The rate of the flow of different gases under constant pressure through capillary tubes into a vacuum, constitutes the capillary transpiration of gases. These rates bear a constant proportion to one another, but they are singularly unlike the rates of effusion. They are independent of the material of the tube; they are not governed by specific gravity; and ‘they appear to be in constant relation with no other known property of the same gases; and they form a class of phenomena remarkably isolated from all else at present known of gases.’ The pores of graphite are so fine that it is incapable either of effusion or transpiration, but it is readily penetrated by means of the molecular or diffusive movements of gases, as appears on comparing the time requisite for the passage of equal volumes of different gases under constant pressure into a vacuum. For oxygen, hydrogen and carbonic acid gas, the times are nearly as the square roots of their densities. The atmolysis or partial separation of mixed gases and vapours of unequal diffusibility, can be effected by allowing the mixture to penetrate through a graphite plate into a vacuum. The amount of separation is in proportion to the pressure, and attains its maximum when the gases pass into a perfect vacuum. One of the results of atmolysis was the concentration of oxygen in atmospheric air. When a portion of air confined in a vessel was allowed to penetrate into a vacuum through graphite or unglazed earthenware, the nitrogen passed more rapidly than the oxygen in the ratio of 1·0668 to 1, and the portion of oxygen is proportionally increased in the air left behind in the vessel. The increase of oxygen actually observed when the air in the vessel was reduced from 1 volume to 0·5 was 0·48 per cent. The diffusion was continued till the air in the vessel was reduced to 0·0625 and the concentration of the oxygen in it amounted to 2·02 per cent. The molecular or diffusive mobility exercises a certain influence on the heating of gases by contact with heated liquid or solid substances. The more rapid the molecular movement of a gas is, the more frequent will be the contact of the molecules and the quicker will be the communication of heat. The greater cooling power of hydrogen compared with that of oxygen or air is probably owing to that cause. ‘Oxygen and hydrogen gas have the same specific heat for equal volumes; but a hot object placed in hydrogen is really touched 3·8 times more frequently than it would be if placed in oxygen gas. Dalton had already ascribed this peculiarity of hydrogen to the high mobility of the gas.’[13] It appears that isomorphic substances such as chloride, bromide, and iodide of sodium, have a similar diffusibility, another of the many analogies between these singular marine substances.