During the progress of this work it soon became apparent that these temporary inhibitions described above do not cut short a hunger period, but simply delay its culmination. The contractions that appear as the inhibition ceases are the continuation of the period temporarily checked by the inhibition. They are not the beginning of a new period. When the tetanus stage of the hunger period is reached a stimulation of the gastric mucosa sufficiently strong to cause prompt cessation of the contractions seems actually to terminate the period, for when the contractions reappear they are not the incomplete tetanus or strong and rapid contractions of the culmination of the period, but the feeble and slow movements characteristic of the beginning of a period. By careful adjustment of the quantity and strength of the material introduced into the stomach during the first part of the hunger period and by renewing the inhibition on reappearance of the rhythm, it is possible to lengthen a 30- to 40-minute period into a 90- to 120-minute period. In other words, the motor mechanisms of the hunger contractions may be compared to the spring of a watch. When the spring is wound up it will run the watch for a certain number of hours, and it makes no difference whether or not these hours are consecutive.

It seems to us that this fact has an important bearing on the question of the primary stimulus to the hunger movements. It seems to point to a primary automatism, peripheral or central, or both, relatively independent of the condition of the blood as well as of the afferent nervous impulses. The fact speaks particularly strongly against the hypothesis that the primary stimulus is to be sought in the condition of the blood. For example, if the primary stimulus is in some condition of the blood, this condition must be present and to a gradually increasing degree from 12:30 to 1:00 p.m. to parallel a hunger period beginning at 12:30 p.m. and ending at 1:00 p.m. And this condition of the blood must be absent from 1:00 p.m. to 1145 p.m., as the stomach is relatively quiescent during that time. The hypothesis seems to be rendered untenable by the manipulations which do not, at least in some cases, involve any change in this hypothetical condition of the blood. The culmination of the hunger period may be delayed till 1130 or 1 .-45 p.m., so that the strongest hunger contractions fall in the time when the blood does not stimulate the gastric mechanism in a way to cause hunger movements.

But what is the significance of this inhibition in the normal work of the stomach ? The inhibition of the hunger contractions by mechanical and chemical stimulation of the gastric mucosa in the normal person prevents the appearance of these contractions during the period of gastric digestion. This negative control of the hunger movements from the stomach cavity is obviously a useful co-ordination. The primary or actual stimulus to the hunger contractions is therefore to be sought in the vagus tonus, in some condition of the blood, or in a primary automatism of the gastric neuromuscular mechanism. We have some evidence that the latter is the essential factor and that extrinsic nerves and the condition of the blood only modify the primary automatism. If this is the case the hunger contractions ought to appear as soon as the stomach is empty of food or other substances capable of stimulating the nerve-endings in the mucosa. We should also expect these contractions to be more or less continuous as long as the stomach is empty, at least in young and vigorous individuals, and when the condition of the individual as a whole does not lead to increased activity of the extrinsic inhibitory nerves (splanchnics). On this hypothesis the gradual tonus contraction of the gastric fundus pari passu with the progress of the gastric digestion represents the algebraic sum of the inherent automatism and the inhibitory effects from the gastric cavity. A gradual fatigue of the inhibitory mechanisms is probably also a factor, as we have abundant evidence (in man and dog) of such "escape" of the stomach from inhibitory nervous processes.

We should probably look for the closest parallelism between the gastric hunger contractions and the absence of stimulation of the gastric mucosa in infants and young children, that is, before cerebral (and possibly gastric) habits relative to feeding have been established. We have made a close study of a healthy (bottle-fed) infant touching this point. It is well known that, other things being equal, the more food put into the stomach the longer is the time required for the completion of gastric digestion. If this infant (five months old) is given only 4 ounces of food he calls for more after about 2 hours. If he is given 7 to 8 ounces of the same food the call for more food is delayed for 3 to 4 hours. If he is given 5 ounces of the food at 6:00 p.m. he nearly always wakes up and calls for more at 12:00 or 1:00 o'clock, while if he is given as much food as he will take (7} to 8j ounces) at 6:00 p.m. he rarely wakes up and calls for food until 3:00 or 5:00 o'clock the following morning. There is evidently a close parallel between the time of the emptying of the stomach and the appearance of the hunger contractions. The more frequent calls for food during the day are obviously due to the fact that the gastric hunger contractions must reach a certain degree of intensity before they cause the soundly sleeping infant to wake up. This is certainly true in the case of dogs. A dog may sleep on peacefully and quietly during gastric hunger contractions of moderate intensity. When these contractions become very intense the dog moves or moans in his sleep and sometimes wakes up.

While we have made no observations on the action of acids, alkalies, and alcoholic beverages on the gastric movements of digestion in man, it is well known that these substances do not inhibit these movements to the extent that they inhibit the hunger contractions. The movements of digestion are primarily concerned with the pyloric region, while the hunger contractions involve the cardiac and fundus region. Evidently these two regions of the stomach react differently to local chemical stimulation of the gastric mucosa.

In view of the fact that acids as well as normal gastric juice inhibit the gastric hunger contractions, one might expect that persons having gastric hypersecretion should experience little or no true hunger sensations or pangs of gastric origin. At the same time we must consider, in cases of prolonged hypersecretion, the possibility of a readjustment of such a character that the acid stimulation of the mucosa causes less inhibition than is the case in the normal stomach.