46 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxxi. 



second; and in the type of G. willemoesi, 136 mm. long, slight^ beyond 

 the first segment. In the larva before leaving the marsupium, as has 

 been said, it reaches to the middle of the telson, and thus the length of 

 this spine entirely depends upon age. 



2. The second is the most important character of G. sarsi, and I 

 find it in all the }^ounger individuals at hand. The lower lateral keel, 

 and also the marginal keel or rim, curve upward near the postero- 

 inferior angle of the carapace; but the actual margin of the carapace 

 extends behind the point, where the marginal rim begins to curve up, 

 and runs for a short distance straight back; then it forms a right 

 angle, extending toward the dorsal spine. Thus there is, behind the 

 marginal rim, a "rectangular lamina" as described by Wood-Mason. 



Sars does not mention such a structure, neither in G. willemoesiwoY in 

 G. zoea, he only says that the lower lateral keel curves upward before 

 reaching the postero-inferior corner, and that the latter, in G. wille- 

 m.oesi, is eventy rounded off. He does not mention the fact, that the 

 marginal rim curves upward before reaching the' posterior margin, 

 and that there is a "lamina" behind the marginal rim. Such a lamina, 

 however, is -distinctly seen in Sars's figures of G. vjillemoesi and zoea 

 (Plate V, fig. 1, and Plate VI, fig. 6). This is the more important, and 

 clearly establishes the presence of this lamina in Sars's specimens, 

 although he did not pay much attention to this feature, he gave a fair 

 representation of it in the figures. The lamina, however, in both 

 cases, is not rectangular, but evenly rounded off. 



Looking at the specimens at hand, I find that the largest, a typical 

 willemoesi, represents this character as described and figured by Sars, 

 only the lamina is somewhat wider than in his figure; but it is evenly 

 rounded off. Exactl} 7 the same condition obtains in our second largest 

 individual, 88 mm. long. From the Hawaiian Islands 1 have mentioned 

 two specimens of G. willemoesi, which I identified chiefly according to 

 this character, which measure 73 and 52 mm. The largest individual 

 observed by myself among the Hawaiian material, possessed a rectan- 

 gular lamina, and consequent^ was recorded under 4G. sarsi. It 

 measured 62 mm. The smallest measured 34 mm. 



Considering that Wood-Mason's G. sarsi was 75 mm. long, and that 

 Sars's specimens of G. zoèa, which have apparently a rounded lamina, 

 were 70 mm. and less, the conclusion is reached that all specimens 

 hitherto observed that are over 75 mm. long, have this character 

 developed according to the willemoesi type ; all specimens smaller than 

 52 mm. have it corresponding to the sarsi type; specimens between 52 

 and 75 mm. may possess either a rectangular or a rounded lamina. 



But it can not be said positively that this character is due only to 

 age. It ma} r be that the rectangular lamina becomes rounded with 

 advancing age, and that this transition takes place at a different period 

 in different individuals, in the average, when they are about half grown 



