One more remarkable fact observed by M. Moser may be quoted. He finds that the action of light upon the Daguerreotype plate is of an _alternating_ kind: it first gives an affinity for Mercury, and then removes it. "If light acts on Iodide of Silver," he says, "it imparts to it the power of condensing mercurial vapours; but if it acts beyond a certain time, it then diminishes this power and at length takes it away altogether." This is precisely in accordance with phenomena observed also in the Collodion process, where the deposit of metallic Silver is sometime less marked than usual if the plate has been exposed in the Camera beyond the proper period of time. A curious perversion of the developing process is occasionally met with, in which on the application of the Pyrogallic Acid, the deposit of Silver takes place upon the _shadows_ of the picture, and not upon the lights; hence on viewing the image by transmitted light, the usual appearance is reversed. This may perhaps be explained by an alternating action of the light as above suggested. A phenomenon at first sight even more remarkable has occurred, in which, on developing the plate, _two_ images start out instead of one. The secondary image in such a case is probably the remains of a previous impression which, although apparently removed by washing, had nevertheless modified the surface of the glass so as to affect the layer of Iodide of Silver; and if the glass were _breathed_ upon before again coating it with Collodion, there is every reason to suppose that the outlines of the accidental image would be seen.[10] [10] Since writing the above, the Author has perused with pleasure a paper by Mr. Grove on the production of latent images by electricity, with a mode of fixing them. In the experiments described, a plate of glass, electrized in certain portions only, was breathed upon, or exposed to the fumes of Hydrofluoric Acid. In either case the vapour settled exclusively upon the non-electrical part of the glass, thus developing a latent image. When the plate was first submitted to electrization, and then coated with Iodide of Silver upon Collodion, and exposed to light,--solution of Pyrogallic Acid produced a reduction of Silver only upon the parts of the glass corresponding to those on which the breath settled in the previous experiment; thus indicating that the electricity neutralized the effect of light upon the sensitive Iodide of Silver. CHAPTER IV. ON FIXING THE PHOTOGRAPHIC IMAGE. A sensitive layer of Chloride or Iodide of Silver on which an image has been formed, either with or without the aid of a developing agent, must pass through further treatment in order to render it indestructible by diffused light.