522 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



Carpenter states that " the spines of the cirri are present from the first, but 

 those on the calyx, arms, and pinnules do not appear till after the pentacrinoid 

 stage." This indicates that the ossicles of these pentacrinoids are smooth. But so 

 far as we know the development of spines in the comatulids is always much more 

 marked in the very young than in the adults, the spinosity becoming less and less 

 conspicuous and often disappearing altogether in later life. As these pentacrinoids 

 bear no spines it is almost certain that the species of which they are the young 

 is also devoid of spines. 



The extraordinary stout cirri are obviously more like those of Crotalometra 

 porrecta than like those of Thalassometra multispina. Bearing in mind that the 

 relative length of the cirrus segments always decreases with age. it is evident that 

 it would be difficult to derive the long segmented cirri of Thalassometra multispina 

 from the relatively short segmented cirri of these pentacrinoids, while the cirri of 

 Crotalometra porrecta are very easily derived from them by the simple addition 

 of segments distally, plus the shortening of the segments which always accompanies 

 increase in size. 



Thus there seems to me to be no doubt that these pentacrinoids are the young 

 of Crotalometra porrecta and not of Thalassometra multispina, as supposed by 

 Carpenter. 



As noted by Carpenter, the youngest of these pentacrinoids is remarkable for 

 the shortness of the column, which has only 14 segments, including the centrodorsal, 

 and terminates in a large and expanded terminal stem plate. His figure shows a 

 pentacrinoid 7.5 mm. long, the crown measuring 3.5 mm. and the column 4 mm. 



The centrodorsal is a thin plate but little larger than the columnals below it, 

 and the rudiments of three radial cirri have appeared upon it, the position of 

 the other two being indicated by imperfect sockets. There are three similar very 

 short discoidal segments beneath the centrodorsal, and the longest columnals (the 

 seventh-ninth from the centrodorsal) are scarcely more than twice as long as broad. 



The basals are very short, while the radials have almost attained the perfected 

 form. The elements of the IBr series are so broad as to be just in lateral contact 

 when the arms are closed. There are 12 or 13 brachials, but no pinnules. Very 

 large and conspicuous covering plates, which are supported by imperfect side 

 plates, are present. 



In the next stage, in which the crown is 5.8 mm. long, the first pinnules have 

 appeared on the eleventh or twelfth brachials and are provided with large covering 

 and side plates. The five radial cirri which were first formed are well developed, 

 so far as can be judged from their basal segments, which are all that remain, while 

 there are one or two slight indications of the second whorl of cirri, the positions 

 of which alternate with those of the first. The centrodorsal has increased consid- 

 erably in thickness, as have also the segments below it, which have similar reenter- 

 ing angles, just as is the case in the infranodals in the column of the pentacrinites. 



In the oldest larva the basals are concealed by the centrodorsal, which has now 

 reached a considerable size, with the second whorl of cirri well developed and even 

 traces of a third, while there is only one discoidal segment below it. 



