264 DISCOVERY REPORTS 



Specimen A which is only 51-5 mm. long, also has been figured by Kolliker (1880, PI. 10, fig. 396). 

 His description runs: 'Four polyps, all lateral with a free end of the rachis, are visible in fig. 39, C 

 The letter C must be due to a slip of the pen. The specimen is figured in dorsal (in Kolliker's termino- 

 logy 'ventral') view. It seems to have escaped Kolliker's attention that a small but undeniable 

 autozooid bud is placed ventrally between the two figured median autozooids (PI. I, fig. 3); the 

 cluster thus has 5, not 4 autozooids. This, however, does not alter Kolliker's implication in the text 

 that no primary polyp can be traced. On the other hand, the obvious, free end of the rachis strongly 

 recalls the young stages from the Swedish Deep-Sea Expedition, although in these a primary polyp is 



clearly to be seen (Broch 1957). 



The Challenger collection in the British Museum (Nat. Hist.) also contains two more young stages 

 of the species lindahli, viz. the two smaller specimens assigned by Kolliker to U. huxleyi (see p. 255). 

 Their quadrangular axes diflFer from this species, and also their other features show that they belong 

 to the same species which Kolliker thoroughly described under the name of U. carpenteri. 



Kolliker (1880, p. 24) instituted another, new species, U. magniflora. Here the rachis is bowl- 

 shaped, and as far as can be discerned in the type specimen, which is in a rather bad state, the axis 

 enters from below centrally. It is impossible to see where it terminates, and it is also not possible to 

 tell which of the autozooids might be the primary polyp. The arrangement of the autozooids (text- 

 fig- 3. 3) is very like that in a specimen from 'Discovery' St. 371 with a total length of 46 cm. 

 (Table, p. 266 no. 5). Kolliker's type has an outer whorl of 9 autozooids along the brim of the ' bowl ' 

 and in the centre, within it, one central autozooid of the same size as the marginal autozooids; 

 finally, there is also one polyp bud below it as indicated in text-fig. 3,3. Because of the brittleness of 

 the specimen I did not venture upon a more thorough investigation. This specimen is without doubt 

 only an individual variant, or growth form, of ' U. carpenteri ', a variant rather common among the 

 specimens from subantarctic-antarctic waters, which are a little larger than those examined and 

 mentioned by Hickson (1907) and Jungersen (1907) under the name of U. carpenteri. The size of 

 Kolliker's specimen (740 mm. total length) corresponds with the smaller specimens of the ' magni- 

 flora ' stage from the Discovery collections. 



A third antarctic Umbellula was announced in a preliminary paper by Kiikenthal (1902) as U. en- 

 crinus var. antarctica. In his paper on U. carpenteri from the Belgian Antarctic Expedition, Jungersen 

 (1907) hints at the possibility of all three antarctic Umbellula species being synonyms: 'es kommt mir 

 deshalb als sehr moglich vor, dass diese Art nicht nur noch grosser wachsen, sondern hochst wahr- 

 scheinlich die Zahl der Polypen bedeutend vermehren und die ganze Form des Schopfes andern mag, 

 mit anderen Worten sich den Typus der U. encrinus (oder deren antarktischen Vertreters der 

 U. magniflora= U. encrinus var. antarctica Kkth.) mehr nahern kann.' Kiikenthal and Broch (191 1), 

 and Kiikenthal (191 5) raised the variety to a separate species U. antarctica, and Kiikenthal dis- 

 tinguished between three antarctic species, viz. U. carpenteri, U. magniflora and U. antarctica. 



The following table gives a list of the specimens from the Discovery collections: 



Station Depth (m.) Remarks oti the specimens 



180 160 One specimen : total height 67 cm. 



1 1 March 1927 3-4 irregular whorls of autozooids. The axis enters the centre of the bowl-shaped basal 

 Palmer Archipelago part of the rachis. Siphonozooids in all intervals between the autozooids, some of them 



with half extended tentacles. In spite of a very thin stem it is a typical antarctica specimen. 



181 160-335 ^ specimens. 



12 March 1927 i and 2, total height 7-5 cm. 



Palmer Archipelago Both have only one pair of fully developed secondary autozooids, but whereas i only 



shows the first traces of a lower pair, in 2 this is seen as small buds (text-fig. 3, 4). The 



