3 6 DISCOVERY REPORTS 



The colony from the Burdwood Bank is on the whole very coarsely constructed like the colonies 

 from St. 2200 (PI. IV, figs. 3, 4 and 6). Colonies which are comparatively heavily built are almost 

 always rather regularly flabellate, sometimes also a little incurved on the anterior side. On the other 

 hand, colonies of more delicate construction and with marked ramification in many cases exhibit a more 

 irregular growth and tend to have a more bushy appearance, although a main plane can still be distin- 

 guished (also from St. 2200, PL IV, fig. 5). An excellent figure of this type has been given by von 

 Marenzeller (1903, pi. i, fig. 1) in his description of E. gracilis. 



The branching is dichotomous or subdichotomous. However rich the branching may be, anastomoses 

 are very seldom found. 



The colour of the colony (in spirit) varies from a vivid brilliant carmine or brick red to almost white 

 or yellowish, or occasionally light brown. Dead colonies may be greyish white, but very soon they 

 attain a dark dull brown colour, probably owing to the attack of foreign micro-organisms. (Dead 

 colonies are of course also exposed to attacks of boring animals. In living colonies, sections have 

 revealed the occurrence of small, parasitic crustaceans lodged in 'ampullae', evidently caused by the 

 parasites, but not visible to external examination. The identity of the crustaceans has not been 

 established.) 



The surface of the coenosteum after cleaning with sodium hypochlorite generally exhibits the 

 peculiarly reticulated structure reminiscent of a lump of sugar ; but in some colonies the structure is 

 more vermiculated and feebly glossy. The differences are so small and gradual that they cannot be 

 regarded as of taxonomic importance. 



In a previous paper (1942) I distinguished provisionally a form kergiielensis from the typical form 

 from Cape Horn. The scanty material suggested small differences in the surface structure of the 

 coenosteum which, however, fell well within the range of variation. Another small difference was 

 observed in the gastrostyles, which in the Cape Horn specimen were more slender and spine-like. 

 This, however, must be looked on as an anomaly. Several gastrostyles from different colonies have 

 been examined, and in the main they coincide with the fir-cone-like gastrostyles of the Kerguelen 

 colony (1942, fig. 12), which are characteristic of E. antarctica. On the other hand, E. spongiosa 

 Broch, 1942 may exceptionally show some external likeness to the coarsest colonies of antarctica. 



Among the features separating these two suggested forms of E. antarctica the ampullae were also 

 mentioned. In the Cape Horn specimen they protrude 'almost hemispherically on the surface', 

 whereas in the Kerguelen fragment they ' are not visible on the surface of the branchlet ', although 

 belonging to the superficial layers of the coenosteum. Both fragments were from dried specimens, and 

 it was not possible to ascertain the sex; however, I hinted that possibly the Cape Horn specimen 

 might be a female. 



In the present abundant material, several colonies occur in which ampullae cannot be seen externally, 

 whereas in others they are very obvious, protruding like sections of a ball, and, moreover, in vividly 

 red-coloured specimens they are emphasized by a yellow or whitish colour. Fractures reveal, however, 

 that ampullae are present in great numbers in the peripheral layers of the coenosteum of the first- 

 mentioned colonies. An investigation of the contents reveals that these colonies are males, whereas 

 those with protruding ampullae are females. 



All earlier descriptions have been based on dried specimens, and I took the opportunity to examine 

 the soft parts in the rich material to hand. 



The gastropolyps have from three to six (mostly four) small, thread-shaped tentacles, which exhibit 

 pronounced extension and contraction. 



In antarctica as well as in some other Errina species the dactylopores are of two types, viz. the niche- 

 or slit-like dactylopores (grooves) of the grooved spines, and the small circular pores of the coenosteum 



