284 PHYSIOLOGIC GENETICS 



BASIC CONCEPTS 



The unit of behavior. — The fundamental unit of behavior is the behavioral pattern, 

 which may be defined as a regular combination or series of actions having a particular 

 adaptive function. Behavior can, of course, be almost infinitely subdivided into 

 minutely perceptible movements, but such units themselves have no special function 

 and any one may be part of many different patterns of behavior. To take a familiar 

 example, the crowing behavior of a rooster is a behavioral pattern. Analyzed in 

 terms of bodily movements, components are found which are parts of eating, drinking, 

 visual investigation, and the like. 



Behavioral patterns having a similar function can be combined into a loose 

 behavioral system. For example, fighting mice show several patterns of behavior: 

 attack, defensive posture, running away, and complete passivity, and all these may be 

 considered as parts of the agonistic behavioral system. For purposes of genetic analysis 

 it is usually more rewarding to study specific patterns rather than to search for a general 

 trait such as aggressiveness. 



Behavioral analysis becomes even more difficult when behavior is measured in 

 terms of performance or adaptation. When behavior is studied on the level of the in- 

 dividual organism, its primary function is adaptation to the environment. Consequent- 

 ly, an important dimension is the degree of adaptation achieved. The vast majority 

 of scientific tests of intelligence and performance are based on a scale of this sort. 

 Obviously this is not a scale found in the physical sciences, nor is it directly related to 

 the action of single genes. The person who is interested in performance is measuring 

 the results of behavior, not behavior itself. When such results can be achieved in 

 only one way, genetic analysis should be reasonably simple. However, in the majority 

 of practical situations, adaptation can be achieved through a variety of behavioral 

 patterns or combinations of patterns, and the test measures only the end result. In 

 short, measures of this sort usually involve extremely complex interaction, and the 

 genetic analysis of differences in performance may therefore be extremely difficult and 

 unclear. For example, we devised a test for dogs in which food had to be pulled out 

 from under a box with a string. Some breeds of dogs would do this with their teeth 

 and others with their paws. The end result was the same, but the motor aptitudes 

 involved were quite different. Most other methods of measuring performance, from 

 the currently popular bar-pressing technique for measuring behavior of rats to human 

 intelligence tests, are subject to the same difficulty. 



In a recent paper, Ginsburg 432 has argued that the best natural unit for analysis 

 of behavior is the effect of a single gene. This definition may be of great practical 

 importance in such syndromes as phenylpyruvic oligophrenia. However, the natural 

 units of behavior proposed above do not, for the most part, directly correspond to 

 genes or other natural units on the genetic level of organization. Actually, genes and 

 behavioral patterns are separated by complex pathways of physiologic events. One 

 may start at either end of the path and try to find the way through. In either case, there 



