t till very late--or
early. But I feel sure I've got a strangle-hold on this mystery. If I
get those papers from Riley in good time to-morrow I shall invite you
and several others to a grand demonstration here to-morrow night. Don't
forget. Keep the whole evening free. It will be a big story."

Kennedy's laboratory was brightly lighted when I arrived early the next
evening. One by one his "guests" dropped in. It was evident that they
had little liking for the visit, but the coroner had sent out the
"invitations," and they had nothing to do but accept. Each one was
politely welcomed by the professor and assigned a seat, much as he would
have done with a group of students. The inspector and the coroner sat
back a little. Mrs. Parker, Mr. Downey, Mr. Bruce, myself, and Miss
La Neige sat in that order in the very narrow and uncomfortable little
armchairs used by the students during lectures.

At last Kennedy was ready to begin. He took his position behind the
long, flat-topped table which he used for his demonstrations before his
classes. "I realise, ladies and gentlemen," he began formally, "that I
am about to do a very unusual thing; but, as you all know, the police
and the coroner have been completely baffled by this terrible mystery
and have requested me to attempt to clear up at least certain points in
it. I will begin what I have to say by remarking that the tracing out
of a crime like this differs in nothing, except as regards the
subject-matter, from the search for a scientific truth. The forcing of
man's secrets is like the forcing of nature's secrets. Both are pieces
of detective work. The methods employed in the detection of crime
are, or rather should be, like the methods employed in the process of
discovering scientific truth. In a crime of this sort, two kinds of
evidence need to be secured. Circumstantial evidence must first be
marshalled, and then a motive must be found. I have been gathering
facts. But to omit motives and rest contented with mere facts would be
inconclusive. It would never convince anybody or convict anybody. In
other words, circumstantial evidence must first lead to a suspect, and
then this suspect must prove equal to accounting for the facts. It is my
hope that each of you may contribute something that will be of service
in arriving at the truth of this unfortunate incident."

The tension was not relieved even when Kennedy stopped speaking and
began to fuss with a little upright target