i64 DISCOVERY REPORTS 



upwards to form a dorsally directed flange which is continuous with a similar one on the endopodite. 



The endopodite (Text-fig. 107) is nearly twice as long as it is broad. 



Distribution. The specimens were collected at 49'" 28' 25" S., 66" 29' 00" W., that is to say, in an 

 area of shallow water outside the Antarctic Convergence, off the coast of the southern part of South 

 America. 



Subfamily Glyptonotinae Miers, 1881 



This subfamily was instituted by Miers (i 881) to contain the single genus Glyptonotus with its five 

 species: G. antarcticiis Eights, G. entomon Linn., G. sabini Kroyer, G. caeciis Say and G. tuftsii 

 Stimpson. 



The diagnostic characters of the subfamily were : ' Sides of the head emarginate or cleft and laterally 

 produced beyond the eyes, which thus are situated upon its dorsal surface. The three anterior pairs of 

 legs with the penultimate joint or propodus dilated and forming with the reflexible dactylus a pre- 

 hensile hand. Species more or less ovate.' 



At the present time the only species from the above list remaining in the genus is G. antarcticus, 

 but a new species G. aciitus Richardson was added in 1906, and new genus Symmiiis, with one species 

 S. cmidatiis by Richardson in 1904. The diagnostic character, 'sides of head emarginate or cleft ', no 

 longer holds for this subfamily. Although Collinge (1916, p. 114) used this character as diagnostic in 

 separating the two genera, in a later paper (191 8, p. 64) he pointed out that the margins are entire 

 both in the existing species of Glyptonotus and in Symmius caiidatus. 



d 



Text-fig. 1 1 . Glyptonotus antarcticus. (a) Ventral view of coxa and plate of second pereiopod and a portion of the tergum, x 2. 

 {b) Posterior view of {a), {c) Ventral view of sixth coxa and plate, x 2. {d) Posterior view, showing the relationship with the 

 tergum. 



Nordenstam (1933, p. 103), in his diagnosis of the subfamily, mentioned that the coxal plates are 

 'marked oflF by dorsal sutures on the last three pereion segments'. It is strange that he makes no 

 reference to the coxae of the anterior segments until his discussion on the affinities of the subfamily 

 Macrochiridotheinae ; here, on p. 1 10, he says 'the coxae on these segments are very small in Glypto- 

 notus, though distinctly delimited from the segments '. This is an important point, since most authors 

 state that the coxal plates of the anterior thoracic somites are completely fused with their respective 

 somites. Actually, the terga of the second to the fifth thoracic somites inclusive are produced into 

 broad pleura which extend laterally over the bases of the limbs (Text-fig. 1 1 b). On the ventral surface 

 of these somites each coxal joint is clearly defined, its outer margin is ring-like, and except for that of 

 the second, is separated from the ventral surface of the pleuron by a suture. The inner part of the 

 joint is produced as a coxal plate which partly covers the ventral integument and extends almost to the 

 mid-ventral line (Text-fig. 11 a). In the hinder three somites the pleura are replaced by coxal plates 



