99 



the original points of difference between the two accounts of Asterid 

 development given by Goto and myself. 

 I asserted that in Asterina gibbosa : 



1) The rudiments of the aboral surface of the adult make their 

 appearance partly on the right side and partly on the posterior aspect of 

 the larva, and consequently the plane of the disc of the adult makes 

 an angle of from 60° to 70° with the frontal plane of the larva, and no 

 correspondence exists between the planes of symmetry of the larva 

 and adult. 



2) The coelom of the larva consists of an anterior portion filling 

 the praeoral lobe and sending back two tongues lying one at each 

 side of the gut which eventually meet one another below so as to form 

 a ventral mesentery. 



A transverse mesentery or septum makes its appearance first on 

 the left side of the gut, and then on k the right; and in this way the 

 hinder portions of the two coelomic tongues referred to above, become 

 separated off and form closed sacs denominated by me left and right 

 posterior co el oms respectively. The unpaired portion of the 

 coelom together with the anterior portions of the two 

 tongues constitute the. anterior coelom of the larva. 



Subsequently two openings — a dorsal and a ventral 5 — are formed 

 in the left transverse septum, but these are temporary and become 

 closed before the end of the metamorphosis. 



3) The sac situated, bof;h in the larva and adult, close to the madre- 

 poric pore, is a rudimentary water-vascular system or hydrocoele belon- 

 ging to the right side; so that at first the water-vascular system is re- 

 presented by a pair of rudiments, the left of which gives rise to the 

 water-vascular system of the adult, whilst the other remains a small 

 sac. This sac arises in the larva as a very thick-walled outgrowth 

 from the hinder wall of the anterior coelom to the right of the middle 

 line: its cavity is at first a mere slit which becomes at once shut off 

 from the anterior coelom. The cavity subsequently enlarges so that the 

 right hydrocoele becomes a thin walled sac but never again is in com- 

 munication with the anterior coelom or with the axial sinus which is 

 derived from the anterior coelom. 



4) The radial perihaemal canals and the outer perihaemal ring- 

 canal which connects them with one another, originate as 5 interradial 

 evaginations of the coelom, four arising from the left posterior coelom, 

 and one from the anterior coelom. 



5 These openings are shown, the ventral in Fig. 1, the dorsal in Fig. 3, in the 

 present memoir. 



