test he did the same thing four times under similar conditions.
There now began that period of alternation which has happened in previous series in which first one then the other was successful. The first method of Oriole 2 was to stand on the lower string and strike the one just above. Several times Oriole 1 was chased away just as he was on the point of pulling the threads. In the following trials Oriole 2 was successful by flying up on strings near the top. It is evident that if he did imitate it was not for long. Flying up and pushing the top string is certainly not an imitation of Oriole 1 in his reaching through and pulling the threads. The number of efforts required of Oriole 2 at this stage would indicate that he had not yet singled out the single part and only that which must be worked.
The 62d test gave interesting proof of the statement just made. Oriole 2 alighted on the post just above the outer end of string C. In an apparently accidental manner he slipped from the post and struck this string and the door opened. Would he use this method in the following trials? I was watching for him with camera focused and the accompanying figure, 17, was the result. His erroneous association and the efforts which he made to operate the strings by merely alighting on this post was what allowed Oriole 1 to step in and be successful during the following two trials. This, too, in face of the fact that Oriole 2 was the first at the box. In the later trials the latter would often go to the post and thence up to the top string.
Usually in learning tests with animals it is unfortunate if the fastenings of the food-box fail to work as easily or in the same way as they have in the past. Yet as we shall see directly for these Orioles and later for the Crows, such an accident in tests on imitation is really a favorable occurrence. Reference to Table XI will show that the time is greatly lengthened for the 70th and 72d tests. It was evident that there was something wrong with the box. Oriole 1 was successful only after fifteen attempts, all save one of these being on the threads. Oriole 2 made eight attempts at hopping on the post and Oriole 1 made one such attempt. Apparently in imitation of Oriole 2, and in a very sudden and impulsive manner. In the 71st trial Oriole 1 opened while at the same time Oriole 2 was hanging to the door. What was the effect of this in the next trial? The length of time in this next trial is due partly to the fact that Oriole 2 must hop on the door five times. In fact he left the door to go directly to string C, Fig. 18. Again something entirely unnecessary becomes to this bird an important link in the association series because his hopping on
