ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 61 



The fact already mentioned (sect. 38), that in two cirrhi in which the number of 

 segments is different, the character of the first six or seven is the same, would seem to 

 point to the same conclusion. 



(iii.) The Centrodorsal Plate. 



(§ 41) In all the Actluometrm with which I am acquainted, the external appearance 

 of the " Knopf," or centrodorsal piece, is very characteristic. Like the cirrhi which it 

 bears, it is far more constant throughout a considerable range of species from very 

 various localities, than it appears to be in the individual members of a single species, 

 both of the European and of some of the foreign Antedons, even when existing in the 

 same locality. 



Thus, for example, the centrodorsal piece of Ant. celtiea may either be very shallow, 

 flattened, and bluntly rounded off at the base, where it was originally united to the 

 joint of the stem next beneath it, with only two rows of sockets ( V) for the attachment 

 of the cirrhi (PI. IV. fig. 1) ; or it may be deep and nearly hemispherical, with three or 

 four alternating rows of cirrhus-sockets, and terminating inferiorly in a flattened circular 

 or rudely pentagonal base (fig. 6) ; or, finally, it may have the form of a five-sided 

 pyramid, the apex of which is directed downwards and very slightly truncated, while 

 the sides bear three or four alternating rows of indistinct cirrhus-sockets (PI. IV. fig. 8). 

 This last character indicates that the animal had attained a considerable age ; and as all 

 the specimens of this kind which I have seen have been smaller than normal specimens 

 of Ant. celtiea, and exhibit slight differences from them in the characters of other parts 

 of the skeleton, I am disposed to regard them as dwarfed varietal forms rather than as 

 young and incompletely developed individuals of the ordinary type. 



Again, of the two specimens of Ant. macrocnema, Val., in the Paris Museum, both of 

 which were brought from New Holland by Quoy and Gaimard in 1829, one has a large 

 and hemispherical centrodorsal piece, while in the other it is a short pentagonal or 

 nearly circular column, on which the cirrhi are disposed in three or four alternating 

 rows. 



These instances, which might be greatly multiplied, suffice to show that the centro- 

 dorsal piece of Antedon may vary very considerably in its external appearance. In Acti- 

 nometra, so far as my experience goes, it is almost invariably a flattened circular, or 

 rudely pentagonal disk, somewhat hollowed in the centre of its dorsal surface, and with 

 low sloping sides marked out into distinct sockets for the articulation of the cirrhi 

 (PL V. figs. 1, 6, 15, and PL VI. figs. 1, 2, 7, 14, 16, 20). It generally conceals more or 

 less of the first radials which rest upon it (PL II. figs. 9-11, cd), but it may sometimes 

 be very irregularly extended so as to conceal a considerable part of one or more of the 

 second radials (PL II. fig. 8, c d). 



As a general rule, only one row of cirrhus-sockets can be traced ; but in the large 

 Act. robusta (PL V. fig. 15) I have found two alternating rows of sockets to be distinctly 

 visible, and even traces of a third row, so that the dorsal surface of the plate becomes 

 dightly convex, though by no means to the same extent as in most Antedons. In fact, 

 the flattened plate-like condition of the centrodorsal piece, and the existence upon 



9* 



