290 PHYSIOLOGIC GENETICS 



stimulus intensity can vary over a rather wide range without changing the difference 

 between the strains. It is clear, however, that such a measure tells only part of the 

 story and that a more dynamic concept of behavior is obtained when we consider the 

 quantitative stimulus-response relationship over a wide range of intensities. That is, 

 A and B may be compared with respect to the slope of a curve of learning or extinction 

 or in terms of a psychophysical function. Such measures may have complex genetic 

 elements, but they are apt to be more meaningful in relation to the adaptive function 

 of behavior. 



Genotypic versus phenotypic orientation. — The preceding remarks point up a major 

 difference in the methodology of behavioral genetics, genotypic versus phenotypic 

 orientation. In the former, an investigator starts with a known genotypic difference 

 and studies its effects upon behavior. In such a situation a gene, a chromosome, or a 

 whole genotype is analogous to a treatment applied to an organism. The genotypically 

 oriented investigator does not usually stop with demonstrating a correlation between 

 genotype and behavior. He is also interested in tracing the path between gene and 

 character through intervening physiologic mechanisms. 



In using mammals he has available two general methods. The technique of 

 strain comparison makes use of populations with differences at a large number of loci, 

 some of which are known but the majority are unknown. The experimenter using 

 this method has an excellent opportunity to find behavioral differences, but genetic 

 analysis is apt to prove difficult except in terms of a coefficient of heritability. Some 

 success has been obtained in analyzing results of crosses between strains by standard 

 biometric techniques using transformed scores, 140, 410 but these methods sometimes 

 yield ambiguous results. 



More precise genetic analysis is possible when a single genie difference can be estab- 

 lished between the experimental populations. One of us 1182 used the method of 

 repeated backcrosses to eliminate genes linked with brown and white eye color in 

 Drosophila. The gene bw appeared to have no important effect upon phototaxis. 

 The white-eye gene did affect phototaxis. Earlier work on the effect of genes on the 

 behavior of Drosophila often failed to take into account the numerous loci in addition 

 to the one with conspicuous morphologic effect. 



The preparation of stocks of this type in mammals is laborious, but the behavioral 

 geneticist can find a number of inbred strains of mice maintained in forced heterozygosis 

 at a particular locus. The difficulty of such methods is that the genes in question may 

 have no behavioral effects, or, at least, no effects detectable on a background of en- 

 vironmentally induced variability. One can, of course, force the issue by using genes 

 which produce major defects in sensory and motor organs, but the behavioral effects 

 in this situation are not of great psychologic interest. 



Here it may be pointed out that behavioral techniques potentially provide a 

 sensitive method for the detection of cryptic mutations — those without readily de- 

 tectable morphologic defects. When heritable differences in behavior appear between 

 recently separated sublines of established inbred strains, it is conceivable that a single 



