756 J. STANLEY GARDINER. 



Echinopora. I am aware that they are veiy doubtful, at least 500 or 1000 specimens 

 being required to establish any case, but I am of opinion that the phenomenon is much 

 commoner in corals than has been previously supposed. 



A species in one locality may show definite discontinuous variations in one or more 

 characters, but it by no means follows that the same discontinuous variations will be shown 

 by the same species in a different locality. The specimens may, indeed, show none of the 

 same discontinuous variations, though others may be clearly indicated. It is not improbable 

 that a large collection of one and the same species from many localities would appear to vary 

 in a perfectly continuous manner, and thus all idea of its true variability would be lost. 

 Work on a collection from a single locality is best calculated to bring out the phenomenon, 

 but one requires a collection from some other locality for comparison. This I have to some 

 extent had in my Rotuma and Funafuti collections, but unfortunately they are not nearly 

 large enough for the purpose. 



Mr Bernard's method of cataloguing the corals of the British Museum (Vol. iv.) appears 

 to me to be well calculated to throw light on discontinuous variability, but we require a still 

 further development of his system of tables, comparisons of two, three, or as many characters 

 as possible, those particularly that are capable of being expressed in numerical terms'. 

 Mr Bernard was in my opinion unfortunate in starting on Goniopora, though he has got 

 results of great value, because there are relatively few specimens of the genus in the Museum. 

 The fact that Mr Bernard is able to place the specimens of each single locality under so 

 many forms, whereas the species are in his present collections impossible to determine, points 

 to the probability of discontinuous variability being, as I have above inferred, from my own 

 collections. Later on, no doubt the species will be determinable with their various discontinuous 

 variations. It will then be possible to see how far the environment and the physical conditions 

 of the localities have affected the parent corallites to cause them to vary their gametes. For 

 this purpose sedentary organisms and particularly corals are peculiarly favourable, because they 

 can, so far as we are aware, in no way make choice of their environment, and must all be 

 affected in any region by the same physical conditions. Larvae of neighbouring localities, 

 such as Ceylon and the Maldives, would certainly be interchanged to some extent, but the 

 immensely greater number of coral colonies in a locality would presumably be produced from 

 larvae of the same locality. Yet, unless there is a real action of the environment in each 

 region in producing its peculiar discontinuous variations in the gametes of the new immigrants 

 it will be obvious that the species must remain approximately constant. 



The second class of variation referred to is the continuous. That it exists in a very definite 

 form is shown in some species below, but it is exceedingly complex and almost impossible 

 to satisfactorily distinguish from the vegetative. The physical conditions of the Red Sea, 

 Maldives, and Rotuma differ very markedly, and if their action is of any great importance 

 in natural selection, we should not unnaturally expect to find considerable differences between 

 the corals of the three regions. Yet the opposite is the case — more especially between 

 the Red Sea and the Maldives, which are the most different — very many species being 

 precisely the same, i.e. showing no dissimilarities in their specific characters. Yet, there 

 does appear to be a real difference. The Red Sea specimens of Astraeidae, considered in 

 their entirety, have larger corallites and calices, and are smoother with less exsert and less 

 toothed septa than the Maldivan, which in the same characters lie intermediate to those 



' Anyone working at the question could to a large extent do this for himself from Mr Bernard's last catalogue. 



