358 BULLETIN 82, UNITED STATES NATIONAL MUSEUM VOLUME 1 



longer than broad. The terminal brachials are somewhat elongated. P! and P 2 are long 

 and flagellate, composed of numerous short segments, the basal wide and slightly 

 carinate, sometimes with dorsal processes, the later serrate; P 3 little more than half 

 as long as P 2 , with much fewer segments, of which the basal are stouter and the 

 remainder mostly much longer than wide. The following pinnules are more massive, 

 with square segments which gradually become elongated. The two basal segments of 

 the distal pinnules are flattened and trapezoidal in some forms, but only slightly 

 modified in others. The ambulacra are not plated, but the disk sometimes bears a 

 number of small calcareous granules. 



Carpenter here repeats the discussion of this form given in 1884, adding that the 

 radials are not entirely concealed by the centrodorsal but appear above it as short 

 bandlike plates, and the IBri have more sloping sides than in the smaller forms of 

 glacialis so as to be trapezoidal in general outline, while the axillaries have a blunter 

 distal angle and the articular tubercles are much less developed. 



He writes that the relatively long quadrate shape of the brachials immediately 

 after the third syzygy is less marked in the Challenger specimens of quadrata (from 

 station 48), the southernmost ones known, than it is in the two obtained farther 

 north by the Triton and in those from the Arctic Ocean; but the middle and outer 

 brachials of the two species are always distinguishable, those of glacialis being short, 

 generally triangular, and much wider than long till quite near the end of the arm, 

 while those of quadrata are obliquely quadrate and the length is more nearly equal to 

 the width. 



Regarding the other special mark of quadrata, he notes that in large examples of 

 glacialis from the Arctic Ocean P 3 is of almost exactly the same length as P 2 , but in 

 the western Atlantic representatives of quadrata (from Challenger station 48) it is 

 distinctly shorter, and the southern forms of the two therefore approach one another 

 in the characters of the pinnules just as in those of the brachials although the more 

 northern varieties are entirely distinct in both respects. 



He remarks that not only is P 3 in quadrata altogether smaller than P 2 , but its 

 component segments, while fewer hi number, are also different in their relative 

 proportions; the basal segments are stouter, as in the following genital pinnules, and 

 their successors are distinctly longer than wide, indications of which appear in P 2 . 

 But there is no sign of this in glacialis, the segments of P 3 being as wide as or wider 

 than long; furthermore there is generally less trace in quadrata of the modification of 

 the two lowest segments in the distal pinnules which is usually so marked in glacialis } 

 though it is extremely well-developed in a specimen taken by the Varna in the Kara Sea. 



He gives measurements of P 2 and P 3 in a specimen of quadrata and in another of 

 glacialis of equal size, P 2 in quadrata being 14 mm. long with 31 segments and P 3 

 8 mm. long with 17 segments (four-sevenths as long as P 2 ), while P 2 in glacialis is 

 15 mm. long with 39 segments, and P 3 12 mm. long with 28 segments (four-fifths as 

 long as P 2 ). 



He points out that quadrata has been dredged at 1 1 stations altogether, but at only 

 5 of these was it found in association with glacialis; the Triton, Alert, Valorous, Tegetthoj 

 and Varna obtained examples of quadrata at localities where glacialis did not occur, 

 and in the last four cases there were only single individuals. He says that these facts 

 would seem somewhat improbable if quadrata were merely an immature stage of glacialis, 

 as supposed by Levinsen ; and further that it is a common experience of Arctic dredging 



