392 Mr. P. H. Carpenter. Preliminary Report upon [Mar. 6, 



in an. alteration of the relations of the different surfaces of the radials 

 to one another which takes place during their growth. This is the 

 conclusion to which I have been led by an examination of the 

 separated radials of young and adult examples of Act. Jukesii ; but it 

 entirely fails to account for the stellate form of the centrodorsal in 

 Act. stellata and in Phanogenia. This feature appears to me to be a 

 further development of a condition which I have already described in 

 Act. pectinata ;* but I do not expect to get abetter understanding- of it 

 until I am able to separate the parts of the calyx, and also to make 

 sections through it. Both of these modes of research are at present 

 unavailable, owing to want of material. 



The appearances presented by the dorsal surface of an isolated first 

 radial are very different among the different species of Comatula. In 

 many Antedons such as Ant. Eschrichtii, the whole of this surface rests 

 upon the centrodorsal, and except for the edge separating it from the 

 distal articular surface there is no external indication of the presence 

 of a first radial at all, as the second seems to be in direct contact with 

 the centrodorsal. The superior or ventral surface of the latter slopes 

 downwards from its circumference towards the centre. 



In such species as Ant. macrocnema, however, and in most Acti- 

 nometrce a dorsal view of a first radial shows two surfaces inclined to 

 one another more or less obtusely. One of these appears externally 

 and is the true or outer dorsal surface of the radial. It is often 

 marked by a median dark line which extends outwards over the other 

 radials far on to the arms. The other, or inner dorsal surface, is the 

 surface of synostosis with the centrodorsal plate, and may be at right 

 angles to the outer surface when the ventral face of the centrodorsal 

 is perfectly fiat as in Ant. macrocnema. But in Actinometra it is 

 always placed at an obtuse angle to the outer surface, for the ventral 

 face of the centrodorsal on which it rests slopes downwards and out- 

 wards from the centre to the circumference. I have examined the 

 separated radials of two specimens of Act. Jukesii, one young with a 

 centrodorsal still marked by cirrhus sockets, and the other full grown 

 with a large discoidal centrodorsal within the radial pentagon, and 

 below the level of its outer surface when viewed from its dorsal aspect. 

 There is a considerable difference in the relative sizes of the inner and 

 outer portions of the dorsal surface of the radials in these two cases. 

 The absolute length of the outer dorsal surface seems to increase very 

 little after a certain stage of growth is reached, for it is nearly the 

 same in the large specimen as in the small one, but the inner or 

 synosteal surfaces of the two differ very greatly in size. This surface 

 is not only absolutely, but also relatively larger in the older specimen, 



* See cap. vi, sect. 61, of my memoir on Actinometra, now in course of publica- 

 tion in the ' Transactions of the Linnean Society." 



