196 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



and patula must be referred to Glyptometra, and so must lata, crassa, and perhaps also 

 other species now referred to Perissometra. 



In 1927 Gislen recorded specimens of Perissometra aranea, which he had described 

 in 1922, and described P. carinata and Monachometra mortenseni, all three of which are 

 based upon specimens of Parametra orion, a species of the family Thalassometridae. 

 In 1928 after examining the Challenger material in the British Museum he published 

 various notes on the classification of the Charitometridae. He said that the genus 

 Parametra must be referred to the Charitometridae, repeating the error I had made in 

 my original diagnosis of the genus Charitometra to which I referred Antedon compressa 

 and A. orion, both species of Parametra. He noted that Carpenter's Antedon distincta, 

 which I referred to the genus Pachylometra, must be referred to Crossometra, as it has the 

 cirri arranged in 10 columns. Under Perissometra angusticalyx he said that Perissometra 

 is indistinctly separated from Crossometra; all Perissometras with the IIBr series 4(3 + 4) 

 should be referred to Crossometra, leaving in Perissometra 10-armed forms and those with 

 the IIBr 2 patula, timorensis, lata, flexilis, and aranea (the last= Parametra orion). 

 He noted that two widely different forms are included under Perissometra patula. One 

 specimen with only 10 arms, as is often the case the larger one, has the centrodorsal 

 discoidal, the cirri stout, arranged in 12 columns, the segments with a dorsal spine and 

 keel, and a simple opposing spine. In the other, which seems to be the same type as 

 described and figured by Carpenter, the centrodorsal is sharply conical with the cirri 

 in 10 columns, the cirrus segments without dorsal spines, the opposing spine forked, and 

 the brachials collar-shaped; the arms, he said, were very similar to those of his Dio- 

 dontometra= Chlorometra) , but the arrangement of the cirri is different; the cirrus sockets 

 are striated, just as in some fossil comatulids. He said that the first type should be 

 referred to Monachometra. Under Perissometra flexilis he said that the subdivision of 

 the genus Perissometra into species with more or less than 13 arms is not suitable. Un- 

 der Monachometra robusta he said that this species, considered by me to be a Peris- 

 sometra, must be referred to Monachometra. He remarked that Charitometra and 

 Poecilometra are closely related. The flanges on the proximal pinnules are rather simi- 

 lar. The difference of the basicurvity is, however, noticeable. He suggested the fol- 

 lowing differences: (a) Arm bases well separated, Poecilometra; (b) Arm bases flattened 

 and close to each other, Charitometra. He said that Strotometra parvipinna ought to be 

 referred to Perissometra and is perhaps the same as his P. aranea. 



In conclusion Professor Gislen said he thought it evident from the examples given 

 that the family Charitometridae is still not satisfactorily subdivided. It seemed to 

 him that I had laid too much stress on the ornamentation of the arms. He believes 

 that the cirri offer quite as good or even better differential characters. In many cases 

 they are exceedingly characteristic, and evidently they are much less subject to varia- 

 tion than the ornamentation. He said that there are at least two very distinct types of 

 cirri within the family: (1) Stout cirri lacking dorsal spines, and (2) Slender cirri with 

 distal dorsal spines. The former may have or lack an opposing spine which is, besides, 

 often transverse or even bisected. Though in the former case the cirri are devoid of 

 dorsal spines the segments may possess a faint longitudinal crest. He believed that 

 by considering these characters we may find it much easier to subdivide the family so 

 as more correctly to indicate interrelationships. 



