These partial photographs are also useful in the analysis of rapid movements, because they allow of the number of attitudes represented being multiplied. At the same time, as a man's leg is rather large, its reproduction cannot be multiplied very often, owing to confusion by superposition. I have therefore sought to diminish the size of the images, so as to an admit of repetition at very short intervals. The method consists in attiring a walker in a black costume having narrow bands of bright metal applied down the length of the leg, thigh, and arm, following exactly the direction of the bones of the limbs. This plan permits the number of images formerly produced to be increased at least tenfold; thus, instead of ten photographs per second, one hundred may be taken. To do this it is not necessary to change the speed of rotation of the disk, but instead of piercing it with one aperture, ten holes are made equally disposed around the circumference.[3] [Illustration] The figure here shown is from one of the negatives projected on the screen from the lantern. The dotted lines have been filled in to form direct lines. The figure shows the successive phases of one step in running. Only the left leg is represented; the lines correspond to the thigh, leg, and foot; the dots to the joints at the ankle, knee, and hip. This diagram shows pretty clearly the alterations of flexion and extension of the leg on the thigh, the undulating trajectories of the foot, knee, and hip, and yet the number of images does not exceed sixty in a second. A revolving shutter pierced with more holes would give more perfectly the angular displacements of the leg on the thigh, and the positions of the three joints. The finer the dotted lines expressing the direction of the limbs, the more the images may be multiplied; but in the present case, sixty times in a second more than suffice to show the displacements of the limbs when running. In this photographic analysis the two factors of movements--time and space--cannot be both estimated perfectly; knowledge of the positions the body has occupied in space requires that one should possess complete and distinct images; in order to obtain such images, a sufficiently long space of time must elapse between the two successive photographs. If, on the contrary, it is desirable to estimate time more perfectly, the frequency of recurrence of the image must be greatly increased. To bring these two exigencies as closely together as possible, lines and points must be chosen for the partial photographs which best show the successive attitudes of the body.