26 C. McLEAN FRASER 



12 tentacles, but these tend to degenerate. Though I found no- 

 specimen entirely without them, the number in some cases was. 

 reduced to 3 or 4. A mouth appears to be present in all cases. 

 The female sporosacs become of large size and remain quite glob- 

 ular. Each contains a large number of ova. The male sporosacs 

 are not nearly so large as the female and are oval in shape. In 

 some cases at least nematocysts are present, but instead of being 

 definitely grouped, they are somewhat irregularly scattered over 

 the surface both distal and proximal to the tentacles. I did not 

 find any of these on the young forms with the numerous tentacles,, 

 and it may be that they are not developed until the tentacles are 

 to some extent degenerated. 



Distribution. — Departure Bay, San Juan Archipelago. 



The generative and nutritive zooids are not scattered promis- 

 cuously. Each kind has its own definite locality with no inter- 

 mixing except at or near the limit where the areas meet. All 

 specimens were found on gastropod shells, inhabited by hermit 

 crabs. The nutritive zooids are restricted to an area extending 

 some distance from the inner border of the lip, while the gener- 

 ative zooids occupy the remainder of the surface of the shell, and 

 hence are many times as numerous as the nutritive zooids. The 

 conical spines arising from the common basal expansion have a 

 very extensive development. They may appear with the regular 

 conical shape, common to Hydractinia polyclina, they may retain 

 their diameter throughout to form columns or they may appear 

 as a ridge continuous for some distance. These ridges may even 

 join to form a network over a large portion of the surface of the 

 shell. The ridges instead of being simply jagged, are in the ma- 

 jority of cases provided with small sharp-pointed spines. When 

 a network of ridges is formed, it provides a very efficient protec- 

 tion for the developing zooids that are packed so closely together 

 that one can scarcely see through to the ccenosarc at any point. 

 This is especially noticeable when the sporosacs are developing 

 on the generative zooids. Though I have examined a large 

 amount of material, I have seen no indication of dactylozooids. 

 It may be possible that as Bergh says in regard to E. carica^ that 

 none are present in any case. This might account for the strong 

 development of the protective spines. 



3 Goplepolyper fra Kara Havet, 1887, p. 331. 



