102 CORA D. REEVES 



shows that two out of three sunfish discriminated while 

 from the text it appears that the third, large sunfish, 

 discriminated accurately in later trials not included in 

 the table. Three dace discriminated, but one did not. 

 Its failure is perhaps adequately explained on p. 54. The 

 table might be interpreted as showing response to unequal 

 energies if only the slit method of varying the red had 

 been used in equating. Since both methods were used, 

 energies were equated in a considerable but unknown 

 number of trials and in these the fish should have failed 

 if they were reacting to energy differences. The high 

 percentage of correct choices for five (or six) of the fish 

 indicates that Doctor Reeves is right in holding that they 

 were not responding to total energy differences. 



Undoubtedly the tests should be repeated and in equating 

 energies of the red and blue plates the light for both should 

 be passed through filters of the identical water used in the 

 experimental aquaria and as thick as the layer of water 

 through which the fish ordinarily see the stimulus patches. 



The recent work of Pietenpol gives reliable data on the 

 absorption of rays of different wave-lengths by the waters 

 of Wi-sconsin lakes. It is clear that absorption is unequal 

 and that either the red or the blue or both may be absorbed 

 more rapidly than other rays. The total-energy ratio of 

 two stimuli of unequal wave length probably could not 

 be constant for a fish unless its choice is made always 

 at the same distance from the stimulus plates. This was 

 by no means the case in Doctor Reeves' experiments. As 

 judged by the behavior of the fish (change of direction) 

 choices were made at all possible distances and with all 

 possible variation in the energy ratios. This seems to me 

 to show that choices could not have been made on the basis 

 of energy differences nor of brightness differences, but 

 only as the result of quality difference. But the truth 

 is that no value of the red variable was found at which 

 the fish did not learn to discriminate the longer and shorter 

 wave lengths. Amongst the values tried was one which 

 showed temporary failure of the fish (slit width 0.9-1.0 

 mm. for red). Whether we assume that at this point we 

 have to do with matched brightness or equated energies 



