of a rose, the leaves of which are sonietunes keeled, seated on and projecting directly above 

 the radiated basal expansions. In some cases there may be but two of these foliated expansions, 

 or even but one; but they can always be easily distinguished by the fact that their surfaces 

 are smooth, instead of being notably tubcrculate as are the basal expansions. The resulting 

 spicule, then, consists of a basal set of branched or foliated, densely tuberculated expansions 

 extending downward and outward in a radiate manner; and a distal exposed portion set imme- 

 diately above the basal part, and consisting of erect inflated lamin;e which are often curved or 

 crimped, but the surfaces of which are smooth. This type is very characteristic of the genus 

 Echi?togo>'gia, and will be called the "■ Echinogorgia type" . (See plate XXI, figs. 9, « & h). 



A fourth form of what I call generic types of spicules is quite different from either of 

 the preceding. A simple, rather heavy, tuberculate spindle developes rounded projections on its 

 upper side only. These projections finally become quite pronounced and sharp-pointed, forming 

 a more or less regular row of tooth-like points on one side of the heavy spindle. The opposite, 

 or under, side is simpl}^ tuberculate, without pronounced tooth-like processes. These heavy 

 unilateral spindles are characteristic of the genus Tkcsca, and will be called the " Tliesea type' . 

 (See plate XXI, figs. 2 & Sa). 



A fifth form, is carefully described by Wright and Studek^ as follows: — 



"The polyp spicules are remarkable; from the spindle-like bases one or two foliar expan- 

 sions arise. A somewhat bent spindle is furnished with a dentate foliar expansion, from the 

 margin of its convex side, the long axis of which slants a little obliquely to the long axis of 

 the spindle. In most cases the spicule has become triradiate by the shooting out of several 

 projections from about its middle. 



From the place wdiere these rays centre one or too dentate foliar expansions arise, these 

 latter stand at an acute angle to one another, and their long axes always fall into the angle 

 between two of the rays. These spicules are so placed on the wall of the polyp body that the 

 stellate rays are always inserted into it, while the foliar expansions project, scale-like, over 

 the surface". 



This type, evidently allied to the Echinogorgia tjpe, will be called the '^ Acaiiipfogorgia 

 type\ (See plate XXI, fig. 12, rt' & b). 



A sixth form is based, like the others, on the primitive spindle; but is produced by a 

 flattening and lateral expansion into a disk-like or scale- like form which is regular in outline 

 in some species, but exceedingly irregular in others. 



It consists of a flattened centre from which complexly branched and irregularly foliate 

 lateral expansions arise, giving the typical " Stachelplatten" of authors. 



These occur in their most characteristic form in the genus Placogorgia, and will constitute 

 the "■ Placogorgia type" . (See plate XXII, figs. 5 & 10). 



A seventh type is probably the most peculiar and distinctive found in the Jlliiriceidce. 

 It is essentially radiate in structure, and consists in a basal expanded, or disc-shaped, or radiately 

 branched part; and an upper or distal expanded knob or disc, the two being united by a very 



' Wright and Studkr. Challenger Report, the -Mcyonari.i, i88g, p. 115. 



