ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 31 



ciliated epithelium of the floor of the groove lies a fibrillar structure, to which I have 

 given the name of the " subepithelial band " l , and to which a nervous character has 

 been attributed by myself and by all the other observers who have described it. Each 

 side of the ambulacral groove is bounded by an elevated fold of perisome, the edge of 

 which is not straight, but cut out into a series of minute valvules, the cresceutic or 

 respiratory leaves (Wyv. Thomson), or " Saumlappchen " of the German authors. 



At the base of each leaf, and to some extent protected by it, is a group of three 

 tentacles, one of which, the more distal one, is larger than the other two. This trind 

 group of tentacles and the cavity of the respiratory leaf adjacent to them receive a 

 common branch from the radial water-vessel. These groups of tentacles alternate on 

 the opposite sides of the ambulacral groove from the base to the tip of each arm, and are 

 distributed in the same manner at the sides of the ambulacra of the disk, though they 

 are not so markedly developed, especially near the peristome, where every lateral branch 

 of the water-vessel supplies only one tentacle. The crescentic leaves at the sides of the 

 groove are also far less distinct than in the arms, the edges of the folds of perisome 

 bounding the groove being only marked by a faint wavy line, and not distinctly cut out 

 into " Saumlappchen." 



In many Actinometrce, however, the above description only applies to the arms of the 

 two anterior radii, A, B (PI. II. figs. 2, A), and to more or fewer of the antero-lateral 

 arms, C^ and E^. The arms of the posterior radius, D, and of the posterior divisions of 

 the lateral radii, C 2 and E l5 are often entirety devoid of tentacles ; and in many of them 

 the ventral perisome not only exhibits no ambulacral groove, but is, on the contrary, 

 convex, as in the oral pinnules of Antedon (PL II. figs. 5, 6). 



We have just seen that in Act. pohjmorplia, as in all Actinometrce with an interradial 

 mouth, the anal area is bounded by two large aboral groove-trunks, which start from the 

 posterior angles of the peristome, and form a horseshoe-shaped curve, tbe limbs of which 

 are unequal in size (PL II. fig. 2). The smaller right limb is formed by the right lateral 

 ambulacrum, C ; while the larger left limb represents the posterior ambulacrum, D, com- 

 bined with part or the whole of the left lateral ambulacrum, E. In neither of these limbs 

 are the tentacular groups and crescentic leaves so well developed as they are in the 

 two anterior ambulacra. After the branches to the two antero-lateral primary arms, 

 C^ and E 2 , have been given off, or sometimes even sooner (PL I. figs. 13, 15), the ten- 

 tacles at the sides of the two aboral groove-trunks become more and more insignificant, 

 and finally disappear altogether, while the position of the crescentic leaves is only 

 indicated by a very faint wavy line at the edge of each groove. 



In small specimens with but few arms (PL I. figs. 6, 9) the grooves of the posterior 

 (D) and postero-lateral arms (C 2 , E,) may remain in this condition; but in larger speci- 

 mens with many arms all trace of the crescentic leaves disappears, and the two edges of 

 the groove meet and unite so as to produce the condition represented in PL II. figs. 5 & 6, 

 where the ventral surface of the arms and pinnules is convex, and does not show the 

 least trace of a groove of any description. 



" Kcmarks on the Anatomy of the Arms of the Crinoids. Part I.,'' Journ. of Anat. and Physiol, vol. x. p. 57!'. 



