MONOGRAPH OF THE EXISTING CRINOIDS. 219 



regarded as an exactly similar body covering the development of which is induced 

 by the atrophy or cessation of development of the primary plates. The reason 

 that the perisomic plates are very rarely so perfected as the primary plates is prob- 

 ably to be found in the much greater general mobility of the areas where they are 

 formed, which tends to separate the mesodermic lime depositing cells into small 

 groups, thus inducing the formation of a great number of individual spicules, 

 instead of allowing these cells to lie in broad areas and thus permitting a few 

 spicules to develop into large plates. 



If this is so, then the perisomic plates represent phylogenetically the most 

 primitive type, and we are led to the hypothesis that the original covering of the 

 crinoid body essentially resembled that of the holothurians (figs. 782. p. 366, and 

 790-792, 796, p. 372), and that this covering developed into a complete calcareous 

 investment comparable to that of the echinoids; in the later types wherever and 

 whenever the latter is suppressed the former immediately reappears. 



Woodland states that in the disk of Antedon bifida, which he studied at 

 Plymouth, the majority of the spicules lie toward the inner side of the soft integu- 

 ment and are of two kinds, the imperf orate thin " glasplattchen " of comparatively 

 wide diameter and concentrically marked, and the ordinary perforated plates 

 which, however, often assume a very irregular shape and occasionally become mere 

 branching structures. All transitions are to be found between the " glasplattchen " 

 and the ordinary perforated plates; the former, however, are much more common 

 and exist in great numbers. The imperforate plates each originate as a spherical 

 granule contained within a single cell ; this granule gradually becomes larger and 

 early assumes the plate-like form. Thus there is no rod or triradiate stage in the 

 growth of these plates, and the plate is never of the perforate type. The nucleus 

 of the original mother scleroblast divides at an early stage of growth of the spicules, 

 and the number of nuclei present at different stages of growth (a score or more on 

 fully developed plates) is strictly proportionate to the size attained by the spicule. 

 These large, thin, squamous spicules of Antedon exist in great numbers. They are 

 very easily decalcified in virtue of their extreme thinness. The dark granules so 

 common in the scleroblasts of many Cucumariidse are also present in the sclero- 

 blasts of Antedon. 



P. H. Carpenter noted that the margins of the ambulacra of the disk, arms, 

 and pinnules of the comatulids are rarely, if ever, perfectly free from any traces 

 of calcareous structures, which may take the form of simple short spicules almost 

 entirely limited to the marginal leaflets, or may be forked and branching spicules 

 of various shapes and sizes. 



The spicules in the ventral integument of Comactinia echinoptera were long 

 ago figured by Carus. 



Present within the integument in all of the comatulid groups the perisomic 

 spicules are often, as in Antedon mediterranea and in the Macrophreata generally, 

 of small size and quite invisible exteriorly. When this is the case the species is 

 said to have an " unplated " disk. They may, however, become enlarged and stand 

 out from the integument as calcareous plates or nodules of various sizes and shapes, 



