I now replaced the bait and gave the monkey a second chance to obtain it.  Almost immediately he climbed up as far as the second box, but although he could reach the banana only from the uppermost box, he deliberately shoved it off to the ground and sat down upon box 2.  As he was unable to obtain the banana from this, he soon began to gnaw and pull at it, and as he was succeeding all too well in his efforts to tear the box to pieces, he had to be returned to his cage.

The most important features of his behavior were, first, his stealthy and indirect manner, and second, his failure to use other means of obtaining the bait than that supplied by the observer.  Instead of looking straight at the experimenter, or at the object which he wished to obtain, he apparently looked and attended elsewhere.  For this reason it was often difficult to decide whether or not he had noticed the bait or the boxes.  Finally I was led to conclude that he usually knew exactly what was going on and had in his furtive way noted all of the essential features of the situation, and that his manner was extremely indicative of his mental attitude of limited trust.  Both Julius and Skirrl went to the opposite extreme in the matter of directness, or as we should say in human relations, frankness.  They would look the experimenter directly in the eye, and they usually gazed intently at anything, such for example as the bait, that interested them.  Sobke, even when very hungry, instead of going directly toward the bait, and trying to obtain it, usually did various other things as though pretending that he had no interest in food.

On the following day, June 15, the three boxes were again placed nearly under the banana, but this time the two smaller boxes, numbers 1 and 2, were pushed to the extreme end of the lower box and so far from the bait that it could not be reached from box 1.  It was necessary then for the animal to push boxes 1 and 2 along on box 3 until they were nearer the bait.

Sobke, when admitted to the cage, evidently noticed the banana, but as formerly, he made no immediate effort to obtain it.  After wandering in search of food and quarreling with the other monkeys for several minutes, he went to the boxes, pushed the topmost one, number 1, off on to the floor, and then carried it into his cage where he quickly tore one side off.  He next returned to the large cage, climbed up on box 2, and he was able, by jumping, to reach and obtain the banana.

As Sobke was very good at jumping, his new method rendered the box stacking experiment of uncertain value, since it was next to impossible so to arrange the spatial relations of bait and boxes that he should be neither discouraged by too great a distance nor encouraged to jump by too small a distance.  Evidently it would be more satisfactory to simplify the conditions by trying to discover, first of all, whether he would use a single box as a means of reaching the reward.