APPENDIX 



unreliable approximation. It is more than an even wager that it differs 

 from the true value by as much as T V 



0.6745 12400 



When a hundred years are used, the probable error is \ 



100 \ 100 



0.033, which shows that the result T % derived from a hundred years is 

 much more significant than that which was obtained when only ten years 

 were used. 



The following table will show how, with increasing numbers, the probable 

 error in the determination of the probability decreases. 



Remarks. In conclusion, it should probably be said that we have, in 

 the foregoing brief discussion of statistical methods, touched only " the 

 fringe of a great subject." We have for the sake of simplicity confined 

 ourselves to the normal curve of distribution ; but it is to be hoped that 

 we have given a general view, in this short space, of the methods by which 

 the formulas are established which are now being commonly used in the 

 quantitative study of evolution, and that the reader may come to see the 

 proper place of statistical methods in solving problems of this character. 



Furthermore, it is hoped that the results as here presented will be found 

 practical in the sense that they may be of use to the non-mathematical 

 reader, and pave the way for further investigations along this line. 



