104 Heredity and Environment 



tor may affect the entire body is less likely to do harm than to 

 state that each factor affects only a particular character." And 

 again he says, "It cannot too insistently be urged that when we 

 say a character is the product of a particular factor we mean no 

 more than that it is the most conspicuous effect of the factor" 

 (Morgan, 1916, p. 117). 



Lethal Factors. Morgan and his associates have also demon- 

 strated the existence of a considerable number of lethal factors 

 in Drosophila that cause the early death of those gametes or 

 zygotes in which this factor is not balanced by a normal one. 

 This phenomenon greatly modifies expected Mendelian ratios 

 for only heterozygotes survive, and all individuals that are homo- 

 zygous for a lethal factor usually die so early that they are never 

 seen. Nevertheless their existence can be determined by indirect 

 methods that will be mentioned in the next chapter under "link- 

 age." Such lethal factors greatly complicate the study of genetics 

 but they do not destroy its fundamental principles. 



What are factors? Inheritance factors are probably complex 

 chemical substances which preserve their individuality in various 

 combinations, just as groups of atoms or radicals do in chemical 

 reactions ; they may be dropped out or added, substituted or trans- 

 posed, just as chemical radicals may be in chemical compounds. 

 To this extent they maintain continuity and independence, but they 

 are not absolutely independent for they react upon one another 

 as well as to environmental changes, so that the characters of 

 the developed organism are the resultants of all these reactions 

 and interactions. 



Some progress has been made, in identifying certain structures 

 of the germ cells with certain hereditary units, but quite irrespec- 

 tive of what these units may be and where they may be located it 

 is possible, by means of the Mendelian theory of segregation of 

 units in the germ cells and of chance combinations of these in fer- 

 tilization to predict the number of genotypes and phenotypes 

 which may be expected as the result of a given cross. 



