Color Discyiiiiiiiafioii in the Gray Sitaf'f'cr. 281 



coincide at the 50 per cent level. No such tendency appears in the diagram. 

 Such a tendency of the two lines to approximate each other and the 50 per 

 cent level more in the final series of the experiment than in the earlier series 

 would then be evidence that the snappers rapidly learn the equal palatability 

 of the blue and white atherinas. In only one respect do the lines seem to 

 afford possible evidence that the snappers tend to discriminate less between 

 the white and blue fish toward the end of the day's experiment. This is in 

 the crossing of the solid and broken lines at a point which is nearer the ordi- 

 nate of percentages in series III and IV than in series I and II. The two 

 lines come together more rapidly in the third and fourth series than in the 

 first and second. The series of July 19, 21, and 22 have also been plotted. 

 The series of July 19 and 21 do not afford the evidence apparently shown in 

 that of July 20 of a decrease of discrimination during a single day, nor does 

 the series of July 22 when compared with those of earlier date show any 

 decrease of discrimination from day to day. If the snappers could be iso- 

 lated from the living atherinas which they have constantly in view, and if 

 they could be fed continuously on blue and white fish in equal numbers, they 

 might soon take the one in each place as frequently as the other, but with the 

 living white fish constantly in sight as part of the natural environment, and 

 with the blue supplied only at intervals and for a short time, this result is 

 not to be expected. 



Experiment j/; Blue and red color discrimination. On July 21 two 

 series of blue and white discrimination trials were made, the first at 8'" 7" 

 a. m., and the second at 2" 40' p. m. Three hundred and sixty atherians 

 were used in addition to those involved in two preliminary feedings of 

 normal fish. The results, as plotted in table 10. series I and II, do not 

 differ from those shown in table 9. Immediately after the second series of 

 12 trials with blue and white fish there followed a series of 10 trials (in- 

 volving 100 atherinas) in which red fish were substituted for the white, so 

 that, whereas the snappers had before had a choice between white and blue 

 and had become familiar with blue, they now had a choice between red and 

 blue. The red atherinas were prepared as in experiment 2 and had there- 

 fore the same acid taste as the blue. The results are plotted as series III 

 of the experiment of July 21 and are shown in table 10. The blue fish are 

 represented as before by the broken line, while the red fish are represented 

 by a line of dots and dashes. It is seen that the broken line occupies in 

 series IV the position of the solid line in series I and II. The line of dots 

 and dashes representing the red fish occupies, on the other hand, the position 

 taken in series I and II bv the broken line which represents the blue fish. 

 In other words, the blue fish are now taken more frequently than the red 

 in the first four orders. Blue is taken first in 70 per cent of the trials, while 

 red is taken first in but 30 per cent. It is further to be noted that not only 

 is the position of the broken line reversed in series III as compared with 



