t with number
three, dated the eighteenth of September last year."

"They look to me identical," said I, after a careful comparison.

"So they do to me," said Thorndyke. "Neither of them shows the change
that occurred later. But if you look at number two, dated the sixteenth
of September, you will see that it is in the later style. So is number
four, dated the twenty-third of September; but numbers five and six,
both at the beginning of October, are in the earlier style, like the
signature of the will. Thereafter all the signatures are in the new
style; but, if you compare number two, dated the sixteenth of September
with number twenty-four, dated the fourteenth of March of this year--the
day of Jeffrey's death--you see that they exhibit no difference. Both
are in the 'later style,' but the last shows no greater change than the
first. Don't you consider these facts very striking and significant?"

I reflected a few moments, trying to make out the deep significance to
which Thorndyke was directing my attention--and not succeeding very
triumphantly.

"You mean," I said, "that the occasional reversions to the earlier form
convey some material suggestion?"

"Yes; but more than that. What we learn from an inspection of this
series is this: that there was a change in the character of the
signature; a very slight change, but quite recognizable. Now that change
was not gradual or insidious nor was it progressive. It occurred at a
certain definite time. At first there were one or two reversions to the
earlier form, but after number six the new style continued to the end;
and you notice that it continued without any increase in the change and
without any variation. There are no intermediate forms. Some of the
signatures are in the 'old style' and some in the 'new,' but there are
none that are half and half. So that, to repeat: We have here two types
of signature, very much alike, but distinguishable. They alternate, but
do not merge into one another to produce intermediate forms. The change
occurs abruptly, but shows no tendency to increase as time goes on; it
is not a progressive change. What do you make of that, Jervis?"

"It is very remarkable," I said, poring over the cards to verify
Thorndyke's statements. "I don't quite know what to make of it. If the
circumstances admitted of the idea of forgery, one would suspect the
genuineness of some of the signatures. But they don't--at any rate, in
the case of the later will, t