That afternoon a systematic search of the entire house was made. The cellar room in which the assault upon Oakes had occurred was thoroughly lighted and examined. The heap of rubbish which Mike had been investigating at our previous visit proved to be composed of plaster and bricks. The wall in which the door was cut was found to be about three feet thick, and one of the foundations of the house. It was solid, save for a chimney-like opening which had been trapped with the door. Above, at the level of the dining-room floor, the great wall ceased. From one edge was continued upwards the original partition between that room and the next--the parlor; but it was thin, and had evidently been recently strengthened by another wall, slightly thicker, and built from the opposite edge of the foundation, leaving a space between the two. Into this space entered, at a certain point, the opening from the cellar room below. It was a peculiar arrangement. As Oakes remarked, the new wall had been made with no regard to the economizing of space; for, had it been built immediately back of the old, considerable room would have been saved for the parlor. One of the "carpenters" thought that the original idea had been to utilize the space for closets. The only other possible use for it, so far as we could discover, was the one which Oakes had surmised--ventilation for the cellar. Still, to our ordinary minds, a chimney would have answered that purpose quite as well. A little further investigation, however, showed the top of the foundation wall to be covered with cement well smoothed, and the walls themselves were plastered. It was generally conceded, therefore, that the first idea had been to use it as closet room, which could easily have been done by cutting doors through the walls. As Oakes said, the notion had evidently met with opposition and been abandoned, so communication had been made with the cellar instead, and the roof opened to afford ventilation. The opening into the cellar was large. A man could easily enter it, and, standing, reach the top of the foundation wall; then, by a little exertion, he could raise himself into the intermural space. Oakes, Moore and I proved this by actual experiment and found that the passage was quite wide enough to accommodate a man of average proportions. I have said that the dining-room was finished in oak panels. These had been reached from our side of the wall by removing the bricks and mortar--the same stuff evidently which helped to form the rubbish heap in the room below. One of the larger panels had been made to slide vertically. It had been neatly done and had escaped detection from the dining-room because of the overlapping of the other panels. Some débris still remained between the walls. "The fellow we are after knew of the space between the walls and worked at the panel after the repairs were completed," was Oakes's remark. "How do you know that?" asked Moore.