EXPERIMENTS ON THE RABBIT 173 



except on the supposition that to a certain extent a habit had 

 been formed of pushing the door which carried the darker paper, 

 irrespective of the absolute brightnesses involved. Yet the 

 absolute brightnesses were apparently not wholly without effect. 

 In the latter part of this combined series the rabbit began to 

 show signs of confusion. Further, as has just been noted, if the 

 habit of reacting to the darker impression as such had been per- 

 fectly formed, the introduction of grey-white tests ought not 

 to have had an unfavorable effect on succeeding red-grey tests. 

 That such an effect probably was exerted is shown by the 

 results of the next series. Series 37 lasted from November 13 

 to November 24. 



Series 38. Red and grey number 15. Red opens. This series 

 lasted from November 25 to December 6. There were 105 

 experiments in the series, made, with the exception of one day, 

 at the rate of ten a day. The total percentage of correct choices 

 was 62. The percentages on successive days were 60, 90, 30, 60, 

 70, 35, 80, 70, 56, 80. The rabbit's behavior suggested that the 

 difficulty lay in the fact that the grey, whose absolute character 

 was somehow recognized, sometimes set off the impulse to push 

 it, as in the grey -white tests of the preceding series, and some- 

 times inhibited the impulse, as usually in the red-grey series. 

 When the animal started in the direction of the red, she ran to 

 it and pushed without hesitation ; when she started for the grey 

 she commonly drew back or stopped on the way to it, either 

 finally pushing it after doing so, or turning to the other side and 

 pushing the red. 



Series 39. Red and Hering grey number 46. Red opens. 

 The paper called " Stoelting black" in preceding series was, 

 as has been said, nearer in brightness to Hering grey number 46 

 than to any other of the Hering greys. This last series was 

 undertaken to find whether Polly, working with red and gray 

 number 46, would get results like those obtained by Dark Nose, 

 Light Nose, and Abednego in their experiments with red and 

 Stoelting black paper. There were 180 tests in the series, twelve 

 experiments being made on a day. The series lasted from Decem- 

 ber 7 to December 22. The total percentage of correct choices 

 was 52. The daily percentages were 58, 58, 58, 58, 50, 58, 50, 

 58, 66, 50, 50, 66, 42, 50, 66. This series of figures may fairly 

 be compared with those from Series ^2, with red and grey number 



