathbun who had occasion to examine the typical specimen 

 of Dana, first regarded *) it as a subspecies of Gr. strigosus, that differs in somewhat more 

 elongate propodites, in meropodites narrowing more distinctly, in the more enlarged abdomen 

 of the Q (in the d 1 the abdomen is equilaterally triangular) and in the front being less advanced. 

 Afterwards 5 ) this author raised the subspecies to the rank of a separate species and added 

 some more characters, the most important of these being that the front is wider than in 

 Grapsus strigosus. 



1) Arm. Sc. Nat. (3) t. 20, 1853, p. 167 — 170. 



2) Buil. U. S. Fish Comm. for 1903, v. 23 pit 3, 1906, p. 838. 



3) This is the very specimen mentioned by de Man (Notes Leyden Mus., v. 5, 18S3, p. 159). This author afterwards had the 

 opportunity to examine another specimen from Ternate and to compare it with the typical gracilipes of Milne-Edwards (Abhandl. Senckenb. 

 Gesellsch., Bd 25, Heft 3, 1902, p. 5 02 )' From the measurements taken (the Ternate specimen is only 1 mm. broader than that of the 

 Leiden Museum) we must conclude, that the length of the propodite of the penultimate pair of legs varies somewhat individually, for in 

 the Ternate specimen it is 25 mm., in my specimen 23.5 mm.). 



4) L. c. p. 838, textfig. 4, pi. 8, f. I. Grapsus subquadrattts Stimpson (Proc. Acad. Nat. Sc. Philadelphia, 1 858, p. 103; Smithson. 

 Misc. Coll., v. 49, 1907, p. 119, pi. 16, f. 4) is added as a synonym. 



5) Mem. Mus. cornp. Zool. Harvard Coll., v. 35, n° 2, 1907, p. 28. 



72 



