The two preceding sections have been devoted to a study of the centre which controls the level of sugar in the blood. Its location in the brain and the changes in carbohydrate metabolism resulting from its stimulation have been examined. The present section will be devoted to consideration of the results of certain experiments designed to throw light on the mechanism which is involved in this nerve control. The problems to be considered are: (1) the nerve pathway through which the centre discharges its impulses, and (2) the particular mechanism upon which these impulses primarily act. Investigation of the first of these would be greatly simplified if a definite answer could be given to the second, but, as we have seen, this is not possible, since functions of varying types located in different organs become involved. Indeed, the widespread nature of the disturbances set up by stimulation of the centre suggests the possibility that they must be due to the action of some hormone, or hormones, whose secretion into the blood is regulated by the nerve stimulus. If this be so the problem narrows itself down to finding the nerve pathway passing from the centre to the glands secreting the hormone. Two glands, the adrenal and the isles of Langerhans, have to be considered, and there is much evidence that the hormone which each secretes, adrenaline and insulin respectively, is under nerve control. But the nerves concerned apparently belong, in the one instance to the sympathetic (adrenal) and in the other, to the parasympathetic system (isles of Langerhans), so that the impulses resulting from piqure might act through either pathway; through the spinal cord and splanchnic nerves to the adrenal gland so as to stimulate an increased secretion of adrenaline or through the vagus so as to inhibit the secretion of insulin. In its rapidity of onset the hyperglycsemia which follows piqure resembles much more closely that which follows the injection of adrenaline than that supervening upon pancreatectomy, so that most workers who have supported the hypothesis that the nerve control acts through hormones have paid attention only to the adrenal gland. Our attention in the present paper will first of all be given to the effects of pontine decerebration on adrenalectomised animals.