330 CRUSTACEA. 



third joint of the outer maxillipeds short and broad, Pacliygrapsi 

 excluding, however, Edwards's Cyclograpsi. It is important then to 

 ascertain what are the natural groups among these species. In ex- 

 ternal form, there are two groups of very obvious and trenchant 

 characters : the first having arcuate sides, and a front not as long as 

 half the breadth of the carapax, as the pictus, strigosus, and variegatus; 

 the second having straight sides, often convergent backward, and a 

 front longer than half the breadth of the carapax, like the cruentatus, 

 messor, and plicatus. These groups appear to be natural and of equal 

 importance; and the latter forms a transition to Sesarma. In the former, 

 the outer antennae come out either side of the front; in the latter, often 

 from directly beneath the front. May we retain these as true genera, 

 or are other subdivisions indicated by the outer maxillipeds ? As to 

 the length of the third joint of these organs, we would observe, that 

 there are the oblong and the short forms, in both of these groups. In 

 the former group, the two varieties connect by very gentle gradations. 

 In the pictus, the third joint is but slightly shorter than the second ; 

 in the strigosus, the difference is a little greater ; in the plani/rons, a 

 species from Valparaiso, so resembling the variegatus that it has ap- 

 parently been called by this name, the length does not exceed the 

 breadth. The variegatus, having a nearly horizontal front like the 

 planifrons, has the same joint oblong as in the >tcfr, the latter species 

 with a vertical front. The variations in the form of this joint hence 

 do not correspond with any peculiar character in the front ; and they 

 occur among species that have the same general outline, that is, ar- 

 cuate sides and a short front. Again, among the square Grapsi, there 

 are both kinds : the cruentatus has the third joint of the outer maxil- 

 lipeds as long as the second joint ; in the plicatus it is a little oblong ; 

 and in the messor and Thukujar it is as broad as long. If we follow 

 then the peculiarities of this joint, we must either divide in two or more 

 parts each of the groups pointed out, or associate the species without 

 reference to what seem to be characters of quite as great importance, 

 the character of the front and general form of the species. For our- 

 selves, we deem it best to disregard the form of this maxilliped joint, 

 since we have found that even in the genus Xantho (for example), it 

 may be both transverse and oblong. We therefore divide the genus 

 Grapsus into GRAPSTJS proper, in which the sides are arcuate and 

 front narrow ; and GONIOGRAPSUS, with the sides straight and front long. 

 The latter group is not the Goniopsis of De Haan, nor the Pachygrapsus 



