258 DISCOVERY REPORTS 



comparatively large interspaces, especially on the lower part of the rachis, whereas all interspaces, 

 as well as the tapering part of the rachis below the autozooids, are packed with comparatively 

 prominent siphonozooids. 



The autozooids are remarkably slender and long, in spite of the basal halves being filled with eggs. 

 The length of the polyp body is up to 35 mm., but the breadth never exceeds 3 mm. The pinnulae do 

 not show any features of taxonomic importance, their lengths evidently varying with the degree of 

 contraction. 



The salient feature of U. spicata as compared with U. huxleyi is the remarkable length of its rachis. 

 This is very obvious on comparison of specimens in the photographs on PI. i, the autozooids of which 

 are filled with eggs or larvae. The photographs also show that the reticulation of the surface of an 

 autozooid does not depend on its being ripe, as was maintained by Hickson (1937, p. 120). 



In the British Museum specimen the rachis occupies about 17% of the entire height, but in other 

 specimens found hitherto it does not seem to be more than 10 %. This long and slender rachis is thus a 

 specific character, which however does not appear until, with growth, more autozooids have developed 

 (cf. Kiikenthal and Broch 1911, PI. 16, fig. 19, and PI. 20, fig. 51). In quite young specimens the 

 autozooids show a bilateral arrangement, which however very soon disappears with the irregular 

 budding of secondary autozooids on all sides of the rachis. 



The British Museum specimen also presents other features which are of special interest. The upper 

 part of the axis is round in section, but about the middle, or a little lower down, the axis is almost 

 square, and this continues till the upper limit of the holdfast, whence the circular section continues to 

 its lower end. It is remarkable that in the region with the square section the axis has an obviously 

 loose structure and that calcareous incrustation seems to be completely wanting. This structure 

 strongly recalls that of U. huxleyi and is evidently connected with the great flexibility of these speci- 

 mens compared with species whose axes are rich in lime incrustation. The central core also of the 

 axis in U. spicata is much looser in structure than the surface layer, and in some places below the 

 middle length the axis even seems to have a central lumen. It is probable that such axes may become 

 partially compressed into a square shape, by strong contractions of the sheath, e.g. when irritated by 

 the unusual environment in the trawl, net, or when placed in preservative. This distortion has 

 probably led some investigators to disregard the value of the normal shape of the axis as a taxonomic 

 character. In U. spicata the British Museum mature specimen shows that a round axis with an 

 extraordinarily loose or almost spongy structure in part, or even with a central lumen, is one of the 

 specific characteristics. 



It is probable that U. elongata, U. radiata and U. pendula Thomson and Henderson (1906) must 

 be included among the synonyms of U. spicata, but the descriptions are so incomplete and expressed 

 in such general terms, and the drawing of U. eloiigata is so schematic, that it is impossible to deter- 

 mine their identity without examination of the original material. 



Group B 



Axis quadrangular, encrusted with lime. 



In the second group of Umbellula without spicules the axis is quadrangular and strongly encrusted 

 with calcareous matter. A long series of species has been described from all waters, together with 

 the type species of the genus, the arctic U. encrinus (L. 1758), which seems to be confined to the deep 

 waters of the Arctic Ocean, north of the ridges separating them from the Atlantic. U. lindahli 

 KoUiker 1874 ^^^ been reported from the Atlantic Ocean, southwards from Greenland; U. ambigua 

 Fischer i88g (Marion 1906) from the central Atlantic; and more recently, U. carpenteri KoUiker 1880 

 has been shown to have a circumglobal habitat (Broch 1957). KoUiker described carpenteri from 



