THE atomic weight of the element osmium has been redetermined by Prof Seubert. The necessity for this redetermination has been felt ever since the principle of periodicity began to take firm root in the minds of chemists; and the more recent values arrived at for the atomic.weights of iridium, platinum, and gold have tended to render this necessity even more imperative. The natural sequence, according to their chemical and physical properties, of the metals of the platinum group is generally accepted as—osmium, iridium, platinum, gold. Now, the atomic weight of iridium as determined in 1878 by Seubert is 192'5, that of platinum as fixed hy the same chemist in 1881 is 194'3, and that of gold as estimated last year by Thorpe and Laurie, and by Kruss, is 196'7, while the recognized atomic weight of osmium as given by Berzelius in 1828 is so high as 198'6. Obviously, if the grand conception of Newlands, Mendelejeff, and Lothar Meyer is correct, the atomic value of osmium required most careful revision. Such an undertaking, however, is endowed with peculiar interest owing to the dangerous nature of work with the osmium compounds, and many chemists who have been interested in this subject have been deterred by the knowledge that accidental contact with the fumes of the tetroxide, which are so frequently evolved by the spontaneous decomposition of many osmium compounds, might deprive them of the use of their eyes forev\lr. Prof. Seubert has happily succeeded without accident in establishing the validity of our” natural classification” by means of the analysis of the pure double chlorides of osmium with ammonium and potassium, (NH^OsCl. and KjOsCl,. Both these salts were obtained in well-formed octahedral crystals, of deep red color while immersed in their solutions, but appearing deep black with a bluish reflection when dry, and yielding bright red powders on pulverization. The method of analysis consisted in reducing the double chlorides in a current of hydrogen; in case of the ammonium salt, the spongy osmium which remained after reduction was weighed, and the expelledammonium chloride and hydrochloric acid caught in absorption apparatus, and the total chlorine estimated by precipitation with silver nitrate. In case of the potassium salt, the expelled hydrochloric acid was absorbed and determined, and the metallic osmium left after removal of the potassium chloride by washing was weighed. The mean value yielded by all these various estimations is 191'1, thus placing osmium in its proper place before iridium, and removing the last striking exception to the “law of periodicity.”