The Echinoderm Fauna of South Africa. 355 



of Astrophiura on the Agulhas Bank, off the coast of Cape Colony 

 in 175 fms. ; 5 specimens were taken, one of which was figured, 

 but not named or described, by Chun (1900, Aus den Tiefen des 

 Weltineeres, p. 488). In 1913, Matsumoto was so fortunate as to 

 be able to describe as a new species, a fine specimen of Astrophiura 

 from Okinose, a submarine bank in the Sag-ami Sea, Japan. Finally 

 Koehler in 1915 (I. c.) gave full descriptions of the five specimens 

 taken by the VALDIVIA, which he considered different from both the 

 Madagascar species and the Japanese. Matsumoto (1917, Mon. Jap. 

 Oph., pp. 245-246) fails to realize that it is Chun's specimens upon 

 which Koehler's species is based and hence he writes as though 

 there were four species of Astrophiura known. 



There is no doubt that the Japanese species (A. kawamurai) is a 

 well-marked form ; it needs no further discussion here. But when 

 one begins to compare the South African and Madagascar species, 

 difficulties arise. In the first place, there is but one specimen 

 known of the latter (per in Ira') and it is obvious from Sladen's figures 

 that it is either an aberrant individual or the dorsal surface has 

 been injured and more or less regenerated. In the second place, 

 no two of the five specimens of cavellae are exactly alike in the 

 arrangement of their dorsal plates. It is true that no one of them 

 agrees with permira but it is hard to see that they differ more from 

 that species than they do from each other. In the third place, the 

 two specimens in the PIETER FAURE collection, measuring 9 and 

 10 mm. in diameter of entire body, agree closely with each other 

 but differ from both permira and cavellae in certain particulars, 

 although they were taken very near the type-locality of cavellae. 

 Both specimens have large tubercles on the five largest radial plates, 

 and a central cluster of five erect, peg-like spinelets or tubercles 

 crowded at the center of the centrodorsal plate ; the height of these 

 is about one-half the radius of the centrodorsal. No such cluster is 

 recorded for any specimen of Astrophiura as yet described. Again 

 the first circle of plates surrounding the centrodorsal is made up, 

 not of five plates as in typical cavellae, but of ten nearly equal 

 plates, arranged in five radial pairs; there is a minute tubercle, at 

 the center of more than half these plates. One of Koehler's spec- 

 imens had ten plates in this first series but these were very une- 

 qual and so arranged as to give three large plates in each inter- 

 radial series, besides the extramarginal triangle. In one of the 

 PIETER FAURE specimens, there are three such plates in one interra- 

 dins but this is due to the horizontal division of what is typically 

 the uppermost interradial. There is no trace of a tubercle on the 



