The McndtUan Inheritance of Certain Chtmieal Characters in Maiz.-. 25 



and unequivocal manner the existence of Mendclian segregation in a 

 series of characters of a kind which has not hitherto been directly 

 shown to follow Mendel's law in inheritance. While in a general way 

 all characters of any organism obviously depend upon chemical inter- 

 actions and relations, it is another thing to find that purely ciiemical 

 characteristics not directly correlated or associated with any visible 

 character of the organism are also inherited in accordance with 

 Mendelian principles. Such a result at once extends the scope and 

 generality of those principles. Such an extension is not only inter- 

 esting and important from a theoretical standpoint, but quite as much 

 so from the standpoint of the practical breeder, because it is with 

 exactly the sort of chemical characters here discussed that much of 

 the effort of breeders towards the amelioration of agricultural plants 

 has had to do. To find that these characters are inherited strictly 

 according to Mendelian principles even in a single plant species gives 

 a new and helpful point of view in regard to breeding work in this 

 particular field. 



Improvement of plants in respect to protein content, sugar content, 

 or similar characters has been held as an instance, par excellance, 

 of the gradual accumulation of minute favorable variations by con- 

 tinued selection. Results such as those set forth in this paper, however, 

 do not lend particular support to this prevailing view. 



They suggest that probably what has taken place when improve- 

 ment follows selection for one of these chemical characters is an 

 isolation or picking out of favorable genotype combinations rather 

 than a gradual, additive building up of the germ plasm with respect 

 to the character, which is the theoretical implication involved in the 

 older selectionist view. That this is actually the case with reference 

 to one of the classic pieces of experimentation along this line, is 

 shown by Surface's (11) analysis of the published pedigrees of the 

 high protein strains of corn developed at the Illinois Station (Smith 

 [i]), wherein he makes it clear that all of the Illinois "high protein" 

 corn during the last two generations is the progeny of one single ear 

 (presumably carrying "high protein" genes in preponderant number). 

 Surface furtiier shows b}' his analysis that the improvement in the 

 protein content of the Illinois corn was really by definite and con- 

 siderable steps or saltations (as would be expected if what had been 

 unconsciously done was the picking out of genotypes) rather than 

 gradually accumulative. Surface's results and those of the present 

 paper, though quite independently worked out, fit together in a 



