MODIFICATION OE SELECTION. 177 



stance, in the composition of the water of a lake or a river, will 

 not affect the fauna inhabiting it equally and as a whole, but 

 will act on individuals; some will bear the change without 

 being in any way aflfected by it ; others will die, whUe others 

 a^aui will survive, but their habits of life will be changed, and 

 at the same time their structure will be modified, as in the 

 case of BrancMpus and' Artemia. Thus the constancy of the 

 aquatic fauna of any spot depends on the constancy of all the 

 external conditions of life prevailing there, and every change, 

 however small, in these wUl effect, a selection among the old 

 forms, facilitate the introduction of new ones, and sometimes 

 even determine the transformation of one species into another. 

 On this last and most important point we at present certainly 

 know very little; but the old experiments of Beudant, the 

 more recent ones of Plateau, and, above all, those of Schman- 

 kewitsch, show that this absence of information cannot be 

 adduced as a convincing argument against the assumption that 

 careful experiments directed to this question must tend to 

 prove that stagnant water and the substances contained in it 

 can exercise a far more direct transforming influence than has 

 hitherto been considered possible. 



9 



