In the years 1846-1849 Graham resumed this inquiry; he now distinguished between _diffusion_, or the passage of gases through porous plates, and _transpiration_, or the passage of gases through capillary tubes. He showed that if a sufficiently large capillary tube be employed the rate of transpiration of a gas becomes constant, but that it is altogether different from the rate of diffusion of the same gas. He established the fact that there is a connection of some kind between the transpiration-rates and the chemical composition of gases, and in doing this he opened up a field of inquiry by cultivating which many important results have been gained within the last few years, and which is surely destined to yield more valuable fruit in the future. Returning to the diffusion of gases, Graham, after nearly thirty years' more or less constant labour, begins to speculate a little on the causes of the phenomena he had so studiously and perseveringly been examining. In his paper on "The Molecular Mobility of Gases," read to the Royal Society in 1863, after describing a new diffusion-tube wherein thin plates of artificial graphite were used in place of plaster of Paris, Graham says, "The pores of artificial graphite appear to be really so minute that a gas _in mass_ cannot penetrate the plate at all. It seems that molecules only can pass; and they may be supposed to pass wholly unimpeded by friction, for the smallest pores that can be imagined to exist in the graphite must be tunnels in magnitude to the ultimate atom of a gaseous body." He then shortly describes the molecular theory of matter, and shows how this theory--a sketch of which so far as it concerns us in this book has been given on pp. 123-125--explains the results which he has obtained. When a gas passed through a porous plate into a vacuum, or when one gas passed in one direction and another in the opposite direction through the same plate, Graham saw the molecules of each gas rushing through the "tunnels" of graphite or stucco. The average rate at which the molecules of a gas rushed along was the diffusion-rate of that gas. The lighter the gas the more rapid was the motion of its molecules. If a mixture of two gases, one much lighter than the other, were allowed to flow through a porous plate, the lighter gas would pass so much more quickly than the heavier gas that a partial separation of the two might probably be effected. Graham accomplished such a separation of oxygen and hydrogen, and of oxygen and nitrogen; and he described a simple instrument whereby this process of _atmolysis_, as he called it, might be effected.