44 PROCEEDINGS OF THE NATIONAL MUSEUM, vol. xxxi. 



specimen (from Station No. 2723, about 105 mm. long), which undoubt- 

 edly belongs to G. willemoesi according to Sars's conception, and these 

 young ones (Plate II, fig. 2a) have the postero-dorsal spine well devel- 

 oped, and comparatively much longer than any specimens ever 

 described, extending to about the middle of the telson. Thus the 

 length of the postero-dorsal spine depends without question on the 

 age of the individual. 



2. The denticulations or serrations of the posterior margin of the 

 carapace, the pòstero-dorsal spine, the spines of anterior margin of 

 carapace, and of the rostrum are generally in this genus more distinct 

 in younger individuals than in older ones. I have called attention to 

 this above (under G. calcarata). In the present case the large indi- 

 vidual from Station No. 2723, which is surely G. willemoesi, has the 

 margin of the carapace not ''decidedly glabrous," as Sars states, but 

 there are a number of fine denticulations, less distinct than in young 

 individuals, but easily seen. Faxon (1895) says that in G. willemoesi 

 there are denticulations along the margin of the dorsal spine. Thus 

 this character does not hold. 



3. That the relative length of the rostrum, like that of the spines of 

 the carapace, changes with age is now well established. In the } T oung 

 specimens extracted from the pouch of the mother, the rostrum is 

 decidedly longer than the carapace (Plate II, fig. 2a). If the rostrum 

 becomes shorter with age it is not astonishing that the denticulations 

 become less pronounced, and this is entirely in keeping with what I 

 have shown in the second character. Thus the length of the rostrum 

 does not possess any s} T stematic value. 



4. The fourth character needs special attention, but I think I am 

 able to prove that it also is influenced by age. In young specimens 

 the spine of the outer margin of the antennal scale is longer than the 

 terminal lobe, and it is slightly serrated on the outer edge. With 

 increasing age it becomes slightly shorter than the terminal lobe, and 

 the serrations disappear. The following ma} 7 be said in support of 

 this view: 



a. The specimens representing the original G. zoea are small or of 

 medium size (not longer than 70 mm.), while the specimens upon which 

 G. tvillemoes i was founded are very large, one measuring 136 mm., 

 and the other being "somewhat smaller;" that is to say, they were 

 about double the size of G. zoêa. 



b. A large specimen (Station No. 2723) is about 105 mm. long, and 

 has the antennal scale of G. willemoesi; another (Station No. 4306) 

 is 88 mm. long, and has the antennal scale intermediate between G. 

 zoea and willemoesi; the spine is about as long as the lamellar portion 

 on the left side and very slightly longer than the latter on the right 

 side, and it has on the outer margin very indistinct indications of ser- 

 rations, visible only under the microscope. The latter specimen is 



