374 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



calcareous processes are so developed as to form a spongy calcareous mass entirely 

 fillinn; the funnel-shaped cavity of the radial pentajjon, resulting in the formation 

 of a comparatively dense central plug (fig. 11, p. 65). 



Unless the central plug is so fully developed as entirely to obscure the internal 

 and ventral faces of the radials, the funnel-shaped interior of the radial pentagon 

 is seen to be marked with live furrows, interradial in jjosition, which lie in the 

 interradial sutures (fig. 441, p. 351); between them, in the midradial line, there 

 are usually five broader and shallower furrows, which run to the intermuscular notch 

 (figs. 435 and 445a, p. 351), and often through it, traversing the joint face nearly to 

 the central canal. They are extended outward in a similar position over the skele- 

 ton of the rays and arms. These are known as intermuscular midradial furrows. In 

 some species they are represented by low broad ridges, or merely by a gi-eatcr density 

 of the calcareous structure; often they are not present at all, the midradial portion 

 of the radials not being different from the lateral portions. The midradial furrows, 

 when developed, serve to lodge the proximal portion of the coeliac canals. They 

 are well shown in Tropiometra picta, Cyllometra manca and in Nemaster lineata. 



At the inner margin of the ventral face the midradial furrow turns downward 

 and passes (when developed) directly into a nearh' vertical furrow, occupj'ing the 

 median axial line of the jiroximal or internal face, and becomes more or less com- 

 jiletely converted into a canal by the union of irregular processes (forming part 

 of tlie outer portion of the central plug), which extend themselves from the side 

 to meet the spoutlike processes of the rosette. As it descends toward the dorsal 

 face and passes between the inner raised edges of the two apertures of the central 

 canal (lodging the secondary basal cords of the dorsal nervous sj^stem), this axial 

 radial furrow becomes a complete canal, for its edges are closely applied to the 

 inflected margins of one of the five radial spoutlike processes of the rosette. 



These axial canals are therefore the proximal ends of the five coeliac canals 

 of the arms and their extensions into the pinnules, and they thus inclose portions 

 of the body cavity which Carpenter called the radial coelom. As a general rule 

 they become closed up by calcareous tissue and do not reach the dorsal surface 

 of the radial pentagon, which presents no real openings except the central one 

 occupied by the rosette; but they sometimes open on the dorsal surface of the 

 radial ]>entagon, as in Antedon, Stenometra and CyTlometra, by five large holes 

 that correspond with five more or less distinctly marked circular depressions placed 

 interradially on the ventral surface of the centrodorsal around the margin of its 

 central cavity, and the canals end blindly in these depressions. In Antedon these 

 depressions are usually shallow pits of considerable size, but they are variable in 

 their development, and are sometimes, though rarely, absent altogether. This 

 condition, in which there are no radial depressions on the ventral surface of the 

 centrodorsal, is the normal one in Leptometra. Here, as described by Carpenter, 

 the margin of the central opening is usually almost circular (fig. 287, p. 262), though 

 sometimes bluntly stellate as in Antedon (figs. 2S0, 281 , 283, p. 261) ; at the same tune 

 the five openings on the dorsal surface of the radial pentagon are but little devel- 

 oped or even entirely absent. The absence or slight development of these open- 

 ings in Leptometra is considered by Carpenter to be i)rincij)iUIy due to the fact 



