6 VICTOR E. SHELFORD AND W. C. ALLEE 



reflexes of the fishes and notable peculiarities of behavior were 

 recorded in full at the left of the graph. The fishes were observed 

 continuously for from twenty to ninety minutes. In many 

 cases they were caught at the end of such a period of observa- 

 tion in a small hand net and replaced in the small dishes. Ob- 

 servers then changed places and transferred the fishes from one 

 tank to the other. Thus the control fishes of the first experiment 

 were observed in the experimental tank and the experimental 

 fishes were observed in the control tank for a time equal to the 



first test. 



III. EXPERIMENTAL RESULTS 



Most of the fishes studied reacted negatively to various concen- 

 trations of carbon dioxide; to little oxygen; to boiled water 

 with the removed oxygen restored; and to boiled water with 

 acetic acid or carbon dioxide added (see table 3, p. 20). The 

 behavior of the fishes when giving a negative reaction usually 

 possessed prominent features. They tried the modified water a 

 number of times and then began to turn back in the lower and 

 lower concentrations of the gradient, or to spend shorter and 

 shorter time in the modified water with each visit. 



Usually these modifications did not show a uniform gradual 

 decrease in time spent in the modified water or an increasing 

 tendency to turn in lower and lower concentrations of the gra- 

 dient which extended throughout the experiment. The response 

 was rhythmic rather than cumulative. This may be seen in 

 the charts especially in Expt. 10, chart 1, p. 9, and in controls 

 78 and 83, chart 2, p. 11. A number of successive trials of the 

 modified water resulted in either a lessening of the time spent 

 in that water or in turnings in the gradient, or both. This led 

 to spending more time in the untreated water. After some time 

 in this water there was again a tendency to enter the treated 

 water with the same results as before. That is, there was a 

 rhythm of reaction, which, while not perfect, was present to a 

 recognizable degree in the majority of cases. 



The number of trials of the modified water at the beginning 

 of the experiment was in most instances inversely proportional 

 to the degree of stimulation as indicated by the special activities, 

 except where the concentration was great enough to cause 

 " staggering," or other abnormal reaction. The control fishes 

 on the other hand, went back and forth quite symmetrically. 



