and the prevention of escape of the subjects by completely covering the
apparatus and alleyways with a heavy wire netting.

Each of the eighteen doors of the multiple-choice boxes, and in addition
doors 11, 12, and 15 of the runway D, were operated by the observer from
his bench C by means of weighted window cords which were carried by
pulleys appropriately placed above the apparatus. Each weight was so
chosen as to be just sufficient to hold its door in position after the
experimenter had raised it. For the convenience of the experimenter in
the rapid operation of the twenty-one doors, the weights for the doors
of runway D were painted gray, those for the entrance doors, white, and
those for the exit doors, black.

In each entrance door, as is shown in figure 15 of plate IV, a window
was cut so that the experimenter might watch the animal after it had
entered a given box, and especially note when it left the box after
having received its reward. This window was covered with wire netting.
No such windows were necessary in the exit doors, but to them were
attached heavy galvanized iron flanges which served to cover the food
receptacles. One of these flanges is labelled o in figure 17. The food
receptacles were provided by boring holes in a 2 by 4 inch timber
securely nailed to the floor immediately outside of the exit doors. Into
these holes aluminum cups fitted snugly, and the iron flanges, when the
doors were closed, fitted so closely over the cups that it was
impossible for the animals to obtain food from them.


[Illustration: FIGURE 17.--Ground plan of multiple-choice apparatus in
experiment room A. Scale 1/60