Kennedy blew a cloud of smoke.  “There are three ways of tracing down a crime, aside from the police method of stool pigeons to betray the criminals and the detective bureau method of cross-examination under pressure, popularly known as the third degree.”

“What are they?” Mackay asked, unaware that Kennedy needed little prompting once he felt inclined to talk out some matter puzzling him.

“One is the process of reasoning from the possible suspects to the act itself—­in other words, putting the emphasis on the motive.  A second is the reverse of the first, involving a study of the crime for clues and making deductions from the inevitable earmarks of the person for the purpose of discovering his identity.  The third method, except for some investigations across the water, is distinctly my own, the scientific.

“In all sciences,” Kennedy went on, warming to his subject, “progress is made by a careful tabulation of proved facts.  The scientific method is the method of exact knowledge.  Thus, in crime, those things are of value to us which by an infinite series of empiric observations have been established and have become incontrovertible.  The familiar example, of course, is fingerprints.  Nearly everyone knows that no two men have the same markings; that the same man displays a pattern which is unchanging from birth to the grave.

“No less certain is the fact that human blood differs from the blood of animals, that in faint variations the blood of no two people is alike, that the blood of any living thing, man or beast, is affected by various things—­an infinite number almost—­ most of which are positively known to modern medical investigators.

“In this case my principal scientific clue is the blood left upon the portiere by the man who took the needle the night following the murder.  Next in importance is the fact, demonstrated by me, that some one at the studio wiped a hypodermic on a towel after inoculating himself with antivenin.  Of course I am presuming that this latter man inoculated himself and not some one else, because it is obvious.  If necessary I can prove it later, however, by analyzing the trace of blood.  That is not the point.  The point is that whoever removed the needle pricked himself and yet did not die of the venom—­unless it was a person not under our observation, an unlikely premise.  Therefore, because of this last fact, and because again it is obvious, I expect to find that the same individual inoculated himself with antivenin and removed the needle from the portiere; and I expect to prove it beyond possibility of doubt by an analysis of his blood.  A sample of the blood from this person will be identical with the spot on the portiere, and—­much the easier test—­will contain traces of the antitoxin.

“With that much accomplished, a little of the, well—­third degree, will bring about a confession.  It is circumstantial evidence of the strongest sort.  Not only does a man take precautions against a given poison, but he is proved to be the one who removed the needle actually responsible for Miss Lamar’s death.