OPHIUBANS OF UNITED STATES NATIONAL MUSEUM. 131 



The brachial articles are elongated. The upper brachial plates do not show in 

 the American specimens such a marked division as in the two types. Only on two 

 specimens from station 2668 and on that from station 2667, where the diameter of 

 the disk reaches 5 mm., do I observe this dividing, which is limited to the first two 

 or three dorsal brachial plates ; each of these plates offers a triangular chief portion 

 with a truncated proximal angle and a proximal fragment which is larger on the 

 first article, but very small on the following ones, where it is reduced to a small 

 circular plate. The other upper plates are always whole; generally, there is left 

 between their proximal angle and the distal margin of the preceding plate, a little 

 space covered by the tegument and which is never beset by a plate. The dorsal 

 brachial plates always remain a little longer than wide, and their proximal angle is 

 rounded. 



The first under brachial plate is large, trapezoidal, longer than wide. The 

 following ones are pentagonal, but in the American examples, they are compara- 

 tively smaller than on my 1909 drawing, and they get separated beyond the second 

 one; they are a*t first a little wider than long. 



The lateral plates, very much elongated owing to the length of the articles, 

 carry five short, thick spines which are beset with strong spinules in their terminal 

 part; these spines do not exceed half the article, except the last dorsal one which 

 is a little longer. These spines always remain more or less closely applied against 

 the lateral plates. 



I had not noticed the tentacular scale in 1909. There is really one such scale, 

 but it is very small, more or less buried in the tegument and very difficult to recog- 

 nize; I have been unable, besides, to notice it on all the articles; it is spiniform. 



I have said above that 0. granulatus is very near 0. armatus, and one might 

 first suppose that the latter is but a young stage of the former, but it is not so and 

 it is easy to grasp the differences which separate the two species. The most obvious 

 is supplied by the brachial spines which are much longer in 0. granulatus and which, 

 instead of being applied against the lateral plates, are on the contrary, off-standing 

 and diverging from them. The upper plates of the disk, which are larger and less 

 numerous than in 0. armatus, are beset with strong and thick granules, provided 

 with strong spinules which never get elongated into small, conical spines, but always 

 remain cylindrical, and almost as long as wide. The dorsal brachial plates show 

 no dividing up, except on the first two or three brachial articles, and they are much 

 wider than long; lastly, the brachial articles are shorter. The oral papillae may 

 become fairly numerous in 0. granulatus, and in the large specimens they may 

 amount to six or seven. It seems also that 0. armatus always remains rather small, 

 while the dimensions of 0. granulatus are great. 



The differences which I have just indicated appear plainly when specimens of 

 equal size are compared, such as a little specimen of 0. granulatus which I have 

 represented on plate 14, figure 4, compared with the examples of 0. armatus repro- 

 duced on the same plate. 



