NEW RESEARCHES 87 



when fixed paleontologically become the strongest characters for the delimita- 

 tion of genera. That in some of the fossil stages the radianal plate became 

 fixed in its primitive or early position, and did not follow the distal end of the 

 gut in its growth, might on first thought seem a singular fact, but it is no more 

 so than the fact that the two structures part company when the radianal is left 

 behind in the larva. 



Recalling now the descriptions of the origin of the radianal, or the so- 

 called anal plate, in Antedon and Comactinia as given in the foregoing pages, 

 which I have called the "Antedon stage " and the "Actinometra stage," re- 

 spectively, it will be observed that we have in these two types stages of anal 

 development comparable to those of important contrasting fossil groups both 

 in the Flexibilia and the Inadunata: — the Antedon stage to that of the median 

 anal series of the Cyathocrinidae and some of the Silurian and Carboniferous 

 Flexibilia, and the Actinometra stage to that of the more primitive Dendrocri- 

 nus, Protaxocrinns and Lecanocrinns groups. Apparently the Antedon stage 

 is characteristic of a considerable group among the Recent crinoids, since we 

 have proof of it in at least two widely different species. The Actinometra 

 stage is as yet strictly known in only one species, the development of Promach- 

 ocrinus (which is classed among the Antedonidae) being a little different in 

 detail from that of the others. 



The relations of the medially situated anal plate in Antedon are by no 

 means clear, and there is strong reason for doubting if it can be placed in the 

 same category as the migrating plate in Comactinia. Its first appearance is in 

 an interradial position, where by growth from above it truncates the posterior 

 basal distally. Its position and mode of growth are not (except for its evident 

 connection with the movement of the gut) essentially different from those of 

 the interradials in Comactinia and Pvomachocrinns, or from those of the inter- 

 radials in fossil forms like the Reteocrinidae or Rhodocrinidae, or some excep- 

 tional Flexibilia, in which they extend downward to a connection with the 

 basals. The radianal of Comactinia, on the other hand, is of a strictly radial 

 origin, and by shifting toward the left and upward it truncates the posterior 

 basal first laterally and afterward distally. Therefore we may have in these 

 two existing types of anal structures representatives of the two principal types 

 in the fossil forms, viz., that with median plates only, and that with asym- 

 metric radianal. 



Generalizations upon these conditions, or attempts to point out definite 

 homologies, must be made with much reserve. Nevertheless, according to the 

 foregoing facts, we have in the ontogeny of this living crinoid an unusually 

 close recapitulation of the phylogenetic history of some of the Paleozoic groups 

 of the class. 



