A MONOGRAPH OF THE EXISTING CRINOIDS 53 



in number; and the basals are usually strongly reduced except in certain Himero- 

 metridae and in the family Tropiometridae of the Tropiometrida. 



Gislen said that most of the primitive features are shown in the Zygometridae, 

 certain Himerometridae, and the Mariametridae. Such features are: A syzygy 

 between the elements of the IBr series in the Zygometridae; an indefinite number of 

 components in the division series (in the Zygometridae and certain Himerometridae) ; 

 few connecting fibers between the parietal and visceral layer of the coelome sacks in 

 the disk so that the disk is easily thrown off (especially marked in the Zygometridae) ; 

 the disk covered by calcareous granules; few syzygial septa (Mariametra and the 

 Zygometridae); and undifferentiated proximal pinnules (Zygometridae and certain 

 Himerometridae and Mariametridae). 



The whole of the suborder Mariametrida, with the exception of the small genera 

 in the Colobometridae, has a rather coarse and clumsy structure, a small and shallow 

 centrodorsal cavity, and sparse syzygies. Among these features the two first men- 

 tioned and probably also the last are primitive. Besides, there are solanocrinid 

 characters in the discoidal centrodorsal with rather stout cirri, as well as in the shallow 

 and often meandering branched furrows that lodged the dorsal coelome between the 

 centrodorsal and the radial pentagon. 



In regard to the structure of the centrodorsal, and in the Comasterida and Maria- 

 metrida of the radial pentagon too, there seems, on the whole, to be evident a striving 

 toward a similar ideal type. The further the comatulids progressed in their develop- 

 ment, the closer they approached this goal as a rule, and it is therefore often more 

 difficult to distinguish later forms of different groups than is the case with earlier 

 ones. Gislen said that this explains also why the centrodorsal and radial ring in 

 certain recent forms of Comasterida and Mariametrida, although hi many instances 

 easy to distinguish from each other, in some cases can not be diagnosed with certainty. 

 The same is true with a number of younger fossil forms. There are types among the 

 forms from the Upper Cretaceous known only from the centrodorsal that can only 

 doubtfully be referred to the Comasteridae, or to the descendants of the solanocrinids. 

 It seemed to Gislen that the parvicavus group of the genus Glenotremites may be as- 

 sumed to belong to these latter because of the size of the cirrus facets. Presumably 

 the angelini group may be referred here too. Gislen believed that in Cypelometra 

 there is found a specialized descendant of the solanocrinids which as regards the 

 appearance of the cirrus facets has reached the mariametrid stage. 



The evolution oj the notocnnid and conometrid types. Gislen said that these two 

 form series may be treated most suitably in connection with each other. 



In the notocrinid series the most primitive type is Loriolometra, represented 

 among the species belonging to Glenotremites by G. arnaudi and the essensis group. 

 Loriolometra is continued by Sphaerometra , represented among the species of Glenotre- 

 mites by the paradoxus and possibly the rotundus groups. With the genus Semio- 

 metra the type has acquired so many new features that Gisle'n believed it necessary 

 to place it in the family Paleantedonidae. The series had originally rather few large 

 cirrus sockets arranged in columns, with a weak transverse crest, often peripherally 

 striated, protruding basals, broad free margin of the dorsal side of the radials, rela- 

 tively large radial cavity, a more or less distinct dorsal star, and, most important of 

 all, the specially characteristic deep radial pits in the centrodorsal for the reception 



