THE CYSTWEA 41 



in which neither ciliated food-grooves, though perhaps present, 

 nor radial ambulacral vessels, have left any trace on the skeleton, 

 in which the porous structure of the stereom is indefinite, and in 

 which no stem is differentiated. Modifications of this soon 

 appeared in many directions. In one direction arose an antero- 

 posterior flattening of the theca and the extension of food-grooves 

 along two lateral articulated spines, with a peculiar and character- 

 istic arrangement of stereom ; this was accompanied by develop- 

 ment of a stem (Anomalocystidae, pp. 49, 52, Figs. XI.-XIIL). In 

 another direction was an extension of the theca downwards to 

 form a stem, and upwards from the mouth to form a single jointed 

 process for the support of a ciliated groove (Dendrocystidae, 

 p. 47, Fig. IX.). Neither Anomalocystidae nor Dendrocystidae 

 proceeded very far, and they may conveniently be grouped with 

 Aristocystidae and a few other primitive forms into an order, 

 AMPHORIDEA, distinguished from the re^st chiefly by absence of 

 radial symmetry in food-grooves and ambulacra. 



A very different modification was that which produced a theca 

 flattened horizontally, with five ciliated grooves passing from the 

 mouth between its plates (" endothecal 1 "), and protected by distinct 

 covering-plates ; ambulacral vessels lay beneath or within the 

 grooves, and podia from them passed between the adjacent thecal 

 plates. So different is this type from those of other Echinoderma, 

 that such forms have here been separated as a class, EDRIO- 

 ASTEROIDEA (Chapter XII.). 



Returning to the primitive Amphoridea, we find a difficulty in 

 distinguishing some of them from their immediate descendants, 

 owing to the very slight traces left on the theca by the originating 

 extensions of the food-grooves. Those forms in which such traces 

 are perceptible may almost from the outset be grouped under two 

 heads. One group includes those in which the grooves wander 

 outwards from the mouth over the thecal plates, which gradually 

 become arranged regularly on either side of the grooves, while still 

 further extensions ascend from the " epithecal " grooves on small 

 "exothecal" processes called " brachioles." In the other group 

 the grooves do not tend so much to stretch over the theca as to be 

 raised away from it on relatively larger brachioles, arising in 

 the immediate neighbourhood of the mouth. 



At the same time, a difference manifests itself in the structure 

 qf the thecal plates. From the indefinite relations of stereom and 

 stroma noticed in earlier Amphoridea arise two types of structure 

 (Fig. I.). The canals traversing the stereom, more or less per- 

 pendicularly to the thecal surface, either cease to be simple 

 (" haplopores ") and become connected in pairs (" diplopores ") still 

 perpendicular to the surface ; or they come to lie parallel to the sur- 

 face and at right angles to the sutures. In the latter case we may 



