i893- THE PROBLEM OF VARIATION. 285 



conditions. For instance, the metamorphosis of tadpoles can be 

 prevented for a long time by certain appropriate conditions, which 

 would not be the case if the changes were entirely determined by 

 modification of the germ-plasm. 



A case which I have myself recently investigated experimentally 

 seems to me to support very strongly the theory of the inheritance of 

 acquired characters. 1 have shown that in normal flat-fishes, if the 

 lower side be artificially exposed to light for a long time, pigmen- 

 tation is developed on that side ; but when the exposure is com- 

 menced while the specimens are still in process of metamorphosis, 

 when pigment-cells are still present on the lower side, the action of 

 light does not prevent the disappearance of these pigment-cells. 

 They disappear as in individuals living under normal conditions, but 

 after prolonged exposure pigment-cells reappear. The first fact 

 proves that the disappearance of the pigment-cells from the lower 

 side in the metamorphosis is a hereditary character, and not a change 

 produced in each individual by the withdrawal of the lower side from 

 the action of light. On the other hand, the experiments show that 

 the absence of pigment-cells from the lower side throughout life is 

 due to the fact that light does not act upon that side, for when it is 

 allowed to act, pigment-cells appear. It seems to me the only 

 reasonable conclusion from these facts is that the disappearance 

 of pigment-cells was originally due to the absence of light 

 and that this change has now become hereditary. The pigment- 

 cells produced by the action of light on the lower side are 

 in all respects similar to those normally present on the upper 

 side of the fish. If the disappearance of the pigment-cells 

 were due entirely to a variation of the germ-plasm, no external 

 influence could cause them to reappear, and, on the other hand, if 

 there were no hereditary tendency, the colouration of the lower side 

 of the flat-fish, when exposed, would be rapid and complete. 



Weismann admits that hereditary colour-changes, such as a 

 darker colour in butterflies, are produced by climatic influences, but 

 explains this on the supposition that the climatic influences have 

 simultaneously modified in an individual the determinants of the 

 pigmented cells in the integument, and the determinants in the germ- 

 cells of that individual which belong to the corresponding pigmented 

 cells of its offspring. Now it is impossible to suppose that the light 

 acting on one side of a flat-fish would modify the determinants 

 belonging to one side in the ova of that individual, and even if it did 

 it would not cause the pigment-cells to develop symmetrically on 

 both sides of the larva, and then to disappear from one side of the 

 perfect fish. 



It is impossible or inconceivable on Weismann's theory that 

 changes of somatic cells should affect the determinants of the germ- 

 cells contained in the soma ; there is no connection between the 

 two ; but, on the other hand, it is equally inconceivable on his 



