280 BULLEnN 82, UNITED STATES NATIONAL MUSEUM. 



the terminal clinv in its general relationships than to the doreal spines proper; it 

 is, however, intermediate between them. It is ])resent and well developed in many 

 species (as in all those of the Comactiniina') in wliich no dorsal spines are ever 

 developed (figs. 394, 395, 397, 398, 400, 401, p. 309), and it exliibits the perfected 

 acutely conical shape and erect median position in many cases where the processes 

 on the preceding segments are as yet in the primitive stage of a broad serrate 

 transvci-se ridge. In structure it is more dense than the dorsal spines, and it 

 possesses a tliick vitreous cortical layer of condensed stereom resembling that on the 

 terminal claw, though never quite so weU developed. In young and in regenerating 

 cirri it is verj' early in making its appearance, being well developed before the cirrus 

 segments have lost their original short cylindrical form. 



The opposing spine may make but a sUght angle with the median axis of the 

 'penultimate segment (fig. 314, p. 273), or it may be quite erect and at right angles 

 to that axis (fig. 317, p. 273). The proportion of declination is correlated with its 

 position; if it is tcrmally situated it makes the minimum angle with the median axis; 

 it is not erect unless its position is at the center of the dorsal surface of the penulti- 

 mate segment (fig. 352, p. 291). The degree of declination is in direct inverse 

 ratio to its distance from the distal edge. 



The opposing spine functions as a hUt for the sharp and dagger-like terminal 

 claw, preventing the cirrus from sinking too deeply into, and thereby becoming hope- 

 lessly entangled with, the substance of the organism to which the crinoid is clinging. 



Typically the opposing spine reaches a height about equal to the distal trans- 

 veree diameter of the penultimate segment, though it is often less, especiaUj' in those 

 species in which the preceding segments bear no dorsal processes; in the oligo- 

 phreate species it is commonly triangular, arising from the entire dorsal surface of 

 the penultimate segment, thus being considerabh- broader basally, and also longer, 

 than the processes on the preceding segments (fig. 318, p. 273); in the macrophreate 

 forms, as well as in certain of the oligophreate, however, the base is usually shorter, 

 and the spine arises from the outer part only of the penultimate segment (figs. 395, 

 396, p. 309); this is always the case if dorsal processes are not developed on the 

 preceding segments. 



Tlie origin of the dorsal spines and of the opposing spine was probably some- 

 what as follo\vB: The central canal through the cirrals is at first central in position; 

 after the middle of the cirrus it moves slowly and gradually ventralward (fig. 587, 

 pi. 13). This results in a difference in size between the doi-sal and the ventral hga- 

 ment bundles by wliich the cirrals are articulated (the two sets at first being similar 

 and equal), the former becoming progressively larger and stronger and the lalter 

 correlatively smaller and weaker. In consequence of the normal state of balanced 

 tension of the ligament fibers the cirri assume a curved shape, the curve being very 

 gradual at firet , but increasing toward the 1 ip, the radiusof curvat ure being evciywhere 

 proportionate to the difference in strength between the dorsal and the ventral liga- 

 ment bundles. The calcareous elements of wliich the cirrals are composed are 

 deposited as rings or cyhnders within the sarcode of the growing cirri; normally 

 they increase in length by the addition of calcareous matter equally all around 

 their margins; where the ligament bundles are equally balanced tliis occurs, but 



