2^0 DISCOVERY REPORTS 



Kerguelen. Cette curieuse affinite merite d'etre remarquee.' (By a slip of the pen Marion has written 

 grandiflora instead of magniflora as his particulars prove.) There are no differences between Marion's 

 and Kolliker's descriptions, and the Discovery specimens also coincide with them. It might be 

 objected that Marion's specimen has a much shorter stem than the ' Challenger ' specimen (measured 

 in the same way as Stephens's specimens, the stem of Kolliker's specimen is some 690 mm.). Stephens, 

 however, in her table gave measurements of a variation of ' ambigua ' quite parallel to that of U. thom- 

 sonii (Broch 1957 ' U. guntherV), additional evidence against the use of length of the stem (or total 

 length of the specimen) as compared with the numbers of autozooids in the clusters for specific 

 identification. 



Finally, it is impossible to trace any distinct limit or difference between the typical U. carpenteri 

 and the species U. lindahli (comp. KoUiker 1880, Jungersen 1904, 1907, and Broch 1957). 



It is also necessary to discuss the relations between U. lindahli KoUiker (1874) and U. encrinus 

 (L. 1758). Jungersen (1904, 1907) dealt partly with the question but akhough he considered them 

 to be different species, he omitted to define their characteristic differences precisely and only hinted 

 at their varying dimensions. 



Summing up the characters stated by Kiikenthal (191 5) the following differences appear: 

 (i) t/. lindahli has a very long and thin stem, whereas the stem of U. encrinus is not especially thin. 



(2) The autozooids of U. lindahli are gathered in a slender cluster and occur in smaller numbers than 

 in the much more robust clusters of U. encrinus, but the arrangement of the polyps otherwise agrees. 



(3) The tentacles of the autozooids of U. lindahli are twice as long as the polyp body, in U. encrinus 

 of about the same length as the body. (4) In U. lindahli siphonozooids constitute two longitudinal 

 rows along the thin part of the stem, in U. encrinus they are arranged all round it, and the coenosarc 

 is thicker and more fleshy than in U. lindahli. The differences under (2) and (3) can at once be dis- 

 carded as being due to contraction ; (4) is correlated with a more or less copious individual develop- 

 ment of the coenosarc, a fact clearly evident in the specimens from the Discovery collections. 



Jungersen (1904) in his first paper came to the conclusion that U. encrinus must be regarded as a 

 distinct species apart from U. lindahli, although differentiation was difficult. After the examination of 

 new supplementary collections (19 16 6, 191 9) he decided that lindahli and encrinus represent two 

 varieties of one species, the older, Linnean name, encrinus having priority. This view was also accepted 

 by myself (Broch 1929) in my review of the Arctic Octocorals where I mentioned the two groups as 

 U. encrinus forma typica and forma litidahli. 



Although there are no definite limits between the other ' species ' amalgamated here under the name 

 of U. lindahli, a somewhat greater gap apparently separates U. encrinus from the rest. It is of course 

 necessary to make reservations when dealing with characters which may be correlated with the di- 

 mensions of the coenosarc, or which are largely due to contraction. The obvious parallel development 

 in both cold-water areas must be considered: in extreme cases theoretically this should lead up to 

 U. encrinus from U. lindahli. It would then be correct to consider encrinus as an ecologically distinct 

 geographical form of lindahli. However, although at present it is difficult to define the limits exactly, 

 I prefer to treat encrinus as a separate species. Probably the gap between the two ' species ' emphasized 

 by Jungersen (1904) is due to the limits and differences between the Atlantic and the Arctic deep-sea 

 areas. Their size differs so greatly that it is generally easy to distinguish between specimens of the 

 two 'species'. 



