MODIFICATION OF THE BEHAVIOR OF FISHES 27 



'11). The failure of the fishes to react to ammonia accords with 

 the known effects of the drug. According to Cushny, when 

 ammonia or its common salts are absorbed by the blood, it is 

 not rendered more alkaline but the ammonia is rapidly changed 

 to urea and excreted. The effects of the kation on the common 

 frog is to paralyze the terminations of the motor nerves. While 

 the effect upon the sensory ending appears not to have been 

 investigated, paralysis appears to have occurred in the fishes 

 studied. Frogs and mammals usually die from ammonia 

 poisoning in tetanic convulsions as did the fishes used in the 

 experiments. 



It may be noted that the stimuli which give rise to the modifi- 

 cations most quickly are those commonly encountered by fishes 

 in nature. Ammonia, which is rarely encountered in any 

 considerable quantity, did not give rise to a modification and not 

 even to an avoiding reaction. Again the darters are swift 

 stream fishes, depending upon mechanical conditions to main- 

 tain themselves in a suitable environment. They rarely encoun- 

 ter carbon dioxide or low oxygen and failed to react to them 

 at all quickly, although they were affected by both. Advan- 

 tageous reactions appear to be confined to stimuli commonly 

 encountered in the normal life of the animal. By this we mean 

 merely to imply that whatever the processes of origin and sur- 

 vival may have been in detail (Mathews, '13), there is correlation 

 between the conditions of existence and types of irritability 

 (Henderson, '13). 



It thus seems probable that associative memory does not 

 necessarily play any role in the process of modification des- 

 cribed. Since the experiments were conducted with a view to 

 eliminate any possible effects of learning, only incidental evidence 

 was acquired. The treatment of the stock of fishes was as fol- 

 lows: There were a number of individuals of each species in the 

 aquaria, and individuals were drawn at random for each experi- 

 ment. A given series of experiments were run and the fishes 

 were returned to the aquaria and not used for experiments 

 until several days had elapsed. In nearly all cases the same 

 fishes were used only by chance. When it was necessary to repeat 

 an experiment on the same day, different fishes were used except 

 in one or two cases not included here. 



Apparently, in experiments thus conducted, evidence of reten- 



