600 BULLETIN 82, UNITED STATES NATIONAL MUSEUM VOLUME 1 



Notes. — A reexamination of the type specimen of tenuipes showed that I had made 

 a mistake in descrihing the arrangement of the cirrus sockets on the centrodorsal. 

 They are not in 2 columns in each radial area but in 2 converging lateral columns with 

 an incomplete median column between the proximal halves of the two lateral columns. 

 This discovery at once suggested reexamining the description and figures of Hartlaub's 

 Antedon arcana, which previously I had tentatively assigned to the genus Coccometra. 

 I find that Hartlaub's description appUes perfectly well to tenuipes. In his figure 1 on 

 plate 9 the arrangement of the cirrus sockets in 3 columns is shown, but the median 

 column is complete. The delineation of the centrodorsal is not quite correct, however, 

 as is seen by a comparison with the photograph reproduced as figure 5 on plate 15. 

 This photograph shows a specimen more robust and more mature than the tj'^pe of 

 tenuipes, though undoubtedly belonging to the same species. 



Antedon arcana was thus described by Dr. Clemens Hartlaub: The centrodorsal is 

 slender-conical with the apex blunt, 2.5 mm. high. There are about 70 to 80 strikingly 

 flat cirrus sockets. The rim of the centrodorsal is deeply concave under each radial. 

 According to Carpenter's drawing reproduced by Hartlaub the cirri are arranged in 

 15 crowded columns. 



The cirrus stumps preserved are confined to the proximal two-thirds of the centro- 

 dorsal. The very weak impressions on the distal third make it apparent that from these 

 sockets the cirri were early lost. The cirri decrease in size toward the apex of the 

 centrodorsal, and already in the middle they are only half as stout as those at the base. 

 No entire cirrus is preserved, only a few stumps. Carpenter's figures published by 

 Hartlaub may show cirri belonging to the type specimen. One of the cirri shown has 

 19 segments which in the proximal half are much elongated, the fourth to seventh being 

 nearly or quite three times as long as broad; the distal end of the elongated segments is 

 expanded and overlaps the base of the segments following. The other cirrus sho\vn has 

 35 segments and a much larger number of short segments, which are broader than long, 

 in the distal half. Hartlaub said that in both cirri the distal segments are markedly 

 shorter than in Poliometra prolixa or in Hathrometra tenella, in which also the segments 

 are not so strongly set off from each other. 



The radials are visible, and are short-discoidal. The IBri are also short, especially 

 in the middle, where they are deeply incised for the reception of the proximal process 

 from the axillary; there is a weak synarthrial tubercle on the articulation between the 

 IBri and the axillary. The IBrj (axillaries) are approximately square with the sides, 

 especially the distal, concave. The anterior angle separates the bases of the first 

 brachials. The IBr series are laterally free, but the IBr, shows laterally a trace of 

 "wall-sidedness." 



The 10 arms are estimated as about 50 mm. long; the disk reaches as far as the 

 seventh brachial. The brachials are of average length and have a smooth surface. 

 The first brachials are similar to the IBr, and are deeply incised by the strong pro.xunal 

 process of the second. The second brachials are markedly longer than the first, about 

 four times as long exteriorly as interiorly, the inner side being very short, and are not 

 in contact with those on neighboring arms. They are distorted rhombic with arched 

 sides, prox-imally elongated and drawn out to a point. The first syzj-gial pair (com- 

 posed of brachials 3+4) is shorter and is longer interiorly than exteriorly. Then follow 

 a number of approximately discoidal, or more precisely trapezoidal, brachials which are 

 broader than long and are in contact only at two points, on one side ^vith an alternating 



