384 DISCOVERY REPORTS 



i8° 37' E, 150 fm. = 275 m.). The nearest Discovery stations at which soundings were 

 taken were St. 102 at 35 29' 20" S, 18 33' 40" E, in 1800 m., and St. 407 from 

 35 13' S, 17 50J' E to 34 57' S, 17 48' E in 2822 m. It is thus evident that Chal- 

 lenger St. 142 must have been on the slope of the continental shelf. Hydrological data 

 are rather scanty, but the 'Discovery' took a series of temperatures at St. 102 which 

 shows a uniform temperature down to 50 m. after which the temperature falls, the 

 maximum fall being between 150 and 200 m. In this connexion it is worth noticing 

 Deacon's reference (1937, p. 57) to a northward-flowing current of sub-Antarctic water 

 at 200 m. in the region of the subtropical convergence south of South Africa. These are 

 mere hints, but they suggest that hydrological conditions off the continental shelf in the 

 South African region may be such as to account for the occurrence there of specimens 

 of C. darwinii of more southerly type. 



To get a rough idea of the relationship of the whole fauna at Challenger St. 142 the 

 list given by Murray (1895, p. 416) was examined. At that time about 45% °f the 

 species were not known elsewhere. Of the remainder, more than a third were also taken 

 at sub-Antarctic stations. 



The occurrence of specimens approaching the Antarctic type at Possession Island 

 (Challenger St. 148) may perhaps be related to the greater depth at which they were 

 obtained, the dredgings being made at 210-550 fm. as compared with 20-127 fm. at 

 Kerguelen and 50-150 fm. at Prince Edward Island (Murray, 1895, pp. 439, 458, 475). 

 The relationships of the other Cellularine Polyzoa of these islands are discussed on 

 p. 483. 



Where two or more types are found in the same region, there are some indications 

 of a correlation with depth, despite the small range of depth represented (see, for 

 example, Tables 4 and 5, pp. 487, 490, showing the range of depth at South Georgia and 

 in the Patagonian region), but it is very inexact, as might be expected in view of the 

 complicated depth relations of the different types of water. 



The smaller zooecia, flat branches, greater number of spines, and less specialized 

 vibracula, scuta and avicularia, are all features of C. minima which can be regarded as 

 juvenile, and one does in fact find some gradation in these characters as one passes from 

 the base to the tip of some colonies. I have not, however, found the whole range in 

 any one specimen, and the correlation with geographical distribution is enough to show 

 that it is not simply a case of the gradual assumption of the adult characters as the colony 

 grows. The distribution of certain species of Amastigia, discussed on p. 325 is comparable. 



Ancestrulae. Ancestrulae attributed to Caberea darwinii were obtained from Marion 

 Island and Prince Edward Island (Sts. 1562, 1563, 1564, 7. iv. 35, 12 ancestrulae), the 

 Burdwood Bank (St. WS 87, 3. iv. 27, 1 ancestrula), South Georgia (St. 42, 1. iv. 26; 

 St. WS 33, 21. xii. 26; St. WS 42, 7. i. 27, 3 ancestrulae), off Oates Land (St. TN 194, 

 22. ii. 11, 2 ancestrulae), Ross Sea (St. TN 339, 24. i. 12, 1 ancestrula), and McMurdo 

 Sound (National Antarctic Expedition, 28. ii. 02, 2 ancestrulae). 



The ancestrula is vase-shaped, with spines all round the obliquely terminal opesia 

 and the colony is slung by rootlets (Fig. 23 A-C). The first two, or more, zooecia have 

 more than the typical number of spines, and no frontal avicularia. Scuta, with a well- 



