12 VICTOR E. SHELFORD AND W. C. ALLEE 



Chart III 



Showing modification by contact with boiled water with 50 cc. per liter of car- 

 bon dioxide added. For further explanation see Chart I. 



In these experiments the fishes tried the gradient of the entire tank a number 

 of times and then either remained in one end except when driven as indicated 

 by arrows, or turned back often before the center was reached. For example 

 Hybopsis (experiment 56) entered the modified water once or twice and then 

 came to rest in the low concentration, invading the high only when driven with 

 the exception of a single excursion which followed a disturbance. The graph 

 of Ameturus shows one of the very few cases in which the fishes turned back before 

 the strongest stimulus had been encountered. Still the graph brings out the same 

 general fact of modification. The gradient in experiment 76 was established with 

 some difficulty as the nitrogen was only 93% of the gas available the rest being 

 oxygen. Boiled water had to be used at both ends and sufficient oxygen added 

 at one end to balance the oxygen added with the nitrogen at the other. Even here, 

 with the factors involved somewhat in doubt, the same modification is suggested. 



