BI-PARENTAL REPRODUCTION 83 



143. But now suppose the plant we are considering is 

 removed to entirely new surroundings, to a garden, for 

 instance. The conditions of survival are immensely altered. 

 Selection ceases, or, as artificial selection, changes its direc- 

 tion. Unrestrained by Natural Selection, sexual reproduction 

 causes, in many parts, rapid regression, which, in a few 

 generations, becomes easily appreciable. Restrictions on that 

 very ancient heritage of the race, 1 the tendency to vary, are 

 thus removed. Not only regressive, but progressive, variations 

 make their appearance, if not in increased numbers, at least 

 in forms of such magnitude as to be noticeable. The culti- 

 vator, not realizing that his hypothesis controverts the doctrine 

 of Natural Selection, leaps, of course, to the conclusion that 

 the alteration has been produced by the direct action of the 

 altered environment on the germ-plasm. He is in error. 

 One truth cannot contradict another. Species are closely 

 adapted to their environments; the only possible cause of 

 adaptation known to scientific men is Natural Selection. 

 Natural Selection could not bring about adaptation if the 

 environment, acting directly on the germ-plasm, were more 

 than a very insignificant cause of variations. 2 



144. In exactly the same way as sexual reproduction 

 planes away useless variations against tuberculosis or malaria, 

 so it eliminates every other kind of useless variation or organ 

 in every species of plant and animal in which the reproduc- 

 tion is bi-parental; whereas useful variations are left un- 

 affected. At any rate all useful variations which are of 

 common occurrence, such as those of an established structure, 

 are unaffected. Great abnormalities, or " sports " as they are 

 termed, though said by some biologists often to be prepotent, 

 tend to disappear, even if useful. 3 They are sometimes pre- 



1 See 160 and 163. 2 See 72-9. 



3 Of course, owing .to blended inheritance, a great abnormality is 

 liable to persist for more generations than a smaller variation. More- 

 over, in the case of great abnormalities, owing, presumably, to incom- 

 patibility in the germ-plasm, the inheritance is often exclusive. Thus 

 the offspring of Ancon sheep, inherited exclusively from one parent 

 or the other. (Animals and Plants, vol. ii., p. *70.) Conceivably 

 a great abnormality might be of such a kind as to be useful. But 

 this is highly unlikely. No doubt man has found one or two which 

 are useful to /iim, as in the case of the short-legged Ancon sheep above 

 cited, which could not leap fences ; but, as far as I am aware, no instance 

 is known of such an abnormality being useful in a state of nature to the 

 organism that produced it. The chances are almost infinity to nothing 

 against a large and elaborate new structure, a deformity, increasing the 

 adaptation to the environment. Apart from tumours most abnormalities 

 are examples of regression, not of progression. Probably even tumours 



