194 CRUSTACEA MALACOSTRACA. III. 



cesses in the anterior transverse row on the head. Two or three pairs of submedian dorsal processes 

 on first abdominal segment. Marsupium with lamellse at fifth pair of legs very distinct. 



a. Penultimate joint of the antennal peduncles produced at the end into an acute tooth. First 

 abdominal segment with two pairs of submedian dorsal processes, and first pair much smaller 

 than second P. hystrix G. O. S. 



b. Penultimate joint of the antennal peduncles without any tooth at the end. First abdominal seg- 

 ment with three pairs of submedian dorsal processes in two longitudinal rows, and all nearly 

 similar in size P. Murdochi Ben. 



B. End of abdomen triangular, subacute or somewhat obtuse. Penultimate joint of the antennal pe- 

 duncles with some spiniform processes on the upper surface towards the lateral margin. Four spini- 

 form processes in the anterior transverse row on the head. Only a single pair of submedian dorsal 

 processes on first abdominal segment No marsupial lamellse at fifth pair of legs. P.frigidum n. sp. 



136. Pleuroprion hystrix G. O. Sars. 

 (PI. XV, figs. 6 a- 6 b). 



1877. Arcturus hystrix G. O. Sars, Arch. Math, og Naturv. Vol. II, p. 350. 

 ! 1885. G. O. Sars, North-Atl. Exp. p. 104; PL IX, figs. 2326. 



Female. The elaborate description and the figures published by Sars are on the whole good, 

 but several points must be mentioned, especially for procuring characters between this species and the 

 two following forms both unknown to Sars. 



The spiniform processes on the body are slender, most of them long, and all long processes 

 are arranged in transverse rows, excepting on the abdomen. The head has two transverse rows, and 

 three such processes in the row behind the front margin ; Sars says that only two spines are found in 

 this row, but according to my experience the median spine is always present but varying extremely 

 in length, being from very short to much longer than the sublateral pair. The dorsal surface of each 

 thoracic segment with a transverse row, excepting fourth segment, which has two rows, and so far 

 the segments agree with the figure of Sars, but in adult females the fourth segment has an incom- 

 plete or complete additional transverse row of much smaller spines close behind one of the normal 

 rows, or even one behind each normal row, and sometimes also two or three similar spines are found 

 behind the row on third segment -- Sars' figures do not convey a correct picture of the abdomen, as 

 he has drawn the line between the two abdominal segments in a wrong place; especially when the 

 animals are seen from the side, it is easily seen that the first segment has two pairs of dorsal spines, 

 furthermore that the second pair of these spines, which are much larger than first pair (fig. 6b), in 

 reality are placed on the first and not, as figured and described by Sars, on the second, large segment. 

 This segment has on each lateral margin a long and proportionately broad process, but no smaller 

 spiniform processes, while on its dorsal surface the spiniform processes vary considerably in number, 

 but only a single pair are rather long; the end of abdomen is moderately to very deeply emarginate 

 between the triangular, acute points. 



