194 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



that have proved to be unreliable, some of them of at most only specific significance, 

 some even inconstant within a single species. 



The first genus established, Charitometra (genotype Antedon incisa), was essen- 

 tially the equivalent of the family Charitometridae as now understood, except for the 

 exclusion of the species of the genus Poecttometra (genotype Antedon acoela) established 

 at the same time, and the inclusion of Antedon compressa and A. orion which are now 

 assigned to the genus Parametra in the family Thalassometridae. The genus Poectto- 

 metra was separated on the basis of the concave sides of the elements of the IBr series, 

 their produced and flangelike dorsolateral edges, and the much expanded genital pin- 

 nules. In separating Poecilometra from Charitometra I was influenced by the fact 

 that Carpenter had made his Antedon acoela the type of a special group, the Acoela 

 group, in which he also included Antedon discoidea, a species of the family Calometridae. 

 He defined the Acoela group as having "The rays not flattened laterally. Pinnule 

 ambulacra well plated." 



In 1909 I established the subfamily Charitometrinae and gave a key to the seven 

 included genera, five of which were diagnosed for the first time. The first division 

 included genera in which the IBr series and lower brachials have produced or everted 

 edges forming a thin flangelike border. Poecilometra was distinguished by having the 

 dorsal edges also produced, the genital pinnules greatly and abruptly expanded, and 

 the IBr series and brachials rounded dorsally without ornamentation. In Glyptometra 

 (genotype Antedon tuberosa) the IBr were said to have only the dorsolateral edges 

 produced, the genital pinnules very regularly expanded and evenly tapering, and the 

 IBr series and lower brachials with the dorsal surface rugose or tubercular. The 

 second division included genera in which the IBr series and lower brachials do not have 

 the dorsolateral edges produced. In one section the included genera have the third 

 and fourth segments of the genital pinnules broad and nearly flat on the outer side, but 

 the fifth segment smaller. This section included Strotometra (genotype Antedon hep- 

 burniana) with less than 12 cirrus segments and the IBr series and arm bases diverging 

 at a wide angle, and Charitometra with more than 15 cirrus segments and the IBr 

 series and arm bases diverging at a relatively small angle. The other section included 

 genera with the genital pinnules evenly and only slightly expanded. There were 

 three divisions of this section. In one the lower pinnules were said to be twice as long 

 as the middle pinnules; this included Crinometra (genotype Comatula brevipinna). 

 In the other section the lower and middle pinnules were said to be of approximately 

 the same length. This section included Pachylometra (genotype Antedon distincta) in 

 which the cirri were said to be large and stout with 18 or more segments, and Chlo- 

 rometra (genotype Antedon garrettiana) in which the cirri are weak and slender with 

 16 or fewer segments. 



In his excellent memoir on the comatulids of the Blake Expedition published in 1912 

 Dr. Clemens Hartlaub had considerable difficulty with the interrelationships of the 

 species of this family. He used Carpenter's classification instead of my new classifica- 

 tion then in process of development for the reason, as he told me, that he was simply 

 completing a work begun by Carpenter and therefore felt himself bound to follow Car- 

 penter's original plan. He identified two specimens from Blake station 157 as Antedon 

 angusticalyx, but said that they stand between A. angusticalyx and A. inaequalis. 

 But in the larger of these specimens P D is 11 mm. long with about 30 segments, PI is 



