pistol was cocked when it was in the manikin’s belt, for on the half-completed picture it was so painted by Mandel on the last day of his life. Furthermore, the position of the right index finger of the manikin can also be plainly seen in the picture; for the artist, not having a model to hold the weapon, had roughly rubbed in the angular fingers of the lay figure, preparatory to finishing the hand from life. The pistol then, being loaded and cocked, needed but the pressure of the finger to discharge it. That pressure was given by the rain on the night of the death of my friend. The lieutenant will find, on reference to his note-book, that on the morning when he examined the studio there had been quite a serious leak in the ceiling, and that the water had fallen directly on the manikin. He will find also in his notes the exact position of the manikin in reference to the divan on which the corpse lay. Now, it is clear that when the wrist of the manikin was bent, and the index finger was placed on the trigger of the pistol, only a very slight motion of the whole was necessary to give the pressure required to fire a pistol. The weapon was braced against the inside of the thumb of the hand, and thus held firmly there as it stuck in the belt ready to be drawn and fired. When the water first fell from the ceiling, it soaked the covering of the wrist and hand, and swelled the wrist joint so that it became absolutely immovable. Next the moisture extended to the tip of the fingers, the hand being held somewhat downward. In the manikin we have here, the exact construction of the fingers and the movement of the joints of the hand and wrist can be plainly seen. In my working model I have imitated the mechanism of one finger, so arranging it that the least deflection of the finger from the straight line will be measured on a scale of millimetres. The joints are so constructed that any elongation of the pieces of wood will curve the line of joints away from the straight line which I have drawn on the board. I propose to experiment with this model so as to make it perfectly plain that my friend’s death was accidental. If the experiment were tried on the manikin, and with a flint-lock pistol, it would doubtless fail ninety-nine cases out of a hundred. In the accident which caused my friend’s death everything happened to be perfectly adjusted. If my model works, of course the manikin might have worked in exactly the same way.”