The Echinoderm Fauna of South Africa. 257 



of the inferomarginals is like that of the upper series. Actinolateral 

 plates in about eight series; the first (next the adambulacrals) 

 extends from the oral plates to about the fourteenth inferomarginal 

 (counting from interradius): the second series extends to the eighth 

 inferomarginal and the third reaches the sixth ; remaining series 

 irregular and made up of somewhat smaller plates than the first 

 three ; each actinolateral plate carries a marginal series of 7-9 angu- 

 lar granules, more widely spaced than on the abactinal plates, and 

 a single central granule, or rarely two; there seem to be no pedi- 

 cellariae on these plates. 



Adambulacral plates about 56 on each side of the furrow; they 

 are distinctly wider than long and their armature is in three very 

 sharply defined parallel series; the furrow series consists of 4, rarely 

 5, subequal, almost cylindrical, blunt spines, over a millimeter long; 

 the second series consists of 3, rarely 4, very similar but somewhat 

 more prismatic spines of about the same size; the third and outer- 

 most series is made up of 3 angular spinelets not much larger than 

 the granules on the adjoining plates. Oral plates not at all conspic- 

 uous and little swollen; their outlines are quite indistinct; proximally 

 there are 5 spines on each side, the ones at tip of jaw longest 

 (about 2 mm.) ; these spines are very strongly compressed, with 

 widened and rounded tips; on the surface of each plate are a dozen 

 or more smaller and more prismatic spines, the distalmost much 

 like the actinolateral granules. Colour, brownish-yellow. 



P.F. 18183. Cape Point, N. by E., 9 miles. 81-87 fins. Gr. m. 

 and s. 2 specimens; adult. 



P.F. 18230. False Bay, 21 fins. Fne. s. 2 specimens; adult. 



Holotype, South African Museum, No. A 6422, P.F. 18230. 



Examination of the internal anatomy confirms the evidence of the 

 external characters, and proves this to be a true Mediaster: The 

 internal radiating ossicles of the abactinal skeleton are well developed 

 and rudimentary superambulacral plates are present. As regards the 

 latter feature, however, I do not place very much confidence in its 

 value, for unless these plates can be shown to have a real morpho- 

 logical value in some group of sea-stars, I must doubt their phylo- 

 genetic significance, and their presence in a rudimentary condition, 

 or their absence, would not seem a matter of any real importance. 

 Their position is such with reference to the ambulacrals and adam- 

 bulacrals that their independent origin in totally unrelated groups 

 would appear to be highly probable. 



There is no doubt that Mediaster capensis is very nearly related to 

 M. australiensis H. L. C. but 1 think the differences in the paxilliform 



