2^2 DISCOVERY REPORTS 



autozooids which are evidently much more extended than those of Kiikenthal and Broch (again re- 

 produced by Kiikenthal 1915), which were drawn on board the 'Valdivia' from living specimens. 

 The autozooid cluster strongly recalls that of U. hiixleyi. The autozooids are small and crowded on to 

 the short distal part of the rachis, and as Hickson (1937) pointed out, their arrangement varies so 

 much from specimen to specimen, that in most cases it is impossible to trace any regularity. In larger 

 clusters the numbers of autozooids commonly amount to some 25 to 30. In several specimens a 

 naked narrow dorsal field along the axis gives an impression of certain bilaterality in the pompon. 



The specimens from the German ' Valdivia ' Expedition had a somewhat thicker basal holdfast, the 

 length of which according to Kiikenthal (191 5) measured one-sixth to one-quarter of the entire stem. 

 In most of the specimens from the John Murray Expedition, this part is so contracted transversely 

 that it is impossible to trace an upper limit of the holdfast. 



Young specimens of U. pellucida are probably difficult to distinguish from those of U. Undahli in 

 some cases, but developmental stages of U. pellucida have hitherto escaped the collectors ; probably 

 their dimensions will be very tiny. 



In his review Kiikenthal (191 5) stated that there were no siphonozooids between the autozooids of 

 the cluster, but that they only occurred on the swollen part of the rachis below the cluster and in 

 small numbers on the thin stem. External examination of some of the specimens from the John 

 Murray Expedition, the siphonozooids of which are very obvious on the rachis below the cluster, 

 showed that no siphonozooids could be detected between the autozooids. This feature seems to be 

 constant and therefore pellucida differs from the following species, Undahli, where comparatively 

 large siphonozooids are present between the autozooids of the cluster. 



Thomson and Henderson (1906) described an U. purpurea with round axis from the Indian Ocean. 

 A 'cotype' of it had been deposited in the British Museum (Nat. Hist.). This specimen has evidently 

 been regarded as a younger stage of purpurea, described from a type specimen with no less than 

 53 autozooids arranged in 4 whorls + 7 'at the tip that had no definite arrangement'. The cluster 

 of the ' cotype ' has only 1 1 autozooids, which measure some 5-6 mm. in length (without the tentacles). 

 Eight of them, including the primary, largest polyp, are placed in one whorl, the last three in the 

 central upper field within this whorl. All the autozooids are directed towards the ventral side of the 

 specimen and give the entire cluster a decidedly bilateral appearance. The polyps are semi-pellucid 

 and in all characters including the quadrangular axis recall U. pellucida ; the axis obviously differs 

 from the description of Thomson and Henderson. This cotype is evidently not purpurea but a rather 

 young stage of U. pellucida. A renewed investigation of the type specimen of purpurea is needed to 

 settle its identity. 



Umbellula Undahli Kolliker 1874 (PI. I, fig. 3, PL II, figs. 5, 6) 



U. triiniacea+ U. pallida Lindahl 1875, pp. 12, 13, Pis. 1-3. 



U. Undahli Kolliker 18746, p. 11. 



U. carpenteri+ U. magniflora Kolliker 1880, pp. 23, 24, Pi. 10, figs. 38-40; Pi. 11, figs. 41, 42. 



U. Bairdii Verrill 1884, p. 219. 



U. Bairdii Verrill 1885, p. 509, PI. i, figs, i, 2. 



U. arnhigua Fischer 1889, p. 37. 



U. encrinus var. antarctica+ U. Kdllikeri+ U. rigida Kiikenthal 1902, pp. 595, 596. 



U. Undahli Jungersen 1904, p. 75, PI. 3, figs. 37-46. 



U. ambigua Marion 1906, p. 146, PL 15, fig. 22. 



U. carpenteri Hickson 1907, p. 12, PL i, figs. 1-7. 



U. carpenteri Jungersen 1907, p. 5, PL i, figs. i-io. 



U. encrinus var. ambigua Stephens 1907, p. 21. 



