"At least we may accept it as a working hypothesis. We will suppose, then, that something was burned in each case which produced an atmosphere causing strange toxic effects. Very good. In the first instance--that of the Tregennis family--this substance was placed in the fire. Now the window was shut, but the fire would naturally carry fumes to some extent up the chimney. Hence one would expect the effects of the poison to be less than in the second case, where there was less escape for the vapour. The result seems to indicate that it was so, since in the first case only the woman, who had presumably the more sensitive organism, was killed, the others exhibiting that temporary or permanent lunacy which is evidently the first effect of the drug. In the second case the result was complete. The facts, therefore, seem to bear out the theory of a poison which worked by combustion. "With this train of reasoning in my head I naturally looked about in Mortimer Tregennis's room to find some remains of this substance. The obvious place to look was the talc shelf or smoke-guard of the lamp. There, sure enough, I perceived a number of flaky ashes, and round the edges a fringe of brownish powder, which had not yet been consumed. Half of this I took, as you saw, and I placed it in an envelope." "Why half, Holmes?" "It is not for me, my dear Watson, to stand in the way of the official police force. I leave them all the evidence which I found. The poison still remained upon the talc had they the wit to find it. Now, Watson, we will light our lamp; we will, however, take the precaution to open our window to avoid the premature decease of two deserving members of society, and you will seat yourself near that open window in an armchair unless, like a sensible man, you determine to have nothing to do with the affair. Oh, you will see it out, will you? I thought I knew my Watson. This chair I will place opposite yours, so that we may be the same distance from the poison and face to face. The door we will leave ajar. Each is now in a position to watch the other and to bring the experiment to an end should the symptoms seem alarming. Is that all clear? Well, then, I take our powder--or what remains of it--from the envelope, and I lay it above the burning lamp. So! Now, Watson, let us sit down and await developments."