204 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



more complicated than the rough and general sketch just presented might lead 

 one to suppose; but it has certain ontogenetical and phylogenetical facts in its favor, 

 and does not involve the formation of two separate plates by simple post-larval 

 division of a primitive single plate a process the existence of which is as yet un- 

 proven among the echinoderms. 



A crinoid attached by the adherence of the central plate to some solid object 

 upon the sea floor would be subjected to a certain amount of strain from wave 

 motion, or from the unequal movements of its own arms, as well as from the pas- 

 sage of other organisms. This strain would be felt along the suture connecting 

 this central plate with the basals (or the infrabasals), and along the interbasal 

 sutures. There are two ways of meeting this condition: (1) the basals (or the in- 

 frabasals or both) may become more upright and more nearly parallel with each 

 other and fuse solidly with the central plate (now become a tlu'ck stalk); this has 

 occurred in Holopus; (2) a second central plate, exactly similar to the original 

 one, may be formed within that is, ventral to it, leaving a ligamentous articula- 

 tion between them by which the strain is taken up, and this process may be con- 

 tinued indefinitely until a long articulated stalk is formed. 



A column formed by this process would of necessity be composed of very 

 numerous and very short columnals, for the columnals would be attached to each 

 other not by true articulations but by loose sutures; the amount of possible accom- 

 modation at a loose suture is far less than that at a true articulation, in which an 

 articular fulcral ridge is developed and the ligament fibers have become segre- 

 gated into two bundles one on either side of it, and therefore many such loose su- 

 tures must be developed in a given length of column to do the work of a single 

 articulation. 



This explanation derives the crinoid stem from the original central plate equally 

 well with the first, while at the same time it indicates the formation of the col- 

 umnals from their first inception by a continuous twinning or reduplicative process, 

 each columnal being formed by an original ossification of the same type and in the 

 same place, thus making each individual columnal, as well as the entire column, 

 the equivalent of a single calyx plate a state of affairs which, so far as we can 

 see, is probably very near the truth. 



From what we know of the formation and development of the columnals in 

 the recent crinoids it would appear that they are derived from an apical calyx plate 

 after the phylogenetical formation and fixation of that plate, in other words, 

 after the perfection of the skeletal investment of the calyx. Moreover it is only 

 by such a supposition that we are able to bring into phylogenetical agreement and 

 to reduce to a common and logical starting point such diverse apical conditions 

 as are found in Marsupites and Uintacrinus, Eolapus, the pentacrinites and the 

 comatulids. 



But there is another possibility which, however remote, should not be over- 

 looked. The apical area of the crinoidal ancestor may have been merely a pliable 

 integument filled with primitive spicules and dissociated plates, as we see it in 

 the earlier and many of the later echinoids, the sum total of which is the equiva- 

 lent of the apical plate, later formed or assembled. The animal may have become 



