OPHIUKANS OF THE PHILIPPINE AND ADJACENT WATERS. 307 



tremely reduced, their length equaling almost one-fifth of the radius 

 of the disk. In the figure published in 1882 (pi. 24, fig. 14), on 

 the other hand, the plates become larger at the periphery of the 

 disk and in the vicinity of the radial shields; the primary plates 

 are larger, and the radial shields are also larger, their length ex- 

 ceeding a third of the radius of the disk; in both cases the same 

 specimen from Challenger station 232 is the subject. I may add 

 that I find among my specimens the two alternatives figured by 

 Lyman. 



The plates of the dorsal surface of the disk are always very 

 numerous and very small, but with some variations; an arrange- 

 ment which may be considered as the mean is that shown by the 

 specimens from stations 5582 and 5585 (pi. 76, figs. 1, 2) ; these 

 plates are a little larger in the specimen from station 5619 (fig. 

 10). On the other hand, in that from station 5348 they are of an 

 extreme minuteness (fig. 4), and they become only a little larger 

 toward the borders of the disk and in the vicinity of the radial 

 shields; the six primary plates are also extremely small, though 

 they are more evident. The dimensions of the radial shields also 

 vary; in the specimen from station 5619 (fig. 10) they are elon- 

 gated and rather slender, three times as long as broad, pointed in- 

 wardly, and their length is equal to a third the radius of the disk. 

 In the specimen from station 5582 (fig. 2), and especially in that 

 from station 5650, these shields are shorter and relatively a little 

 broader, their length scarcely reaching a quarter the radius of 

 the disk; they are still more reduced in the specimen from station 

 5585 (fig. 1), and especially in that from station 5348 (fig. 4), in 

 which their length scarcely reaches a fifth the radius of the disk; 

 in the last they are small, triangular, scarcely longer than broad, 

 with the proximal angle rounded and the internal border convex. 



The keel on the ventral surface of the arms is generally only 

 slightly marked; the specimen from station 5583 alone (fig. 3) 

 shows a fairly developed keel. This is scarcely indicated in those 

 from stations 5348, 5585, and 5650; it is not all evident in that 

 from station 5619 which, it is true, is smaller than the others, the 

 diameter of the disk not exceeding 16 mm. 



It is only in the specimen from station 5582 (fig. 3) that the 

 mouth shields show the broadened and low form described by 

 Lyman; these shields arc a little higher and less broad in the two 

 from stations 5585 (fig. 7) and 5619 (fig. 5), while in that from 

 station 5348 (fig. 8) they are still broader than long, but their 

 proximal angle no longer exists, being replaced by a convex border 

 in such a way that the shields become semicircular with a small 

 lobe on the middle of the distal border. In the specimen from 

 station 5650 (fig. 6) they are simply triangular and pronouncedly 



