14 CORA D. REEVES 



ray filter (Red 3.65 C 3-48 A) secured fr6|m the Eastman Kodak Company. The 

 red is a piece of copper-flashed ruby glass. 



4. Detailed Description of Methods 



To make clearer some methods employed during experimentation the following 

 details of procedure are included. That the fish might behave normally care was 

 taken to avoid frightening them. They were allowed to live for months in the 

 experiment aquarium without being lifted from the water or being moved in the 

 aquarium. The partitions were moved very slowly. Often when they were shifted 

 from one compartment to another, the fish showed absence of fear by swimming 

 up to the partition as if to nibble its lower margin. It may be that the partition 

 when thus moved resembles a clod which gradually settles to the bottom or slides 

 along the stream margin and offers the fish a good place to feed, because worms 

 previously covered are exposed. Because the fish exhibit little fear of these moving 

 partitions, they were used exclusively when changing the fish about in the aquarium. 

 A third partition, T in figure 6, was often put into the position shown, in order to 

 separate from the other fish the one that it was desired to test. 



Then after swinging the partitions A and B, figure 6, out to positions A' and B' , 

 the fish to be tested would swim quickly into the discrimination compartment, 

 and with A and B brought back into place, was ready for the raising of the sliding 

 door, D, when a test was to be made. If when the door was raised, the fish swam 

 toward the plate determined upon for the food association, that is, toward the 

 positive stimulus plate, then the bar with the food was allowed to fall. After time 

 had teen allowed for feeding, the fish was brought back into the discrimination com- 

 partment by lifting the partition A with the door closed and slowly lowermg it so 

 that the fish was on the side away from the stimulus plates. The oartition was 

 then slowly returned to its former place. If the fish made a wrong choice it was 

 not fed but brought back into the discrimination compartment without releasing 

 the food-bar. The latter was not set free until the fish had come within 15 cm. 

 of the positive stimulus plate. It was thus possible for the fish to change 

 direction of response within the discrimination compartment or at any point 

 between it and the distance of 15 cm. from the positive plate. 



It was determined by the use of a mirror (see p. 46) that when the fish was swim- 

 ming low in the aquarium, a reflection of the slit over the red plate on the surface 

 of the cylindrical lens was sometimes visible to the fish when within 15 cm. of the 

 stimulus patches and might guide it in its choice. Choice must therefore be made 

 at a point farther from the plates. If the fish entered the negative half of compart- 

 ment 2 but while still more than 15 cm. from the stimulus plates turned in a straight 

 course toward the positive plate, then the food-bar was released and the choice 

 recorded as correct. 



It was found that caution was required, for when a fish swam up toward the 

 positive plate, there was a tendency to drop the bar too quickly. The fish soon 

 learned to go up to one side of the aquarium and wait, and if the bar was not dropped, 

 to swim promptly to the opposite side. Only by very careful attention as to whether 

 the fish was more than 15 cm. from the positive plate was this error avoided. If 

 the fish was more than 15 cm. from either plate, the food-bar was not dropped and 

 either side might be correct. Waiting for the food-bar could not indicate need for 

 change of direction when the fish was at a greater distance from the plate. 



That there might be no regular order for the side on which the positive stimulus 

 was presented, six slips of paper were prepared, each with one of these inscriptions 

 upon it: 



