As a result of numerous analyses it is concluded, first, that the oxygen exists normally in the atmosphere of soils in large proportion; second, very probably the gaseous atmosphere of arable soils, to a depth of sixty centimeters, contains scarcely one per cent of carbon dioxid and about twenty per cent of oxygen; third, the highest percentages of carbon dioxid correspond to epochs of highest temperature and periods of greatest calm; fourth, the proportion of carbon dioxid increases ordinarily with the depth at which the samples are taken. This disposition of the carbon dioxid would appear almost necessary, since near the surface the internal atmosphere is almost constantly diluted by external air by virtue of diffusion. Fifth, from one epoch to another the composition of the atmosphere of the soil can undergo considerable variation. =287. Determination of Diffusion of Carbon Dioxid in Soil.=—The method proposed by Hannén[188] is a convenient one to use in studying the rate of diffusion of carbon dioxid in soils. A large Woulff’s bottle with three necks serves for the reception of the gas. The two smaller outer necks of the bottle carry two glass tubes bent outwards and provided with stop-cocks. One of these passes to near the bottom of the bottle and the other just through the stopper. The middle tubule of the bottle is of a size to give in section an area of about twenty-two square centimeters. It is made with a heavy rim two centimeters wide and plane ground. This rim carries a plane-ground glass plate with a circular perforation in one-half of it, of the size of the opening in the central tubule of the bottle. A glass cylinder, carrying a fine wire-gauze diaphragm near the lower end, fits with a ground-glass edge air-tight, over this aperture, being held in position by a brass clamp. The ground-glass plate moves air-tight between the cylinder and the bottle, so that the cylinder can be brought into connection with the bottle or cut off therefrom without in any way opening the bottle to the air. The plate and all ground movable surfaces should be well lubricated with vaseline. The experiment is carried on as follows: The glass cylinder is filled with the soil to be tested, closed above with a rubber stopper carrying a gas tube, and then by moving the perforated-glass plate brought into connection with the bottle. The side tube, with short arm inside the bottle, is then closed, and carbon dioxid introduced through the other lateral tube until the gas passing from the tube at the top of the cylinder is pure carbon dioxid.