154 Memoirs of the Indian Museum. [Vol. IV, 



telson are distinguishable at a glance from G. chiragra or any of the other species already- 

 mentioned. It is true that one particular telson pattern in G. chiragra, in which the 

 flukes of the ' anchor ' are separated from the distal end of the median carina and form 

 short additional carinae on either side of it, seems to represent a type intermediate be- 

 tween that species and G. glabrous. But such specimens are invariably extremely small 

 in size, and even if it were impossible to separate the two forms at this early age, which 

 is not the case, there would still be no justification for uniting the two as Lanchester 

 has done. In addition to the distinctions afforded by the telson, G. glabrous and gra- 

 phurus may be distinguished from G. chiragra by the small size of the dorsal processes 

 of the ophthalmic somite and by the presence of a small tooth at the base of the outer 

 spine of the bifurcate process of the uropods ; but these characters are of little value in 

 the case of very small specimens. 



G. graphurus is separated without difficulty from G. glabrous by the presence of the 

 fine incised lines or grooves which traverse the first five abdominal somites in the 

 former species. The character is an obvious one and, as yet, there is no evidence of the 

 existence of intermediate forms. G. glabrous seems to be the most abundant of the two, 

 and exhibits a very considerable amount of variation in the development of the keels 

 on the telson. The keels may be narrow and well separated or swollen and touching one 

 another and the three median keels, as well as those on the last somite of the abdomen, 

 may or may not terminate in spines. There can be no doubt that the variation, as in 

 the case of G. chiragra, is continuous and the same objections that have already been 

 made must be urged in this instance also against the use of special names to designate 

 particular varietal phases. G. graphurus seems to exhibit a range of variation closely 

 similar to that of G. glabrous. 



The species belonging to the chiragra section of Gonodactylus form a very homo- 

 geneous group. It seems probable that they are more primitive than the others in- 

 cluded in the genus, and it may be inferred that all the other more highly specialized 

 forms were derived from the common ancestor of this section. 



The abundance of species of Gonodactylus in the Indo-pacific region and their 

 poverty in the Atlantic may perhaps indicate that they have originated in the former 

 area, while the very slight distinctions that exist between G. chiragra and G. oerstedi 

 may be taken as evidence that their extension to Atlantic waters is of compara- 

 tively recent date, though it is possible that unknown factors, absent in the Atlantic, 

 may have induced a more speedy differentiation of species in Indo-pacific waters. 



Of the species with a three-segmented mandibular palp, G. chiragra, in the simplicity 

 of its structure appears the most primitive, while G. graphurus is perhaps the most 

 highly specialized. 



As has already been pointed out, the vast range of variation in these species is the 

 more remarkable when the singular constancy of slight specific distinctions in other Sto- 

 matopoda is considered. I am of the opinion that four, or perhaps five forms in the 

 Indo-pacific, and two in the Atlantic, are sufficiently isolated to merit specific recogni- 

 tion ; but from the numerous varietal phases it is impossible, in the present state of 

 our knowledge, to draw any satisfactory conclusions. That the ceaseless production of 



