A shutter combining these qualities will not yet be perfect. It is necessary, according to the time and the light, that the time of exposure shall be capable of being varied. In a word, it is necessary that the apparatus shall be _graduated_ and permit of taking views more or less quickly. The different velocities might be given to the shutter by means of weights, rubber, or springs. The latter seem to be preferable, since they permit in the first place of operating out of the vertical; moreover, they are less fragile, and, through different tensions, they permit of these graduations that we consider as indispensable. For the current needs of practice 1/100 of a second is a limit that seems to us sufficient as a maximum of rapidity. In order to know the time of exposure obtained we employ the following method, which permits of graduating an apparatus rapidly and with extreme precision: A band of smoked paper is fixed upon the shutter, then a tuning-fork provided with a small stylet resting against the paper is made to vibrate. Better yet, a chronograph which vibrates synchronously with a tuning-fork, whose motion is kept up by electricity, is put in the same place. Fig. 3 shows the arrangement to be employed. We then let the shutter fall, when the little stylet will inscribe a certain number of vibrations. Knowing the number of vibrations of the tuning-fork, and counting the number of those inscribed upon the paper, it is very simple to deduce therefrom the amount of the time of exposure. The results of one of these experiments we have reproduced in Fig. 4. The tuning-fork gave 100 double vibrations per second. Six vibrations are included between the opening and closing of the apparatus. Each vibration estimated at 1/100 of a second. The exposure was 6/100 of a second in round numbers. This is the amount of the total time of exposure. As for that of the effective time, that is just as easily ascertained. It suffices to know the number of vibrations comprised between the moment at which one point of the objective has been completely uncovered and that at which it has begun to be covered again. The time is equal to 2/100 in round numbers. In the experiment in question, with an aperture equal to twice the diameter of the diaphragm, we have, then, 1/3 of the half-open exposure; and the amount of the effective time is 1/3. The difference that we have in practice is due to the fact that the velocity is uniformly accelerated. In order to increase the amount of the effective time, it will be only necessary to increase the aperture of the shutter and apply again the method that we have just pointed out. [Illustration: FIG. 3.]