282 PROCEEDINGS OP THE ACADEMY OF [1885. 



Among the earlier groups of the Neocrinoidea, the perisome is 

 only known in the Jurassic Extracrinus and in Marsupites, in 

 both of which it consists of small irregular pieces, forming a 

 rather substantial pavement, which frequently encloses the lower 

 pinnules. Among Palseocrinoidea, a perisome has rarely been 

 observed, but it was preserved among several groups, and we 

 have been led to the conclusion that the perisome was probably 

 subtegminal in the Camarata, the Articulata, and the earlier 

 Inadunata, but external in the later Pistulata, our former Cyatho- 

 crinidte, which we make a branch of the Inadunata. 



The perisomic skeleton of the Camarata is frequently pre- 

 served in Baiocrinus, Eretmocrinus, Physetocrinns and Dory- 

 crinus, in which it probably attained a more substantial form 

 than in any of the other genera. It is there composed of rather 

 distinct plates, placed parallel to those of the test or tegmen 

 calicis, and arranged in a similar manner. Each plate of the 

 outer test has a corresponding plate in the ventral perisome ; 

 the plates of the latter, however, are not connected by suture, 

 but disposed loosely, leaving an open space at each angle, which 

 was perhaps tilled by membranous substance. In their usual 

 preservation the plates are frequently dense, owing to incrusta- 

 tions of inorganic matter, but in their natural state they consist 

 of a fine network, and resemble similar plates in recent Crinoids. 

 They form a sort of internal lining, wdiich extends from the 

 second primary radials and first intenadials, uninterruptedly, to 

 the central piece, or near it, underlying the proximals and enter- 

 ing the anal tube. This structure is well shown in the specimen 

 (PL 5, fig. 6), in which the sutures between central plate and 

 proximals are visible. Most of tin 1 perisomic plates, along their 

 median port ions, are connected with the plates of the test by 

 small pillars or partition walls, leaving between them open 

 chambers, evidently for the free circulation of water. The water 

 probably entered from without by means of the respiratory pores, 

 which we described in Part I, p. 11, and which jointly may have 

 performed the functions of a madreporite. There are, however, 

 no pillars between tin- plates along the radial regions, which take 

 the shape of closed galleries or corridors, formed by groove- 

 along the inner floor of the test, and closed from below by peri- 

 some. These passages diverge toward the arm bases and contain 

 the ambulacral tubes. The perisome of other genera was prob- 



