550 COMPARATIVE ANATOMY CHAP. 



cannot be regarded as primitive forms of Holoth\n-iut<l<>. They are, on the contrary, 

 highly specialised forms, which, in adaptation to the limicolons life, have lost the 

 tube-feet and the radial canals, which, however, still occur ontogenetically. 



We have thus seen how the Pentacta>a might become modified in the direction 

 of certain Cystids and of the Ho/ntlun-ioidea. 



By remaining attached, the Pentactrea might develop in another direction. 



The body carried by the stalk might remain small, but become drawn out into 

 processes or arms in the directions in which the primary tentacles travel from the 

 mouth, the ends of these arms being always marked by the possession of the primary 

 tentacles. Secondary tentacles then rose out of the radial vessels which ran along 

 the arms (the tentacle canals of the primary tentacles) in the way above described ; 

 the food grooves and their nerve ridges also formed in the same way. A still more 

 complete adaptation to the attached manner of life was attained by the branching of 

 the arms and the formation of pinnula?. In this way the surface for capturing food 

 was continually increased. 



The direction of adaptation here indicated might be called the Crinoid direction, 

 the Crinoids having, in fact, gone furthest in this direction. 



The development of the crown of anus on a body which remained small had 

 necessary consequences. The body (calyx, disc) and the stalk (should this latter 

 develop) would have to gain the necessary stability for carrying the growing arms. 

 This was provided for by the formation of the more or less firm carapace of plates. 

 The stalk attained firmness by the development of joints ; the calyx, by that of the 

 dorsal cup, and here all the facts seem to indicate that in the racial form of 

 the Crinoids, the dorsal cup had a definite composition, viz. five infrabasals, five 

 basals, five radials arranged in the typical manner, and the anals. For the 

 protection of the mouth, five orals were added, forming together a pyramid which 

 could be opened and closed. For the support of the arms, and in connection with 

 the developing capacity for unfolding and closing the crown of arms, the jointed 

 brachial skeleton formed. 



As the arms grew out from the small body, the co-loin was produced into them, 

 and processes of the single rudiment of the gonad (the axial organ) spread in one 

 way or another into them. They became fertile more or less far from the calyx (or 

 disc) and yielded the gonadial bundles, each of which opened outwards through one 

 or more special apertures. 



In this point also, the Crinoids are the most extreme forms. 



The Echiiioidea, Ophiuroidea, and Asteroidea appear also to belong as lateral 

 branches to this Crinoid development. 



First, and probably very early, the Echinoids seem to have branched off. They 

 became free, used their tentacles for locomotion, and took in food direct through the 

 mouth, the food grooves, with the nerve ridges, becoming the subepithelial radial 

 nerves. The arms were again incorporated into the enlarging calyx, or test ; in that 

 the apical skeleton of the arms degenerated, and thus brought the (ambulacral) ends 

 of the arms close up to the continually decreasing apical capsule. As this latter 

 was free, the anus could shift into its centre. 



"We have come to hold this view of the derivation of the Echinoidea from attached 

 ancestral forms with arms, a view which, as far as we know, has never before been 

 published, chiefly for the following reasons : 



The Echinoidea possess five pairs of gonads, which are at first connected with the 

 axial organ by means of an aboral circular strand. 



This important distinction from the Holothurioidea, with which the Echinoidea 

 are usually compared in other points, can only be explained by the assumption that 

 the Echinoidea originally possessed arms which contained the fertile outgrowths of 

 the central genital rudiment. That the gonads now lie interradially, presents no 



