122 'Journal of Comparative Neurology and Psychology. 



bites. We then made two tests with unbaited forks, the usual 

 arrangement of apparatus being maintained. The first of these 

 rather indicated that, in this new order of things, the fish was 

 being somewhat influenced by the sight or smell of the food, for 

 in the absence of food it relapsed and bit ten times at the red 

 before it tried the green. In the second "no-lood" test, however, 

 the subject barely touched the red, swam to the green and bit three 

 times. And in three more baited experiments it bit first every time 

 at the green; while in the first of the final two tests without food 

 the subject swam past the red, turning away from it, snapped at a 

 small object floating in the water near the red, then at the red, and 

 then, more vigorously, at the green; and in the last unbaited test 

 it bit first at the green, then at the red, then twice more at the green. 

 The fish's behavior throughout this entire series was of the utmost 

 interest as illustrating the process of animal learning. On the 

 following day, the older habit reasserted itself, at the first trial, 

 and the fish bit three times, vigorously, at the red before it tried 

 the green; the second time, it swam straight as an arrow to the 

 green, although it entered on the opposite side and had to pass the 

 red. From this time on, with occasional rather marked relapses, 

 the new "association" between green and the biting impulse 

 shows growth, until in the last series, made on August i8, and con- 

 sisting of eight experiments with food and four without, our sub- 

 ject bit first at the green every time, except in the first experiment 

 of the series. The following table (II) shows, like the preceding 

 one, the color first bitten at in each test. 



We had some difficulty, in this latter part of our work, on 

 account of the fish's diminishing appetite. The confinement of 

 its life undoubtedly told upon its digestive powers. The series of 

 August 14 and 15 were, for this reason, incomplete, and after that 

 we attempted only one feeding a day, instead of a morning and an 

 afternoon feeding as had been our custom. The " no-food " experi- 

 ments of August 16 and 17 again furnished some indication that 

 the sight of the food in the green forceps had partly influenced the 

 fish in the feeding experiments and that when both forceps were 

 empty it had a tendency to relapse into the old "red-food" habit. 

 On August 16, after six feeding experiments in which the animal 

 had only once bitten first at the red, w^hen confronted with the two 

 pairs of empty forceps, it bit nine times at the red, went down to 

 the bottom of the tank, rose, and bit five more times at the red. 



