86 ELIZABETH LOCKWOOD THOMPSON 



corresponding wave showing fewer errors might follow. It is 

 more probable however, that the greater irregularity evident 

 in the first thirds of the curves was due in part to disturbance 

 following the change of apparatus and that the slight increase 

 in error toward the end was the result of a possible rough 

 preference linked with a too long delayed punishment. 



3. The tliird scries of labyrinth tests was planned to test 

 the capacity of the snail to form associations on the labyrinth 

 that it had failed to solve. The Y labyrinth was used, but all 

 traces of roughness were removed. The short arm was shortened 

 still more making the interval between the warning stimulus 

 and the punishment less than in the preceding tests. The 

 warning stimulus adopted was the irritation caused by stroking 

 the dorsal head region and tentacles of the snail with a hair. 

 That this stimulus was an irritation was evidenced by the fact 

 that its use caused the snail to withdraw its tentacles,, and at 

 times to partially retract its head toward the edge of its shell. 

 Check tests, however, showed that the irritation was not severe 

 enough in itself to act as a punishment. The labyrinth as before 

 offered no choice of right or left. 



The curves (Plate VIII) of error percentage show much irreg- 

 ularity but a slight downward tendency. In 15.6% of the total 

 of 930 trials, the snails changed their course from the wrong to 

 the right path after contact with the warning stimulus but before 

 the shock (punishment) was received. Since the warning stimu- 

 lus was known to irritate the animal, there could have been no 

 preference for the wrong side of the labyrinth to offset the 

 associative effect of the shock (punishment). The 15% of 

 correct-by-association trials, then, represents the strength of 

 the association formed between the warning stimulus and the 

 punishment, no part of which could have been masked through 

 preference for the signal stimulus used. This series of tests 

 thus shows the formation of a weak association between two 

 stimuli, one used as a warning of the punishment to follow if 

 the course be not changed. Physa then, profits by experience, 

 but cannot be said to solve a labyrinth. The work with the 

 modified apparatus is merely confirmatory of that by the method 

 of simultaneous stimuli. It shows no greater capacity on the 

 part of the snail than was made evident by that method except 

 that to form an association between stimuli separated by a 

 small time interval. 



