their influence should disappear in the total impression received by the plate, even after a very short exposure. I contrived, besides, to eliminate even still more completely this cause of perturbation, by repeatedly alternating the experiments, as I will proceed to show.
Fig. 4 represents a horizontal section of the apparatus employed. AB is the photographic plate, 13 cms. wide; E is the spark enclosed in a cardboard box, FGHI, open only on the side facing the plate, and allowing the spark to act on one half, OB, of the plate only; CD is a lead screen wrapped in wet paper, rigidly connected with the frame which holds the plate. The "N" rays, proceeding from any source, form a pencil, having the direction NN'. With this arrangement the "N" rays are arrested by the screen CD; the spark, while it acts on half-plate OB, is sheltered from the rays.
Now impart to the frame containing the plate a translation to the right equal to half its length (Fig. 5); the other half, AO, of the plate takes the place formerly occupied by
