SUMMARY OF CAUSES 105 



The experiments of Prof. W. L. Tower, in particular, suggest that 

 important external changes may provoke changes in the germ-cells 

 without necessarily affecting the parental body. He subjected full- 

 grown potato-beetles (Leptinotarsa) to peculiar conditions of tem- 

 perature and humidity during the time when the eggs were maturing, 

 and found that " mutations " occurred in a certain proportion of the 

 offspring. The parents were not affected, having passed the plastic 

 stage; and some of ih^ eggs were not affected at all. Moreover, 

 the same environmental peculiarity did not always produce the 

 same mutation in the offspring. But what Tower's experiments 

 forcibly suggest is this : that deeply saturating environmental changes 

 may serve to pull the trigger of germinal variability. 



2. It is also to be remembered that if the chromosomes stand 

 in some definite causal relation to heritable qualities, as seems 

 practically certain, then the maturation reduction of the chromo- 

 somes to one half their original number offers an opportunity 

 for variation. 



3. It is likely that fertilisation or amphimixis the intimate 

 and orderly union of two sets of hereditary contributions which 

 have often had very different histories will promote variation. 

 It is difficult to believe that it does not bring about new permuta- 

 tions and combinations. 



4. It is possible that variations may also arise in a less con- 

 ceivable fashion " bathmically," as the phrase goes for un- 

 known internal reasons. It is not absurd to suppose that the 

 germ-plasm grows from generation to generation, and, in growing, 

 changes because it is its nature so to do. 



Apart from variation of internal origin and positive modifica- 

 tion of external origin, we must remember that the offspring may 

 differ from its parents through non-expression of certain items 

 of its inheritance, the non-expression being due to the absence 

 of the appropriate liberating stimulus. This kind of deviation 

 may of course be obliterated next generation, when the full en- 

 vironment allows the latent character to re-express itself. 



