98 THE VOYAGE OF H.M.S. CHALLENGER. 



pores divided into two chambers, an upper and a lower, by a constriction of their 

 walls. Opening between the chambers rendered horse-shoe shaped by the projection 

 across it, in the direction of the tips of the branches, from that side of its margin 

 placed nearest the base of the branches, of a tongue-like excrescence. Ampullse in 

 the male stocks in a ring around the cyclo-system masses ; none scattered on the 

 branches. Dactylozooids, when at rest, doubled down within the upper chambers of the 

 gastropores. Gastrozooids flask-shaped, devoid of tentacles, with numerous basal canals. 

 Gonangia in the male stocks containing a central mass of cells from the surface of which 

 are developed as buds numerous pedicellate lobular sacs, in which the spermatozoa are 

 produced. Female stocks unknown. 



12. Cnjptohelia, Milne-Edwards and Haime. 



Ccenosteum closely resembling that of Astylus in all respects, excepting that the cyclo- 

 system masses are not so prominent, that the opening between the upper and lower 

 chambers of the gastropores is circular in outline, and that a lid-like lamina of calcareous 

 matter is directed horizontally across the mouths of all the cyclo-systems. The lids are 

 supported on stout columns arising from the margins of the cyclo-systems and inclined 

 over them. They spring from the sides of the systems nearest the bases of the branches, 

 and are directed towards the tips of the branches. In female stocks only a single 

 ampulla and gonangium developed in relation with each cyclo-system. No ampullae on 

 the connecting branches. In the males several ampullae in the walls of each cyclo- 

 system. Soft structures as in Astylus. In female stocks numerous gonophores present 

 in each gonangium in all stages of development. Spadix cup-shaped, developing, as 

 in Errina, into a fringed network at the margin. A solitary ovum developed in relation 

 with each spadix. Planula very long and worm-like. 



Pedigree of the Hydeocoeallin^e. 



The line of descent of the various genera of the Stylasteridaa from a parent form seems 

 to be traceable with especial clearness. All gradations are present by which simple 

 circular mouthed pores sporadically scattered over the ccenosteum become grouped and 

 modified into cyclo-systems of the most symmetrical and complex character. Since styles 

 appear in some genera in the dactylopores as well as in the gastropores, it seems probable 

 that in the ancestral form or "Archistylaster " styles were present in both forms of pore. 

 If the Milleporidaa prove closely related to the Stylasteridae when then gonophores 

 have been investigated, it will follow that the two families have had a common ancestor, 

 and that Hydroids have developed a calcareous support only once in their history 

 and not in two separate instances. This common ancestor may be presumed to have 

 had a hydrosoma composed as throughout the sub-order Hydrocorallinas ; with its 



