54 BULLETIN 84, UNITED STATES NATIONAL MUSEUM. 



lateral sides, which are almost straight, meet at rounded angles. They are about 

 as long as wide. 



The lateral plates, fairly protruding, bear seven and sometimes even eight 

 spines at the base of the arms. The first under spine is strong, thick, and conical 

 with a blunt point; it is at least as long as the article and sometimes even a little 

 longer, and then the length of the spines decreases by degrees down to the last 

 dorsal one. From the first ventral spine upward, all are flattened, especially from 

 the third or fourth one upward, and the last ones take a somewhat lanceolate shape. 

 The spines are located chiefly on the upper part of the arms, and when looking 

 at the animal from the upper face, one perceives at least four spines on each side; 

 an arrangement which accounts for the upper brachial plates being comparatively 

 narrow. The surface of the first ventral spine is rough and the rugosities become 

 coarser near the point, which is rounded. The second spine always has a truncated 

 end; it displays aU along its edges very small denticulations which are more developed 

 at the end where they form a fairly regular little row; the last denticulation, which 

 occupies the distal angle, is more developed than the others and constitutes a little 

 transverse, hyaline hook, which is triangular, but always very short. Sometimes, 

 a hke formation, which always remains less important, appears at the proximal 

 angle, so that the spine has a tendency to become bihamuled; it wiU also be seen 

 that the small lateral denticulations of the spine are often more developed on the 

 distal side than on the proximal side. The next two spines, that is to say, the 

 thu'd and fourth ones, display a structiu-e similar to the second one, but the distal 

 hook alone is developed and the lateral denticulations do not appear except on the 

 distal side of the spine. FinaUy, the other spines simply remain rough at their 

 end, which is rounded. 



The tentacular scales, two in number, are at a right angle and middle-sized; 

 the external scale is rounded and it is a little larger than the internal one, which is 

 elongated and less wide. 



By comparing, as I have been able to do, the four specimens gathered by 

 the Fish Hawk with those of the Travailleur and the Talisman, to which I had given 

 in 1907 the name of A. pahneri, I have been convinced that the latter belong 

 certainly to the same species as the former and that they also ought to be called 

 A. Mnhergiensis. In fact, the under face of the disk is completely bare, and the 

 other characters offer a striking hkeness to those I have just pointed out in A. Icin- 

 iergiensis. I reproduce here two photographs of one of these examples, which 

 will sufficiently illustrate that hkeness (pi. 4, fig. 4, and pi. 5, fig. 2). The only 

 difference I can note regards the radial sliields which are comparatively a little 

 shorter in the examples of the Talisman, but we have seen that the shape of these 

 shields varies in the American examples; the lateral teeth of the spine appear chiefly 

 near the end. These specimens of the Talisman were dredged at a depth of 290 

 fathoms. 



Connections and differences. — A. Mnhergiensis is extremely near A. latispina 

 Ljungman, from which it differs chiefly by the presence, on the lateral spines, of 

 denticulations which develop principaUy at the end of the spines and which replace 

 the large bent hook which terminates the second ventral spine of A. latispina. 

 The under brachial plates are more elongated in the latter, the shape of the mouth 



