and a curious feeling of relaxation
would ensue. I have made a series of tests to determine the most
favorable degree of tension in expectation. It was possible to
distinguish with certainty, three degrees of tension besides the state
of utter relaxation,--all of which I measured by means of the
differences in the sensations of tension. In cases of tension of the
first degree (greatest concentration) the responses were usually
correct, a few, however, were lacking by one unit. There was therefore
in the latter instance a premature release of inner tension. In cases of
tension of the second degree all answers were correct except a very few
which were too great by one unit. In cases of tension of the third
degree, many answers were wrong, and usually by several units too many.
I wished to have the horse tap 10, with the lowest degree of
concentration. He tapped 13, then in a repetition of the test, 12. I
thereupon increased the tension, Hans then tapped 8. I decreased the
tension once more, but so that it was somewhat greater than at first.
Hans tapped 10 correctly. At another time I tried to have him tap the
number 5, with a low degree of tension. He tapped 6. I intensified
expectation and Hans tapped 4. I again decreased it, and he tapped 5,
_comme il faut_. Apparently, therefore, the most favorable degree of
tension was one between the first and second,--the latter being the
least favorable. After some practice a lesser degree than was used in
the beginning sufficed to evoke adequate reactions. The flow of nervous
energy to the motor centers of the brain evidently became facilitated
through practice. It will be easy to understand why the first days of
experimentation caused intense headaches, which later never occurred.

Whenever, in the foregoing, we spoke of a certain degree of
concentration which had to be attained, it is not to be understood that
the same tension had to be maintained throughout the test, from the
horse's first tap to his last. But rather, that it began with a low
degree, and gradually increased as the final unit of the count was being
approached. It may best be represented by a curve whose maximum
represents that degree of tension which we have been discussing. The
rise to this maximum which, when attained, was followed by a sudden
fall, did not always occur in the same manner. Three types of curve may
be distinguished, which were first discovered in purely empirical
fashion, and later reproduced v