292 PHYSIOLOGIC GENETICS 



with Mendelian inheritance, without attempting to measure the effect of segregation 

 on variance. 



Scott 1183 has suggested a method of analysis of segregation independent of variance, 

 based on the point of maximum separation between the two parental strains and a 

 comparison of ratios between the two backcrosses. The method assumes the existence 

 of a threshold, plus the additive effect of genetic and environmental factors which lead 

 to crossing the threshold, and it can be used where such assumptions are justified. It 

 will give an estimate of the number of genetic factors involved and can be used to 

 predict an expected ratio in which correspondence with actual data can be tested 

 by the usual statistical methods. 



FUTURE DEVELOPMENTS IN BEHAVIORAL GENETICS 



A major problem in behavioral genetics is the investment in manpower needed 

 to determine a behavioral phenotype. A long period of testing under carefully control- 

 led conditions may be required to obtain a score which can become part of a genetic 

 analysis. Mechanization of tests may in the future make possible more elaborate 

 experiments. Hirsch and Boudreau 584 devised a geotactic maze for Drosophila which 

 sorted out the population automatically. Bruell 140 has automated various pieces of 

 apparatus for the study of murine behavior. It might be hoped that the operant con- 

 ditioning procedure (which is almost the limit of automation) would prove useful 

 for behavior genetics, but in actuality the long period of time required for shaping 

 behavior in the operant apparatus seems to preclude such applications, and procedures 

 so far used have been designed to minimize genetic variations. It does, in fact, seem 

 doubtful that the more complex functions of learning can be automated to the extent 

 that large numbers of subjects can be run through tests with little involvement of 

 humans. Many behavioral patterns are more accurately and reliably recorded by the 

 human eye than by elaborate apparatus. Nevertheless, ingenuity and the desire for 

 new tests may well help the investigator with his problems of manpower. 



The demonstration of differences in behavior between strains should continue, 

 but it is to be hoped that such demonstrations will merely be the prelude to more ela- 

 borate genetic analyses. The modern tools of biometric genetics have actually been 

 applied only sporadically to behavioral characters. Experiments by Bruell 140 and by 

 Broadhurst 136 indicate a trend toward more sophisticated designs. 



Possibly the most significant advances will be made in the relation of genes to 

 behavior through biochemical pathways. Ginsburg 431 has outlined the general 

 scheme for such studies. The relationship between investigations of this sort and the 

 search for a biochemical factor in mental disorder in man 704 ' 705 is obvious. Behavioral 

 phenotypes are more difficult to measure than ordinary phenotypes, but a large number 

 of investigators have demonstrated that the problems are not insuperable. In fact, 

 the difficulties in studying the genetics of behavior are fairly similar to those encountered 

 in studying the inheritance of complex physiologic characters. The methods, too, 



