284 Proceedings of the Royal Physical Society. 



externally by the coenosarc [polypary], thus reminding us 

 of the sclerobasic corallum of some of the Actinozoa." This 

 doctrine had been previously promulgated by Quatrefages 

 (Ann. des Sc. Nat., xx. 232), who considered the polypidom 

 to be an endoskeleton deposited in the substance of the 

 polypary, like the solid axis of Gorgonia. If this view were 

 correct, it would not only remove Hydractinia from the 

 Tubulariadse, but would segregate it from the whole of the 

 Hydroid Zoophytes, not one of which is destitute of an 

 investing polypidom. 



In the " Edinb. Phil. Journal " for April 1857, I stated, in 

 a paper on Hydractinia, my conviction of the incorrectness 

 of Quatrefages's opinion, and that the mode of secretion of 

 the polypidom of Hydractinia did not differ from that of 

 the rest of the Tubulariadee, as was seen in the development 

 of its young and its propagation by stolons. Since then I 

 have come to the following conclusions, after the examina- 

 tion of a very large number of specimens, some hatched 

 from the egg and adherent to glass, others removed as cut- 

 tings from adult specimens and transplanted on glass, to 

 which they readily grow, and others removed entire from the 

 shell of the Pagurus by acid, and put up in spirit or balsam. 



The polypidom and polypary are found in the following 

 forms, all of which are frequently combined in the same 

 specimen : — 



1. An open network of delicate chitinous tubes without 

 spines, enclosing a polypary composed of several combined 

 endodermal tubes surrounded by a single layer of ectoderm. 

 Found in very young specimens, or in old ones growing on 

 protected parts of the shell. (Analogous to Clava repens 

 (mihi), the G. discreta of Allman.) 



2. An open network as in the last ; the tubes of thick 

 brown chitine, with single hollow spines rising from a single 

 tube, or from the confluence of four tubes. 



3. A close reticulate plate, as in Clava cornea (mihi) and 

 G. membranacea (mihi), formed from states 1 or 2 by the 

 continual nlling-up of the meshes by anastomosing branches, 

 with or without spines. 



4. A fleshy plate of ectoderm permeated by a network of 



