606 INVEETEBEATA chap. 



as we have seen, and by the premandibular cavities of liigher 

 vertebrates, from the walls of which most of the eye muscles arise. 

 Now in the Tornaria larva we find at the apex of the proboscis an 

 apical plate of nervous and sensory cells on which two eye-cups are 

 situated; and in the embryos of the frog the optic areas can be 

 recognized as pigmented spots in the region of the fore-brain whilst 

 this is still an open plate. 



It appears probable, therefore, that the ectoderm covering the 

 proboscis of the Enteropneusta corresponds to the fore-brain of 

 Vertebrata, whilst the collar region of the Enteropneusta corresponds 

 to the region of the mid-brain of Vertebrata and what lies beneath it, 

 viz. the mandibular cavities, from the walls of which jaw muscles 

 and the superior obhque muscles of the eyes arise. The mandibular 

 cavity would then correspond to the collar-cavity of Enteropneusta 

 and to the region of the first myotome in Amphioxus. How far the 

 ventral extension of the first myotome, which is so clear and distinct 

 from the splanchnocoele in the early larva of Amphioxus, corre- 

 sponds to the system of cavities in the Hps of the adult, called by 

 Van Wijhe (1901) collectively the stomocoele, is a matter requiring 

 further investigation. 



The connective tissues and peritoneal membranes of the larvae of 

 Amphioxus are difficult to preserve and stain, only mixtures con- 

 taining osmic acid seem to confer on them the power of taking up 

 stains so as to be clearly distinguishable. The sUt-Hlie coelomic 

 cavities are apt to collapse, and then it is impossible to be sure of 

 their extension and hmits, as their adherent walls look hke single 

 membranes. The best results are obtained by staining the sections, 

 after prehminary stain in bulk with haematoxylin or eosin, with 

 a 1 per cent aqueous solution of picronigrosin. 



But the working out of the larval development under con- 

 ditions hke these is still a task for the future. If we assume that 

 the hp-cavities of the adult Amphioxus arise from portions of the 

 collar-cavities of the embryo, and that the collar -cavities are homo- 

 logous structures in Amphioxus and Enteropneusta, then we may 

 note that the ciliated oral cirri, which aid Amphioxus in obtaining 

 its food, are outgrowths of the lip-cavities. These cirri may then, 

 perhaps, be compared to the cihated tentacles which grow out from 

 the collar-cavities in Ceplialodiscida, and to the radial canals, and 

 the tube feet, which are branches of these canals, in Echinodermata. 

 These canals and their branches are outgrowths of the left hydrocoele, 

 which we may compare to the left collar-cavity of Amphioxus and 

 Enteropneusta. 



When we endeavour to answer the second of our two questions 

 we must admit that the larvae of Amphioxus are so extraordinary as 

 to present a serious problem for those who, Hke ourselves, believe 

 that the larval stage represents, in however modified form, an ancestral 

 condition of affairs. Balfour considered them so abnormal that, when 

 they were first described by Kowalevsky, he thought they must be 



