172 W. C. ALLEE 



and physiological effect (Hober, s. 308). Again the simple 

 broken line shows the percentage of isopods that were made 

 more positive at some time during the treatment with each 

 ordinate space worth 5 per cent. The dot-dash line gives the 

 average difference between the percentage of positive rheotaxis 

 before and at the most positive point after treatment with each 

 ordinate space equal to 2 per cent. The unbroken line gives 

 toxicity, each ordinate space representing one minute, and the 

 dotted line shows the toxicity -depression ratio with the ordinate 

 spaces worth 5 per cent. 



Experiments with potassium sulphocyanide in N/5 solution 

 showed that it was more toxic than any other anion tried at 

 this strength. In N/20 solution it caused some increase in 

 positive rheotaxis but the evidence of the experiments at 

 hand is contradictor^" and there are too few instances to locate 

 this anion definitely. Potassium sulphate (2N/7 and N/lo) 

 increased the positiveness of the response, especially in the weak 

 strength where it was almost as effective as potassium chloride 

 in N/10 solution but more toxic. 



As the figure shows the relative toxicity was Br > NO,-? > I > 

 CI > Ac, CIO3, and the effect on positive rheotaxis was Br < 

 NO3 < I < CIO3 < CI < ^c if one considers the effectiveness 

 in causing some increase in the positiveness. The same order 

 holds true when the extent of the increase is considered except 

 that Ac is less effective than CIO3 and CI. I am not prepared 

 to state why the order found here differs from the usual relative 

 effect shown by the order of arrangement in the figure or why 

 the more toxic salts should vary in their toxicity and their effect 

 on rheotaxis. It is possible that more experiments might change 

 the order in one or two places, but enough tests have been made 

 to demonstrate conclusively that potassium chloride is the most 

 effective salt so far as amount of change is concerned, and the 

 regularity with which bromine takes the usual place of iodine 

 at this concentration indicates that the order found is not a 

 mere accident. A number of the anions that failed to show 

 a marked increase in N/10 solution were tested in N/20 and 

 N/40 concentration. No marked change occurred except in 



