506 Cresswell Shearer 



it is very irregularly formed, and in the latcr stages it is always 

 in free commiinication witli the haemocoelic space of the collar 

 region, and for these reasons I am inclined to consider it with 

 ÖELYS LoNGCHAMPS RS a true haemocoel. It certainly on the other 

 hand bears a striking resemblance to the preorai cavity of the 

 Enteropneusta and Ccphcdodiscus in its general configiiration, and in 

 the manncr in which the muscle cells develop in relation with it. 



For the sanie reason it may be argued that the preseptal 

 coelom makes its appearanee late in development, because it 

 represents the small preseptal coelom of the adult, iucluding the 

 lophophore and the tentacles, which are only reciuired late in devel- 

 opment. In this connection Ikeda has observed that the part of 

 the larvai tentacle thrown off during metaraorphosis is that in 

 which the preseptal coelom is wauting. 



Caldwell was of opinion (3) that the truuk body-cavity 

 took its origin in the masses of cells derived from the nephridial 

 diverticula — "in a paired mass of cells Avhich grows out from the 

 first formed sacs (posterior diverticula) and remains separated from 

 the latter by a septum" (pag. 376). This view is now impossible 

 for reasons which I bave already mentioned. For it will be seen 

 that at the time the body-cavity arises the nephridial diverticula are 

 some distance from this structure in a ventral direction. I bave 

 never observed anything in sections that would tend to support 

 this view. 



Hatschek in bis Text hook of Zoology gives a diagrammatic 

 figure of a young Actinotrocha larva, in which the trunk cavity is 

 represented as consisting in a pair of coelomic sacs one on either 

 side of the gut. This figure lias given rise to cousiderable comment 

 in the recent literature of Phorouis; unfortunately it was published 

 without any description. In certain views of the larva, especially 

 ventral views, tlie two lateral portions of the coelom certainly give 

 a vcry misleading Impression that the coelom is paired as shown 

 in this figure. For instauce, fig. 58 which is taken from a recon- 

 struction model of Ä. hranchiata. Here tlie two lateral portions of 

 the coelom which are about to meet on the ventral line certainly 

 look like separate sacs ;fig. 57). I believe Hatschek was misled 

 in considering the coelom paired from the examination of the larva 

 from the ventral surface. Masthrman (18) shows the trunk coelom 

 arising from two lateral masses of cells on the dorsal side of the 

 archeuteron. These masses are at first solid and form part of the 



