CEPHALOCHOBDA. 357 



position, and has not yet taken up its permanent location on the ventral side of the pharynx, 

 which shews that ■ a certain amount of torsion still has to take place. It now extends over 

 and beyond the club-shaped gland for some distance. The two arms of the latter organ have 

 become histologically differentiated, the upper arm having a large lumen with tall glandular 

 cells, the lower arm being drawn out into a tube with a small lumen and much smaller cells. 



The preoral pit now leads right do^vn to the front end of the larval mouth as a narrow 

 gutter, open to the exterior throughout its whole length. The buccal hood has not yet begun 

 to grow out. 



In this larva the gill slits numbered 31. As in the younger form all the openings lie 

 in a series one behind another in the mid-ventral line. Plate XVIII. fig. 5 shews the ventral 

 view of this larva. The appearance of the slits in the mid-ventral line is curious, for the 

 reason that the bridges between the openings are very long and narrow. Owing to this they 

 have become much twisted, and their shape somewhat obscured. Their form can, nevertheless, 

 be traced out, and it is seen that they have a somewhat rectangular shape. There is no sign 

 of any crescent-shaped infolding to form a tongue bar. There is absolutely no trace of any 

 secondary gill slits breaking through in the manner usually described — that is, on the right 

 side and above the primary slits. Normally the secondary slits should shew, at their first 

 origin, as six oval thickenings situated on a continuous ridge of tissue, in the right wall of 

 the pharynx above the primary slits, and this should take place when the larva has fourteen 

 primary slits fully formed. As has already been seen the first larva possessed the requisite 

 number of primary slits without signs of the appearance of any secondary ones. Further, 

 in the normal development, the formation of the secondary slits, after they have once started, 

 is very quick, and the number of slits on each side becomes even. A certain number of 

 primary slits close up and a stage with eight on each side is reached. 



Indeed, the phenomena found in these larvae cannot be brought into any uniformity with 

 the plan of A. lanceolatus. In none of the specimens obtained is there any sign of secondary 

 gills, although possessing from 15 to 31 'primary' slits. It seems impossible that after such 

 a late stage has been reached, the secondary slits should arise by perforation in the usual 

 place, especially now that the metapleural folds have been well formed. What precisely 

 does occur it is impossible to say owing to the lack of the later stages which only could throw 

 light on the subject. For the same reason one cannot say whether any of the slits present 

 close up as do some of the primary slits in the larva of A. lanceolatus. 



Such evidence as there is, points to some different method of origin of the gills to 

 that which is found in A. lanceolatus, the only species in which the full development has 

 been thoroughly observed. There is no reason why all species should have precisely the same 

 way of development, but unfortunately the facts here are hardly sufficient to warrant any 

 deductions as to how the gills arise in the present peculiar form. 



The fact that the larvae were taken right away from the grounds where the two Maldivan 

 species of Gephalochorda were found, and also that they were taken in heavy weather close to 

 the channels leading fi:om the sea, makes it possible that they may be the larvae of some 

 pelagic form, whose altered mode of life may account for the differences of larval growth 

 mentioned above. 



46—2 



