of objects merely described by another person. The crystal-gazer may know
the inquirer so intimately as to have a very good guess at the subject of
his meditation. Again, a man is likely to be thinking of a woman, and a
woman of a man, so the field of conjecture is limited. In answer to the
first objection I may say that the crystal-gazer was among strangers, all
of whom, myself included, she now saw for the first time. Nor could she
have studied their histories beforehand, for she could not know (normally)
when she left home, that she was about to be shown a glass ball, or whom
she would meet. The second objection is met by the circumstance that
ladies were _not_ usually picked out for men, nor men for women. Indeed,
these choices were the exceptions, and in each case were marked by
minutely particular details. A third objection is that credulity, or the
love of strange novelties, or desire to oblige, biases the inquirers, and
makes them anxious to recognise something familiar in the scryer's
descriptions. In the same way we know how people recognise faces
in the most blurred and vague of spiritist photographs, or see family
resemblances in the most rudimentary doughfaced babies. Take descriptions
of persons in a passport, or in a proclamation sketching the personal
appearance of a criminal. These fit the men or women intended, but they
also fit a crowd of other people. The description given by the scryer then
may come right by a fortuitous coincidence, or may be too credulously
recognised.

The complex of coincidences, however, could not be attributed to chance
selection out of the whole possible field of conjecture. We must remember,
too, that a series of such hits increases, at an enormous rate, the odds
against accidental conjecture. Of such mere luck I may give an example. I
was writing a story of which the hero was George Kelly, one of the 'Seven
Men of Moidart.' A year after composing my tale, I found the Government
description of Mr. Kelly (1736). It exactly tallied with my purely