194 PHYSIOLOGIC GENETICS 



demonstration of strain differences is therefore no more than the starting point of the 

 investigation. 



There are, broadly speaking, four sorts of genetic investigations of a quantitative 

 character that may be made. The first two — estimation of the degree of genetic 

 determination and of the heritability — are concerned with the relative importance of 

 genetic and environmental factors as determinants of an individual's phenotypic value 

 of the character. The third is a description of the effect of selection applied to the 

 character. This may be of practical interest, but the additional genetic conclusions 

 that can be drawn are rather limited. The fourth is a description of the effect of 

 inbreeding or of crossing inbred lines, and again the conclusions that can be drawn 

 are rather limited. 



None of these investigations can be made without a noninbred, or genetically 

 heterogeneous, strain. The most satisfactory strain for this purpose is one that has been 

 maintained by random mating among a large number of parents over many generations, 

 so that its genetic properties have had time to become stabilized. But to advocate 

 the use of such a strain is obviously a counsel of perfection. In the absence of a random- 

 bred strain the most convenient form of genetically heterogeneous strain is an F 2 of a 

 cross between two inbred lines, or the third generation of a 4-way cross of four inbred 

 lines. These synthetic strains have the advantage over a random-bred strain that they 

 can be reconstituted at will from the original inbreds. Their disadvantages are that 

 the conclusions about their genetic properties have less generality and that their genetic 

 properties are in some ways necessarily unnatural. The lack of generality can be 

 expressed through the inbreeding coefficient. Nobody would claim generality for 

 any biologic property discovered from the study of a single highly inbred line. The 

 F 2 of a 2-way cross is 50 per cent inbred, and the third generation of a 4-way cross is 

 25 per cent inbred; in other words, a strain derived from a 2-way cross is equivalent 

 to the progeny of a single self-fertilized individual, and a strain derived from a 4-way 

 cross is equivalent to the progeny of a single pair of full sibs. To this extent, therefore, 

 synthetic heterogeneous strains resemble inbred lines in lack of generality. The un- 

 natural features of the genetic properties of synthetic strains are the restricted range 

 of genie frequencies, nonrandom linkage associations, and the absence of lethal and 

 severely deleterious genes. The disadvantages of synthetic strains, however, are only 

 relative. Complete generality is an unattainable ideal; no random-bred strain of mice, 

 however carefully constructed and maintained, is truly representative of all mice, 

 just as no laboratory mammal is truly representative of all mammals. 



DEGREE OF GENETIC DETERMINATION 



If there are genetic differences, then how important are they? This is the nature- 

 nurture question: what is the relative importance of heredity and environment in 

 determining the value of the character in an individual ? (For the sake of convenience 

 all nongenetic differences are referred to as environmental, so that the genotype and the 



