1Q PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxxvi. 



group. In S. hiunguicidatus Stimpson, in the forms pachymerls Cou- 

 tiere, lophodactylus Coiitiere, and pescador'ensis Coutiere, these hairs 

 are disposed in tufts formed in line, directed obliquely downward or 

 even perpendicular to the surface of the mobile finger {S. lopho- 

 dactylus Coutiere) ; at the same time, the chela terminating the sec- 

 ond pair is provided with a large number of tufts of long hairs, some 

 carried by the jjalm, especially on its lower face, others by the mov- 

 able finger, these last arranged regularly in a brush. 



Judging by the sketch of an Australian specimen at the British 

 Museum, taken some time previously, S.spinige?' (Stimpson), with re- 

 curved hook, would also fall in this group, but this is not significant 

 at the present time. 



No American species, unless it heS.spinifrons (H. Milne Edwards), 

 belongs to the Biunguiculatus group ; this species can be equally 

 claimed by the L^evimanus group, this group, as I have pointed out, 

 being the continuation, pure and simple, of the preceding one, the 

 characters of which it extends to the extreme limit, especially those 

 which concern the reduction of the antennal scale. The character 

 which essentially distinguishes the two groups, in spite of its con- 

 stancy and its importance, is itself only the variation in the cleansing 

 apparatus present in the species of the Biunguiculatus group; the 

 bristles of the chela of the second pair, especially those of the movable 

 finger, have not persisted in the L.evimanus group, while the brush 

 of the little chela of the first pair has acquired the quite remarkable 

 development that I have described above. If one tried to express 

 these facts in plain language one would say that the experiment of 

 the numerous cleansing appliances had been abandoned in the de- 

 scendants of certain of these species, and that a single apparatus, 

 much more perfect, had been substituted for them. 



The continuity of the two groups is so evident that it leads to this 

 conclusion : If the species of the Biunguiculatus group are not rep- 

 resented on the American coasts, it is because they have all undergone 

 the variation toward the L.T5vimanus group, with the exception, per- 

 haj^s, of S. paraneptnnvs Coutiere, a species particularly instructive 

 because of the much more feeble development of the brush of bristles 

 of the small claw. 



In the Indo-Pacific region, on the other hand, this variation ap- 

 pears to be very rarely realized, since the only species which presents 

 it up to the present time is S. sladeni Coutiere, which has not, how- 

 ever, altogether the aspect of the American species of the Ljsvimanus 

 group. 



A rather similar condition is observed in the relations between the 

 Paulsoni and Brevicarpus groups, with two differences: First, the 

 species of these two groups are still found side by side on both 

 American coasts; second, the known distribution of the L^evimanus 



