Thus the process of repeated segmentation continues, with the little particles flitting towards each other and the larger segments shifting to and fro, commonly for more than half an hour without cessation. From the beginning to the end of a period of segmentation the food is seen to have changed its position in the abdomen to only a slight extent; whether this change is a passing of the food along the loop, or a movement of the loop itself, it is impossible to tell from the shadows on the screen. The change of position, however, is much less conspicuous than the lively division and redivision which the mass suffers so many times from the busy, shifting constrictions. From this typical form of rhythmic segmentation there are several variations. Sometimes, and especially when the mass of food is thick, the constrictions do not make complete divisions and are so far apart that the intermediate portions are relatively large. Moreover, the constrictions do not take place in the middle of each portion, but near one end; thus each portion is constricted, not into halves, but into thirds. If a little pointer is placed at the middle of a segment, when the segments are completely divided into halves, in a few seconds the pointer will be in the middle of the clear space between two segments; but in a few seconds more the first phase will return and the pointer will again be indicating a segment,--two operations intervene between similar phases. When, however, the portions are constricted into thirds, the indicator shows it, since three operations intervene between similar phases. The manner of these changes is made clearer by reference to the diagram in Figure 3. That each portion is constricted into three pieces is proved also by watching the gradual reduction of the portion at the left end of line 1 through lines 2 and 3, and also in the gradual formation of a full-sized portion at the right end of lines 2, 3, and 4. When food undergoing this process is watched, it appears to be affected by a series of constrictions, each of which starts at one end of the mass and marches through to the other end, leaving its impress at short intervals along the length. The progression of the dotted lines from right to left in _a_, _b_, _c_, and _d_, etc., Fig. 3, gives a notion of these advancing constrictions. Another variation of the segmentation is shown in Figure 4. In this type there are evidently divisions and subdivisions, _i.e._, one more operation between the appearance and the reappearance of the same phase than is present in the simple division of the small segments in a long string of food (Fig. 2). This form of segmentation is fairly typical for the constrictions seen in food advancing through the intestine. Sometimes the divisions occur in the middle of a long string of food and leave the ends wholly unaffected.