538 W. T. JAMES 



appear, however, in dealing' with this reaction, especially 

 when a large number of dogs are studied. In the first place, 

 many of the animals do not form the conditioned salivary 

 response, yet observation of their behavior during the pro- 

 cedure gives some indication of their nature. In addition, 

 variations in size among the dogs must be taken into con- 

 sideration. Some small animals having many characteristics 

 of the excitable dog, give comparatively less salivary reaction 

 than would a larger animal of the same type, since the small 

 animal consumes less food, resulting in a correspondingly 

 weaker unconditioned salivary secretion. 



Although the salivary reaction itself is significant in a 

 classification of types, an additional important factor is the 

 modification in behavior during the course of the performance, 

 from the beginning to what may be termed the logical end. 

 Thus, the elimination of unnecessary response, substitution 

 and correlation of action, changes in the emotional aspect 

 of the performance, etc., and all factors involved in habitua- 

 tion are significant modifications. Because of the novel situa- 

 tion in the laboratory, each dog begins the training above 

 its normal outdoor level of excitability. As the adjustment 

 progresses, numerous conditioned and unconditioned patterns 

 of action appear, until finally a normal level of action is 

 reached. When there is a minimum of change from day to 

 day, the training may be said to have reached its logical 

 end. The dogs differ when this point is reached, just as at 

 tlie beginning of the training. In some cases, by the time 

 150 conditioning signals have been applied, the dog has com- 

 pleted his performance and reached a level of habituation 

 and stereotyped performance. In others, it occurs in less 

 than 100 applications of the signal. Other dogs never reach 

 the point at which they react in a definite, habitual, and 

 stereotyped manner even though over 300 conditioning signals 

 have been applied. These differences did not take on a proper 

 perspective until a large number of animals had been studied. 

 Behavioral type is a relative matter, and can be determined 

 only by comparing many dogs in the same situation. 



