1886.] NATURAL SCIENCES OF PHILADELPHIA. 155 



of difference about which we are enabled to speak with certainty." 

 This character, however, is not restricted to Encrinus; we find it 

 more or less among all Poteriocrinidffi, but exclusively there and 

 in the Encrinidae. The two brachials are alwa3'S joined by suture, 

 not by articulation, and this is the case in Encrinus, in which the 

 two pieces are united by s^'zj^gy and form actually one plate. 

 The compound brachial is scarcely more than a specific character, 

 and hence has no value as a distinction between two orders. 



Carpenter takes not only Encrinus, but also Erisocrinus and 

 Stemmatocrinus to be Neocrinoidea, because they have no anal 

 plate. Anal plates, he states, " are absent in Erisocrinus as in 

 Encrinus, and since a ventral tube or sac like that of Cyatho- 

 crinus is always found associated with a system of anal plates, 

 the lowest of which is intercalated between two radials, it seems 

 rash to postulate its presence in the symmetrical Erisocrinus.''^ 

 That our conclusions were not so " rash " as Carpenter suggested 

 is emphatically proved by the fact that an anal plate has actually 

 been discovered by Dr. White in Erisocrinus. It is located above 

 the radials, being enclosed among the interradial plates, where we 

 should have expected it from analogy with other groups. Anal 

 plates are absent also in two genera of the Cyathocrinidse, in 

 Godiacrinus Schultze, and Lecythiocrinus White, which both 

 have a perfectly symmetrical dorsal cup. But does that make 

 them Neocrinoidea ? What more do we know of the ventral struc- 

 ture of the Poteriocrinidse than of Encrinus and Erisocrinus? 

 If Encrinus is a Neocrinoid, why not all Poteriocrinidae ? No- 

 body ever saw the ventral covering in any Poteriocrinoid we 

 have only found the ventral sac. Yet from analogy with Gyatho- 

 crinus and allied forms in which a vault has actually been ob- 

 served, it was generally supposed that vault plates were present 

 also in the Poteriocrinidse. We think the width of the brachials, 

 which in this family frequently occupy the entire width of the 

 radials, gives a satisfactory explanation why the vault plates are 

 not preserved in forms like these. The plates either rested against 

 the sloping edges of the radials as in Stemmatocrinus Traut- 

 scholdi, or only against their articular extensions as in Symbatho- 

 crinus. It is even very possible that the interradials were par- 

 tially or wholly resorbed by the muscular processes of the 

 radials, as these became developed in the growing Crinoid. 



The subdivisions of Poteriocrinus, which we proposed, have 



