LEARNING OF MAZE BY AMBYSTOMA 31 
99 which is the average of the third series, although two days 
intervened between the first and second series of trials. After 
this the average time interval remains fairly constant until the 
sixth series of trials which was made on the seventh day, when 
the experiment was discontinued by reason of the sluggishness 
200 
10 
1545617182884 1578 3023) 33 
Fig. 2. A B, Curve based on the averages of three trials in successive 
experiments of all the animals in the first series; A C, Curve based on 
the averages of the three trials in successive experiments of all the 
animals in the second series; A D, Curve based on the averages of three 
trials in successive experiments of larva No. 2 in the second series. 
The ordinates indicate the time in seconds; the abscissae, the days on 
which the trials were made. 
of the remaining larvae due to their approaching metamorpho- 
Sis. 
The second series consisted of six larvae. For this series 
three partitions of screen were used in the aquarium as illus- 
trated in fig 1. The first (hi), with an aperture at the bottom 
2.5 em. square, and 8.5 em. from one end, was placed approxi- 
mately 10 em. from the unshaded end. The second (jk), with 
