548 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



In 1888 in the Challenger report on the comatulids Carpenter said that the leading 

 character of Comaster, according to its proposer, depended upon the number of divi- 

 sions in the arms, and was rightly disregarded by Goldfuss who thought more of the 

 presence of basals on the exterior of the calyx as a generic distinction. He said 

 that Miiller adopted the genus in the sense in which it was understood by Goldfuss, 

 but that he seems eventually to have abandoned it altogether. This will doubtless 

 prove to be its ultimate fate, as it has not been seen by any naturalist since the time 

 of Goldfuss, whose original specimen of it was dissected and has since disappeared. 

 If his account of it is correct, Comaster must have been a very remarkable type 

 differing in many respects from all other recent comatulids. But Carpenter was 

 now inclined to believe that the apparent peculiarities are merely due to the want of 

 knowledge respecting the internal structure of the calyx of comatulids which was 

 prevalent at the time of Goldfuss, and that Comaster is in reality nothing but a large 

 Antedon (in the broad sense in which the term was used by Carpenter) or Actinometra 

 (comasterid). He placed Comaster of Agassiz, 1835 (really 1836), in the synonymy 

 of Actinometra. Comaster as used by Dujardin and Hupe he placed under both 

 Antedon and Actinometra, although the only included species mentioned by them was 

 the Comaster multiradiatus of Goldfuss. Comaster as used by Lundgren, 1874, he 

 placed in the synonymy of Antedon. 



Carpenter did not discuss bennetti in detail, since it was not secured by the 

 Challenger, but he gave its specific formula, several times mentioned the chief fea- 

 tures of the arm division, and inserted it in the key to the species of the Parvicirra 

 group of Actinometra. Here it fell in the section characterized by the presence of 

 IIIBr 4 (3 + 4) series, and also I VBr 4 (3+4) series, and by having the cirri XL-L, 25. 

 It was contrasted with parincirra, regalis, schlegeli and peroni, in which the cirri 

 were said to be X-XXX. 



Among the localities for this species he included the Sooloo (Jolo) Sea, this 

 record being based upon some arm fragments collected by the United States Exploring 

 Expedition and now in the Museum of Comparative Zoology at Harvard University. 



Of peroni he gave a specific formula, and mentioned the specimen which he had 

 found among the 3 individuals included by Lamarck in his Comatula multiradiata. 

 But the only locality he gave for this form was Ceram. He inserted it in the key 

 to the species of the Parvicirra group, where it was distinguished by having 30 cirrus 

 segments in contrast to parincirra, regalis, and schlegeli, in which the cirrus segments 

 number 10-20. He mentioned elsewhere that the cirri are very long. 



Dr. Clemens Hartlaub in 1891 recorded and published notes upon 4 specimens 

 which had been collected by Dr. J. Brock at Amboina. One of these had the charac- 

 ters described by Carpenter for peroni, and Hartlaub said that he believed the separa- 

 tion of these 2 supposed species could no longer be maintained. 



He wrote to Carpenter expressing this opinion, and Carpenter replied that he 

 agreed that the 2 forms are in reality identical. 



Hartlaub reexamined the specimen at the Gottingen Museum from the Loyalty 

 Islands which had been described by Bolsche in 1866, and in this he found additional 

 evidence that bennetti and peroni are synonymous. The locality he gave us Uca, 

 which is a slip for Uea. 



