328 BULLETIN 82, UNITED STATES NATIONAL MUSEUM 



spicules of limestone, but there is never anything like a definite skeleton. The color 

 is extremely varied. Some specimens are dark reddish purple or light yellowish brown 

 all over. Others have alternating bands of these two colors, each band covering two 

 or three brachials. In others, again, the bands are quite narrow, while some individ- 

 uals have a more or less mottled appearance, with the brown occasionally replaced 

 by white. In his discussion of comatulids in general he described the median groove 

 on the ventral surface of the radials, the form of the centrodorsal, the unusually sub- 

 stantial basal rays, and the articular faces of the radials, and said that the ambulacral 

 plating is reduced to small and irregular spicules without definite arrangement. 



The description of Antedon carinata given by Carpenter in the Challenger report 

 seems to have been based entirely upon specimens of pida collected at Bahia. It 

 reads as follows: 



The centrodorsal is a thick roughly circular disk with the dorsal surface bare, bearing marginal 

 cirri. 



The cirri are XX-XXX, 20-30. The segments are stout, and are all broader than long, the basal 

 ones especially so. The penultimate segment sometimes, but rarely, has a small opposing spine. 



The radials are partially visible. The IBri are short and oblong. The IBrj (axillaries) are 

 triangular, twice as long as the IBri, and forming with them a more or less distinct median tubercle. 



The 10 arms are about 125 mm. long. The first two brachials are wedge-shaped (the first least 

 so) with sharp outer edges and a median elevation similar to that on the IBr series in their line of 

 union. The arms consist of about 160 short brachials which are broad and nearly oblong till the 

 second syzygy, after which they are narrower and more triangular, gradually becoming more oblong 

 again and finally square at the ends of the arms. The middle of the distal edge of each brachial in 

 the lower part of the arm is slightly raised, and gradually develops into a keel or crest curving slightly 

 forward. This may follow immediately after the median tubercle between the first and second 

 brachials, or not begin till after the twentieth brachial, and varies much in its development, gradually 

 becoming less marked toward the ends of the arms. 



Syzygies occur between brachials 3-1-4 and from between brachials 9-1-10 to between brachials 

 13-1-14, and distally at intervals of 3-9 (usually about 5 or 7) muscular articulations. 



The lower pinnules are all of tolerably equal length. Pi is about 12 mm. long, rather slender, and 

 consists of some 20 trihedral segments most of which are longer than broad. The pinnules following 

 have rather broader and more flattened segments with a sharp dorsal edge, and they gradually 

 increase in stoutness to P3, which is the largest pinnule on the arm. The next few pinnules decrease 

 slowly in stoutness, but increase in length, their outer segments becoming relatively longer, but the 

 basal ones remaining broad for some distance. The terminal pinnules are long and filiform. 



The disk is 11 mm. in diameter, and is naked. Sacculi are very abundant on the disk, arms, and 

 pinnules. 



The color (in alcohol) is very variable, light brown, purple, or various combinations of the two, 

 either mottled or in broad or narrow bands; other specimens are mottled purple and white. 



Carpenter said that the centrodorsal of this species is very characteristic. It is a 

 thick disk with a single or partially double row of marginal cirri, but its dorsal surface 

 is smooth and free of cirri, though in young individuals it is more convex, with only a 

 small cirrus free space at the dorsal pole. The ventral surface is marked by an 

 indistinct pentagonal impression, which corresponds to a similar marking on the 

 under surface of the radial pentagon, and is of interest from its foreshadowing to a 

 certain extent the deeper bilobed impressions in the corresponding positions on the 

 centrodorsal and radials of Heterometra quinduplicava. He remarked that this species 

 has a well-developed basal star, and the articular surfaces of the radials, though 

 relatively wider than is usually the case in the endocyclic comatulids, are considerably 

 inclined to the vertical axis of the calyx, while there is a wide central funnel the opening 



