30 PROCEEDINGS OF THß NATIONAL MUSEUM. vol. xxxi. 



4. GNATHOPHAUSIA CALCARATA G. O. Sars. 

 Plate I, figs. 2a, 2b. . - X 



Gnathophausia calcaraia G. 0. Sars, Forh. Selsk. Christiania, 1883, No. 5; Rep. 



Challenger, XIII, 1885, p. 35, pi. iv.— Ortmann, Bull. U. S. Fish Comm. 



for 1903, Pt. 3, 1905, p. 968. 

 Gnathophausia bengalensis Wood-Mason, Ann. Nat. Hist, (6), VIII, 1891, p. 269. 



Specific characters. — Aside from the group characters (see a in the 

 key), the following are to be considered as of specific value: 



(1) The subovate, not lanceolate, shape of the antennal scale. 



(2) The presence of two pairs of oblique keels on the upper surface 

 of the carapace. 



(3) The shape of the epimera of the second to fifth abdominal seg- 

 ment, both lappets of which are pointed and spiniform. 



(4) The bifid points of the epimera of the sixth abdominal segment, 

 with the inner point much shorter than the outer (in old specimens 

 only). 



Description. — Carapace with dorsal, upper, and lower lateral keels. 

 Dorsal keel interrupted in the middle part. Lower lateral keel ending 

 in a spine at the postero-inferior angle of the carapace. On upper 

 face of carapace, between median and upper lateral keels, there are 

 two oblique keels on each side, converging posteriorly, the anterior 

 pair running toward the anterior end of the posterior section of the 

 dorsal keel, but not joining it; the posterior pair running almost par- 

 allel to the first pair, their hind ends not joining the dorsal keel. Ros- 

 trum of various lengths, according to age, about as long as the rest of 

 the carapace in very young specimens. In older ones, the part in front 

 of the supraocular spines is about one-third of the length of the rest 

 of the carapace. Supraocular spines very small, sometimes obsolete. 

 Antennal spines small, but well developed, the most constant spines 

 in size. Branchiostegal spines quite large and well developed in } T oung 

 specimens, and longer than the antennal spines. In old specimens they 

 are not only relatively, but absolutely smaller, and become shorter 

 than the antennal spines. Postero-dorsal spine of various lengths, 

 according to age, but the variation is not very great; it is always well 

 developed, but shorter than the postero-inferior spines. Spines of 

 postero-inferior angle greatly varying in length with age; very long, 

 almost half the length of the carapace (excluding the rostrum) in } T oung 

 specimens, and distinctly diverging and spreading out in a postero- 

 lateral direction. In old specimens they are much shorter, even abso- 

 lutely shorter, and are as short as about one-seventh of the carapace 

 (without rostrum); they are not divergent, but directed straight back- 

 ward. Branchiostegal, póstero dorsal, and postero-inferior spines, 

 when well developed, with more or less distinct serrations, which 

 become indistinct with age, and even disappear entirely. 



