30 BULLETIN 84, UNITED STATES NATIONAL MUSEUM. 



are irregular; in fact, only tloree of them are well shaped, the fourth one is rudi- 

 mental, and the fifth one completely lacking. Outwardly there is another circle of 

 five larger interradial plates. In each radial space there is a row of three plates ; the 

 first one is large and polygonal, even more developed than the centro-dorsal one; 

 the two succeeding ones separate the two radial shields of each pair, but the last 

 one is very small. In each interradial space there is also a row of three plates 

 succeeding the primary interradial one; the first one is about as large as the said 

 primary interradial plate, the second one is a httle more elongated, the third one, on 

 the contrary, which is located at the margin of the disk, is transversely widened. 

 Very small plates are sometimes observed between the first and the second inter- 

 radial plates which I have just indicated. All the plates of the upper face of the 

 disk are plainly polygonal with well-shaped angles,, whereas, in Lyman's drawing 

 they are quite rounded, the centro-dorsal one excepted. The radial shields, wliich are 

 larger than the other plates, are triangular, one and a half times longer than wide, 

 their length being a little inferior to the radius of the disk. They are diverging, 

 close lying or contiguous distaUy, but separated on their whole length. The arrange- 

 ment of the upper plates of the disk which I have just described is repeated exactly 

 on the two specimens at the Jardin des Plantes. 



The under face of the disk offers, out of the large mouth sliield, but one large 

 median plate with a wide proximal side, a fairly opened distal angle and two convex 

 lateral sides; it is a little longer than wide. This shape is very different from that 

 in Lyman's drawing, which shows it to be transversely oval (75, pi. 1, fig. 6). On 

 one of the specimens of the Jardin des Plantes, this median plate does offer the 

 shape which I have just described, but on the other one the proximal side is a httle 

 convex, the distal angle is roimded, as well as the lateral sides, so that the outhne 

 of this plate tends to become oval, and it is then just as long as wide; in one of the 

 interradial spaces it becomes even sUghtly wider than long, and it then resembles 

 Lyman's drawing. Two very small plates, which succeed the large median plate and 

 are placed on the margin of the disk, complete the covering of the under face between 

 the genital plates. The latter are very large, elongated, and wide. The genital 

 slits are very small, scarcely distinguishable, and they only appear on a part of the 

 distal side of the adoral plates. 



AU the upper and under plates of the disk bear rounded granules which are 

 conspicuous and widely separated. 



The upper brachial plates are relatively well developed on the specimen from 

 the Albatross, and they are larger on Lyman's type and also on the two specimens 

 from the Jardin des Plantes, which, in this respect possess the arrangements de- 

 scribed by Lyman. These plates extend the whole length of the arm; the first 

 two, especially, are large, but their size rapidly decreases. 



The brachial spines are generally two in number; and yet I sometimes find 

 three, not at the base of the arms but about the middle, and these are fouiid both 

 on the specimen from the Albatross and on the two specimens of the Paris Museum. 



The type of 0. testudo was found by the Eassler at Barbados at a depth of 100 

 fathoms ; the species was found again by the Blake at several stations in the West 

 Indies between 73 and 508 fathoms. 



