ses, they probably
have normal vision. But we are told that many horses--and according to
some authors all--have an innate imperfection which detracts
considerably from the clarity of vision. This imperfection consists in
an irregular formation of the sclerotic coat and of the lens of the
eye.[38] The two organs do not have the same refraction in all parts. As
a result, objective points are not imaged as points upon the retina.
(Hence the name: astigmatism, i. e., "without points", for this
disorder.) The retinal image of the object is not only vague, but also
distorted.[X]

[Footnote X: There is no justification for the wide-spread belief
that the horse which on account of the greater size of his eye (more
correctly, on account of the greater focal distance) receives larger
retinal images of objects than does the human eye, for that reason
also sees objects, larger than we do. Horses' shying is often
explained in this way. But the conclusion just mentioned is
erroneous. The retinal image is not the perceptual image. It
undergoes many transformations within the nervous system itself.]

Many will doubt whether with such imperfect images an animal can react
to directives so minute, as we have asserted to be true in the case of
Hans. In considering this question we must distinguish between the
directives for pointing out colors and the directives for tapping and
for head movements on the part of the horse. In pointing out and
bringing forth pieces of colored cloth there is involved the perception
of an object at rest, viz.: the direction of the questioner who is
standing quietly; whereas in the case of responses by means of tapping
the stimulus is the horse's perception of the questioner's movements.
Now, the construction of the horse's eye, as described above, is not
favorable for the perception of objects (so-called acuity of vision).
This may partly account for the slight success of the horse in those
tests in which he was required to select a piece of cloth of a
designated color, in so far as these commands were not accompanied by
calls or exhortations. Where human observers averaged eighty per cent
correct responses (page 135), Hans, under similar conditions was
successful in only one-third of the tests. In his errors he was also
wider of the mark than were the human observers (page 82). The object
perceived, to be sure, is a large one, viz.: the questioner, and he at
close range. We