Now commences the tedious part of the process. The metallic ingot is attacked with hot nitro-hydrochloric acid until no more iron is dissolved. The bulky residue consists chiefly of graphite, together with translucent chestnut-coloured flakes of carbon, black opaque carbon of a density of from 3·0 to 3·5 and hard as diamonds--black diamonds or carbonado, in fact--and a small portion of transparent, colourless diamonds showing crystalline structure. Besides these there may be carbide of silicon and corundum, arising from impurities in the materials employed. The residue is first heated for some hours with strong sulphuric acid at the boiling-point, with the cautious addition of powdered nitre. It is then well washed and for two days allowed to soak in strong hydrofluoric acid in cold, then in boiling acid. After this treatment the soft graphite disappears, and most, if not all, the silicon compounds have been destroyed. Hot sulphuric acid is again applied to destroy the fluorides, and the residue, well washed, is attacked with a mixture of the strongest nitric acid and powdered potassium chlorate, kept warm--but not above 60° C., to avoid explosions. This treatment must be repeated six or eight times, when all the hard graphite will gradually be dissolved and little else left but graphitic oxide, diamond, and the harder carbonado and boart. The residue is fused for an hour in fluorhydrate or fluoride of potassium, then boiled out in water and again heated in sulphuric acid. The well-washed grains which resist this energetic treatment are dried, carefully deposited on a slide, and examined under the microscope. Along with numerous pieces of black diamond are seen transparent, colourless pieces, some amorphous, others with a crystalline appearance. Fig. 21 B shows one of these crystalline fragments. Although many fragments of crystals occur, it is remarkable I have never seen a complete crystal. All appear shattered, as if on being liberated from the intense pressure under which they were formed they burst asunder. I have singular evidence of this phenomenon. A fine piece of artificial diamond, carefully mounted by me on a microscopic slide, exploded during the night and covered the slide with fragments. Moissan’s crystals of artificial diamond sometimes broke a few weeks after their preparation, and some of the diamonds which cracked weeks or even months after their preparation showed fissures covered with minute cubes. I have explained that this bursting paroxysm is not unknown at the Kimberley mines. So far, all such artificial diamonds are microscopic. The largest artificial diamond is less than one millimetre across. [Illustration: FIG. 21. ARTIFICIAL DIAMOND MADE BY THE AUTHOR FROM MOLTEN IRON.] [Illustration: FIG. 22. MOISSAN’S ARTIFICIAL DIAMONDS. To face p. 120.]