240 PERCY SLADEN TRUST EXPEDITION. 



alternative occurrence, in a species wMcli in other cases undergoes metamorptiosis at an 

 earlier stage : " Man kann annehmen, dass junge Larven vor der Metamorphose auf die 

 hohe See verschlagen warden und so an der Metamorphose verhindert, ibre larvale 

 Organisation weiterlnldeten." But this is a suppositioa which the remarkable constancy 

 of the number of gill-slits, as of other characters, in Amphioxides, would alone make it 

 difficult to accept ; and now that animals in no way differing from their pelagic 

 relatives have been taken in the immediate neighbourhood of land, I do not think that 

 it can be any longer maintained. There can be little doubt that the different kinds of 

 Amphioxides represent fixed and normal stages in the life-history of the species to which 

 they belong. Possibly the great length of the larval period is an adaptation to a more 

 prolonged pelagic existence, originally developed in the species in response to the 

 same need wliich Goldschmidt supjioses to have evoked it in individuals ; but it is also 

 cimceivable that it is a primitive condition. No doubt the period of pelagic life became 

 progressively shortened, as the Cephalochordate ancestor became adapted to life in the 

 ^aud: the species of which Amphioxides is the larva may well represent an earlier 

 stage in the process than By^aiiciiiostoma, without any implication of the primitiveness of 

 the larval organisation. 



We have now to consider the question to wiiat, if any, known adult form we can 

 regard Amphioxides as belonging. Certain featui-es in the structure of the larvse give 

 slight clues as to their affinities. Thus the occurrence of gonads on the right side only, 

 in A. pelagicus and A. valdivice (two of Goldschmidt's specimens), precludes them from 

 belonging to the genus Bratichiostoma. They must therefore belong to Asymmetron, 

 Heteropleuron *, or some unknown genus. In both the genei*a named gonads occur on 

 the right side only, and the right metapleure is continuous with the ventral fin ; 

 Asi/mmet?-on is further marked by the presence of a urostyloid process and the absence 

 of ventral fin-ray chambers ; but in one species of Heteropleuron [II. agassizii), described 

 by Parker (1904) from a single specimen, ventral fiii-ray chambers also appear to be 

 absent. Their non-occurrence in any species of Amphioxides would suggest the alliance 

 \\ ith Asymynetron or with H. mjassizii rather than with other species of Heteropleuron. 

 It is true that the ventral fin-ray chambers cannot be seen in the Branchiostoma larva t 

 in its oldest pre-metamorphic stage, but neither can the dorsal ones : since the latter are 

 extremely well developed in Amphioxides, it is hardly likely that the absence of the 

 former is merely a larval character. Another feature allying Amphioxides pelagicus, 

 but not vahUvice, to Asymmetron, is the absence of distinct dorsal fin-ray chambers over 

 tlie most anterior portion of the nerve-cord. This is a feature not elsewhere found 

 outside the genus Asymmetron, so far as I am able to discover. 



Unfortunately, the essentially larval character of Amphioxides gives us few points of 

 general comparison with known adult Cephalochordates. We must expect it to differ 

 as much from its adult as the larva does from the adult of JBranchiostoma lanceolatum; 



* I follow Parker (1904) in provisionally regarding Branchiostoina, Heteyopleuron, and Asijminetron as three 

 distinct genera. 



t I have determined this point on mounted larvte kindlv lent to me by JI. I.egros. 



