The results of thirteen control experiments, designed to show the number of glomeruli in the rabbit's kidney open to the circulation under the chosen experimental conditions without intentional interference, indicate the "normal" range to be from 42 to 100 per cent. Since ten of the thirteen results fall within the figures 56 and 89 per cent, we may take these figures as the chief basis for our discussion. Three experiments only were made in which renal vasodilatation was produced by caffeine and salt. The percentage of open glomeruli found was in every case higher than any control except one. The results show without ambiguity that in rabbits, as in frogs, renal vasodilatation by caffeine is accompanied by increase in number of patent glomeruli. Our prime interest lay in the general question rather than in the action of individual substances in the group of vasodilators; hence this series was not extended further. Among the experiments designed to test the effects of renal vasoconstriction are to be found nine in which adrenalin was injected, two in which CO2 was inhaled, two in which the splanchnic nerve was stimulated, and four in which hemorrhage was induced. Not all are of equal value for our present purpose, inasmuch as the degree of certainty with which we can assume that renal vasoconstriction was actually produced is not the same in all. Enough experience with the action of adrenalin on the kidney of the anesthetized rabbit is available to permit the assertion that the dosages used in the nine adrenalin experiments were sufficient to insure constriction of renal vessels. Similar certainty exists in the experiment in which high concentration of CO2 was used; less in the case of 10 per cent CO2. Stimulation of the splanchnic nerve in rabbits so frequently fails to produce results typical of direct constriction of renal vessels that we may regard the production of this effect in the two experiments in which this was done as doubtful. In the perfusion experiments by Richards and Plant (7) the reactions of the renal vessels to stimulation of the splanchnic nerve resembled those to intravenously injected adrenalin rather than to the direct excitation of constrictor fibers. In six rabbits in which Livingston subjected the nerve to varying degrees of electrical stimulation, in one only was distinct constriction of the renal vessels produced, and in this a latent period of 45 seconds occurred between the beginning of stimulation and the production of effect. In one of our experiments a rabbit was used in which the superior cervical sympathetic ganglion on the left side had been extirpated months before. During the stimulation of the splanchnic nerve the left pupil was observed to dilate, showing increased adrenalin secretion. Hence, we are inclined to regard the two attempts to produce vasoconstriction by this means as having been largely unsuccessful. In one experiment only of the four in which hemorrhage was induced was there unmistakable evidence of compensatory vasoconstriction which may have involved the renal vessels. The blood pressure curve of this animal showed rhythmically occurring waves of the Traube-Hering type and dyspnea was noted. In this the estimate of open glomeruli was 28 per cent. In the other three experiments no such change occurred and no dyspnea was seen. In pithed frogs, hemorrhage alone of moderate extent is less apt to lessen the number of patent glomeruli than are any of the other constrictor agencies tried by Richards and Schmidt. In this discussion, therefore, we lay little weight on the results obtained in the two experiments in which the splanchnic nerve was stimulated and on those of the first three hemorrhage experiments. The chief conclusion to be drawn from these experiments is that which was anticipated from the observations of Richards and Schmidt on frogs, and of Khanolkar on rabbits; viz., that in the rabbit, renal vasodilatation and renal vasoconstriction are usually associated with increase and decrease respectively in number of glomeruli through which blood flows (see Text-fig. 1). Analogous changes apparently occur in the capillary pathway in individual glomeruli. Hence renal function in mammals may be altered by changes in the extent of glomerular filtration surface to which the blood has access. Other conditions remaining the same, it is obvious that changes in extent of filtration surface must result in proportionate changes in urinary output. The figures for rate of urine elimination at the time of injection of the dye in these experiments are in substantial agreement with these statements (see Text-fig. 2). Exceptions to our chief conclusion as stated have been encountered. In Experiment 57,86 per cent of the glomeruli were open in a constricted kidney which was excreting no urine: in Experiment 30, 16 per cent were open in the kidney which was eliminating seven drops per minute: the outputs of the kidneys in the caffeine experiments were far higher than those of control kidneys in which comparable numbers of glomeruli were open. In considering these exceptions, account must be taken of the fact that other conditions do not commonly remain constant. When a renal vasodilator is introduced we conceive not only of possible increase in extent of accessible glomerular surface, but also of increase in glomerular pressure and increased rate of renewal of fluid in contact with glomerular membranes. Hence the response is greater than can be accounted for by any one factor alone. A basis of experiment exists in support of the belief that usually sufficient differences in physiological state exist among the small arteries and arterioles of the kidney so that a constrictor influence, exerted equally upon all, elicits various degrees of response (1). Closure of some, continuing patency of others, results. Blood flow and blood pressure in the glomeruli which are supplied by the vessels which remain open may be decreased, increased, or unchanged...