3. On the 8th August and 2nd September of the same year, I examined, north of Spitzbergen, in 80 deg. N.L., and 13 deg. to 15 deg. E.L., the layer of snow that there covered the ice. The nature of this layer is shown by the accompanying woodcut, in which 1, is new-fallen snow; 2, a layer of hardened old snow, eight mm. in thickness; 3, a layer of snow conglomerated to a crystalline granular mass; and 4, common granular hardened snow. Layer 3 was full of small black grains, among which were found numerous metallic particles that were attracted by the magnet, and were found to contain iron, cobalt, and possibly nickel also. 4. On the melting of 500 gram. hail, which fell in Stockholm in the autumn of 1873, similar metallic particles containing cobalt (nickel) were obtained, which, in this case, might possibly have come from the neighbouring roofs, because the hail was collected in a yard surrounded by houses roofed with sheet-iron painted red. The black colour of the metallic particles enclosed in the hail, their position in the hail, and finally, the cobalt they contained, however, indicate in this case too, a quite different origin. 5. In a dust (kryokonite), collected on the inland ice of Greenland in the month of July, 1870, there were also found mixed with it grains of metallic iron, containing cobalt. The main mass consisted of a crystalline, double-refracting silicate, drenched through with an ill-smelling organic substance. The dust was found in large quantities at the bottom of innumerable small holes in the surface of the inland ice. This dust could scarcely be of volcanic origin, because by its crystalline structure it differs completely from the glass-dust that is commonly thrown out of volcanoes, and is often carried by the wind to very remote regions, as also from the dust which, on the 30th March, 1875, fell at many places in the middle of Scandinavia, and which was proved to have been thrown out by volcanoes on Iceland. For, while kryokonite consists of small angular double-refracting crystal-fragments without any mixture of particles of glass, the volcanic Haga-dust[191] consists almost wholly of small microscopic glass bubbles that have no action on the polarisation-planes of the light that passes through them. Similar investigations have since been made, among others, by M. TISSANDIER in Paris, and during NARES' English Polar Expedition.