14 THE VOYAGE OF H.M.S. CHALLENGER. 



movable limb y' of the forceps, which is furnished with a large muscle which runs up into 

 its hollow base, while its tip is spoon-shaped on its external surface which is opposed to 

 the third portion, the fixed limb, e, of the forceps. This is a raised ridge on the flat 

 surface of the appendage, with its tip free, and, in most cases, bent into a hook, which is 

 curved towards the spoon-like hollow at the tip of the movable limb. 



In Gonodactylm (PI. XV. fig. 4) the terminal joint A is divided into two lobes, a 

 and h, by a marginal notch, but the inner lobe is not separated from the outer one by a 

 suture. Pseudosquilla has justly been I'egarded, by nearly all writers on the Stomato- 

 poda,' as very closely related to Gonodactylus, and the modified appendage (PI. XV. 

 fig. 10) is very similar indeed to that of the latter genus, as will be seen by comparing 

 the figure with that of Gonodactylus graphurus (PI. XV. fig. 8). In both, the inner 

 lobe of the terminal joint is larger than the outer one, and separated from it by a marginal 

 notch but not by a suture. 



The close relationship between Squilla and CJdorideUa is unquestioned, and I have 

 shown that the species which have been divided between the two genera really form a 

 single phylum or stem which cannot be divided into two genera. A comparison of the 

 modified abdominal appendage of the male Squilla leptosqudla (PI. I. fig. 2), or that 

 of SquUla quinqiiedentata (PI. II. fig. 6), with that of Squilla (Chloridella) chlorida, 

 will show that there is the closest resemblance. In all these the outer lobe a of the 

 terminal joint is separated by a suture from the inner lobe, the tip of which is elongated 

 considerably in advance of the tip of the outer lobe, although the more primitive rank of 

 Squilla chlorida is shown by the fact that the separation between the two lobes is much 

 less pronounced than it is in the more typical and highly modified Squilla. I have 

 shown that the species of Lysiosquilla and Coronis also form a single phylum, which it is 

 not practicable to divide into two genera, and the character of the modified appendage 

 is as similar in Lysiosquilla macidata (PL X. fig. 6) and Lysiosquilla (Coronis) 

 excavatrix (PI. X. fig. 12) as it is difi"erent from that of Gonodactylus and Squilla. The 

 terminal joint B is subtriangular, and the suture c which separates it from the proximal 

 joint is transverse, while it is oblique in the other forms. The outer lobe a is very 

 much larger than the inner h and triangular, while the inner lobe is very small and 

 nearly circular, and separated by a movable suture from the outer. The movable limb 

 of the forceps y is greatly elongated and is not bent at the tip, while the fixed limb e is 

 very small and rounded at the tip. 



While no classification can ever be accepted as final, the one which I have expressed 

 in the preceding diagram is at least an approximation to the truth, and gives, so far as 

 our present knowledge enables us to trace it, the genealogy of the species of Stomatopoda, 

 so far as this can be established from the study of the adults. 



In most animals we look ro the larvse or embryos for evidence of genetic relationship, 

 more satisfactory and intelligible than any which is furnished by the adults, for it is a 



