294 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



(in which there apparently are never ventral deposits o| any kind) , Calometridse, 

 ThalassometridfB, and CharitometridiB had a few spicules in the proximal portion : 

 17, in the Comasteridse, Zygometridpe, Himerometridse, and Mariametridse showed 

 two conspicuous bands, a distal and a proximal, becoming confluent in the outer 

 portion; 11, 9 in the Caloraetridai and 2 in the Charitometridse, had the tip pro- 

 tected by a more or less developed plate, and usually spicules in the outer portion ; 

 4, all belonging to the Comasteridse, had a few scattered spicules; 1, belonging 

 to the Comasteridse, had numerous scattered spicules in the proximal portion; 

 and 1, belonging to the Thalassometridse, had scattered spicules in the distal portion, 

 but no terminal deposit. 



SUBAMBCL-ICBAL PLATES. 



P. H. Carpenter noticed that deep ambulacral grooves with strongly plated 

 sides are met with on the disk of Comatula solans. In this species the disk may 

 be nearh' naked, or plated very completely. The whole of the large interambu- 

 lacral area in which the anal tube is situated is usually covered with more or less 

 scaly plates which become stouter and more granular in the neighborhood of the 

 subcentral anal tube. The sides of the deep ambulacral grooves are bounded by 

 numerous smaller plates without any definite arrangement; but they are .strictly 

 limited to the disk, not extending onto the arms. Much of this plating extends 

 beneath the water vessels of the disk, and a similar plating is found in a similar 

 position in the arms, taking the form of more or less regular plates, though with 

 no definite arrangement, being practically only a portion of the general calcareous 

 plating beneath the upper surface of the disk; this corresponds to what Miiller 

 called the subambulacral plating in Isocrinus. 



In Comatula rotalaria the large anal area is often beset with niunerous irregu- 

 lar plates many of which bear nodules of variable size. They are smaller on the 

 base and sides of the anal tube, and there are few or none in the small interambu- 

 lacral spaces between the edge of the disk and the marginal ambulacra, which are 

 themselves devoid of supj^orting plates. 



VISCERAL SKELETON". 



P, H. Carpenter employed the term " visceral skeleton " to denote the numerous 

 spicules and networks of calcareous matter which occur more or less plentifully in 

 the bands of connective tissue that traverse the visceral mass of the comatulids. 



He noted that the simple spicules and thin networks of calcareous matter which 

 occur in the less heavily plated disks are especially abundant in the visceral layer 

 of the peritoneum. This is well seen in those crinoids in which there is but little 

 connection between this visceral layer and the parietal layer lining the interior of 

 the cup, so that the visceral mass is readily detached from the calyx. Unlike that 

 which lines the cup, the oral perisome is usually very closely adherent to the visceral 

 mass and can not be separated from it without some trouble. The peritoneal 

 covering of the latter also contains limy deposits, so that sections through the upper 

 part of the disk show two layers of plates and spicules. The upper one is in the 

 perisome itself and belongs to the anambulacral system, while the lower belongs to 

 the upper surface of the visceral mass. All of the recent crinoids have more or less 



