342 THE VOYAGE OF H.M.S. CHALLENGER. 



speak positively as to their nature. But they must certainly belong either to Actino- 

 metra quadrata or to Actinometra parvicirra. 



Remarks.— This latter variable and much be-named species is a somewhat isolated 

 one. It is separated from Actinometra trichoptera and Actinometra japonica by its 

 smaller number of cirri ; while Actinometra regalis and its allies have many more arms 

 which are united more or less completely by interradial plating (PI. LXVIII. fig. 2). 

 The only type which approaches Actinometra parvicirra at all closely is Actinometra 

 quadrata, which seems to be distinguished from it by the shape of its middle and later 

 arm-joints (PI. LXII. fig. 1). There are, however, one or two forms among those collected 

 at Samboangan by the Challenger which appear to approach Actinometra quadrata, and 

 it may be that the latter name will have to be abandoned. 



When describing the series of specimens which I called Actinometra polymorpha, I 

 put down the number of arms as ranging from thirteen to thirty-nine ; and among all 

 the many specimens of this type which I have examined during the last twelve years I 

 have found but one in which those limits have been exceeded. Bell, indeed, described 

 his single individual of Actinometra annulata as having forty arms, which would mean 

 that it has the full complement of ten distichal and twenty palmar axillaries ; but I have 

 found on examination that this is not the case in reality. Post-palmar axillaries occur in 

 but few individuals that I have examined ; and even when they are present I have only 

 once found the number of arms to exceed thirty-nine, owing to the absence of palmars 

 and even of distichal axillaries in other parts of the rays, as is well shown in PI. LXVII. 

 fig. 3. In this individual three of the primary arms do not divide at all, i.e., there are 

 only seven instead of ten distichal axillaries, and the deficiency of arms arising from their 

 absence is partly compensated by the presence of three post-palmar axillaries. The same 

 is true, though in a less degree, of two more Challenger specimens from Samboangan 

 (PI. LXI. fig. 5), and also of one from Batjan which I have seen in the Berlin Museum ; 

 but in one example from Banda which I must provisionally refer to this type, all the 

 distichal and palmar axillaries are present, together with four post-palmars in addition. 

 There are five more individuals in the Challenger collection which have distichals and 

 palmars but no post-palmars, and eight more in which palmars are not present at all, a 

 condition which also occurs in three of Semper's eleven examples, one of which has only 

 thirteen arms. So great a variation in the number of arms as this is certainly unusual, but 

 I have found myself quite unable to draw any fixed line of separation, often as I have 

 attempted it. It does seem, however, as if forty were the usual limit of the number 

 of arms in this species, even though post-palmars may sometimes be present ; and I 

 am inclined to lay more stress upon this as a character of systematic value than upon the 

 presence or absence of palmar or post-palmar axillaries. 



The potential dimorphism in the characters of the arms of Actinometra is very 

 well shown in this species. It presents itself in two large individuals dredged by the 



