The discovery of a small picture fallen from the wall of the drawing-room was quite sufficient to satisfy Liddy that the alarm had been a false one, but I was anything but convinced. Allowing for my nerves and the fact that small noises magnify themselves at night, there was still no possibility that the picture had made the series of sounds I heard. To prove it, however, I dropped it again. It fell with a single muffled crash of its wooden frame, and incidentally ruined itself beyond repair. I justified myself by reflecting that if the Armstrongs chose to leave pictures in unsafe positions, and to rent a house with a family ghost, the destruction of property was their responsibility, not mine. I warned Liddy not to mention what had happened to anybody, and telephoned to town for servants. Then after a breakfast which did more credit to Thomas’ heart than his head, I went on a short tour of investigation. The sounds had come from the east wing, and not without some qualms I began there. At first I found nothing. Since then I have developed my powers of observation, but at that time I was a novice. The small card-room seemed undisturbed. I looked for footprints, which is, I believe, the conventional thing to do, although my experience has been that as clues both footprints and thumb-marks are more useful in fiction than in fact. But the stairs in that wing offered something. At the top of the flight had been placed a tall wicker hamper, packed, with linen that had come from town. It stood at the edge of the top step, almost barring passage, and on the step below it was a long fresh scratch. For three steps the scratch was repeated, gradually diminishing, as if some object had fallen, striking each one. Then for four steps nothing. On the fifth step below was a round dent in the hard wood. That was all, and it seemed little enough, except that I was positive the marks had not been there the day before. It bore out my theory of the sound, which had been for all the world like the bumping of a metallic object down a flight of steps. The four steps had been skipped. I reasoned that an iron bar, for instance, would do something of the sort,—strike two or three steps, end down, then turn over, jumping a few stairs, and landing with a thud. Iron bars, however, do not fall down-stairs in the middle of the night alone. Coupled with the figure on the veranda the agency by which it climbed might be assumed. But—and here was the thing that puzzled me most—the doors were all fastened that morning, the windows unmolested, and the particular door from the card-room to the veranda had a combination lock of which I held the key, and which had not been tampered with.