A MONOGRAPH OF THE EXISTING CRINOIDS 407 



and in number, but there are the added complexities arising from the varying develop- 

 ment and distribution of the pinnule combs. In one of the specimens of Comanthus 

 timorensis described by him the combs reached as far as P,,. In another specimen 

 from Java combs occur on every other pinnule at least to P35. This last he said has 

 not been described, and possibly belongs to a new species because of the very long, 

 slender, and well-separated arms. 



Gisl6n said that it is possible that the two species of the subgenus Comanthiis 

 have become a rubbish heap for various species difficult of solution, or types incon- 

 venient for the systematist. In such an event it would be possible to bring order 

 into the now confusing multitude and to get a comprehension of the existing types 

 by a detailed description of specimens. Characters that might perhaps be applied 

 would be the number, or the presence or absence, of cirri, the width of the arm bases, 

 the appearance and form of the brachials, of the division series, of the proximal and 

 distal pinnules and their component segments, and of the disk, and also the features 

 presented by the calcareous spicules. In other words, Gisl^n advocates a critical 

 examination of all available characters in a large amount of material. 



He suggested that possibly existing small species form hybrids between them- 

 selves and also cross with related species in the genera Comaster and Comantheria 

 and in the subgenus Cenolia. He said that if these hybrids are fertile they will, 

 after a couple of generations, be split into countless different types. 



Supposing, for instance, that a form without cirri, with the IIIBr and IVBr 

 series 4 (3 + 4), and with a comb of the Comanthus type, restricted to the proximal 

 pinnules, were crossed wath a form with cirri, with the IIIBr and IVBr series 2, 

 and with combs of the Comaster type — short high combs extending on to the distal 

 pinnules. Further, supposing that such a hybrid were fertile, and therefore in the 

 second generation was normally split up into the different possibilities of recombina- 

 tion. Then one would get forms that might be referred to all the known genera 

 within the subfamily Comasterinae. If, for instance, the first 3 genes as heterozy- 

 gotic might be represented with different forms — rudimentary cirri, a mixture of 2 

 and 4 components in the IIIBr and IVBr series — one would get 108 different possi- 

 bilities. The great variability in the division series, even in the same specimen, 

 according to Gisl^n, seems to be proof of the probability of the heterozygotical 

 nature of certain individuals of Comanthus parvicirra. 



Gisl^n regarded the subgenus Comanthus as a very critical group, forming to a 

 certain extent a connecting link between Comantheria-Comanthus (Cenolia) and 

 Comaster. He said that the reason he did not unite all these genera was that he 

 perceived the possibility of the subgenus Comanthus representing a rallying group 

 of hybrids and hybrid splittings with combinations of characteristics from different 

 form circles. He said that he should have hked to put the subgenus Comanthus as 

 a separate genus between Comaster and Comantheria-Comanthus {Cenolia). In this 

 case he would have distinguished it from Comaster by having the IIIBr series chiefly 

 consisting of 4 elements, and from the two last by its having comb-bearing pinnules 

 often occurring far out on the arms. 



But he said that tliis would not be satisfactory if all the forms in the subgenus 

 Comanthus of the comanthipinna type were not transferred to the subgenus Cenolia. 



