GIBSON— CEPHALOCHOEDA : " AMPHIOXIDES." 223 



This is quite distinct in A. valdivice {vide fig. 12) as a tract of cells taller than their 

 neighbours, with elongated, deeply-staining nuclei. Behind the mouth it passes up onto 

 the left lateral fold of the pharynx, at the summit of which it appears to become con- 

 tinuous with the band of modified cells with long cilia noted by Goldschmidt. A baiul 

 of similar cells can be seen on the dorsal side of the right lateral fold, but cannot be 

 traced into continuity with the upper limb of the endostyle. The absence of the anterior 

 portion of the right peripharyngeal band is hardly surprising, in view of the great 

 crowding of the region in which it would be expected to occur. 



The opening of the club-shaped gland into the pharynx is situated in A. pelagicus 

 at the anterior end of its dorsal limb. In A. valdivice the dorsal limb is reduced to a 

 mere vestige, but here, too, the opening obviously belongs to it and not to the ventral 

 limb— a fact of importance for those theories which interpret the gland as the homologue 

 of an anterior right gill-slit, since the ventral limb lies below the endostyle and so to the 

 left of the morphological middle line. 



In the smallest A. pelagicus obtained (3-9 mm.) the gland does not yet possess an 

 internal opening. 



Like Goldschmidt, I never found any trace of an external opening. 



The Pharynx. 



This region of the alimentary canal is, on Goldschmidt's view, of primary and 

 predominant importance for our morphological interpretation of Amphioxides and of the 

 Amphioxus larva. 



Its most notable characteristics are the division by means of lateral folds into dorsal 

 and ventral portions, and the large number and complicated nature of the gill-slits. 

 When Goldschmidt's monograph appeared, the detailed structure of the gill-slits of the 

 Amphioxus larva had not been described, though Lankester and Willey (1890) had alluded 

 to their peculiar folding in the contracted condition. Since then, however, he has him- 

 self (1906) drawn attention to the fact that the contracted larvse — in fact all preserved 

 specimens that have not been carefully killed with osinic acid — present appearances very 

 similar to Amphioxides. The condition described for the latter must, therefore, be 

 regarded as one of contraction, and there can be little doubt that an expanded condition 

 is also possible, in which the slits of Amphioxides will not appear so very different from 

 those of the expanded Amphioxus larvae figured by Willey (1891). 



A few of my specimens of A. pelagicus are, fortunately, in the expanded condition 

 expected *, in spite of the fact that all were killed with formol : PL 15. fig. 4 shows the 

 head of one of these. These expanded specimens were all taken together (with one 

 normal specimen and one A. valdivice) : all exceed 9 mm. in length, being among the 

 largest obtained. Curiously enough, they agree also in the possession of the tuft of cilia 

 in front of the mouth, alluded to above, and not found elsewhere. The largest of them 



* The older larva figured by Cooper (1903) also shows the slits in the expanded condition. I cannot be certain 

 that this liLiYa, is A. pelagicus. The number of gill-slits — 31, in an animal only 6-5 mm. long — and the point at 

 which the dorsal fin-membrane ends suggest rather A. valdivice, but the position of the anua and of the anterior 

 giU-slits exclude it from that species. Cooper's younger larvse were evidently A. pelagicus. 



30* 



