CLASSIFICATION 115 



united to adjacent brachials and to each other, the junction being marked by 

 distinct zig-zag lines. The lateral faces of the brachials which abut on the 

 posterior interradius are angular for a greater or less distance, and exhibit 

 surfaces for sutural attachment to interbrachials of one kind or another high 

 up between the rays. Hence the anal plates form an integral part of the calyx 

 wall, in the same manner as those of the regular interbrachial areas. The pos- 

 terior basal is flush with the other plates, and does not differ from the other 

 basals except in being angular or truncate above, and in supporting other plates 

 which are in sutural connection with it. 



In Taxocrinus the posterior basal is entirely different. It supports no 

 plates at all in direct succession from its distal face, but at a point on the outer 

 surface considerably below the distal margin there is a shallow median depres- 

 sion, or socket, in which rests the first of a series of elongate pieces, curved 

 like arm-plates, forming the support of an anal tube (Pis. XXI, figs. 4, 5; 

 LVII, fig. 5). When all the plates are in position we do not see the entire 

 dorsal side of the posterior basal in its upper half, but only such narrow por- 

 tions of that part of it as are exposed at either side of the tube, usually more 

 to the left than to the right (PI. LVII, figs. 4, 5, 6a). This tube is attached to 

 the perisome, which forms its inner wall, and it is connected with adjacent 

 brachials by the perisomic integument of pliant skin paved with small irregular 

 plates or spicules. The tube plates are not suturally connected with adjacent 

 plates. Above the tube socket the posterior basal is thinner than below, and its 

 distal face becomes a rather thin and rounded edge, without any surface for 

 sutural union. Thus the lateral faces of the brachials bordering the anal area, 

 instead of being angular and having planes for sutural attachment, are curved 

 from the outside inward to the ventral margins where the perisome is attached ; 

 and these margins follow a straight line longitudinally, while the margins fac- 

 ing the other interbrachial areas are zig-zag for their sutural union with the 

 corresponding angular interbrachials. Plence the two can be instantly distin- 

 guished if the lateral margin of the ray bordering the anal area can be seen. 



These two structures prevail to a greater or less degree throughout their 

 respective groups. If we could see the exact details of all the specimens we 

 should probably always have a sure criterion in the condition of the posterior 

 basal. If its distal end is suturally united to a succeeding plate, then the form 

 possessing it does not belong to group B, because the tube does not originate 

 there but at least one plate higher up. When the tube originates on the pos- 

 terior basal it rests in a rounded socket, either at the top or at some distance 

 below the top ; and in this case the upper part of the plate thins out to an edge 

 for attachment of the perisome forming the ventral wall of the tube. This 

 structure is perfectly shown by carefully prepared specimens of Taxocrinus, 



