A MONOGRAPH OF THE EXISTING CRINOIDS 133 



Blake stations 45Ag. and 208) to a smaller thinner disk with a single marginal row of 

 cirrus sockets (type specimen of variety planata; no locality), and to a smaller somewhat 

 convex plate entirely covered with cirri (between Dominica and Martinique, 292-695 

 meters). Generally speaking, its size is proportionate to the size and relative develop- 

 ment of the individual, but there are exceptions as, for instance, a specimen with 

 15 or more arms from Blake station 32Ag., in which the centrodorsal is quite dispro- 

 portionately large. Many of the large centrodorsals are slightly concave in the center 

 (Blake stations 208 and 249). Many more of the larger than of the smaller have the 

 cirri in an almost everywhere single row. Hartlaub considers this the usual arrange- 

 ment of the cirri. A regular single row (as hi a specimen marked "Locality lost") 

 is rare, there usually being a partially developed second row. The types of pulchella 

 have a small centrodorsal, and in both the specimens the cirri are in 2 rows. In its 

 form and in its variations the centrodorsal, according to Hartlaub, resembles that of 

 Comactinia echinoptera. 



The cirri are XII-XXXII, averaging XIX. The highest number, XXXII, is in 

 a specimen from Blake station 45Ag. ; 2 others have XXX and XXVIII, respectively. 

 In 2 from Blake station 189 and Ifrom station 156 there are XXVIII and XXVII. 



Like the centrodorsal the cirri are variable, although they show a fair degree of 

 uniformity, especially in the number and form of their component segments. The 

 length and stoutness of the third segment is variable. In different cirri on the same 

 individual it may be either long or short ; the latter condition is by far the commonest. 

 But in some specimens the third segment is almost as long as the fourth. Sometimes 

 the fifth and sixth segments instead of the fourth and fifth are the longest, as in a 

 19-armed specimen from Blake station 249. 



The usual number of cirrus segments is about? 20. The largest number observed 

 was 24 (Blake station 241), the lowest 12 (no locality; 14 arms) and 9 (10-armed 

 specimen from Blake station 298). 



The longest cirrus measured 28 mm. (Blake station 208), but, on the other hand, 

 cirri with a length of about 12 mm. are not rare (Blake stations 155, 156 or 298, and 

 277). The longest cirri of the specimens from Blake stations 208 and 241 are markedly 

 stout, while the short cirri of the specimen labeled "Locality lost" are slender. Some 

 specimens from Blake station 269 have rather long slender cirri. Stout cirri are 

 apparently much rarer. 



Very characteristic of the cirri are the dorsal spines on the distal segments. 

 While these may be only very poorly developed, they are never absent. 



In the great majority of the specimens studied by Hartlaub the division series 

 are free laterally, often from the IBri outward, or at least from the IBr 2 (axillary) 

 outward. This is true hi the type specimens of pulchella; but other specimens in 

 all other ways similar to these (for instance, those from Blake station 298) have the 

 division series in lateral contact. 



The lateral contact of the division series is most pronounced in a series of speci- 

 mens which are distinguished by having the division series in very close lateral 

 apposition with sharply flattened sides; in this feature, as well as in their very slender 

 build, they vary quite widely from the types. These specimens, designated by 



