176 



Scientific Proceedings (114). 



water to escape at its sides. By a simple arrangement of stop- 

 cocks it is possible to change the inflow of any substance from 

 one trough to the other very quickly. It is also possible to secure 

 conditions similar to those in the Shelford apparatus by using an 

 intake at the end of the large trough, thus furnishing three intakes, 

 two of them parallel to each other. In this case the movable parti- 

 tion is not used, the outflow being in the middle of the larger 

 trough. 



The apparatus used during the summers of 1919 and 1920 is 

 20 feet long. The dimensions of the receiving trough are 10 

 feet in length, 8 inches in depth and 8$4 inches in width. The 

 two tributary troughs are each 10 feet length, 4 inches in depth 

 and 4^2 inches in width. The tributary troughs have been 

 elevated varying degrees so as to cause the water to flow down- 

 wards into the receiving pool. In addition to the use of barrels 

 as containers for the solutions used, in some experiments the inflow 

 has been directly from the circulation pipes of the laboratory. 



The fish used in the experiments included Fundulns heteroclitus, 

 Fundulus majalis, and Clupea harengus. They were tested with 

 sea water of various dilutions and concentrations, ra/10 solutions 

 of NaCl, KC1, MgCl 2 , CaClj, MgS0 4 , various combinations of 

 those salts and also with sewage polluted water. The temperature 

 and the stream flow were varied in some of the experiments. 



The results obtained indicate the following behavior: 



1. The fish used in the experiments (10 fish for 25 trials) 

 responded to both toxic and nontoxic substances, discriminating 

 them readily under the same conditions of stream flow and 

 temperature. 



2. With a more attractive stream flow, usually one slightly 

 more rapid than the control, the fish were lured into solutions 

 that were quickly lethal in experimental jars. No evidence of 

 intoxication in reactions to any particular substance was adduced. 

 The reaction seemed to be one of pure rheotaxis. 



3. Temperature change was readily detected but the optimum 

 temperature varies with the season and the physiological condition 

 of the animal. 



4. In experiments with the apparatus used, errors due to the 

 reaction of the fishes to movements of the observer have been 



