itself anew, a second contraction is followed by a new disk in the same plane as the first tongue, and so on; the disks alternating in the two vertical planes of breadth and depth, as in a paper fidibus. In the photographs reproduced in Figs. 16 and 17, a mirror fixed at an angle of 45° in the vertical plane gives, besides the front view of the vein, its image as seen in profile. A mechanical demonstration can be given of the movement of which the vein is the seat by a parallelogram of basket-work, which we press upon while holding it horizontally, and making it pass, by alternate compression of the extremities, from the oblong to the square shape, and then to the oblong the other way, return to the square, etc. The example of the fidibus is not quite exact. We comprehend at first that, in consequence of the acceleration, the disks will continue lengthening and deviating more and more—we can easily distinguish eight of them, sometimes twelve or more—till at last the molecules of water yielding to this dissociating action of weight group themselves in little cohesion-drops. There is another difference, in that the disks are not superposed as in the fidibus, but are boxed into one another—that is, one begins to form by deviation before the preceding one has done contracting. This is the necessary consequence of a third difference: that the disks are not plainly flat, but are thin in the middle, like the primitive tongue from which they are derived, and are flanked by thick cords on the edges (Fig. 8, xviii). We can, therefore, regard each disk as formed by the shock of the two cords of the disk above it. If this shock is exerted in a horizontal plane, the 
Fig. 12. new disk will spread out equally in every direction, having its center at the point of the shock. Acceleration becoming a factor, the spreading is prolonged farther down than up; the disk is thrown out of center, but does not for that cease to encroach upon its predecessors. The case is like that of the links of a chain, which enter within one another. Nowhere, then, not even at the point of minimum surface, can the jet be cylindrical; at the minimum, it assumes the form of a rounded cross (Fig. 8, xix).
Thus we find that, in the phenomenon so simple in appearance of a fall of water, all the complications that arise one after another under the fullest examination are explained logically down to the most minute details. We can already draw one important conclusion: every mass of falling water, if it is not rigorously cylin-
