26o DISCOVERY REPORTS 



' U. antarctica '. To a great extent development of these covering tissues seems to depend on 



ecological conditions. 



Characters like 'thin' as opposed to 'relatively thick' must be said to intergrade, and lateral or 

 dorso-ventral compression is doubtless due to contraction ; on the other hand in U. pelhicida the shape 

 of the autozooids is to some extent characteristic. 



Hickson (191 6) also discussed the question of distinguishing between the many species of this group. 

 In general terms he gives his view as follows: 



I. With a large number of autozooids (i.e. 40-50 in specimens of from 400-500 mm. in length) the autozooids 



being 20 mm. or more in body length. 



A. Tentacles much shorter. U. antarctica 



II. With an intermediate number of autozooids (i.e. 15-40 in specimens 400-500 mm. in length). 



B. With a quadrangular axis. U. encrinus, U. Weberi 



III. With a small number of autozooids (i.e. 8-15 in specimens 400-500 mm. in length). 



A. With tentacles longer than the bodies. U. Carpenteri, U. Lindahli, U. loma 



B. With tentacles shorter than the bodies. U. Jordani, U. magniflora, U. pellucida 



It is evident that Hickson was of the opinion that the number of autozooids is comparatively 

 constant in any single species of a certain length. Even allowing for the fact that as yet no specimen 

 of e.g. U. pellucida as long as 400 mm. has been observed, the idea of a specific, characteristic, constant 

 proportion between the length of the specimen and its number of autozooids is clearly disproved in 

 my last paper (Broch, 1957 text-fig. 5), by the table of growth in U. gwitheri, and again by the col- 

 lections brought home by the Discovery Investigations. 



In his key, Hickson only emphasizes in one place (II, B) the fact that the axis is quadrangular in 

 two species of the group. Nevertheless, it is quadrangular in all species noted here in his table. The 

 other character, the comparative length of body and tentacles in the autozooids, is of very little value 

 because of the degree of contraction.* 



Later on Hickson (1937) examined a great number of Umbelliila from the Indian Ocean, which in 

 spite of the quadrangular axis he named U. hiixleyi Kolliker. Special attention was paid to the degree 

 of contraction in relation to the sexual development of the autozooids, and Hickson came to the 

 conclusion that 'it is obvious that pregnancy offers a physical impediment to contraction'. 



It is usually quite easy to recognize the ripe from the barren autozooids without dissection, because 

 the gonads in the former are clearly indicated by rectangular areas on the surface of the body wall. 

 Such markings are clearly shown in Kolliker's figure (1880, pi. IX, fig. 37a) of U. huxleyi, and in 

 Kukenthal and Broch's figure (191 1, pi. XVI, fig. 17) of U. spicata. In the picture of U. pellucida 

 (pi. XVI, fig. 20) by the authors of the ' Valdivia' Report the autozooids are strongly contracted, and 

 there are no quadrangular markings on the body wall. No mention is made of the sexual condition of 

 any of the species, but there can be little doubt that the two species, U. spicata and U. pellucida, are 

 only ripe and barren forms of the same species, (according to Hickson). 



It will suffice to point out that Kolliker's type specimen of U. huxleyi is a quite young, unripe 

 U. spicata like Kiikenthal and Broch's specimens of this species. The first known ripe specimen of 

 U. spicata is figured in PI. I, fig. 2 ; its surface is devoid of rectangular areas in the autozooids. The ripe 

 specimen of U. huxleyi figured in PI. I, fig. i« also has smooth autozooids. Several specimens of 

 'U. huxleyi' from the John Murray Expedition (re-examined by me in the British Museum (Nat. 



It might seem questionable, whether U. loma belongs to this group, because Nutting (1909) wrote that it has a few 

 spicules on the swollen part of rachis below the polyp cluster, and one ought possibly to query it here. However, all other 

 features suggest that the specimen is in its right place. 



