22 VICTOR E. SHELFORD AND W. C. ALLEE 



Table 4 shows for the types of modified water shown in detail 

 for all the species and individuals (230 in all). An examination 

 of the table shows the same general relations as have been 

 brought out for the particular species. The average time spent 

 in the two kinds of water is least in the water producing the 

 modification in all cases save one and in that the time in the 

 untreated equaled that spent in the modified water. 



IV. INTERPRETATION OF RESULTS 



The phenomenon discussed and illustrated by the graphs on 

 the preceding pages is clearly one of rapid modification. The 

 behavior of the fishes was different after from one to four entrances 

 into the end where the stimulation was greatest. This modifi- 

 cation is indicated by the two types of behavior suggested above, 

 viz.: (1) Turning in weaker and weaker parts of the gradient; 

 (2) By making shorter and shorter stays in the modified water. 

 This last type of behavior is not considered in the tables, since 

 it cannot be readily tabulated because of its rhythmic nature. 



In general, fishes swim about either continuously or period- 

 ically when in normal water and under uniform light conditions. 

 This tendency under the unnatural conditions of the experi- 

 mental tanks is strong and though some species may rest for con- 

 siderable periods in the control tank, they periodically move 

 from end to end. The fishes move crosswise of the tanks but this 

 was not considered, since it appeared to be a minor matter 

 because of the narrowness of the tanks. Furthermore, it bore 

 no relation to the experimental conditions other than to prolong 

 the time spent in a particular part of the tank. 



The modification, which will be chiefly discussed, consisted of 

 breaking the tendency to swim to the end of the tank and of 

 substituting turnings at points of weaker and weaker concentra- 

 tions of the experimental factor or factors. The tendency to 

 pass to the end of the tank as opposed to turning nearer the 

 center may not be markedly strong, for all the fish occasionally 

 turned and swam back or swam about in circles without crossing 

 the center (see chart 1, column 3, last ten minutes of the control 

 of Expt. 10). Still, on account of the small size of the tanks, 

 this tendency is apparently almost as strong as the tendency of 

 fish to swim anywhere when not especially stimulated. No 

 doubt it is about as strong as the tendency of Mobius' pike 



