1897] KARL PEARSON ON EVOLUTION 51 



simply and naturally under one general law. The reader who is 

 acquainted with the theory of Chance, as it is propounded in the 

 ordinary text-books, will appreciate the great extension of the 

 theory which is necessary in order to treat successfully such statis- 

 tical results as those expressed by the remarkable " curve of infan- 

 tile mortality," or even those relating to the incidence of scarlet 

 fever, enteric fever, and diphtheria, at particular periods of life. 



These illustrations being given in the first essay, the second is 

 devoted to the results of roulette, as played at Monte Carlo, this 

 being chosen as an example of results which are so little capable of 

 prediction by the theory of Probability as to justify the belief that 

 some constant influence other than chance is at work. The point 

 of general interest is the smallness of the discrepancy between the 

 observed result and that given by the theory of Chance which can 

 be used as evidence of some constant disturbing factor. 



The third essay, on " Reproductive Selection," contains a most 

 interesting study of the importance to be ascribed to variation in 

 fertility. The material used consists of two tables, showing the 

 number of children arising from each of a large series of marriages. 

 The first series contains 4390 marriages, which are spoken of as 

 " Anglo-Saxon " ; they are for the most part English and Americans 

 of the well-to-do classes, with some from the Almanach de Gotha. 

 The second series includes over 34,000 Danish marriages. The 

 result is so important that a rough outline at least must be given. 

 The following table shows the number of " Anglo-Saxon " mar- 

 riages which produced any given number of children : — 



No. of Children, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 

 No. of Marriages, 546 656 682 628 496 383 336 228 172 118 63 47 22 8 2 1 2 



The series contains no record of barren marriages ; but for 

 reasons fully discussed in the essay, 320 is assigned as the number 

 of barren marriages likely to have existed in a population with the 

 observed number of fertile unions. The above table may therefore 

 be regarded as representing the offspring of 4710 marriages. 



The total number of children produced is 19,833, giving an 

 average of about four and a half per marriage; but the striking- 

 feature about the table is the demonstration that half of this entire 

 number of children is produced by little more than a quarter (25*8 

 per cent.) of the total number of marriages, so that half the second 

 generation are the offspring of a quarter of the first. Now, suppose 

 this character of excessive fertility to be completely inherited, Pro- 

 fessor Pearson shows that (in the absence of an enormous selective 

 death-rate) ninety-nine per cent of the -sixth generation would be 

 descended from the " superfertile " quarter of the original generation. 

 The effect of selection in checking this result among human beings 



