e hither and
thither in most irregular fashion while the horse was going through his
tapping, and would therefore make the perception of minute movements all
the more difficult.

Nor was there anything in the exhibitions given at the same time in a
Berlin vaudeville by the mare "Rosa," which might have shattered this
belief. For, in the case of this rival of Hans, the movements involved
were comparatively coarse. The closing signal consisted in bending
forward on the part of the one exhibiting the mare, while up to that
point he had stood bolt upright. Most persons were not aware of this,
because this change in posture cannot be noticed from the front. I
happened to sit to the side and caught the movement every time. It was
the same that was noted by Dr. Miessner, another member of the
commission, (see page 256), but concerning which he did not give me a
more complete account. Later I learned through Professor Th. W.
Engelmann that the very same movement was employed not long ago, for
giving signals to a dog exhibited at Utrecht. This particular movement
is very well adapted to commercial purposes, since the spectator always
tries to view the performance from a point as nearly in front of the
animal and its master as possible, thus making the detection of the
trick all the more difficult.

The details of the various experiments made by this commission are given
in an excerpt from the records kept by Dr. von Hornbostel, which I
showed to a small group of persons a few days after the 12th of
September (Supplement III). At that time none of the particulars was
published, because the commission wished to wait until some positive
statement might be made. The public was merely to be assured that a
group of reputable men, from different spheres of life, who could have
no purpose in hazarding their reputation, believed that the case was one
worthy of careful investigation.

I left Berlin on September 17th and did not return until October 3d. In
the meantime Mr. Schillings continued the investigation, and was
assisted in part by Mr. Oskar Pfungst, one of my co-workers at the
Psychological Institute. For the first time a number of tests were now
made in which neither the questioner, nor any of those present knew the
answer to the problem. Such tests naturally were the first steps toward
a positive investigation. The results were such that Mr. Schillings was
led to replace his hypothesis of independent conceptual thinking by on