M. Beeton and K. Pearson 
75 
brothers who frequently leave earlier the parental roof, often go much further and 
are more liable to accident, i.e. to non-selective death. Compare the correlation 
■3322 of two sisters with the •2853 of two brothers. 
Look at the matter as we will it is impossible to place the selective death-rate 
at lower than 50 per cent., and in all probability with the same environment it is 
over 80 per cent. 
(9) The following is, we believe, a fairly accurate statement of the general 
principle of evolution by natural selection combined with heredity: 
(a) With a given environment individuals having certain constitutions are 
fitter to survive than others. 
(b) They thus have a greater opportunity of reproducing themselves and 
rearing their offspring. 
(c) Since the parents' characters are transmitted to their offspring the pre- 
vailing characters of the general population will in every case either be continu- 
ously modified or are only maintained stable owing to natural selection. 
If (a) were not true the death-rate would be non-selective. We have shown 
in this paper that at least 50 and probably 80 per cent, of the death-rate in the 
case of man is certainly selective. Relatives — i.e. men witii similar constitutions — 
have durations of life substantially correlated. In the case of civilised man the 
selection is largely due to the struggle with the physical environment, and not to 
the struggle of individual with individual. We should expect therefore the 
intensity of the selective death-rate in his case to be smaller than it is for many- 
types of wild life. 
In a paper published in the Royal Society's Proceedwgs* we have shown that 
the number of offspring is directly related to the durati(m of life. We have 
thus quantitativel}' demonstrated the truth of (b). 
Finally, in a fairly long series of papers only in part as yet published it has 
been proved that physical and intellectual characters are inherited from the 
parent ; (c) is incontestably true. 
I think, therefore, that we can no longer talk of natural selection as an hypo- 
thesis. It is in the case of man demonstrably at work either changing in a 
quantitatively definite manner his constitution as a whole or else necessary to 
keep that constitution stable. It is not now correct to say as Lord Salisbury 
said in 1894 of natural selection : " No man, so fiir as we know, has ever seen 
it at work." It is sensibly and visibly at work ; a factor in 50 to 80 per cent, 
of the deaths in the case of man is not a slight perturbation, which we must 
seek with very refined analysis as the astronomer might seek for a small 
* Data for Hie Problem of Evolution in Man, V. On the correlation between Duration of Life and 
Number of Offspring, Vol. 07, p. 159. 
