PART A MONOGRAPH OF THE EXISTING CRINOIDS 493 



acknowledged by Mr. Clark to be s3Tionyms), to lower conical, about a3 high as wide 

 at the base in the majority of species, for example Poliometra prolixa, reaching an ex- 

 treme of low rounded conical in a few species such as Cyclometra flavescens. 



At the same time the degree of interradial ornamentation of the centrodorsal is 

 decreased. In Zenometra there is a conspicuous vertical ridge in each interradius. In 

 Balanometra and Adelometra this is replaced by a groove. The various species included 

 in Psathyrometra exhibit grooved, raised or plane interradial areas. In Sarametra and 

 Anisometra the interradial spaces are plane. In the eight other genera there are no 

 distinct interradial spaces between the cirrus sockets. 



The number of columns of sockets may be 10, 15 or 20. In Zenometra, Balanometra, 

 Adelometra and Anisometra there are only two columns in each radius; in Sarametra 

 there are three or four columns, the imique type specimen of S. nicobarica having four 

 columns basallj^, giving way to three towards the apex as the centrodorsal naiTows. The 

 same thing happens in some of the other genera where the centrodorsal tapers towards 

 the apex. There may also be variation in the number of columns of sockets in speci- 

 mens of different sizes of the same species. The two columns of small specimens may 

 gradually separate and one or more sockets of a third or third and fourth columns be 

 introduced between them. The additional midradial column or columns tends to be 

 more irregular than those on each side as well as being more incomplete, so that although 

 in interradial view the columns may appear regular, when seen radially this is not 

 so obvious (see fig. 31, p. 551). On this account it is difficult to decide whether such 

 species as Hybometra senta or Cyclometra flavescens should be referred to the Zenometrinae 

 or to the Bathymetrinae. 



As in some other comatulids a number of the sockets towards the apex of the centro- 

 dorsal may be obsolete with the central canal closed. In the type specimen of Sara- 

 metra nicobarica, for instance, there are about 140 sockets altogether but probably less 

 than 100 of them actually bore cirri at the time of capture. In Zenometra columnaris, 

 where the sockets are much fewer as well as being more widely spaced, nearly half the 

 total number may be obsolete. 



The cirri average about L in number, ranging from XX to about XCV. They are 

 often damaged or lost before specimens reach the museum. 



The number of segments in the peripheral cirri ranges from 20 to 70 in all but two 

 of the known species, the maximum being found in Adelometra angustiradia, while 

 Sarametra triserialis has up to 60 and Poliometra, Zenometra and Leptometra have up to 

 50 segments. It seems to be in the antarctic species that the lowest numbers are foimd. 

 Only one of the five species of Eumorphometra — E.fraseri — has over 30 segments. The 

 cirri of both Anisometra Jrigida and Eometra weddelli are only known from single de- 

 tached ones which were probably not peripheral, these being more likely to be lost; they 

 have 21 and 18 segments respectively. However, in Kempometra grisea the attached 

 cirri have up to only 16 segments, but neither of the type specimens have the arms as 

 much as 40 mm. long. 



In length the peripheral cirri varj' from 23 to about 60 percent of the arm length, 

 averaging 40 percent. The proportions of the segments themselves are very variable 

 but usually the longest ones are three to four times their median \vidths though in 

 some genera such as Caryometra they may be up to six times as long as wide. Con- 

 versely in Eumorphometra they are little longer than wide. The distal segments are 

 usually somewhat shorter. In most of the species there is a tendency for distal en- 



