196 Life and Death, Heredity and Evolution 



The facts of assortative mating, of course, limit the ex- 

 tent to which new combinations of the characters are pro- 

 ducible through mating. After stocks have reached a cer- 

 tain degree of diversity, assortative mating prevents their 

 union, and so prevents the formation of stocks combining 

 their characteristics. But this still leaves a wide field for 

 the formation through mating of varied combinations of the 

 hereditary characteristics of differing stocks. 



We now come to one of the most striking biological re- 

 sults of mating. In this formation of new combinations, 

 characters which were previously in diverse stocks become 

 united in one individual. Sometimes the characters so com- 

 bined are more or less incompatible ; they do not develop 

 well together; or they do not function harmoniously; or they 

 produce secondary characters which do not function well 

 under the particular conditions in which the organisms are 

 found. Individuals resulting from such combinations may 

 be weak or pathological; they may not develop at all; or 

 if they do, they lack vigor, and do not multiply. Many ex- 

 amples of such consequences we saw in our study of the re- 

 sults of mating in Paramecium; many individuals or 

 families produced died out, or were weak, multiplying 

 slowly; or showed hereditary deformities. Other individu- 

 als on the other hand received combinations of characters 



cients of variation were worked out on the basis of but 2, 3 or 4 diverse 

 lines of descent! In only one single case were there as many as eight 

 lines in both the sets compared (conjugants and non-con j ugants ); in 

 this case the variation was much greater in the lines that had conju- 

 gated. But it is obvious that such small numbers cannot give clear and 

 consistent results, particularly when we find that the coefficients of 

 variation brought to light range from to 55.3 per cent. If probable 

 errors had been worked out, they would doubtless have shown the com- 

 parisons of coefficients to be without significance. A study of the ques- 

 tion whether mating actually produces lines with diverse hereditary 

 characters would be of great interest in Didinium, as in any other Pro- 

 tozoa, but the determination of coefficients of variation is a most imper- 

 fect and uncertain index of this matter. 



