e original boy turns up alive, and reports that he had been
washed ashore down the stream and subsequently recovered. We find that
our direct evidence, with numerous witnesses to the actual fact, was
entirely misleading after all, because we had jumped to a conclusion,
without duly considering the attendant circumstances of the case. So
it is always. This is no case manufactured to point an argument. There
is no such thing as positive proof, which does not depend upon
circumstances. The old example may be cited briefly again. If you see
one man shoot at another and see the other fall and die, can you say
without further knowledge, that one killed the other? If you do, you
may find later that the pistol carried only a blank cartridge, and
that the man died of fright.

"It is equally true of circumstantial evidence, that without some
direct fact upon which it depends it is worthless. As an example of
this, I may as well save your time by introducing the case at issue.
If we could show you that the prisoner desired the death of this girl;
that he profited by her death; that he had a secret in connection with
her child which he can keep from the world better, now that she is
dead; that she died under circumstances which made the attending
physician suspect morphine poisoning; that as soon as the suspicion
was announced, the prisoner mysteriously disappeared, and remained in
hiding for several days; that he had the opportunity to administer the
poison; that he understood the working of the drug; and other
circumstances of a similar nature, the argument would be entirely
circumstantial. All this might be true and the man might be innocent.
But, selecting from this array of suspicious facts, the one which
indicates morphine as the drug employed, and then add to it the fact
that expert chemists actually find morphine in the tissues of the
body, and you see, gentlemen, that at once this single bit of direct
evidence gives substantial form to the whole. The circumstantial is
strengthened by the direct, just as the direct is made important by
the circumstantial. The mere finding of poison in a body, though
direct evidence as to the cause of death, neither convicts the
assassin, nor even positively indicates that a murder has been
committed. The poison might have reached the victim by accident. But
consider the attendant circumstances, and then we see that a definite
conclusion is inevitable. It is from the circumstantial evidence