is to propose to oneself one or more provisional explanations or
hypotheses. We have done that, too--or, at least I have, and I suppose
you have."

"I haven't," said I. "There is Jeffrey's will, but why he should have
made the change I cannot form the foggiest idea. But I should like to
hear your provisional theories on the subject."

"You won't hear them at present. They are mere wild conjectures. But to
resume: what do we do next?"

"Go to New Inn and rake over the deceased gentleman's apartments."

Thorndyke smilingly ignored my answer and continued--

"We examine each explanation in turn and see what follows from it;
whether it agrees with all the facts and leads to the discovery of new
ones, or, on the other hand, disagrees with some facts or leads us to an
absurdity. Let us take a simple example.

"Suppose we find scattered over a field a number of largish masses of
stone, which are entirely different in character from the rocks found in
the neighbourhood. The question arises, how did those stones get into
that field? Three explanations are proposed. One: that they are the
products of former volcanic action; two: that they were brought from a
distance by human agency; three: that they were carried thither from
some distant country by icebergs. Now each of those explanations
involves certain consequences. If the stones are volcanic, then they
were once in a state of fusion. But we find that they are unaltered
limestone and contain fossils. Then they are not volcanic. If they were