THE ANALYSIS OF BEHAVIOUR IN ANIMALS 165 



is usually chosen because quantitative measurement of the 

 amount of secretion can be made by using a cannula. 



As Pavlov's work is still somewhat inaccessible to the 

 English reader, the main points will be outlined. We will begin 

 with the formation of conditioned reflexes, as follows : — 



1. Any event in the external world which affects a sense 

 organ may in the intact mammal become a conditioned stimulus, 

 provided that its occurrence coincides with the unconditioned 

 stimulus a sufficient number of times. Even nocuous stimuli 

 such as intense electrical stimulation or burning of the skin 

 may, if systematically accompanied by feeding, cease to evoke 

 their normal consequences and become a signal for salivary 

 secretion. Nocuous skin stimulation may thus be formed into 

 a conditioned stimulus for the unconditioned feeding- reflex, 

 but not for the unconditioned reflex salivary secretion produced 

 by application of acid to the tongue. Nocuous stimulation 

 of the skin over the bones, however, cannot be made a con- 

 ditioned stimulus for either. The response of a conditioned 

 reflex is essentially similar to the unconditioned reflex from 

 which it is derived. If a lighted lamp be made a conditioned 

 signal for food, the dog not only secretes saliva when the 

 stimulus is presented, but makes groping movements appro- 

 priate to food itself. 



2. It is necessary that the indifferent stimulus with which 

 it is desired to form a conditioned reflex should be rigidly 

 isolated ; an unnoticed accompaniment such as an extraneous 

 smell, sound, sight or movement of the experimentalist may 

 otherwise become a new conditioned stimulus and vitiate the 

 interpretation of the phenomena observed. 



3. The indifferent stimulus should operate while the animal 

 is in a quiescent condition with reference to the unconditioned 

 system into which it is to be incorporated, i.e. it should precede 

 by a short interval the unconditioned stimulus. 



From what has been said, it follows that, since the animal 

 is normally subject to an immense variety of stimuli, formation 

 of new conditioned reflexes could only have chaotic con- 

 sequences unless there exist definable factors which tend to 

 inhibit the formation or check the operation of conditioned 



