ii8 



which to grasp it — while nothing may be known of its habitat, 

 its locality alone being broadly indicated. Most collections, 

 too, if they contain many colonial forms, are small in speci- 

 mens, so that variability cannot be properly studied. Lastly, 

 it must be recognised that sedentary animals vary in accord- 

 ance with their surroundings in the same way as do trees 

 and other plants. 



Having determined which are the species in any collection, 

 it is necessary to examine into the question as to whether 

 these species are already known or new. This is a far more 

 perplexing and arduous task.* So far as my experience goes, 

 I find that there are in each genus only a limited number of 

 characters, which vary in a species-making manner. Most 

 of the other characters are those of the family and genus, and 

 require no particular remark. The rest are of but small 

 importance, and belong rather to the individual than the 

 species. They become eliminated necessarily as a larger and 

 larger number of examples is studied. In the earlier descrip- 

 tions the recorded characters often do not include those on 

 which the species are necessarily founded at the present day. 

 Later workers may have recorded these from an examination 

 of the same specimens, or elucidated them from others. In 

 the latter case there must always be some element of doubt, 

 but this is unavoidable. Again, if the variability owing to 

 enmronment vci^.y be expressed as from i to loo, the original 

 specimens — especially if few in number, as is usually the 

 case — may be placed between i and lo or 90 and 100, while 

 the great bulk of subsequent collections will be about 50. 

 Specimens may in the first place be put as species at every 

 tenth, but as more specimens are examined the intermediate 

 forms must be necessarily joined together, until finally the 

 limits of the real species are elucidated. 



The personal element of each worker in the field is — and 

 must always be — enormous, but the remedy lies in the 

 systematic study of the normal variation of different species, 

 particularly in relation to the physical and biological charac- 

 ters of its environment. It may then be possible to separate 

 the variability of the species per se from that due to external 

 causes, by this means possibly arriving at some more accurate 

 ■conception of the formation of the species themselves. 



* Vide *' Some Fossil Corals fiom the Elevated Reefs ot Curacoa, etc.," by T. 

 Wayland Vaughan, Samm. des Gaol. Reichs-Mus. Leiden, ber. II., Bd. II., 

 Heft I. (1901). 



