284 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



are clearly related to the type characteristic of most of the species of Psathyro- 

 metra, and occurring also in T haumatometra, Ifsometra, and Thysanvmetra. 



In the genus Eumorphometra a combination of rods and plates occurs which 

 suggests the conditions found in Sara-metro, inverted, and the conditions in the 

 Pentametrocrinidse. 



The genus Atelecrinus exhibits a very distinctive type of covering plate, but 

 as only one pinnule was available for study its possible variations could not 

 be traced. 



A further development of the same type occurs in the species of the genus 

 Nemaster, and in a still more perfected form in Comatilia. 



In the Comasteridse, Himerometridse, and Tropiometridse the adambulacral 

 armature, when present, is almost invariably of a very distinctive type, which, 

 however, is alike in all three families. 



Most commonly the adambulacral deposits are in the form of simple spicules 

 or rods, which occur either singly or in tandem pairs. These spicules or rods are 

 straight or slightly curved, and measure in Leptonemaster venustus about 0.020 

 mm. ; in Zygometra comata, from 0.015 mm. to 0.025 mm. ; in Eudiocrinus pinnatus, 

 from 0.070 mm. to 0.115 mm.; in Cyllometra albopurpurea, from 0.002 mm. to 

 0.090 mm. ; in Antedon petasus, from 0.025 mm. to 0.050 mm. ; in Perometra 

 diomedece, from 0.025 mm. to 0.060 mm. ; and in Hathrometra prolixa, about 0.300 

 mm. In nearly all of these types they are more frequently absent than present. 



Since in the majority of comatulid types, representing all the comatulid 

 families excepting the Atelecrinidse (of which our knowledge is very limited) the 

 perisome of the pinnules is quite without deposits, contains a vast multitude of 

 extremely minute spicules, or a series of simple spicules, in other words, recapitu- 

 lates more or less exactly the conditions found in the perisome of the disk of which 

 it is the direct continuation ; since the adambulacral deposits never show any trace of 

 ontogenetical degeneration; and since, while all plates of every sort develop origi- 

 nally from minute spicules, and a spicule may increase in size without forming a 

 plate, and there is no evidence, so far as I am aware, that a plate which has once at- 

 tained phylogenetic significance ever degenerates into a solid rod or spicule, we are 

 apparently justified in assuming that the adambulacral armature of the pinnules 

 consists merely of perisomic (secondary) plates strictly comparable with the 

 pcrisomic plates of the disk, but localized and modified as the result of physical 

 limitations and mechanical factors, originating as a consequence of the inherent 

 tendency in the crinoids to form calcareous deposits wherever possible, and of inde- 

 pendent origin within the group and even within small divisions of the group, with 

 no phylogenetic history whatever behind them. The adambulacral plates are thus 

 merely the continuation upon the arms and pinnules of the perisomic plates of the 

 disk, and in those species in which they are highly specialized forms intermediate 

 between them and the plates of the interpalmar regions may always be found along 

 the disk ambulacra. 



The simple spicule may develop in either of two directions: (1) It may simply 

 enlarge and thicken into a stout rod, almost invariably with the outer, or with 

 both, ends swollen and roughened, as in Comatonia cristata, Oligometrides adeonce. 



