on the strings from the floor. As he was very wild, he never realty alighted on the strings. He made many trials that were very similar to those of the female Sparrow. For the reason that she was not in good condition, her attempts to open the box were much delayed. Thus an abundant opportunity to open the door was offered the male.
Both his unsuccessful and his successful attempts, certainly the former, would, according to previous standards of imitation, without doubt be classed as due to imitation. It seems best, however, to delay such a classification until our tests have multiplied in number and other birds of the same and different species have been tried.
It has always seemed to the writer that students of Animal Behavior, particularly those who work with the higher animals, were in danger of overdoing the simple tests which deal only with simple association processes. As Professor Mills has pointed out, our tests should show the power of the animal in question to inhibit old associations and form new ones. Many of the series described will therefore have to do with changes in the location of the strings such that the above powers will be put to the test. Furthermore such modifications will also serve the very useful purpose of giving those birds expected to do the imitating the opportunity to manifest any tendency which the y may have in reality to do so.
Accordingly after a series of seventy-eight tests during the second half of which the errors .of the female English Sparrow were very few, the strings were changed to Place 2 on the opposite side of the door. What, now, is the effect on her behavior? In the part of the table following, "strings to opposite side of door," we see that at first she made numerous errors. But it is of interest to note that four of the nine errors in the first trial are on the new side, and are made because more force is required to open the door now. Of the eleven errors in the second trial, six are on the new side. In the third trial she went but twice to the old side, where there were no strings. In Fig. 6 we see that she came to use the beak in this series as well as in the first. It is only fair to state that about one-half the above errors are clearly due to the fact that her weight was not enough to pull up the latch. In fact by the third trial she had learned to wriggle while seizing the string in her claws. The left side of the door or this part of the box now seemed to be the more interesting to the male English Sparrow.
In that part of the table immediately following "strings on left end at front corner" (Fig. 7) we see that all eleven trials are without error except the first where there is but one. My notes record the fact that she inhibited the impulse to jump on the old positions many times. So far as "errors" are con-
