104 DISCOVERY REPORTS 



" Tridontaster". Not all the lateral teeth have a hyaline tip protruding through the 

 sheath. 



In Odontaster validus there is much variation in the armature of the mouth-plates as 

 regards the size of the unpaired hyaline spine and of the hmermost of the 2 or 3 suborals 

 which stand on either side of this hyaline tooth. 



(i) The unpaired spine may be missing (St. MS 6) and the "bud" of a new one 

 appears in its place. This suggests a strenuous use for the teeth during which they are 

 occasionally broken. The laterals have not superseded it. 



(2) In one angle (St. 164) there are 2 subequal sharp, glassy tipped teeth, as in the 

 genus Asterodoji ; the same variation occurs in Odontaster peiiicillatus forma ^raj'/. From 

 its position I would say that one is the true unpaired tooth and the other a suboral spine. 



(3) In one angle there are 2 enlarged teeth which are certainly the innermost suboral 

 spines. The median tooth appears to have been broken and never to have regenerated. 

 The 2 teeth are blunt, recurved, and hardly hyaline at tip. The other 4 normal mouth 

 angles have only the tip of unpaired tooth emerging from its sheath, while in 2 of these 

 mouth angles the inner suborals are conspicuously enlarged and needed only a little 

 time to become "hyaline" (St. WS 62). 



(4) This last feature is not so important as it would seem since a considerable 

 number of specimens (St. 164, St. MS 10) have the sacculus persisting at the end 

 of some, or all, of the 5 normal teeth which thus have no exposed hyaline tip (see 

 Koehler's genus " Metado7itaster"). 



(5) Finally, there is great diversity in the size and shape of the unpaired tooth. They 

 may be thick, broad, and only half as long as the median suture separating mouth- 

 plates ; or short and slender ; or slender, and of various lengths to a trifle more than 

 three-fourths that of median suture. The truncate base of spine is at the inner end of 

 mouth-plates (M 71) or spaced from the end. These variations in the size and shape of 

 teeth may occur at the same station (164). 



It is obvious from the foregoing paragraphs that there is considerable latitude of 

 variation. In spite of Koehler's resuscitation of his O. tenuis (1920, p. 232) I believe he 

 was correct in his earlier view (191 2, p. 68) when he reduced it to a variety of validus. 



Colour notes (Ridgway's nomenclature). St. 164, Bordeaux purple (apparently no 

 distinction between specimens of A^ and A" of synopsis). St. 173, 22 specimens, pansy 

 purple; 2 specimens, apricot orange. Of these two exceptions one belongs in A^C^, the 

 other in A^B^ of synopsis. St. 1959, very dark brown, almost black, with whitish 

 spinelets dorsally and yellowish white ventrally. 



Type locality. Antwerp Island, Palmer Archipelago (First Charcot Expedition). 



Distribution. Circumpolar, north to South Georgia (including Shag Rocks) and 

 Bouvet Island ; shore to 653 m. 



Koehler in his several memoirs lists numerous stations, including (1908, p. 540) 

 " lies Falkland, 5-20 brasses ". I believe this to be due to a misplaced label. The locality 

 is unlikely. The extensive work of R.S.S. 'William Scoresby' on the Falkland Plateau 

 has failed to provide a single specimen of validus, but numerous examples of other 

 species of Odontaster. 



