MORPHOLOGICAL PART. 133 



which the brachials fill up the greater part^ or all of the distal face of the 

 radials^ it required additional surfaces for the accommodation of the large 

 tube. 



From the structure of the typical Poteriocrinidse we come to that of 

 ZJlocriimSy Graphiocriniis^ Ceriocrinus, Erisocriniis, and Stemmatocrinus^ which 

 we regard as transition forms toward Encriniis. The ventral tube, which in 

 the latter of these forms dwindled to a short cone, did not require as large a 

 support in the dorsal cup, and as the anal plates gradually became obsolete, 

 the posterior radials resumed a symmetrical form. In Ulocrinus (Fig. 17) 

 the plate x was crowded out by the large radianal ; w^hile in GrapMocmiiis 

 and Ceriocriniis (Fig. 16) only the former is represented. In Graphiocrinus, 

 with a wide ventral sac, the anal plate is large ; in Ceriocrinus it is reduced 

 to a small piece, and the posterior basal is considerably elongated. Eriso- 

 criniis (Fig. 18), Stemmatocrinus, and Encrinus have no anal plates at all, the 

 cup being perfectly symmetrical; and the tube rests entirely upon the 

 edges of the radials, whence it started in Hyhocrinns, Ectenocriniis, and 

 allied forms. 



, In view of these facts, it seems to us that Bather's theory of a " brachi- 

 anal'' is based upon a wrong interpretation of some of the plates. If it 

 were true that the plate of locrimis to the left of the supraradial passed 

 down in later forms to the basals, it would mean nothing less than a partial 

 revolution of the entire tube. This, however, is disproved by the structure as 

 well as the pal^ontological development of the tube, which latter is generally 

 composed of longitudinal rows of hexangular pieces, alternating in adjoining 

 rows. In the earlier and simpler forms the tube consists of only five series, 

 one to each interradius, that of the anal side resting upon plate t. Later on, 

 as the tube grew larger, a new row of plates was introduced with plate x sup- 

 porting it. When there are three series, as in Dendrocrinus, the third generally 

 rests upon one side of the left posterior radial. The arrangement of the 

 plates within the rows is so regular that if a sinking of the plate t had taken 

 place, it would certainly be indicated by some disturbance among the lower 

 plates in the tube. In species where the tube has more than three rows, one 

 or more of the primarj^ rows dichotomize at some distance from the cup. 

 The fact that the increase of the tube phylogenetically took place by the 

 introduction of new rows of plates, is a strong argument in favor of our idea 

 that the plate x is also a supplementary piece, and was introduced in the 

 same manner as the plates which it supports. 



