LEARNING OF MAZE BY AMBYSTOMA 29 
treat from the source of the light. In order to determine whe- 
ther the larvae of Ambystoma would find the shaded area at the 
end of the aquarium toward the sun as readily as they would 
a shaded area at the opposite end, individual animals were 
placed in a transverse position at the center of the aquarium so 
that they were free either to turn toward or away from the area 
shaded by the end wall of the aquarium which was toward the 
n k i 
Fig. 1. Diagram of maze with compartments 1, 2, 3, and 4. xz, surface 
of sand; xy, surface of water; hi, jk, ln, partitions with apertures. 
Arrows indicate the shortest route through the maze. 
sun. The first animal used in this experiment turned toward 
the shaded area eight times in the first ten trials, and without 
exception in the next ten trials. This record is typical of the 
records made by all the animals used in this experiment. 
Two glass plates 10 em. wide were now set up near the middle 
of the aquarium parallel with each other and about 2 em. apart 
at right angles to, and in contact with, one of the lateral walls. 
An animal placed between these glass plates with its head to- 
ward the wall would have to move backward before it would be 
free to turn toward either end of the aquarium. This arrange- 
ment would eliminate any directive influence due to manipula- 
tion which might have been present in the previous experiment. 
One of the animals used in the previous experiment when sub- 
jected to a series of trials in this manner turned toward the 
shaded end of the aquarium without exception in the first ten 
trials. The glass plates were now placed in contact with the 
opposite wall of the aquarium in the same manner. The same 
animal was placed between them with its head toward the wall. 
Whereas in the previous series of trials it was necessary, in or- 
