When we dissect out one of these tube-feet, we find that at its base, within the body of the animal, it bifurcates into two branches. One of these branches passes immediately into a closed sac (Fig. 36, _f_), while the other passes into a large tube (Fig. 36, _k_), which runs all the way from base to tip of the ray, receiving in its course similar branches from all the tube-feet in the ray. This common or radial tube itself opens into a circular tube (Fig. 36, _e_) surrounding the mouth of the animal (Fig. 36, _m_). This circular tube therefore receives five radial tubes--one from each of the five rays--and is likewise in communication with a number of membraneous sacs (Fig. 36, _c_, _d_), resembling in their structure (though larger in size) those which occur at the base of each of the tube-feet. The function both of these sacs and of those at the base of each tube-foot is the same, namely, that of acting as reservoirs of the fluid when this is expelled from the tube-feet. Moreover, all these membraneous sacs are provided with ring-shaped muscular fibres in their membraneous walls, which therefore serve as antagonists to the ring-shaped muscles which occur in the membraneous walls of the tube-feet; that is to say, when the muscles of the reservoirs contract (Fig. 36, _c_, _d_, _f_), the pressure in the tube-feet is increased, and when these muscles relax, that pressure is diminished. The animal is thus furnished with the means of varying the head of pressure in its tube-feet, either locally or universally. The circular tube surrounding the mouth communicates at one point with a calcareous tube (Fig. 36, _a_), which runs straight to the dorsal surface of the animal, and there terminates in the madreporic tubercle, to which I have already directed attention (Fig. 32, _m_, and Fig. 36, _m_). Thus it will be seen that all the pedicels of all the rays are in communication, by means of a closed system of tubes, with this madreporic tubercle. It has therefore been surmised that the function of this tubercle is that of acting as a filter to the sea-water which in large part constitutes the fluid that fills the ambulacral system. We have been able to prove that this surmise is correct; for we found that if we injected any part of the ambulacral system with coloured fluid--maintaining the injection for several hours at as great a pressure as the tubes would stand without rupturing--the coloured fluid found its way up the calcareous tube to the madreporic tubercle, on arriving at which it slowly oozed through the porous substance of which that tubercle consists.