the triangular form not becoming pronounced until after the third or fourth; and (6 the seeond 

 segment of the lowest pinnules is very shori ( learly, therefore, Atopocrinus cannot belong to 

 the Pentametrocrinidae. 



Atopocrinus suggests the genus Zenotnetra in certain ways — the arms are stoul and 



ruimst, and the conical centrodorsal is divided into ten sharply dclimitcd areas each with a 

 column of cirrus sockets, but i in Zenotnetra the disk is broader and the ambulacra run 

 direct from its surface to the ventra] surface of the arms; (2) the earlier brachials and the first 

 gial pair are oblong; and (3 the cirrus sockets have no fulcral ridges. 



In : 1 1 the details of the arrangement of the cirri on the centrodorsal, in (2) the details 

 of the structure of the cirrus sockets, in (3) the structure of the disk, in (4) the triangular 

 brachials at the base of the arms, and especially in (5) the triangular firsl syzygial pair, 

 Atopocrinus agrees with Atelecrinus and differs from all other comatulids. 



Atopocrinus differs from Atelecrinus (1) in having no basals, (2) in having only five 

 arms. and (3) in having all of the pinnules present. 



Though very large in . Itelecrinus anoma/us, the external basals become very greatly 

 reduced in such species as ./. sulcalus, the variation in size being so great as to suggest 

 that their presence or absence does not constitute a valid major character in the diagnosis 

 of the family. 



In the family Zygometridae one of the genera, Eudiocrinus, possesses only five arms, 

 so that the same condition in a single genus of another family might be anticipated. 



In the Comasteridae, Colobometridae, Zenotnetrinae, Perometrinae and Pentametrocrinidae 

 species or genera) with deficiënt, and also with complete, pinnulation are found; it is therefore 

 quite natural that there should be a genus of Atelecrinidae with complete pinnulation. 



Since the differences between . Xtopocrinus and Atelecrinus are of relatively small impor- 



while the differences between Atopocrinus and all the other comatulid genera are 

 fundamental, we appear to be abundantly justified in placing . Xtopocrinus in the Atelecrinidae. 



Atelecrinus 1'. H. Carpenter. 

 I\ey to the Species of the Genus Atelecrinus. 



rhe centrodorsal bears 15 columns cirrus sockets (Ha waiian Islands). . conifer 

 a The centrodorsal bears io columns n[ cirrus sockets 



basals very large, the height equal to more than half the width; no raised 

 rim about the cirrus so< cirri with 19 — 20 segments, of which the 



outer are but little long( r than broad, ending in a stout recurved claw 



Mo lm anomalus 



b balais very low, forming a narrow band between the radials and the centro 

 dorsal, or reduced to low triangular areas in the angles of the calyx; a 

 horseshoe-shaped rim borders the cirrus sockets proximally and Iaterally 

 on the surface of the o the columns of cirrus sockets are 



ited in five pairs bj relativelj broad shallow interradial grooves; 



