Now suppose we have laid bare all this mechanism in a living animal and make experiments upon it. If we stimulate the chorda tympani there is a copious flow of thin watery saliva, but if we stimulate the sympathetic there is a less copious flow of thick viscid saliva. Why is this? We find on closer analysis that the chorda contains fibres which dilate the small arteries so that there is an increased flow of blood through the gland; but that, on the other hand, the sympathetic contains fibres which constrict the arteries, thus leading to a reduced flow of blood. This accounts for the fact that “chorda-saliva” is abundant and thin, while “sympathetic-saliva” is scarce and thick. It was thought at one time that the chorda contained fibres which stimulated the gland to produce watery saliva, while the sympathetic contained fibres which stimulated it to produce mucid saliva. This, however, is not the case. Both nerves contain the same kind of secretory fibres: their other fibres differ mainly in that they act differently on the arteries. It might be the case--indeed it was at one time thought that it was the case--that secretion of saliva was simply a matter of blood-flow: an abundant arterial circulation gave rise to abundant saliva, a sparse flow to a sparse saliva. Undoubtedly the secretion depends on blood supply, but not _solely_. If it did, then the whole process might be conceived to be a very simple mechanical one--filtration or diffusion of the saliva from the blood stream through the thin walls of the blood vessels, and the walls of the tubules into the cavity of the gland. If this were the case, then the liquid in the gland would be the same in composition and concentration as the liquid part of the blood--the plasma. But it is really different in composition and it is not so concentrated. Now osmotic pressure--on the action of which so much is based--cannot help us, for the liquid in the gland is less concentrated than that in the blood vessels, so that water ought to pass from gland to blood instead of from blood into gland. Again, if we tie the duct, so that the saliva cannot escape, secretion still goes on, though the hydrostatic pressure of saliva in the cavity of the gland may be considerably greater than that of the liquid in the blood vessels. Yet again, if we stop the blood flow by tying the artery, secretion of saliva may still go on for a time. Therefore the only physical agencies we can think of do not explain the secretion. The latter is actually the work of the individual cells, stimulated by the nerves. If the volume of the gland be measured just while it is being stimulated to secrete, it will be found that the organ becomes _smaller_, yet while it is being stimulated the blood-vessels are being dilated so that the volume of the whole structure ought to become greater. Obviously part of the substance of the gland is being emptied out through its duct as the secretion.