64 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



in serving as a check on conclusions based mainly on the study of the arms and 

 pinnules. 



When attempts are made to overemphasize the systematic importance of the 

 characters presented by the proximal structures difficulties at once arise. The 

 proximal structures of all the comatulids are so very much alike that misinterpreta- 

 tions due to parallelism, convergence, or superficial similarity are very difficult to 

 avoid. An example of this is seen in Ptilometra australis in which the articular faces 

 of the radials (vol. 1, pt. 2, fig. 67, p. 43; pi. 2, figs. 973, 974) are very different in 

 appearance from those of Asterometra macropoda (vol. 1, pt. 2, figs. 69, 70, p. 43). 

 while the cirri, arms, and pinnules in the two types are very similar. In small coma- 

 tulids, also, the articular faces of the radials may differ very widely in general appear- 

 ance from those of larger species in the same group. For instance, the articular faces 

 of the radials in Comatilia iridometriformis (vol. 1, pt. 2, figs. 13, 14, p. 15) or in Lep- 

 tonemaster venustus (vol. 1, pt. 2, figs. 11, 12, p. 6) are, superficially at least, very 

 different from those of Comatella nigra (vol. 1, pt. 2, figs. 1, 2, p. 6) or Capillaster 

 sentosa (vol. 1, pt. 2, figs. 5, 6, p. 6). 



Gisle'n's first three groups, while quite distinct from each other, possess certain 

 features in common (see the keys following) by which they are separated from the 

 fourth group. It is most logical, therefore, to continue to regard them as forming a 

 single group, the Oligophreata, corresponding to his fourth group, the Macrophreta. 



I can see no reason whatever for excluding Comatonia from the Comasteridae, 

 as it possesses an enormous and typically comasterid comb. The comb is the only 

 structure not subject to wide and disconcerting modifications in this family. 



The family Tropiometridae is quite out of place in the Mariametrida. Its pris- 

 matic pinnules, the abrupt termination of the arm tips, and the almost unmodified 

 proximal pinnules emphatically indicate a relationship with the forms included in 

 the Thalassometrida, especially with the Ptilometrinae, while the centrodorsal 

 resembles that of Ptilometra and also those of certain of the Charitometridae, and 

 the cirri recall the cirri of various members of the Charitometridae. 



I can see no reason at all for including the Notocrinidae in the Thalassometrida. 

 Its arm and pinnule structure is certainly that of a macrophreate type, and more- 

 over all of the other comatulids with anomalous reproductive organs or habits are 

 macrophreate forms. 



Gisle'n's family Asterometridae seems to be the same as my subfamily Ptilome- 

 trinae of the Thalassometridae, but he does not say whether Ptilometra is included 

 in it; in fact, he does not mention Ptilometra at all. Though fairly well developed, 

 the side and covering plates in this group are certainly not very highly developed. 



It is difficult to see any reason for Gisle'n's division of the Macrophreata. The 

 Heliometrinae and Bathymetrinae are certainly much more closely allied to the Pero- 

 metrinae than they are to the Atelecrinidae or to the Pentametrocrinidae, while the 

 genera Pentametrocrinus and Thaumatocrinus, making up the family Pentametro- 

 crinidae, are certainly much more closely allied to each other than they are to any 

 other forms. 



