39 



STYLASTERIDAE (H YDROCORALS) FROM SOUTHERN SEAS 

 are present also in older parts of the colonies, although they are more prominent on the thinner, apical 

 or peripheral, branchlets than on more central branches of the meshwork (PI. II, fig. 2; PI. Ill, fig. 1). 

 These features seem to indicate that the spines are gradually absorbed by the growth of the surrounding 

 coenosteum, and that they do not grow materially after the incorporation of the branchlet in the mesh- 

 work of the colony. 



At first glance we might presume dimorphism to exist among the dactylopores, small dactylopores 

 being found as circular holes in the coenosteum in great numbers between the spines (Figs. 3, 4). On 

 the other hand, it is obvious that the spines do not quite correspond either to the nariform or the water- 

 spout-like spines of Errina. In their disposition, especially on the branchlets, they are reminiscent of 

 the nariform spines of Eu-Errina, but it is also evident that they are connected with, or probably 



■ ,;'i. '■ 





""- v --— .--^x £M 



-'. f> 



-'*■■ 



Fi g s - 3"7- Errinopsis reticulum from St. WS 246. (3) Part of the surface of a branch; near the brim of the central gastropore 

 three ' spines ' are seated, the lumpy one with two grooves or conduits from base to summit. (4) Lateral part of a branchlet 

 with full-grown and developing spines. (5) Gastropore from the fracture of a branch showing the entire gastrostyle in 

 side view. (6) Gastropolyp in situ drawn on basis of two slides. (7) Gastropolyp prepared out. (1), (2) and (3) after 

 skeleton having been cleaned with sodium hypochlorite; (4) and (5) from decalcified fragments. (1) x 40; (2>-(s) x °°- 



developed from, spineless pores by growth of the surrounding coenosteum (Fig. 4). The growth is 

 not symmetrical, but at first results in a spine with an obliquely placed, oval pore-opening. In many 

 cases the apical part of the spine grows far above the pore, so that this becomes an oval opening or 

 slit on the side of a conical spine. In other cases the pore continues as a groove or narrow conduit 

 to the very top of the spine. Again, two or three pores may debouch on one spine, and here one may 

 be round, the other more groove-like. In Fig. 3 the lumpy spine is furnished with two grooves, both 

 of which run conduit-like to the very summit of the ' spine '. 



A careful examination of the coenosteum alone does not furnish a safe basis for determining whether 

 the dactylopores exhibit real dimorphism. This question can only be settled by examination of the 

 soft parts of the colony. 



The dactylozooids are throughout simply finger-shaped without any specially developed basal 

 adhesive part; they agree entirely with the smaller dactylozooids of E. antarctica (Fig. 1). The uni- 



