352 DISCOVERY REPORTS 



As emphasized in my paper (1957) I agree with Hickson (191 6) that the Umbellula species fall into 

 two main groups, viz. those with spicules, and those without. On the other hand, I cannot share 

 Hickson's opinion that the shape of the axis (especially above the hold-fast), whether quadrangular 

 or round in section, is only of subordinate value taxonomically. To a marked degree, this shape is 

 correlated with calcification : round axes generally being rather poorly or almost uncalcified, whereas 

 quadrangular axes are as a whole heavily encrusted with lime. This character is doubtless constant in 

 and characteristic of a species. It is necessary to emphasize that we are speaking of the axis, and not of 

 the stem or the stalk with its cover of soft tissues (cf. the remark about Marshall's description of 

 U. gracilis, p. 255). 



Before going into details concerning the species of Umbellula and their synonymy, the following 

 key may clarify the important taxonomic characters: 



Key to the species of Umbellula 



I. With spicules. 



A. Axis round, only feebly encrusted with carbonate of lime and very flexible. Two kinds of spicules, large 



and small, all round in transverse section, feebly fusiform or rod-shaped, often the larger with some- 

 what knobbed ends, and sprinkled with granulations. Umbellula durissima KoU. 



B. Axis quadrangular, rather rigid and strongly encrusted with carbonate of lime. Spicules rod- or needle- 



shaped with three longitudinal ridges or flanges, sprinkled all over with small spines or warts. 



U. thomsonii Koll. 



II. Without spicules. 



A. Axis round, almost devoid of lime, thin and extraordinarily flexible. 



(i) Cluster with rather small autozooids concentrated in a pompon on the distal end of the rachis. 



U. huxleyi Koll. 

 (2) Clusters with large autozooids irregularly placed on a rather long rachis forming a tassel. 



U. spicata Kiikenthal 



B. Axis quadrangular, richly encrusted with carbonate of lime. 



(i) Axis thin but rigid and rather brittle. Autozooids numerous and small, packed in a pompon on the 



short distal end of rachis. Specimens at most some 35 cm. high. U. pellucida Kiikenthal 



(2) Axis variable, but generally rather thick and flexible. Autozooids large, in colder waters generally 

 very large and often very numerous in the tassels and not especially crowded. Full-grown 

 specimens seldom as small as 40 cm. U. lindahli KoU. 



(The arctic U. encrinus L. is more robust in its entire construction. Its axis is thicker than that of U. lindahli 

 which does not seem to exceed 2 mm. in its thinnest part just below the polypiferous rachis. The autozooids are 

 larger than those of U. lindahli. Possibly only an extreme form of this species, see p. 270.) 



SpiCULATE species of Umbellula 



The rich collection of Umbellula from the Swedish Deep-Sea Expedition indicated that as yet only 

 two species with spicules can be distinguished with certainty. These species are above all charac- 

 terized by the shape and structure of their spicules and by the construction of the axis. 



Umbellula durissima KoUiker 1880 



U. durissima Kolliker 1880, p. 16, PI. 8, figs. 32, 33. 



U. durissima + U. dura J. A. Thomson & W. D. Henderson 1906, pp. 92, 93, PI. 8, fig. 9. 



U. durissima + U. dura J. A. Thomson & J. Ritchie 1906, p. 859, PI. i, fig. 5. 



U. eloisa Nutting 1912, p. 43, PI. 6, figs. 3, 3a. 



U. durissima + U. dura+ U. eloisa Kiikenthal 19 15, p. 56. 



U. durissima Hickson 1916, p. 126. 



U. durissima Broch 1957, pp. 360, 363, text-fig. 7, PI. i, fig. 5. 



