THE CAUSE OF VARIATION 91 



facts of heredity in general. In particular may be cited the 

 congruous fact that twins are often much more alike than 

 brothers or sisters born at long intervals of time. It also 

 affords a simple explanation of the evil effects of close in- 

 breeding, since closely related cells would be apt to resemble 

 each other in the excess or defect of particular factors. It is 

 applicable not only to the higher animals and plants, but to 

 all forms of life which are reproduced by cell-division, including 

 the monads, those lowest forms in which the history of the 

 cell is the whole life-history of the race. 



If there is any force in these considerations, it would not 

 be surprising if some check were needed to counteract the 

 strong tendency that we here find towards variation. The 

 germ that nature has elaborated by slow additions through 

 millions of years of unbroken descent in a continuous line from 

 simple beginnings, if any form of matter can be called simple, 

 to its immense present complexity, has to be guarded against 

 a too rapid change in an environment that changes slowly, 

 rather than to be stimulated into variation. To cope with 

 our environment we must come extremely true to a type which 

 is proved by ages of survival to be the fittest. That sex is 

 a device by which such stability is increased is a thesis which 

 I have previously argued elsewhere, and need not here repeat. 

 It is one that appears to me to be supported not only by reason, 

 but by abundant inductive evidence. Briefly the means by 

 which sex effects a reduction of variations is the halving of 

 those differences which we have seen to be established by 

 repeated fission. Except where the two conjugating germ- 

 cells are closely related owing to in-breeding, it is unlikely 

 that the same factors will be much in excess or defect in both ; 

 and a factor that is much in excess in one must therefore 

 be assumed to be at about mean size in the other, so that that 

 excess will be halved in the fertilised cell. This halving appears 

 to be ensured by the " reducing division " of the cells which 

 precedes their conjugation. 



In this differential division of the cell, if it be an agent in 

 producing somatic variation at all, we have a cause which is 

 certainly constant, and so far as concerns creatures which 

 are reproduced by the fission of cells, one which is also 

 universal. It will be worth while to notice into what troubles 

 and fallacies we are led by the refusal to recognise some such 



