DISTRIBUTION AND RELATIONSHIPS 



129 



face waters of the Antarctic and sub-Antarctic zones, passes through it (Deacon, 1937, 

 p. 23 and fig. 5). Kerguelen and the adjacent islands may, however, be regarded as a 

 sub-Antarctic district (see pp. 1 30-131). 



The table shows that seventeen of the nineteen Antarctic comatulids occur in the 

 Weddell Sea sector (in the wider sense, i.e. including outlying islands) ; six in the Indian 

 Ocean sector, including Kerguelen and Heard Island ; five in the Ross Sea sector. Four 

 species, Promachocriniis kergtielensis, Florometra mawsoni, Anthometra adriani and 

 Notocrinus virilis, occur in all three sectors and may therefore be presumed to be 

 circumpolar. They are all large species and therefore the most likely to be taken by 

 collecting expeditions. This and the fact that more collecting, and more careful col- 

 lecting, has been done in the Weddell Sea sector than in any other makes me think that 

 many of the species now known from there alone may later be found elsewhere around 

 the continent ; in other words I think it premature to discuss the distribution of the 

 species within the Antarctic. Of the species that are known to be circumpolar Pro- 

 machocrinus kergtielensis has been found at all the Antarctic localities where collecting 

 has been done except the Shag Rocks ; the other three are, so far as is known, confined 

 to the coasts of the continent and the adjacent islands. 



The most interesting feature of the table is that it shows that two Antarctic species 

 do occur outside the Antarctic and that they occur towards or along the east coast of 

 South America: Phrixometra nutrix on the Burdwood Bank, Isometra vivipara on the 

 Burdwood Bank and on the Patagonian shelf as far north as the River Plate. 



There are, I think, other reasons for believing that the shallower water crinoid fauna 



