A. H. CLARK : THE CBINOIDS OF THE INDIAN OCEAN. 253 



bears only rudimentary pinnules. Infrabasals are present, resembling those of 

 Isocrinus and Hypalocrinus. The stem and cirri are essentially as in Hypalo- 

 crinus. 



Remarks. — The discovery of this extraordinary genus completes the parallel 

 between the comatulids and the pentacrinites. 



The only difference between the pentacrinites and the comatulids is that the 

 former possess a stem, wliile in the latter the young are stalked, but during growth 

 break away from the stem retaining only the topmost columnar attached to the 

 calyx. 



This difference is not nearly so fundamental as it would seem ; for the 

 pentacrinites always discard the distal part of the stem, and are continually 

 dropping it off bit by bit all through their life; and the comatulids, while their 

 calcareous centrodorsal is the last column to be formed, and is never composed 

 of more than one columnar (representing a single pentacrinite nodal), really 

 possess the entire pentacrinite stem enclosed within this columnar ; so that where- 

 as the pentacrinite possesses whorls of cirri at regular intervals (on the so-called 

 nodals) along a greatly elongated stalk , the comatulids are forced to crowd all their 

 cirri together and to extrude them all through the same nodal. 



In its arm structure the genus Eruloxocriniis is exactly similar to such coma- 

 tuhds as those comprised within the families Stephanometridse, Marimetrid<e, 

 Pontiometridffi, etc., while the genus Hypalociinus finds a counterpart in all the 

 ten-armed species, no matter to what family they may belong. The most primi- 

 tive type of arm structure is found in the species of the family Pentametro- 

 crinidae; but this is repeated, somewhat modified, in the genus Metacrinus. In 

 Comastrocrimis we find represented the very highly specialized type of arm division 

 characteristic of the most specialized coraasterid subfamily, the Capillasterinse , 

 and we also find represented both the type characteristic of the genus CapiUaster 

 and that characteristic of the genus Comatella^ the two often occurring side by 

 side in the same individual. In addition we occasionally find the type of division 

 characteristic of the comasterid subfamily Comasterin;e and of the Himeromet- 

 ridse, so that in this genus all the missing data is supplied whereby the arm 

 structure of the comatulids is shown to be exactly parallel to that of the penta- 

 crinites. Moreover, many of the specimens of the various species of Comastrocri- 

 nus are very irregular, and their irregularity is recapitulated in the Capillasterinje, 

 being identical with that seen in the specimen of CapiUaster miiltiradiata from 160 

 fathoms in the Malay Archii)elago described in the Appendix. 



The agreement in arm structure between the various pentacrinite and comatu- 

 lid genera cannot in any sense be interpreted as suggesting a polyphyletic origin 

 for the latter, or as suggesting the origin of certain comatuhd genera from certain 

 definite pentacrinite genera, hypotheses for which there is not the slightest 

 evidence. It is merely a parallelism, the result of the development, under the 

 same conditions, of two divergent branches from the same stock, each of which 

 possesses all the latent tendencies inherent in the other. 



