218 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 105 



evident that Graphularia is very close to Stylatula and should be 

 removed from the Pennatulidae, in which the axis structm'e is clearly , 

 different in detail. 



When the axis structure of all modern pennatulids has been studied, 

 it may be possible to reinterpret the fossil genera and to assign them 

 to their proper positions in the biological system. 



SUMMARY OF AXIS STRUCTURES 



Among the gorgonians with calcified axes, the calcareous material 

 is oriented in two different ways. In the families Chrysogorgiidae, 

 Ifalukellidae, and Primnoidae it permeates the scale-like or lamellar 

 horny sclerodermites, which are concentrically deposited and tightly 

 imbricated in each layer. No trace of radial orientation can be 

 seen either in transverse or in longitudinal sections. In the Elli- 

 sellidae (pi. 8,6) and Isididae, the former with a continuous and the 

 latter with an articulated axis, the calcareous fibers are radially 

 oriented and grouped to form sclerodermites more or less suggestive 

 of those seen in the massive coralla of the Scleractinia. The scleroder- 

 mites forming the calcareous axis internodes of the scleraxonian 

 Parisis (pi. 8,e) are radially arranged immediately adjacent to the 

 axis core and assume a longitudinal direction (but still in radial rows 

 corresponding to the surface grooves) in the secondary thickening of 

 the axis. They are continuous with the spicules of the horny nodes 

 and are nothing more than modified spicules. The axis of Isis (pi. 

 8,c~d) differs only in the predominantly radial arrangement of sclero- 

 dermites, which are not essentially different from those of Parisis, 

 and in the absence of spicules from the horny nodes. The axis 

 structure of the Pennatulacea corresponds very closely with that of 

 the Ellisellidae and Isididae. 



In those forms with radial orientation of the calcareous fibers the 

 axis substance is concentrically stratified and it therefore is apparently 

 deposited by the cells of a continuous tissue, the axis epithelium, just 

 as it is in other gorgonaceans. It seems necessary to suppose that 

 groups of adjacent axoblasts in the axis epithelium are functioning as 

 units for longer or shorter periods, depositing the radially oriented 

 sclerodermites. The relationship of the horny material to the cal- 

 careous substance has yet to be thoroughly investigated, but it 

 appears that any scleroblast has the ability to produce both substances. 



i 



PHYLOGENETIC IMPLICATIONS 



Assuming that the sclerodermites of the Isididae are structurally 

 the same as those of the Parisididae, and from all indications such is 

 indeed the case, the articulate Holaxonia and Scleraxonia are phylo- 

 genetically close rather than far separated as end-points of different 



i 



