side being the one chosen for a reason we cannot now fathom; it 
Fig. 8.—Aristocystis bohemicus; side-view of the theca. The internal structure may be gathered from fig. 7. may be that fixation was facilitated by the presence of the pore on that side, and by the utilization of the excretion from it as a cement. The first result was that which is always seen to follow in such cases—the passage of the mouth towards the upper surface (fig. 7). As it passed up along the left side, the gut caught hold of the left water-sac and pulled it upwards, curving it in the process; this being attached to the left duct from the anterior body-cavity, this structure with its water-pore was also pulled up, and the pore came to lie between mouth and anus. The forward portion of the anterior coelom shared in the constriction and elongation of the preoral lobe; but its hinder portion was dragged up along with the water-pore and formed a canal lying along the outer wall (the parietal canal). As the gut coiled, it pressed inwards the middle of the left posterior coelom of the Dipleurula, and drew the whole towards the mouth, while the corresponding cavity on the right was pressed down by the stomach towards the fixed end of the animal and became involved in the elongation of that region. These changes, which may still be traced in the development of Antedon, resulted in the primitive Pelmatozoön (fig. 7), represented in the rocks by such a genus as Aristocystis (fig. 8). The pear-shaped body is encased in a theca formed by a number of polygonal plates, and is attached by its narrow end. On the broad upper surface are four openings, that nearest the centre being the mouth, 
Fig. 9.—Fungocystis rarissima, one of the Diploporita, in which the thecal plates bordering the food-grooves are not yet regularly arranged. The brachioles are not drawn.which is slit-like, and that nearest the periphery being the anus. The two other openings are minute, and placed between those two; one close to the mouth is almost certainly the water-pore, while that nearer the anus is regarded as a genital aperture. Which of the coelomic cavities this last is connected with is uncertain, for there is considerable doubt as to the origin of the genital glands in the embryonic development of recent echinoderms. It seems clear, however, that there was but a single duct and a single bunch of reproductive cells, as in the holothurians, though perhaps bifurcate, as in some of those animals. The line between mouth and anus, along which these openings are situate, corresponds with the plane of union between the two horns of the curved left posterior coelom, the united walls of which form the “dorsal mesentery.” Since this must have, on our theory, enclosed the parietal canal from the anterior coelom, it is possible that the genital products were developed from the lining cells of that cavity, and that the genital pore was nothing but its original pore not yet united with that from the water-sac. The concrescence of these pores can be traced in other cystids; but as the genital organs became affected by radial symmetry the original function of the duct was lost, and the reproductive elements escaped to the exterior in another way. Aristocystis may have had ciliated food-grooves leading to its mouth, but these have left no traces on the structure of the test. Traces, however, are perceptible in genera believed to be descended from such a simple type, and the majority may 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 further extensions ascend from the grooves on small jointed processes called “brachioles” (fig. 9). 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 close around the mouth (fig. 10).
These two types are, in the main, correlated with two gradual differentiations in the minute structure of the thecal plates. Originally the calcareous substance of the plates (stereom) was pierced by irregular canals, more or less vertical, and containing strands of the soft tissue (stroma) that deposited the stereom, as well as spaces filled with fluid. In the former group (fig. 9) these canals became connected in pairs (diplopores) still perpendicular to the surface, 
Fig. 10.—Chirocrinus alter, one of the Rhombifera, showing the reduced number and regular arrangement of the thecal plates, and the concentration of the brachioles. (Adapted from Jaekel.)and this structure, combined with that of the grooves, characterizes the order—Diploporita. In the latter group (fig. 10) the canals, that is to say, the stroma-strands, came to lie parallel to the surface and to cross the sutures between the plates, which were thus more flexibly and more strongly united: since the canals crossing each suture naturally occupy a rhombic area, the order is called Rhombifera. At first the grooves were three, one proceeding from each end of the mouth-slit, and the third in a direction opposed to the anus; with reference to the Pelmatozoan structure, the anal side may be termed posterior, and this groove anterior. Eventually each lateral groove forked, so that there were five grooves. These gradually impressed themselves on the theca and influenced the arrangement of the internal organs: it is fairly safe to assume that nerves, blood-vessels and branches from the water-sac stretched out along with these grooves, each system starting from a ring around the gullet. At last a quinqueradiate symmetry influenced the plates of the theca, partly through the development of a plate at the end of each groove (terminal), partly through plates at the aboral pole of the theca (basals and infrabasals) arising in response to mechanical pressure, but soon intimately connected with the cords of an aboral nervous system. Before the latter plates arose, the stem had developed by the elongation and constriction of the fixed end of the theca, the gradual regularization of the plates involved, and their coalescence into rings. The crinoid type was differentiated by the extension of the food-grooves and associated organs along radial outgrowths from the theca itself. These constituted the arms (brachia), and five definite radial plates of the theca were specialized for their support. These radials may be homologous with the terminals already mentioned, but this is neither necessary nor certain. In this development of brachial extensions of the theca the genital organs were involved, and their ripe products formed at the ends of the brachia or in the branches therefrom. The remains of the original genital gland within the theca became the “axial organ” surrounded by the “axial sinus” derived from the anterior coelom, and this again by structures derived from the right posterior coelom, which, as explained above, had been depressed to the aboral pole. These last structures formed a nervous sheath around the axial sinus with its blood-vessels, and became divided into five lobes correlated with the five basals (the “chambered organ”) and forming the aboral nerve-centre. Before these changes were complete the Holothurioidea must have diverged, by the assumption of a crawling existence. Thus in them the mouth and anus reverted to opposite poles, and only the torsion of the gut and coelom, and the radial extensions of the nervous, water-vascular and blood-vascular systems, testified to their Pelmatozoan ancestry. The ciliated grooves, no longer needed for the collection of food, closed over, and are still traceable as ciliated canals overlying the radial nerves. At the same time the thecal plates degenerated into spicules. The Edrioasteroidea followed a different line from that of the cystids above mentioned and their descendants. The theca became sessile, and in its later developments much flattened (fig. 11). Mouth, water-pore and anus remained as in Aristocystis, but the five ciliated grooves radiated from the mouth between the thecal plates rather than over them, and were, as usual, protected by covering-plates. The important feature was the extension of radial canals from the water-sac along these grooves, with branches passing between the flooring-plates of the grooves (fig. 12, A). The resemblance of the flooring-plates to the ambulacral ossicles of a starfish is so exact that one can explain it only by supposing similar relations of the water-canals and their branches (podia). On the thinly plated under surface of well-preserved specimens of Edrioaster are seen five interradial swellings (fig. 11, B). These are likely to have been produced by the ripe genital glands, which may have extruded their products directly through the membranous integument of the under side. No other way out for them is apparent, and it is clear that Edrioaster was not permanently and solidly fixed to the sea-floor.
Now comes a great change, unfortunately difficult to follow whether in the fossils or in the modern embryos. We suppose some such form as Edrioaster, which appears to have lived near the shore, to have been repeatedly overturned by waves. Those that were able to accommodate themselves to this topsy-turvy
