26 MADREPORARIA. 
cases of very close work, when, the locality being fixed, the differences in chronological order 
of appearance are the simplest and most certain of the differentiating facts, those which lend 
themselves best to the purpose required. Whether any designation according to horizons will 
ever be required or not, time only can show. The principle on which the method is based, 
viz. the adoption of position on the earth’s surface as a factor in the designation of specimens, 
will doubtless admit of very varied application. No hard-and-fast rule is proposed or even 
necessary, for if the method of application in each case is made clear, the synonymy ought to 
be easy to work out. Actual trial, however, can be the only test. 
Beyond this comment, I have merely to record the results of my further experience of 
the method after applying it to a large, complicated genus like Porites. 
It is obvious, with such a genus as Porites, with its endless minute variations, that the old 
method of grouping into “species” would come in most conveniently, but at the cost of all 
scientific worth.* To sort the specimens into groups of tens or twenties, to which a name 
is given, is easy so long as one is satisfied that one possesses insight into genetic relationship 
in the absence of all evidence. Iam aware that resemblance is often said to be evidence 
enough for the establishment of groups which may be called “species,” so long as we do 
not strain the genetic connotation of the word “ species,” and that works so written will always 
have a classificatory value. But this is openly and avowedly to throw away the modern ideal 
of a natural classification, and is hardly justifiable in an age dominated by the doctrine of 
evolution. It is surely more scientific to hold by this ideal, however uphill the task involved, 
than to attempt short cuts. But this assumed genetic value of resemblance is not justifiable 
in dealing with the more plastic forms of life. In highly specialised groups with complicated 
organs, resemblance is, undoubtedly, in the majority of cases, justification for assuming 
genetic affinity. The chief pitfall to be avoided is that due to what is usually known as 
“convergence.” But with very plastic forms, without any very specialised and complicated 
organ formation, the case is different. In the first place, the resemblances and differences 
are subtle and difficult to define; to the entirely untrained eye the specimens are alike, and 
only as our analysis grows more and more profound do we begin to see the real characters 
of the differences. Here, then, is at once an argument against any attempt to group until we 
feel confident that we have a real insight into the structural principles discoverable in the 
group. But, further, it is just in the case of the plastic forms that convergence is most 
frequently found. They seem to be more easily moulded to the environment, so that 
genetically related forms, dispersed into slightly different conditions, quickly become different, 
and genetically different forms, cast into the same environment, quickly become alike 
(Vol. IV. p. 188). 
* This conclusion has been still further forced upon me. It was during my preliminary efforts 
to arrange this difficult genus that the impossibility of applying to it the usual system first brought 
me to a stand in this matter. With small genera it seemed possible to effect a compromise. But 
not so here. Many remarks will be found scattered up and down the Volume to show into what 
pitfalls coral workers have been led by attempting to establish and to identify species on the old 
method (see, e.g., the discussion of the Red Sea “ species” of Milne-Edwards). 
