AN ANALYSIS OF THE LEARNING PROCESS IN THE SNAIL 83 



obtain air (punishment) and the attainment of air (reward). 

 Table X, p. 48. 



10. The curves based on these trials show irregularity at the 

 beginning and a slightly upward tendency toward the end, indi- 

 cating that the possible right preference was covered in the 

 early part of the series by disturbance such as manipulation 

 and the change in apparatus. Figs. 16-18, Plate VI. 



11. Experiments with Y-shaped labyrinth: roughness of the- 

 path as a warning stimulus; electric shock as punishment. — In 

 a series of 664 tests a Y-shaped path of glass tubing, unlimited 

 by wires, was used. One arm of the Y gave access to the surface- 

 of the water and the other was shortened and supplied with 

 wires over which an electric current could be sent. The space 

 between the point of divergence of the two arms and the wires 

 was roughened. The four snails tested chose the short arm or 

 punishment in 60.3^^ of the trials. This series of tests then,, 

 shows no indication of learning. Table XI, p. 63. 



12. The curves based on these trials show less irregularity 

 than those in the preceding tests (paragraph 10). The latter 

 part hes slightly above the 50% level indicating a possible 

 preference for the rough side, masked by disturbance in the 

 first part of the series, combined with punishment too long 

 delayed after the warning stimulus had been received. Figs. 

 19-22, Plate VII. 



13. Experiments with Y-shaped labyrinth: mechanical stimu- 

 lation as warning signal. — In a series of 930 tests the Y apparatus 

 was used but instead of roughness in connection with the electric 

 shock, the snail was stroked on the tentacles and dorsal head 

 region with a hair, preceding the shock, if a wrong choice were 

 made. Check tests showed that the irritation caused by the 

 warning stimulus was not sufficient to cause it to serve as a 

 punishment. The six snails tested chose the short arm and 

 punishment in 32.5% of the trials, the long arm, by chance 

 in 51.2% and the long arm through association in 15.6%. 

 Modification of behavior through the formation of an association 

 between the two stimuli, one serving as punishment, the other 

 as a warning signal, is thus shown. Table XII, p. 73. 



14. The curves based on these trials show irregularity through- 

 out but a downward tendency toward the latter end, indicating 



