496 THE GRAMMAR OF SCIENCE 



In all cases I have supposed equipotency of both 

 sexes and no sexual selection. Of the results for heredity, 

 I should look upon the law of ancestral heredity in the 

 case of blended inheritance with pangamy as very prob- 

 able. It gives values which, when tested, are sufficiently 

 approximate for most practical purposes. Of exclusive 

 inheritance with reversion, I have as yet discovered no case 

 except possibly coat-colour in dogs. Mr. Francis Galton's 

 investigations on Basset hounds bring, indeed, evidence 

 in favour of the law of reversion, but my own on eye- 

 colour are not in good agreement as far as the direct 

 ancestral relationships are concerned. I should accord- 

 ingly look upon the law of reversion as requiring further 

 observations and experiments before it can be accepted. 



§ 14. — Ott the Inheritance of the Duration of Life. Pro- 

 portions of the Selective and Non- Selective Death-rates 



We have seen in § i 2 how heredity enables selection 

 to establish permanently modified types. The last topic 

 I shall deal with will be an attempt to ascertain what are 

 the relative proportions of the selective and the non-selective 

 death-rates. Now by natural selection we are to under- 

 stand that certain individuals better suited by their con- 

 stitution- — i.e. by the numerical values of the complex of 

 organs and characters which form their individuality — to 

 their environment survive longer, and so are better able 

 to reproduce themselves and protect for a longer period 

 their offspring. To assert that natural selection does not 

 exist is to assert that the whole death-rate is non-selective ; 

 or that it is not a function of the constitution, the char- 

 acters and organs of the individual. Looked at from 

 this standpoint, every medical practitioner, every careful 

 observer of nature, has seen selection at work. It be- 

 comes, indeed, almost a truism. All that really remains 

 for us to do is to determine the relative proportions of 

 the selective and non-selective death-rates for individuals 

 living under sensibly the same environment ; this will 

 enable us to appreciate the quantitative intensity of natural 



