300 DISCOVERY REPORTS 



on hatching than that of Axitis stirhynchus, and it seems never to be an active swimmer. 

 In A. stirhynchus the mouth-parts are normal and functional, whereas they are almost 

 without setae and presumably not functional in Calocaris. 



The large size of the eggs in most of the species in which the eggs have been referred 

 to points to a general prevalence of shortened development, and this is carried to 

 an extreme in Eiconaxius parvus, in which the young leave the egg in almost the adult 

 form (Bate). This appears to occur also in Axius plectorhynchus, according to Hale 

 (1927), but in this species there must be individual differences in this respect, since I 

 find in some eggs which Mr Hale has been good enough to send me a larva not more 

 advanced than that of A. stirhynchus (see pp. 303-4). 



In these circumstances it is perhaps rash to attribute to the Axiidae any larvae which 

 do not show abbreviated metamorphosis and the characters common to the two species 

 known. It is, however, impossible to escape the conclusion that these species are not 

 typical of the family and that some larvae of very different form must be attributed to 

 it. In Calocaris olcocki the eggs are small, according to a figure by Alcock and Anderson, 

 and there may well be other species with a normal larva. 



Larvae are common in plankton from the Atlantic and Indo-Pacific oceans which 

 are evidently Thalassinidea, but certainly do not belong to the Laomediidae nor to the 

 Upogebiidae. As a number of types are represented which are generally found in both 

 oceans the Thalassinidae may probably be excluded. Moreover, as there is but one 

 genus of Thalassinidae with three known species, its inclusion would not widen the 

 field of choice appreciably. There remain only the Axiidae, Axianassidae and Callianas- 

 sidae. The genus Axianassa Schmitt (with one species) is included in the Laomediidae 

 by De Man, but allotted a special family by Schmitt and Balss. As nothing is known 

 about its development it must be left out of account. 



When one is attempting to assign larvae to a position in an adult system it is necessary 

 to attach most importance to those adult characters which are discernible in the larva. 

 For this reason it is the last stages only which can be used, and it is to these stages almost 

 entirely that reference is made in this paper. 



Judging from the species of which the larvae are known we may establish the fol- 

 lowing categories: 



(i) Epipods present; leg 5 with exopod; four pairs of pleopods; telson broadly triangular, with 

 convex margin, median spine and numerous marginal spines, spine 2 reduced to a hair 

 ( = "C. 11" type of telson). Larval life abbreviated v4.wm5 and allies. 



(2) Epipods absent. Leg 5 without exopod; three or four pairs of pleopods; telson of "C. IT" 

 type. Larval life abbreviated C«///fl««^M type II (Gurney, 1937). 



(3) Epipods absent; leg 5 without exopod; three pairs of pleopods ;i telson less than three 

 times as long as wide, lateral spines 1-3 small and near posterior angle, spine 4 long, 

 posterior margin straight or concave, spine formula 8, i, 8 ("C. I" type of telson). Abdomen 

 usually with large spine on somite 2 and somites 3-5 with pronounced ridge. Develop- 

 ment not abbreviated Callianassa type I {e.g. C. subterranea). 



1 There is good reason to believe that Sars' figure (1884) showing four pleopods in C. subterranea is 

 in correct. 



