Evolution and Symmetry in the Order of the Sea-pens. 119 



genus, five species have been described without any axis ; in five species the 

 axis is incomplete or variable, and in three species the axis is said to be 

 complete. In the description of Cavernularia hahereri from Japanese waters, 

 Balss states that, on examining a number of specimens, some were found to 

 have a short axis of variable length, and others to have no axis at all. As 

 the axis in this species is found to be variable, even in specimens from the 

 same locality, it seems probable that the length of axis cannot be taken as a 

 sound specific character even in the diagnosis of a species of this genus. 



As regards the shape of tlie axis in these forms, the evidence is, perhaps, 

 not sufficient to justify any general conclusions. It has been described as 

 round in section, square, square with fluted sides, and in one species at least 

 (Z. hicksonii) it is square above and round in section below. 



The axis then seems to be a character which is variable both in length and 

 shape in the radially symmetrical Veretillidae, but attains to its full length in 

 forms that show an incipient bilateral symmetry, as in Lituaria plialloides, 

 and its full length and definite shape in the higher forms of Pennatulids, or 

 to disappear entirely from all specimens that show an incipient bilateral 

 symmetry, as in the genera Echinoptilum and Kenilla. It is clearly an unre- 

 liable character for systematic purposes in the Veretillidcie, but may be a 

 reliable character in other families. 



The calcareous spicules that occur in the wall of the rachis and in the 

 body wall and tentacles of the zooids have for many years been relied upon 

 to afford good characters for the diagnosis of the species of the Pennatulacea, 

 but the development of our knowledge of these characters shows that there 

 are differences in the degree of variability of the spicules similar to those 

 found in the case of the axis. In the genus Pennatula, for example, the 

 spicules found in the rachis and zooids are always long, narrow spindles, 

 showing three longitudinal curved flanges (the " dreifliigelig " spicule of 

 Kiikenthal). They vary in length according to position, and they vary in 

 number and thickness according to position and according to the species, but 

 they are always three-flanged. In the genus Pteroeides the spicules, although 

 varying in length and thickness according to the position and species, are 

 always smooth, long rods or spindles without the three flanges. Again, to 

 take one more example from the higher bilaterally symmetrical genera, in 

 the genus Scytalium the spicules are always very small, flat discs (about 

 0"05 mm. in diameter), round, oval, or dumb-bell-shape in outline. 



In other families of bilaterally symmetrical sea-pens, we find examples, 

 such as Virgularia, in which there are never any spicules in the rachis or 

 zooids, and Stylatula, in which spicules are always present in a certain 

 position in the colony. In another genus, Umbellula, with less pronounced 



