AN ANALYSIS OF THE LEARNING PROCESS IN THE SNAIL 85 



An inspection of the graphs (Plate VI) shows that the curves 

 are quite irregular and with the progress of time there is a slight 

 tendency to rise above the 50% level. The snails not only do 

 not learn but their percentage of errors apparently increases as 

 the series of trials lengthens. Nothing has been detected in the 

 physical environment to direct the snail to either side of the path. 

 It is in a symmetrical environment as to structure of labyrinth, 

 surrounding tank and house, lighting and temperature. It seems 

 probable that the disturbance due to manipulation and the 

 change in conditions incident to the experimental work probably 

 caused irregularities in the first part of the series which covered 

 the slight tendency to turn to the error (right) side of the path, 

 evident in the latter part of the curves. The graphs show then, 

 a probable masking of the " right " preference in the early part 

 of the experiments, due to disturbance rather than increase in 

 this preference as the series lengthens. There is no evidence of 

 learning. 



2. In the second series of labyrinth tests the U-shaped laby- 

 rinth was discarded and in its place was used a Y-shaped 

 labyrinth of cylindrical glass tubing. Owing to the snail's 

 tendenc}^ to crawl in a spiral such a labyrinth offers no choices 

 of right or left and the snail is free to move in any direction in 

 order to reach the top. One arm of this labyrinth was made 

 rough while the other remained smooth. Thus roughness served 

 as a means of discrimination between the two. At the upper 

 end of the rough arm the snail received an electric shock. The 

 roughness served as a warning signal of the approaching shock. 

 No rough-smooth preference tests were made. 



The curves (Plate VII) again show much irregularity, but not 

 so much as those with the U-shaped labyrinth. There are again 

 days on which all choices are correct and other days on which all 

 are incorrect. Nearly all of these days, however, fall in the first 

 third of the curves, with the result that this part is much more 

 irregular than that which follows. After their first , thirds the 

 curves remain nearer the 50%: level, but on the whole tend 

 to lie above it. In other words, the percentage of error again 

 appears to increase as the series lengthens. It seems possible 

 that a greater number of trials might have brought the curves 

 again to the 50% line or even slightly below it, indicating that 

 the upward tendency was due to a wave of error and that a 



