DEVELOPMENT OF SOME STARFISHES. 561 



postei'ior ccelom, stating that this occurred for certain in inter- 

 radii II. /III., III./IV., and lY./Y. In my former account 

 (1, p. 34) I noted the occurrence of several interradial outgrowths 

 in Solaster endeca, but was unable to say whether such out- 

 growths were present in all the interradii. It may be added here 

 tliat Crossaster papposus falls into line with Solaster endeca as 

 regards the origin of the pharyngeal coelom, except that in Cros- 

 saster the latest formed pouches are IX./X. and X./XI. (See 

 p. 563.) 



2. Origin of perihcemal pouch IX. /I. — Formerly I thought 

 that this pouch took origin from the anterior coelom {1, p. 31), 

 a moiety being possibly contributed by the posterior coelom. 

 Examination of fuller material brings out the fact that the whole 

 of the pouch in question arises from the ventral horn of the 

 posterior coelom. (See p. 556.) 



3 . Sequence in formation of hydrocoele pouches, etc. — In Solaster 

 endeca the usual number of rays is nine, and of these five are 

 primary in the sense that they correspond to the five rays of 

 an ordinary starfish such as Asterina gibbosa or Astei'ias ruhens. 

 To begin with, the five primary hydrocoele rays or pouches are 

 arranged in a crescent with dorsal and ventral ends, pouch I.* 

 being near the dorsal and pouch V. near the ventral end. The 

 formation of the additional rays begins at the ventral end of the 

 crescent, pouches VI., VII., and VIII. being successively added 

 in this region during the progress of metamorphosis. In speci- 

 mens with only eight rays- the hydrocoele ring now closes in 

 interradius VIII./I., and it is in this interradius also that the 

 last remains of the sucker and stalk are to be found. Formerly 

 I believed that in specimens with nine rays (the usual number) 

 pouch IX. was superadded to pouch VIII. at the extreme ventral 

 end of the hydrocoele crescent. However, examination of my 

 new material and re-examination of my former material makes it 

 certain that normally pouch IX. is superadded to pouch I. by 

 outgrowth from the dorsal end of the hydrocoele crescent, and that 

 accordingly the closure of the hydrocoele ring really takes place 

 in interradius VIII./IX., the last remains of sucker and stalk 

 being found either in this interradius or opposite pouch IX. I 

 take this opportunity of correcting my former mistake, which was 

 made somewhat easy by the circumstance that a large proportion 

 of my metamorphosing specimens developed only eight pouches, 

 pouch IX. being absent, and it was from serial sections of 

 these specimens that my description of the details of closure of 

 the hydrocoele ring was chiefly made. However, the correction 

 in no way invalidates my numbering of the rays in Solaster, 

 rays I., II., and V. being fixed by the two great landmarks of 

 starfish asymmetry, the position of the madreporite and of the 

 anus. As r-egards Crossaster pwpposus, we may note here that 

 two or even three out of its six to eight extra pouches are formed 

 by extension fi'om the dorsal q\\<\. of the hydrocoele crescent. 



* Sec footnote on p. 555'. 



