1890.] NATURAL SCIENCES OF PHILADELPHIA. 367 



were exposed, are rooted over by an extension of tlie interainbida- 

 cral plates of .subsequent growth. This incidental observation is 

 very important, as it throws niucli light upon the development of the 

 so-called vault of the Cainerata generally,. It shows that the same 

 system of plates, which in a young specimen is t7(^e?--ambulacral only, 

 may gradually become s«^;?'a-ambulacral in another. 



A somewhat different structure we observed in a finely preserved 

 adult specimen oi Megistocrinus evansi, in which in three of its i*ays 

 two series of large, nodose alternating plates pass out from near the 

 orals in the direction of the ambulacra. The series are frequently 

 interrupted by small, flat pieces, passing out from the interambu- 

 lacral spaces, and these intermingle with the larger ones. At some 

 places the arrangement of the lai-ger plates, which evidently are 

 covering pieces, is as regular as in any Platycrinus, but at others 

 owing to the interference of the smaller plates, quite irregular, especi- 

 ally in the two rays to the right of the anus, in which scarcely any 

 two of these plates are continuous. It is most remarkable that in no 

 two specimens of this species is the arrangement of the covering 

 pieces alike, and in some of them only the five large bifurcating 

 plates, the so-called radial dome pieces, ai-e in view, which are 

 followed by ten others of a second order ; these, however, are always 

 represented. The ventral structure of this species offers not only a 

 most excellent proof that the so-called radial dome plates, as suggested 

 before, are extravagantly developed covering pieces, but indicates 

 also that the "vault" was formed by a gradual extension of the 

 interambulacral pieces toward the ambulacra, either covering the 

 latter entirely, or intruding upon them, leaving the more])rominent 

 plates exposed. 



In Glyptocrinus and Reteocrinns the tegmen is essentially in the 

 same condition as in MeglMoeriims nobilis, but the plates are smaller, 

 and in the absence of orals nearly the whole surface is covered by 

 small granular plates, including the central portions. Toward the 

 periphery there are ridges leading to the arm bases, but these are 

 not found at the middle part, where the surface is evenly convex. 

 Only on top of these ridges, that is, in close proximity to the arm 

 bases, are the covering pieces exposed. They are quite regularly 

 arranged in two rows, and so well defined that we doubt if it is pos- 

 sible that they were continued further up along the surface, as 

 suggested to us by Dr. Carpenter, or they would be recognized in the 

 specimen. It seems to us more probable, and almost certain, that 



