258 PHYLUM ECHINODERAfA. 



intestine between the stomach and the almost central dorsal 

 anus two little outgrowths are given off, perhaps homologous 

 with the "respiratory trees" of Holothurians (Fig. 139, r.t.). 

 Some parts of the food canal are ciliated. 



The ccelom is distinct, though not much of it is left 

 unoccupied either in the disc or in the arms. It is lined 

 by ciliated epithelium, and contains a fluid with amoeboid 

 cells. A few of these have a pigment which probably aids 

 in respiration ; others are phagocytes, which get rid of 

 injurious particles through the "skin-gills"; others con- 

 tinue the work of digestion. 



When a starfish is crawling up the side of a rock, scores 

 ot tube-feet are protruded from the ventral groove of each 

 arm; these become long and tense, and their sucker-like 

 terminal discs are pressed against the hard surface. There 

 they are fixed, and towards them the starfish is gently 

 lifted. The protrusion is effected by the internal injection 

 of fluid into the tube-feet ; the fixing is due to the pro- 

 duction of a vacuum between the ends of the tube-feet 

 and the rock. 



As 10 the course of the fluid, it is convenient to begin with the 

 madreporic plate, which lies between the bases of two of the arms (the 

 biviuni]. This plate is a complex calcareous sieve, with numerous 

 perforating canals and external pores. It may be compared to the rose 

 of a watering-can, but the holes are much more numerous, and lead 

 into small canals, which converge into a main ciliated canal, the stone 

 canal. This, as usual, opens into a ring canal around the mouth. 

 The ring canal gives off nine glandular bodies (Tiedemann's bodies), 

 and five radial tubes, one for each of the arms. Considerations of 

 symmetry suggest that there should be ten glandular bodies, but in the 

 inter-radius containing the stone canal there is only one. In many 

 starfishes there are five or ten little reservoirs (Polian vesicles) opening 

 into the circumoral ring, but in Asterias rubens these are hardly dis- 

 tinguishable from the first ampullse of the radial vessels. These run 

 along the arms, and lie in the ambulacral groove beneath the shelter 

 of the rafter-like ossicles. From them branches are given off to the 

 bases of the tube-feet, but from each of these bases a canal ascends 

 between each pair of ambulacral ossicles, and expands into an ampulla 

 or reservoir on the dorsal or more internal side (see Fig. 134). The 

 fluid in the system may pass from the radial vessels into the tube-feet, 

 and from the tube-feet it can flow back, not into the radial vessel, but 

 into the ampullse. There are muscles on the walls of the tube-feet, 

 ampullse, and vessels. At the end of each arm there is a long unpaired 

 tube-foot, which seems to act as a tactile tentacle, and has also olfactory 

 significance. 



