18 VICTOR E. SHELFORD AND W. C. ALLEE 



peculiar trait seems to be well developed. It avoids mild stimuli 

 by rhythmically shorter stays in the modified water and stronger 

 stimuli by turning back in the gradient. The general behavior 

 when turnings are being given is shown in table 2. 



In our earliest experiment with Abramis we obtained no results 

 because the fishes remained in the tap water and so failed to 

 encounter the treated water. It was noted also that when these 

 fishes were left in the tanks they tended to stay in one end or the 

 other. This led to the following experiment: Three individuals 

 were placed in each tank with the same kind of water flowing 

 into both ends of each. After both groups had developed an 

 apparent preference for one end, the fishes selected for the control 

 of the experiment to follow were disturbed until they tended to 

 go back and forth. At the same time a shadow was thrown over 

 the end of the tank selected for the experiment opposite to that 

 in which the fishes were staying. The experimental fishes were 

 driven into this shadow several times and soon developed an 

 apparent preference for the shaded end of the tank. The fishes 

 to be used for the experiment were left in the shadow for about 

 two hours. The fishes to be used for the control were driven from 

 end to end several times during this period. The apparatus was 

 then arranged for the introduction of treated water which 

 necessitated the removal of the shadow. Even after the shadow 

 was removed and the apparatus was disturbed, the fishes per- 

 sisted in their apparent preference for the end which had been 

 shaded (Edinger, '01; Holmes, '11). Boiled water was intro- 

 duced in the end in which the fishes were staying. The results 

 of this part of the experiment are shown in Expt. 1-A, chart 4. 

 In the control, the time was divided between the two halves in 

 the ratio of 41 to 59 for the hour of the observation. During 

 the same period the experimental fishes stayed in the boiled water 

 end during the first 29 minutes, when they began going back and 

 forth. They clearly stayed a little longer in the tap water with 

 each excursion. After making five trials of the tap water in a 

 trifle more than ten minutes the fishes came to rest in the tap 

 water end and remained there until the completion of the hour's 

 observation. The boiled water introducer was then placed in 

 the newly ' preferred ' ' end and similar results were obtained 

 modified only by the additional activity of a juvenile individual. 



In Expt. 84, chart 4, the same species showed a comparable 



