1890.] NATURAL SCIENCKS OF PHILADELPHIA. 379 



different geuera. In Heterocrinus, Ohiocrinus, locrinns, Merocrmus, 

 Hybocrinus and Hybocystis, the lower portions are considerably larger 

 than the upper; in A)iomaloermus and Dendrocrlnits the two are of 

 nearly equal size; while in Edenocrinus (^Heteroermus) simplex the 

 ujjper ones are three to four times as large as the lower. 



Now, what does this gradual increase of the upper portions, and 

 the disappearance of compound plates in other rays than the poster- 

 ior, tell us? Does it not indicate that in these Crinoids there is a 

 gradual development from three to one compound plate, and from 

 the compound to the simple radial? It certainly looks like it. The 

 idea is further confirmed by the fact that true compound radials 

 do not exist among any of the later Fistulata, in which, as we shall 

 presently show, the lower portion of the right posterior radial, more 

 or less, serves as an anal plate. But what preceded the compound 

 radials? What else, but forms in which the arm-bearing portion was 

 still smaller, and, i)rimarily, those in which it was absent altogether, 

 that is to say, was as yet undeveloped. This, indeed, seems to have 

 been the structure of Baerocrinus. As we understand this genus, 

 it has three well developed radials, and two are in a transition state, 

 only their lower or non-arm-bearing segments being represented. 

 This is further corroborated by the fact that the non-arm-bearing 

 radials occur in the same rays as the compound radials of Anomalo- 

 crinus, for if we take the third plate of our diagram (PI. X, fig. 6) 

 to represent the posterior ray, the other non-arm-bearing plate must 

 represent the antero-lateral one. This would- make Baerocrinus the 

 ancestral form, lower in its development than either Anomalocrinvs, 

 Hophcrinns, or locrinus. 



We now pass to the genera in which the lower segment of the right 

 posterior radial serves as an anal plate. In all these forms, as may 

 be seen in Bather's diagrams, the four other radials are simple, and 

 in all of them the anal plate (X) is represented, and invariably rests 

 npon the basals. The size of the ventral sac had rapidly increased 

 at the close of the Lower Silurian, and the sloping upper faces of 

 the radials were insufficient to support it. This required certain 

 modifications in the structure of the dorsal cup. The posterior radi- 

 als, which theretofore had been in contact laterally, now separated, 

 the posterior basal increased its width, and the plate X was intro- 

 duced to fill the space between the radials. While these modifica- 

 tions were going on, the radianal, or lower segment of the radial, 

 retained its position between the upper sloping faces of two basals, 



