1890.] NATURAL SCIENCES OF PHILADELPHIA, 355 



these were evidently preceded by lower forms, which have so far 

 escaped notice, owing perhaps to their diminutive size. We think 

 this the more probable as the sjjecies throughout this group are ex- 

 ceedingly small, some of them almost microscopic. 



From Huploerinus to Hybocrinus and Hetei-ocrinus there is but a 

 small step. All three have compound radials, and in the two latter 

 the ventral sac, which represents the best character of the Fistulata, 

 is in its first stage of development. The transitions from Symbutno- 

 crinus through the Hexacrinidae to the Actinocrinidae are equally 

 gradual, in the latter the orals being carried inward by sup])lement- 

 ary plates interposed between radials and orals. tSijmbafJiocrinus 

 has an anal plate, and its orals had already that asymmetric ar- 

 rangement — the posterior one larger and interlocking with tlie four 

 others — which they retain throughout the Camerata. The introduc- 

 tion of a single interradial plate between the proximal brachials 

 transforms Allagecrlnus into either Coccoerinus or Culicocrinu.-^. 



We formerly arranged among the Larviformia also the Gastero- 

 comidae, but these will have to be removed to the Fistulata or be 

 placed in a new subgroup. We now think that the Larviformia 

 should be restricted to forms in which the radials are directly 

 followed by the orals. 



The Inadunata Fistulata, like the Inadunata Larviformia, have 

 no interradials in the dorsal cup, the anal piece excepted, but all 

 have interambulacral plates. Four of the interradial spaces of the 

 dome are raised but little above the level of the arm bases, but the 

 posterior area is extended upwards, and formed into a sac or tube 

 of various forms, frequently rising beyond the tips of the arms. This 

 appendage, which in all probability embraced' a large portion of the 

 visceral cavity, must not be confounded with the anal tube of the 

 Camerata, which simply contains the rectum. The ventral sac is 

 generally the only part of the ventral pavement preserved in the 

 specimens, and even this in most cases but fragmentarily. It is 

 either tubular, balloon-shaped, spiral or club shaped, and is generally 

 pierced by pores, which, however, do not penetrate the body of the 

 plates but merely enter the edges. The structure of the four other 

 sides is satisfactorily known only in the Cyathocrinidae and Anom- 

 alocrinidae, and fragmentarily in the Hybocrinidae, Poteriocrinidae 

 and Eacrinidae, enough to indicate that our former definition, giv- 

 ing to all Fistulata but one rinar of interradials, was erroneous. 



