Certain other interesting tricks developed in Julius's behavior. Thus, on July 5, there appeared the tendency to move as though about to enter the right box (feint), then to stop suddenly and promptly enter another box, which was, of course, a wrong one. The reason for the development of this tendency could not be discovered, but in connection with it, there appeared another tendency which possibly can be explained. Julius took to backing into the chosen box so that he could face the experimenter. He would then, after a period of hesitation, come out and promptly enter one of the other boxes. This tendency was apparently due to the fact that during one or two series the experimenter growled at the orang utan every time he made a mistake. The growl startled him and caused him to look around. He evidently felt the need of keeping his eyes on the experimenter,--Hence the backing into the open box. The tendency disappeared shortly after the experimenter gave up the use of the growl as a method of punishing the animal for what were suspected to be careless choices. Curiously enough, it was not until July 10 that direct choice of the right box was made at all frequently. Previously, selection of it had been made almost invariably after approach to other boxes. But in the second series for July 10 there was an extraordinary improvement in method. This developed in the presence of two visitors, and it is therefore all the more surprising. The choices were made not only directly, but with decision and evident certainty that was quite at variance with the previous behavior of the animal. All the while through variation of methods, I was seeking to discover the best means of holding the orang utan to his maximum effort and care in attempting to select the right box. One day it would seem as though forcing him to make round trips with rewards only for correct first choices proved most satisfactory, and the next it might seem equally clear that punishment by confinement for thirty seconds or sixty seconds, with reward for correct choice in every trial, yielded better results. In the end I had to admit that no best method had been demonstrated and that I had failed to develop conditions which served to compel the animal's attention to the problem and to lead him to work without discouragement. There were, it is true, days on which it seemed practically certain that the problem would be solved, but as it turned out, Julius never succeeded in choosing correctly--throughout a series of ten trials.