1894.] ECHI3s T 0DEEMS OF MA.CCLESEIELD BAKE!. 393 



In fact I have had forced on me the conviction that Macclesfield 

 Bank is a nursery ; with a rim submerged 9 fathoms 1 beneath the 

 surface any pelagic larvae that will can enter within its boundary ; 

 being 76 miles long and 36 miles broad, it affords some opportu- 

 nities for the larvae to settle, and the average depth within the rim 

 is from 40 to 50 fathoms. 



I may therefore suggest that it is of great importance that 

 as full collections as possible should always be made in areas 

 resembling the inside of this reef, for not only are questions of 

 synonymy to be by this means settled, but the more interesting 

 study of the changes that occur during growth can only be 

 carried on with extensive series, the relationships of species can 

 be more satisfactorily considered, and material be brought together 

 of immense value to the morphologist. 



It would be quite possible so to arrange the material for this 

 paper as to make it of intolerable length, but I think I can so 

 dispose of it as to bring all I have to say within reasonable 

 compass. 



I propose to give three lists : — (A) of the species from N".W. 

 Australia; (B) of those from the Araf Lira and Banda Seas; and (C) 

 of those from the Macclesfield Bank, in all of which the observed 

 range in depth will be duly noted ; after each species I place the 

 name of the author who first described it. After these lists I give 

 notes and descriptions in systematic order. 



The point of greatest interest is the discovery that the sygygial 

 joints at the bases of the arms of Comatulids by no means 

 exhibit the regularity which is ordinarily believed to be one of 

 their chief characteristics and their best claim to be used as aids 

 and guides to the grouping of the species — that is to say, they 

 have been taken as being very much more valuable than mere 

 specific characters. Antedon bassett-smithi (see p. 399) shows how 

 the syzygies may vary in one individual specimen; while the broken 

 and nameless Actinometra (see p. 402) is positively appalling to a 

 student of Comatulids, for it has no syzygy on either second or 

 third brachial. 



Where there is no large series it is very difficult to speak with 

 any confidence as to the significance of unexpected irregularities, 

 which may, of course, be merely individual, but, in systematic 

 zoology, we have to beware always of the influence of general- 

 izations based on material which is always becoming proportionately 

 smaller. The general acceptation of Dr. P. H. Carpenter's 

 classification removes the " idol " from the category of " idola 

 speeus" to that of "idola fori"; but we must not only remember, 

 we must always keep before ourselves the doctrine of Macleay that 

 " no character is natural until it has been proved to be so." 2 



This is, of course, saying in as many words that there is no 



1 See Bassefct-Smith, Ann. & Mag. Nat. Hist. vi. (1890) p. 356. 



2 See J. D. Macdonald in Trans. Linn. Soo. xxiii. p. 75. Pupils of the late 

 Prof. Eolleston need not be told whence I derived my knowledge of this 

 principle. 



