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



become divergently modified, or that two adults might diverge from each other while 

 the larvae remain alike, yet we should expect a natural or phylogenetic classification of 

 the larvae to stand in some definite and recognisable relation to the natural classification 

 of the adults. 



My attempt to discover a relation of this sort at once brought me face to face with a 

 serious difficulty. In most of the published descriptions little attention is given to any 

 points which are not regarded as diagnostic, and the resemblances, which are of even 

 greater scientific interest than the difi"erences, are often completely neglected ; and 

 careful study of the published figures soon showed that they are untrustworthy so far as 

 relates to points which did not seem significant to the writers. Brevity and exactness of 

 diagnosis is of course desirable and essential to the ready identification of species, but 

 the description and identification of species is only a means for a more important end, 

 the ultimate discovery of the laws of life, and it is therefore desirable that every specific 

 description should consist of two parts, a brief diagnosis for purposes of identification, 

 and a complete description, or brief monograph, giving all the characteristics ; the points 

 of resemblance to allied forms, as well as the distinctive peculiarities. 



The absence of this information renders the establishment of phylogenetic relation- 

 ships very difficult, and I soon found that the characteristics which are most significant 

 and of most scientific importance are by no means the ones which have been selected for 

 diagnosis. The analytical key which Miers^ gives is probably the best which could be 

 devised for ease of identification, and it expresses the general relationship between the 

 genera with sufficient accuracy for the purposes of the systematist ; but whUe most of the 

 genera which are usually recognised are natural ones, the points which are of the greatest 

 value in tracing the relation between the larvae and the adults are entirely ignored in 

 most of the published diagnoses. 



While there can be no doubt that the many diflferences between the Stomatopoda and , 

 the other Malacostraca are of ordinal importance, all the species are included in a single 

 family, the Squillidae, and the difi"erences between the genera are slight. Excluding the 

 genus Leptosquilla, Miers, which is very slightly known, and not represented in the 

 Challenger collection, six genera are usually recognised, Squilla, Chloridella, Lysiosquilla, 

 Cor-onis, Fseudosquilla, and Gonodactylus. 



The study of the Challenger specimens shows the necessity for redistributing the 

 species which have been associated under the generic name Gonodactylus, and the 

 establishment in its place of three genera, Gonodactylus [sensu stricto), Protosquilla 

 n. gen., and Coronida n. gen., and also that it is impossible to draw any natural line 

 between Coronis and Lysiosquilla, or between Chloridella and Squilla, and I therefore 

 recognise seven genera, Protosquilla, Gonodactylus, Fseudosquilla, Coronida, Lysio- 

 squilla (including Coronis), and Squilla (including CJiloridella). My comparison of the 



1 On the Squillidae, Ann. and Mag. Nat. Hist., ser. 5, vol. v. p. 2, 1880. 



