Minerals, such as amphibole, pyroxene, epidote, olivine and garnet (containing iron), may often be separated by an electro-magnet by regulating its magnetic intensity.[146] _Separation by Chemical Means._ Very many different methods may be used, depending on the nature of the work to be accomplished; but they are generally only reliable in the hands of a good chemist. The material should be in the state of fine powder. As an example may be mentioned the treatment with pure concentrated HFl, by which the minerals of a rock are attacked in a certain sequence, the feldspars and related minerals first, then the quartz and finally the ferro-magnesium silicates, such as amphibole, pyroxene, olivine, etc. MICRO-CHEMICAL REACTIONS. The first requisite is to bring the substance to be investigated into solution. This can be done in the case of non-silicates by the ordinary solvents, while silicates can be decomposed and investigated either by the methods of Borichy or Behrens. Both methods rest upon the recognition of the forms, etc., of artificially produced crystals. The size of a fragment for testing may vary, according to circumstances, from that of a poppy seed to a pin head. Good results are recorded from fragments of not more than 0.2–0.7 sq. mm. The substance to be tested is placed on glass, protected by a film of balsam, and covered with a spherical drop of the solvent,[147] which should be allowed to act until all the different elements composing the sample are in solution (in the case of a very small fragment until it has all dissolved). Transfer the solution to another protected object glass, and, after evaporation, the crystallizations characteristic of the different elements will be seen. If the evaporation is too rapid and the crystallizations incomplete, the residue should be redissolved in water, or a very dilute solution of the solvent employed, transferred to a fresh glass and allowed to recrystallize. Borichy’s Method.[148] (Hydrofluosilicic Acid.) This method has the advantage of simplicity of manipulation and relative distinctness in results; but on the other hand these results are only obtained after several hours, and the temperature has an influence on the crystalline forms obtained. It is well to make the tests in a temperature of about 15° C. Put a spherical drop of pure hydrofluosilicic acid[149] on the fragment and leave it for some hours in damp air until the action has been sufficient, then transfer it to a dry air bell-glass and allow evaporation and crystallization to take place. For the microscopic examination the objective (200–300 diams. best for these observations) can be protected with glycerine and a mica disc or thin cover-glass, or the drop can be all evaporated and the crystals covered with liquid balsam and a cover-glass. Crystallizations Obtained by Borichy’s Method.