61 



the third primary radial supports, on each upper side, a single, 

 axillary, secondary radial, which supports, on each upper sloping 

 s'de, two tertiary radial*, except in one of the middle branches 

 there is only one tertiary radial, which gives to this ray four 

 arms. The ray, on the right lateral side, is constructed in the 

 same way, except each of the proximal branches bears only a 

 a single tertiary radial, and it has four arms. The ray on the 

 right of the azygous side bears, on the distal side of the third 

 primary radial, two secondary radials, and, on the proximal side, 

 two secondary radials, the last one being axillary, and supporting, 

 on each upper side, two tertiary radials, which gives to this ray 

 three arms. The left lateral ray is constructed in the same way, 

 except there is one less secondary radial and it bears on one side 

 two tertiary radials and on the other, one; it has three arms. 

 There are, therefore, sixteen arms in this species. The arm 

 formula, in such specimens as are represented by Figures 18, 19 

 and 21, is 3+4+2 +3 +4. But in the specimen represented by 

 Figure 20, the arm formula is 3+3+2+4+4. 



In the specimen represented by Figures 18 and 19, there are 

 two plates in one of the regular interradial areas and one in each 

 of the others. In the specimen represented by Figure 20, there 

 is one plate in one of the areas and two in each of the others. 

 In the specimen represented by Figure 21, there are three plates 

 in each of two areas and two plates in each of the others. In 

 the azygous area there are eight plates. The first one is in line 

 with the first primary radials and of about the same size. It is 

 followed, in the second range, by three plates, and these by three 

 plates, in the third range, above the middle one of which rests a 

 single plate that unites with the plates of the vault. In ore speci- 

 men we notice an additional small plate above the third range. 

 The area is much more depressed between the rays in some speci- 

 mens than in others. 



The vault is more convex in some specimens than in others. 

 In the spreading forms shown by Figure 21, the vault is only 

 moderately convex, while in the biturbinate forms shown in Figure 

 18, the vault is conical; the proboscis is proportionally much 

 smaller, in the spreading forms, than in the biturbinate. 



It will be noticed that the variations in shape, and in the regu- 

 lar interradial areas, are as great in this species as they are in 

 B. variabilis, and it is probable that other specimens may show 

 another variation, in the arrangement of the arms; that is, a four- 

 armed ray, on each side of the azygous area, and a three-armed 

 ray on each lateral side. As a matter of course the affinities of 



