84 CORA D. REEVES 



fish died. This indicates that his other fish were not in normal 

 condition. They may have failed to give responses that 

 could have been obtained from normal fish. 



b. Errors may arise during experiment from uncontrolled 

 environmental factors. — The procedure of Hess (1909) in 

 comparing the sensitivity of dark- and light-adapted fish 

 has been already described. The aquarium containing the 

 light-adapted fish was moved just before testing them, 

 while that containing the dark-adapted fish was not moved. 

 To compare the responses of the two sets of fish under 

 these conditions cannot be regarded as a sound procedure. 

 Again, Hess (1913) in studying the influence of the back- 

 ground in modifying the colors of fish puts several into the 

 same containers or into separate containers so near together 

 that the fish are visible to each other. He thus ignores 

 the fact that the colors of one fish may be influenced by 

 the presence of another fish. I have several times observed 

 the colors of fish when two are in one container. If one 

 chases the other, the pursuer, as observed, may become 

 lighter in color and the pursued darker or at times a re- 

 versed change may take place. In the breeding season, 

 I have found the male rainbow darter to change strikingly 

 in color from second to second and without change of back- 

 ground (Reeves 1907). Colors flash out with the incidence 

 of various stimuli, such as the presence of other fish or the 

 movement of the observer. Similar changes may be seen 

 at other seasons. Yet it appears that both Hess (1913) 

 and Frisch (1911), in some of their experiments, went to 

 the aquarium and stood over the fish to observe whether 

 they changed to match the background. This procedure 

 in itself would cause color change in some of the species 

 with which I have worked. (See also Freytag, 1914). 

 Through neglect of these factors Hess seems greatly to 

 have lessened the value of his earlier experiments. Frisch 

 appears to have avoided putting a number of fish into one 

 container without running water and into such poorly 

 adjusted conditions as to cause the death of his fish, but 

 for the control of all the complex physiological and psycho- 

 logical factors that cause response, his methods are not 

 entirely satisfactory. The recent work of Mast (1916) 



