"a very
small error in perception may sometimes lead to a very large error of
inference," Miss Johnson ends her remarkably interesting paper with two
illustrations--one a hallucination (?)[31] induced by false association
of ideas; the other an incident in her own experience, occurring at a
seance with Eusapia Palladino. Both of these are of importance, and
should be studied carefully.

Count Solovovo on the contrary considers it somewhat in favour of the
hallucination theory that hands were found to melt in the sitters'
grasp, when they were forcibly retained (p. 441). I cannot agree with
this. It is a different thing to say that hallucination might account
for the facts, and saying that the facts tell in favour of
hallucination. Chance might account for an experimental apparition, but
the fact that the apparition occurred does not prove it to be chance.
One must be careful to distinguish facts and inferences, in a case of
this character. Whether or not the hands were hallucinatory will depend,
not upon _a priori_ probability, or the fact they were visible to some,
invisible to others, (for all this might just as well be accounted for
on the opposing theory), but upon the fact that, so far as we know,
there is no analogy whatever between this oft-recorded event and any of
the phenomena of suggestion known to us. If we offer a theory to explain
certain facts, it must not only explain them in a rational manner, but
must dovetail into what we know--into _the known_. That is the whole
method of science. If, therefore, a man advances "hallucination" as an
explanation of such facts as those under discussion, he must show how it
is that hallucination might be supposed to work: he must bring forward
some analogies and examples of somewhat similar instances in order to
have a case at all. In science, we cannot speculate _in vacuo_, but must
connect with what is already known, if we wish to be scientific at all.
What analogies, then, have we that spirit-hands, similar to those
described, can be created by suggestion; and that suggestion can cause a
number of investigators, at various times, in various places, to believe
that these hands melted in theirs while they were trying to retain them?

I venture to think we have no analogies whatever. It is quite possible
that a subject in a hypnotic trance might be induced to believe that he
was holding a hand while in fact no hand was there, and, further, that
this hand melted away in