340 W. C. ALLEE 



factor since pond isopods kept in well aerated water containing a 

 large amount of waste products gave the same increase in the 

 positive reaction, as when running tap water was used. Also 

 the cause of difference in response in isopods from the two habi- 

 tats cannot be due to the mechanical stimulation of the current 

 for experiments with both stocks show that this cannot be true. 

 Pond Aselli kept in still water but in high oxygen, increase in posi- 

 tiveness, while stream Aselli under the same conditions maintain 

 their high percentage of positive responses. The isopods which 

 gave the reactions in table 1 had been in still water all their 

 life and yet gave a positive response of over 89 per cent. From 

 these considerations it appears that oxygen is either the most 

 important environmental factor in determining the rheotactic 

 response or it is the best single factor index of the effect of the 

 complete environment upon this reaction. 



Through the tables one can see evidences of irregularity in the 

 response of the isopods from the same conditions. This irregu- 

 larity was due to two causes. First, the error of the method, 

 which has been shown to be about 5 per cent. This error is due, 

 in part, to the tendency of the isopods to keep on going in the way 

 they may be headed. This would account for an isopod going 

 positive nine times and negative the tenth. The second cause 

 of the irregularity is the fact that an 80 per cent positive response 

 on the part of five isopods may mean that four of them went 

 positive every time while the fifth was entirely negative. At 

 first sight this would appear to be a serious objection to the plan 

 of selecting the animals to be tested at random from the stock 

 under consideration. Conceivably one might pick five isopods 

 that would all give a response opposite to that of the general 

 culture. In practice however this has not occurred and the large 

 number of trials is a sufficient safeguard against such a source of 

 error. However in the later experiments every animal used, both 

 in the control and in the experiment, was tested and more uni- 

 form results were obtained in this way. 



The question at once arises as to why one member of the stock 

 under the same external conditions as the others should give a 

 different rheotactic response. One of the reasons is the state of 



