of the ash of cordite, and obtained a residue among which, under the microscope, crystalline particles could be distinguished. Some of these particles, from their crystalline appearance and double refraction, were silicon carbide ; others were probably diamonds. The whole residue was dried and fused at a good red heat in an excess of potassium bifruoride, to which was added, during fusion, 5 per cent of nitre. (Previous experiments had shown me that this mixture readily attacked and dissolved silicon carbide ; unfortunately it also attacks diamond to a slight degree.) All the operations of washing and acid treatment were performed in a large platinum crucible by decantation (except the preliminary attack with nitric acid and potassium chlorate, when a hard glass vessel was used) ; the final result was washed into a
silica was added to cordite, the residue of the closed vessel explosion contained a much larger quantity of these spheres.