for Systematic Biology. 425 



more difficult than marking out genera, or genera than families, 

 etc. The arduousness of our task of investigating the species is 

 only thrown back a stage further with increase of intricacy. 



The present method does not in truth fit in with all our re- 

 quirements, but I deny that Mr Bernard is in any way providing 

 us with a better. His method takes account of locality alone. 

 The apparent supposition is that the species in any locality will 

 not be able — on account at least of failure of opportunity — to 

 breed with those of other of Mr Bernard's localities. The anti- 

 cipation is that the species of diverse localities may by natural 

 selection, acting on dissimilar variations, have formed and be 

 forming new species. Thus the method would discover in different 

 areas the lines of the evolution. These views are certainly not 

 supported in the Atlantic and Indo-Pacific Oceans by the dis- 

 tribution of those marine organisms, which possess free-swimming 

 larvae. The instances, too, of other forms (not possessing such 

 larvae), varying in dissimilar characters in different regions, are ex- 

 tremely scarce. Further, so far as we know, the normal variability 

 in any one locality, where the conditions of life are luxuriant and 

 diverse, is as great as over the whole of either of our two oceans. 

 That this is the case in corals, I am convinced by my experience 

 of Madreporaria in the Maldives, where nearly every variety of 

 habitat (in which they can dwell) may be found. I may perhaps 

 point to the indubitable norma] variability of Flabellum rubrum 

 (= Flabellum irregulare, etc.), as well as to the probable specific 

 variation of this same coral in a single channel in the Philip- 

 pines 1 . 



" To arrange the organic kingdom in the order of its evolu- 

 tion " and " to arrange each particular form in its order of 

 development above the forms from which it can be derived and 

 below those to which it has itself given rise" are Mr Bernard's 

 ideals of classification. " The facts will themselves reveal the 

 true species in process of time." Once more we seem to be 

 brought back to the mire of a bygone age. We are to deal only 

 with facts, while to a future generation the power of working 

 miracles is to be given. All classification — in its broadest sense — 

 is at present and must, so far as we can foresee, remain theory, 

 but it is to be brought into the realms of creed. From the 

 nature of the case there can be no finality of form to our classi- 

 fication so long as the branches and twigs of our tree of life 



1 Vide Semper, Zeit. f. wiss. Zool. Bd. xxn. pp. 242 et seq. (1872), and papers 

 by the author on the Cape of Good Hope corals in Marine Investigations in South 

 Africa, and on variation, protandry and senescence in Flabellum rubrum and other 

 corals in Proc. Cavib. Phil. Soc, both in the Press. 



In the above paragraph I am dealing with the facts as at present recognised, 

 but it is necessary to observe that the presence of specific variation in the 

 Madreporaria is not as yet generally known. 



