[Illustration: FIGURE 3.--Diagram showing the relations of the portions when they are constricted into three pieces. The dotted lines indicate regions of constriction; the arrows indicate the relationship of the pieces to the portions they subsequently form.] A remarkable feature in the segmentation of the food is the rapidity with which the changes take place. The simplest way of estimating the rate of division is to count, not the number of times the partition of the food recurs in the same place, but the number of different sets of segments observed in a given period. Thus in Figure 4 the appearances of lines 1, 2, 3, 4, etc., would be counted, and not merely lines 1, 4, etc. Repeated observations on different animals have shown that the most common rate of division in long, thin chains of food varies between twenty-eight and thirty times in a minute; _i. e._, there is a change from one set of segments to another set every two seconds, and a return of the same phase every four seconds. In some cases the rate is as low as twenty-three times per minute. The larger masses seem to be associated with a slower segmentation; the operations indicated in Figure 3, for example, occurred from eighteen to twenty-one times in a minute, so that the same phase reappeared only once in eight or nine seconds. The segmentation frequently continues for more than half an hour; in one instance it was seen to persist with only three short periods of inactivity for two hours and twenty-two minutes. At the rate of thirty segmentations per minute it is clear that a slender string of food may commonly undergo division into small particles more than a thousand times while scarcely changing its position in the intestine. I have seen once, in a cat only lightly etherised, the exterior of an intestine which was dividing the food as above described. An hour and a half after a meal of salmon the anæsthetic was given, the abdomen opened, and the flaps raised so as to form walls. Warm salt solution was then poured into the abdominal cavity, and the floating coils left covered with the transparent omentum. The gastric peristaltic waves were running regularly; on the intestine there were visible at various places during the period of observation regions of constriction which had the appearance shown in Figure 3, except that the rings were relatively nearer together. New rings of constriction took place on the same side of all the bulging parts at the margin of the constricted portion (_cf._ dotted lines, Fig. 3). As new rings occurred the old relaxed, but apparently with tardiness, for the contents gurgled as if forced through the narrowed lumen. The constrictions recurred irregularly and at much longer intervals than in the normal animal. The contracted rings were pale and bloodless.