REPORT ON THE CRINOIDEA. 15 



The stellate condition of the centro-dorsal just described in Actinometra paucicirra 

 is sometimes reached by that of Actinometra parvicirra before the cirrus-sockets are 

 entirely obliterated. In one specimen of this variable type which was obtained by the 

 Challenger, the mature cirri have disappeared and are replaced by a few rudimentary 

 stumps, while the sides of the plate are so deeply hollowed by their sockets that its 

 outline is rather stellate than pentagonal. But it is still distinctly above the level of the 

 radials (PL LXI. fig. 3). 



The six examples of the large Actinometra nobilis which were dredged in the 

 Philippines also exhibit a considerable amount of variation in the characters of the 

 centro-dorsal (PI. LXV. figs. 1-6). In the least modified form it is a rounded pentagonal 

 plate distinctly above the level of the calyx, with traces of about ten cirrus-sockets, one 

 of which contains a very rudimentary stump, and a well-marked process at each of its 

 angles (fig. 2). In another specimen it is distinctly sunk below the level of the radials, 

 with which it remains united externally by the interradial processes at its angles ; but 

 its sides are bevelled away, and most of them bear indistinct cirrus-sockets, in one of 

 which a small stump is visible (fig. 3). In the other four examples, however, the 

 centro-dorsal shows little or no trace of cirri, and is distinctly concave on its dorsal 

 surface ; while it is completely enclosed by the radial pentagon, united to it by the 

 interradial processes at its angles, but separated from it by very distinct clefts along its 

 sides. Its shape, however, is more pentagonal than stellate (PI. LXV. figs. 1, 4-G). 



These clefts are rather deeper in Actinometra littoralis, though the centro-dorsal 

 retains its distinctly pentagonal form, and is about flush with the radials, with which it 

 is in contact by its lower angles (PL LXVII. fig. 1). On the other hand, in Actinometra 

 divaricata the centro-dorsal is very markedly stellate, and remains above the level of 

 the radials, the surface of which falls away considerably towards the sides of the centro- 

 dorsal, but not so much so as to give rise to definite clefts (PL LXIII. fig. 6). 



In the Challenger specimen of Actinometra typica, however, in the original type of 

 Phanogenia, and in others which I have seen, the centro-dorsal is both stellate and sunk 

 below the radials, so that there are very distinct clefts between the latter and its incurved 

 sides ; and no one would think from its present appearance that it had ever been a 

 cirrus-bearing joint (PL LVII. fig. 1). But in Loven's specimen the metamorphosis 

 was less complete, for a few cirrus-stumps are figured as still attached to the stellate 

 centro-dorsal, which is slightly above the level of the radials. The facts stated above, 

 however, concerning Actinometra paucicirra, Actinometra nobilis, and other forms, 

 entitle us to assume that cirri were really present in the young Phanogenia, so that the 

 genus ceases to have the extremely anomalous character which Loven not unnaturally 

 attributed to it. 



These clefts which occur at the sides of the centro-dorsal in Actinometra typica and 

 similar species must not be supposed to place the cavity of the calyx in communication 



