630 



SCIENCE 



[N. S. Vol. XXXIX. No. 1009 



In order to accomplish this task, it was 

 necessary to analyze savageness and wild- 

 ness. This meant discovering those acts or 

 •organic processes which, taken together, 

 mean to the observer savageness or wild- 

 ness. The first result of observation was 

 that biting, squealing, struggling to escape, 

 or attacking the experimenter, and process 

 of excretion, appeared as important ele- 

 ments of savageness. The experimenter, 

 relying upon these elements, measured, 

 roughly, the savageness of a large number 

 of individuals, arranging them according to 

 their behavior in six grades, designated 

 to 5. On the basis of this obviously crude 

 preliminary work, certain facts indicative 

 of the mode of transmission of savageness 

 and wildness were ascertained.* 



In yet other observations on rats which 

 involved the comparison of two groups, 

 stock individuals and closely inbred indi- 

 viduals, it appeared that the behavior of the 

 two groups, in the face of certain experi- 

 mentally arranged situations, differed 

 greatly. This, upon careful observation, 

 the experimenter was able to attribute to 

 differences in temperament. The stock rats 

 were rather active, energetic, quick moving, 

 whereas the inbred animals were more 

 stolid, slow and deliberate. In order that 

 the reactions of these individuals in various 

 experimental situations be properly inter- 

 preted, it is essential that the experimenter 

 obtain knowledge of their temperamental 

 character, such, for example, as degree of 

 nervousness or of timidity, of savageness or 

 wildness, quickness of response, persistence, 

 energy and so on through the list of aspects 

 of behavior which, looked at as a whole, 

 might be considered the temperament of the 

 animal. The point which I am trying to 

 emphasize is this. If we are to work effect- 



8 Yerkes, Robert M., ' ' Heredity of Savageness 

 and Wildness in Rats," Journal of Animal Be- 

 havior, 1913, Vol. 3, p. 286. 



ively, with human beings or other animals, 

 we must analyze the concrete behavior of 

 the organism's every-day life into simpler 

 processes and then study these processes, 

 one by one, by means of methods which 

 shall enable us to measure them fairly accu- 

 rately and describe them with correspond- 

 ing accuracy and precision. 



The application of these observations in 

 the work of the eugenics investigator are 

 obvious, for the latter, in dealing with hu- 

 man behavior, first of all observes com- 

 plexes. If he is content to continue to ob- 

 serve these complexes and to try to study 

 their behavior in heredity, he may or may 

 not obtain scientifically valuable results. 

 But in any event, his safer course by far is 

 to deal with part processes, first to analyze 

 his complexes and then to select what seem 

 to be the most important elements and care- 

 fully study their characteristics and their 

 behavior as possible inheritances. Prom 

 his own experience, the writer is inclined to 

 urge that it is always safer to deal with 

 items of behavior than to attempt to deal 

 with behavior in a large or wholesale man- 

 ner — safer, for example, to study capacity 

 for a particular sort of musical expression, 

 singing or violin playing, than to study 

 musical ability in general. 



Were we to present more examples from 

 actual work, we might describe methods of 

 studying distance orientation, visual dis- 

 crimination, other aspects of habit-forma- 

 tion, the permanency of habits, instinct and 

 emotion, in animals, but it will suffice for 

 our present purposes to describe briefly two 

 methods of analyzing behavior which have 

 recently been devised. These methods, un- 

 like those in general use by students of 

 animal behavior, are applicable alike to 

 man and to other mammals, even to birds 

 as well. They were, indeed, planned with 

 the idea that they should make possible the 

 comparison of reaction-types or reactive 



