t Hans had not made an error of calculation
but merely of counting in the process of giving his result, which always
had to be done by the cumbersome method of tapping. As a matter of fact,
the trouble lay in the wrong degree of concentration on the part of the
questioner: In errors of +1, tension was too slight, in those of -1, it
was too great (see page 91). This comes out clearly in a comparison of
the two more extensive series which I took in the case of Mr.
Schillings. During the first series, he was well disposed, and was able
to concentrate effectively, while during the second, he was nervous and
easily diverted. This difference in intensity of concentration in the
case of the two series is attested, not only subjectively by Mr.
Schillings's introspective statement, but may be measured objectively by
means of the number of final taps which the horse gave with his left
foot during these two series. We saw (page 94) that these final taps
were always a sign of intense concentration and, as a matter of fact,
one-half of the horse's responses to Mr. Schillings during the first
series were made in this way; whereas, in the second series, only
one-third were of this sort. (I, myself, was never able to get, without
conscious control, a greater number of this type of response.) We may
therefore say that, in the first series we had a high degree of tension,
or concentration, whereas, in the second series, we had a low degree.
The errors distribute themselves over the two series as follows:

+1 +2 -1 -2
Series I (31 tests)
Correct responses: 87%
Incorrect " : 0% 0% 13% 0%
Series II (40 tests)
Correct responses: 40%.
Incorrect " : 40% 8% 2.5% 0% (and 9.5%
other kinds of errors.)

We find in Series I no "+1" errors, but only "-1" errors; in series II,
on the other hand, the errors are almost exclusively of the "+1"
category, equaling the number of correct responses, and there is only
one "-1" error. A series obtained in the case of Mr. von Osten is almost
as satisfactory an illustration. When he first began to take part in
tests in which the procedure was the one we characterized as "without
knowledge" and had to note their complete failure, he was thrown into
such confusion that the responses in the case of procedure with
knowledge were also incorr