2l6 



served. It is by no means constant, however, some specimens,, 

 particularly those in very shallow water, being devoid of it. 

 The " brilliant purple " colour of this species is particularly in- 

 teresting, and is deserving of further investigation. 



The colour character by itself is not one which would justify 

 us in constituting a new species for the specimens. 



The Mossel Bay Alcyonium, however, differs from other 

 species in a variety of other characters. 



It resembles A. antarcticum in the hemispherial shape of some 

 of the colonies and in the general transparency of the 

 coenenchym. It differs from it in the general character and 

 distribution of the spicules and in the absence of a stalk. It 

 comes nearest, perhaps, to Alcyonium sphaerophorum Ehrb., as 

 described by Klunzinger (8), but differs from it in the greater 

 size of the lobes, the entire absence of a stalk, the absence of 

 smooth calcareous corpuscles, and in the shape and size of the 

 spicules. Ehrenberg described the colour of his species as 

 " pallida, polypis fuscis." 1 his is a good description of the 

 colour of A. purpureum after the purple colour has been 

 washed out. 



In all the specimens the zooids are partially expanded. I 

 have little doubt that if they had been carelessly killed and pre- 

 served, their resemblance to Alcyonium sphaerophorum would 

 have been closer than it is. 



The largest colony sent to me is attached to a Tunicate test. 

 It is roughly hemispherical in shape, 40 mm. in diameter, 20 mm. 

 ni height. It springs from a flattened incrusting base bearing 

 polyps, there being no stalk (sterile Fiiss), as in A. fulvum 

 Forsk. There are fifteen lobes, varying from 15 mm. — 8 mm. 

 in diameter. The zooids are large, being in the specimens nearly 

 I. mm. in diameter at the base of the anthocodiae. The other 

 colonies are much smaller, and consist of only one or two lobes. 



The six ventral mesenteries are well developed. The figure 

 given by Kiikenthal (11), in a recent paper, of Alcyonium fulvum 

 Forsk. has some resemblance to our new species, but in the 

 size and character of the spicules it is clearly distinct. 



The general coenenchym and the body walls of the antho- 

 codiae are singularly transparent. At the surface of the 

 coenenchym the spicules are closely crowded together, but a 

 little way below the surface they are isolated and scattered. At 

 the base of the tentacles there are eight triangular shields, 

 pointed above, of scattered spindle-shaped spicules. The ten- 

 tacles themselves are dark brown in colour and entirely devoid 

 of spicules. The spicules of the coenenchym vary considerably 

 in shape, but are mostly double clubs or double balls with a nar- 

 row neck. Their average size is about o' i mm. The spindles and 

 clubs of the anthocodial body- wall are about 0T4 mm. in length. 

 All the spicules are provided with numerous protuberant 

 tubercles (fig. 18.) 



