OPHIUBANS OF UNITED STATES NATIONAL MUSEUM. 93 



they are, though, a little less bulging, and they become a little longer than wide. 

 They are always separated by a fau-ly long space which, however, remains mferior 

 to theii- length. 



The first under brachial plate is trapezoidal, wider than long, with a wide and 

 convex proximal side, a smaller and rounded distal side toward which converge 

 the lateral sides, which are small and straight; it is separated from the second one 

 by the lateral plates, but sometimes, however, the proximal angle of the former 

 gets elongated so as to reach the distal edge of the first phite. The second ventral 

 plate is generally triangular with a still obtuse proximal angle which is limited by 

 two almost straight sides and a very wide and convex distal side; sometimes, 

 however, it is already pentagonal %\ith two small lateral sides; it is always wider 

 than long. The succeeding plates become pentagonal with an obtuse proximal 

 angle, short lateral sides and a rounded distal side; they always remain much wider 

 than long. The space which separates these plates becomes longer and longer 

 until it finally reaches twice their length; in fact, from the middle of the arms, these 

 plates become very short and, besides, comparatively small, although they always 

 remain wider than long. The lateral plates are fairly widened in their distal part, 

 but they ai-e, however, less tliick than in 0. pentacrinus. They carry seven and 

 even eight spines on the first articles. These spines recall those of 0. pentacrinus; 

 however, the ventral spines are thinner at the ends and their denticulations are 

 stronger and more loosely arranged; besides, their length increases in a more regular 

 manner from the first ventral spine, which is almost equal to the article, to the dorsal 

 spines, the hast two or tliree of which reach about the length of three articles. These 

 spines show denticulations which He rather loosely, are thin and pointed, a Uttle 

 more numerous on the ventral than on the dorsal spines; the rows are very much 

 approximated dorsally. The length of the dorsal spines progressively decreases 

 from the bases of the arms upward, and then- length finally equals one and a half 

 articles at the same time as the number of the spines decreases. 



The tentacular scale, rather small, is spiniform, pointed and fairly rough. 

 Connections and differences.~0. mendionalis is evidently very closely allied to 

 0. pentacrinus; the latter species, as I have stated above, corresponds to Lyman's 

 type, but it is at once distinguishable from it through the ^\idened and flattened 

 external oral papilla, which creates a difference which alone issufl^icient to separate 

 the two species. There are, besides, other distinctive characters. In all the sam- 

 ples of 0. meridionalis which I know the disk is rounded, not excavated in the 

 intcrradial spaces, while in all the specimens of 0. pentacrinus from the Albatross, 

 as well as m Lyman's type, these spaces are not only excavated but they are deeply 

 notched. The upper face of the disk is covered with real little spines, wliich are 

 slender and show on their surface some rugosities or denticulations and are ended 

 by short spinules, while in 0. pentacrinus there are short stumps endmg in three 

 elongated and divei^ng spinules. The first three brachial spines are more pointed 

 and their length increases more progressively than in 0. meridionalis; lastly, the 

 adoral plates are elongated, not very thick, with almost straightand parallel sides 

 in 0. meridionalis, and they separate the mouth shield from the first lateral 

 brachial plate, while in 0. pentacrinus they are thicker, grow outwardly thinner and 

 do not exceed the mouth shield. 



