54 REPORT ON THE PENNATULIDA. 



The rachis is widest at its lower end, where the polype leaves are 

 either absent or very rudimentary (Figs. 1 and 6). As we pass upwards 

 and the leaves get bigger, the rachis at first diminishes in width some- 

 what rapidly (Fig. 1), but having attained a diameter of about 0-045in. 

 it preserves this tolerably uniformly along the greater part of its 

 length, tapering again gradually towards the upper end. It is traversed 

 throughout its length by four main longitudinal canals (Figs. 5 and 6 u), 

 one of which is dorsal, one ventral, and two lateral ; these canals, as 

 noticed above, extending down into the stalk. 



The outer surface of the rachis is an epithelial layer forming 

 the ectoderm ; and the main canals have an epithelial endodermal 

 lining. The rest of the substance of the rachis consists of mesoderm : 

 this is very thin opposite the bases of the leaves, as seen in the 

 left-hand side of Fig. 5 ; but is of some thickness between the leaves, 

 as shown in the right-hand side of the same figure. It is traversed 

 by a network of very fine canals, and contains also definitely arranged 

 muscular fibres. These latter are chiefly longitudinal in direction : 

 they form a well-defined layer, with a crenated outline when seen 

 in transverse section, running along the dorsal surface of the rachis a 

 short distance below the surface epithelium (Fig. 6, / in), and a similar 

 layer along the ventral surface, shown in the sanae figure. In the 

 stalk, according to K5lliker's descriptions and figures, there is a con- 

 tinuous sheath of muscle extending all round ; but in the rachis this 

 sheath is interrupted at the sides by the polypes, and so loses its 

 regular arrangement. The dorsal aad ventral portions remain, as we 

 have just seen, unaltered, but the lateral portions are much changed : 

 they persist in part as the px-otractor and retractor muscles of the 

 polypes (Fig. 6, j))- 



A deeper set of longitudinal muscles is developed in the lower part 

 of the rachis in connection with the inner ends of the polype cavities : 

 it is shown in Fig. 6. 



The polype cavities communicate with the lateral canals, as shown 

 in the right-hand side of Fig. 6 ; but this connection appears only to 

 take place towards the bottom of the rachis. Through its means ova 

 are enabled to pass from the polypes into the lateral canals. 



On the ventral side of the rachis, and along its whole length, there 

 is found a curious system of tubes, which we propose to speak of as the 

 radial canals. These form two lateral masses (Figs. 5 and 6, v c) 

 imbedded in the mesoderm on either side of the main ventral canal, 

 each mass consisting of a number of branching tubes of tolerably 

 uniform diameter, lined by a single layer of short columnar epithelial 

 cells, which stain very readily with logwood or other colouring reagents. 

 At intervals these tubes can be distinctly seen in transverse sections of 

 the rachis to open into the main ventral canal, and such openings are 

 shown in both Figs. 5 and 6. 



Just before reaching the main canal tlie tubes are slightly con- 

 stricted, and their epithelial lining suddenly changes its character, 

 and becomes converted into the much flatter epithelium of the main 



