is indicated on the curve by a dot, while the second or third series on
a given day, although space is allowed for them, are not so indicated.
Consequently, the form of the curve is determined chiefly by the first
series per day. The extreme irregularities of this curve are most
interesting and puzzling, as are also the variations in the daily ratios
of right to wrong first choices. Three times in the course of the
training, this ratio rose to 1 to 9, or higher. The causes for such
extreme variations are not easily enumerated, but a few of the most
obvious contributory causes are variations in the weather, especially
cloudiness or fogginess, which rendered the apparatus dark; variations
in the degree of hunger or eagerness for food; differences in the
activities of the animals in the cages outside of the laboratory
(sometimes they were noisy and distracted the subject), and finally,
differences in the physical fitness and attitude of the animal from day
to day.

The more or less incidental behavior in connection with this experiment
more strongly than the statistical results of the work on problem 2
indicate the existence of imagery. That ideas played a part in the
solution of the problem is probable, but at best they functioned very
ineffectively. The small number of methods used in the selection of the
right box, and the slight variations from the chief method, that of
choosing the first box at the right end and then the one next to it,
apparently justify Doctor Hamilton's characterization of this monkey as
defective.


[Illustration: FIGURE 19.--Error curves of learning for the solution of
problem 2 (second box from right end).]