142 FOSSIL CRINOIDS. 



waiving Poterioa-inus as a transition form — are without pinnules, just as the 

 entire suborder Flexibilia Impinnata is ; although there is an occasional tendency, 

 as in Botryocrinus, to develop pinnules through ramules with short intervals. 

 But of the genera which established the full straight articulation, every one of 

 them has true pinnules ; with doubt as to Cupressocrinus of the Devonian — 

 a highly specialized form, more nearly related to some monocyclic genera. And 

 among these we must include Poteriocrinus, which was struggling hard to get 

 into the more vigorous group. A couple of other instances of intermediate 

 genera show how strongly these two structures correlate : — Most of the typical 

 Silurian to Carboniferous Cyathocrinidae (non-pinnulate) have the posterior 

 side symmetric, with only a single anal plate; while most of the Poteriocrinidae 

 (pinnulate) have an unsymmetric anal side, with a radianal. Now Parisocrimis, 

 although having a Poteriocrinus anal side, has the round facets of the Cyatho- 

 crinidae ; — and with them the pinnuleless arms of that family ; whereas Graphio- 

 crinus, which has the symmetric Cyathocrinus anal side, has the perfect straight 

 facets of the highest developed Poteriocrinidae ; — and along with them it has 

 pinnules in abundance. So there must be a powerful association of these two 

 characters to overshadow the third, which is usually a very important one in 

 classification. 



It is also an interesting fact in this connection that Poteriocrinus has usually 

 very small pinnules, as compared with those of the genera with complete articu- 

 lation, in some of which — e. g. Decadocrinus — they become almost large enough 

 to be called ramules. In specimens of Poteriocrinus as usually found imbedded 

 in firm limestone, where the matrix has to be worked off with tools, we usually 

 do not see the pinnules at all (see PI. IV, figs. 1, 2). When the matrix is soft, 

 they can be exposed, e. g. Plate VI, fig. 1, where it may be seen how small they 

 are compared with the size of the arm. In Plate V, fig. 2, they appear relatively 

 longer, as is always the case, towards the distal end of the arm. They are finely 

 preserved in the young specimens, figs. 4 and 5, Plate IV, showing their dimin- 

 utive size even in that stage, when they are usually relatively large. Thus it 

 would seem that the correlation of the two characters, viz, round facets with 

 non-pinnulation, remains in force in case of modification; and that a tendency 

 towards straight articulation by adding a transverse ridge to the round facets 

 is accompanied by a weak pinnulation. This would seem to hold good now; for 

 in Recent Comatulids those with very long pinnules have the most perfectly 

 developed articulations. 



Hence it may be stated in general: — 



