222 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



the corresponding under arm plate, may be appreciated from the dif- 

 ferent photographs which I include (pi. 51, figs. 1-4; pi. 52, figs. 1, 2). 

 The longest lateral spines may reach and even exceed a length of five 

 segments. These spines are cylindrical, rather slender, with the 

 tip rounded, and they are provided throughout their whole length 

 with rather fine and closely placed denticulations. 



These are the characters of the spines seen on individuals which 

 may be considered as typical, as those of which I include photo- 

 graphs on plate 51, as figures 1, 2, and on plate 52, as figure 1. But 

 in a certain number the dorsal spines undergo a remarkable trans- 

 formation; they shorten and thicken considerably, as in the case in 

 the individual represented on plate 50 as figures 2, 3. This modifi- 

 cation takes place especially on the last dorsal spine, which may 

 become almost globular ; the one before the last is also shortened and 

 thickened, though in a much less marked manner. This trans- 

 formation of the dorsal spines appears only on the twelve or fifteen 

 first arm segments; beyond that the dorsal spines preserve their 

 usual character. The transformation is progressive, beginning on 

 the second or third segment, the spines being at first slightly thick- 

 ened, the thickening becoming rapidly accentuated and the spines 

 retaining the same characters for about twelve segments, then little 

 by little reassuming their normal size and length. The specimens 

 which show this modification thus greatly resemble O. crassispina, 

 which I have just discussed, but they can not be assigned to that 

 species for the reasons which I shall give further on. 



This remarkable thickening of the dorsal spines does not show 

 itself by any means on all the specimens collected by the Albatross. 

 In the 20 which the collection contains, 14 have normal spines, and 

 only four have the dorsal spines modified in the way which I have 

 just described, and which my photographs show. On two others a 

 shortening and thickening of the spines may be also observed, but 

 the modification is very much less marked. It is rather well shown 

 in the specimen figured on plate 50, figure 5, but it only appears on 

 three arms, while the fourth is normal, and the fifth is lacking ; in the 

 other specimen, which is shown on plate 51, figure 1, it may be seen 

 that the dorsal spines are not as long as usual at the base of the arm, 

 and some of them already show traces of a slight thickening. 



I notice that in the regions where the dorsal spines are thickened 

 the corresponding upper arm plates are also slightly modified; they 

 are a little less broad and more projecting than usual, and their distal 

 border is entirely without spinules. 



The first ventral spine does not transform into a hook until some 

 distance from the disk; this hook is rather slender, but elongated, 

 and shows four branches, the last large and stout, the following 

 smaller, and two others very short and very narrow (pi. 98, fig. 3&). 



