If we question the physical possibility of Landois’ explanation, an alternative is still open to us. The late Prof. Graham has applied the principle of Diffusion to the respiration of animals, and has shown how by a diffusion-process the carbonic acid produced in the remote cavities would be moved along the smaller tubes, and emptied into wider tubes, from which it could be expelled by muscular action. The carbonic acid is not merely exchanged for oxygen, but for a larger volume of oxygen (O 95 : CO_{2} 81); and there is consequently a tendency to accumulation within the tubes, which is counteracted by the elasticity of the air vessels, as well as by special muscular contractions.[153] [153] Phil. Mag., 1833. Reprinted in “Researches,” p. 44. Graham expressly applies the law of diffusion of gases to explain the respiration of Insects. Sir John Lubbock quotes and comments upon the passage in his paper on the Distribution of the Tracheæ in Insects. (Linn. Trans. Vol. XXIII.) Whether diffusion or injection by muscular pressure is the chief means of effecting the interchange of gases between the outer air and the inner tissues of the Insect, is a question to be dealt with by physical enquiry. If we suppose two reservoirs of different gases at slightly different pressures to be connected by a capillary tube of moderate dimensions, such as one of the larger tracheæ of the Cockroach, transference by the molecular movements of diffusion would be small compared with that effected by the flow of the gas in mass. But if the single tube were replaced by a number of others, of the same total area, but of the fineness (say) of the pores in graphite, the flow of the gas would be stopped, and the transference would be effected by diffusion only. We may next consider tubes of intermediate fineness, say a tracheal tubule of the Cockroach at the point where the spiral thread ceases, and where the exchange of gases through the wall of the tubule becomes comparatively unobstructed. Such a tubule is about ·0001 in. diameter. If we may extend to such tubules the laws which hold good for the flow of gases in capillary tubes of much greater diameter, the quantity of air which might be transmitted in a given time by muscular pressure of known amount can be determined. Suppose the difference of pressure at the two ends of the tubule to be one-hundredth of an atmosphere, and further, that the tubule is a quarter of an inch long and ·0001 in. diameter. The tubule would then be cleared out every four seconds. Such a flow of air along innumerable tubules might well suffice for the respiratory needs of the Cockroach. Without laying too much stress upon this calculation, for which exact data are wanting, we may be satisfied that an appreciable quantity of air may be made by muscular pressure to flow along even the finer air passages of an Insect.[154]