Saltpetre may be said to be the basis of pyrotechny. There is hardly a formula by any of the writers on pyrotechny up to at least the middle of the nineteenth century which does not contain it. Gunpowder is composed of saltpetre, sulphur, and charcoal, three chemicals which it will have been gathered from the previous pages play a prominent part in very many of the pyrotechnic compositions. In some compositions their proportion is apparently identical with that of gunpowder, yet they do not form gunpowder as they are not milled, and are consequently not so intimately mixed. Compositions containing these ingredients have frequently an admixture of mealed gunpowder, the function of which is to give additional fierceness when required, as is the case in some rocket mixings. These chemicals, as we have seen in the previous chapters, are the components of rockets, turning cases, tourbillions, saxons, Roman candle fuse, and many others. When variation was required in fireworks used to give a simple fountain effect the earliest addition was of metal in finely divided particles, as filings, borings, or the now almost obsolete iron sand. Steel filings were used in what was known as “brilliant fire,” a term which has fallen into disuse since the introduction of other metals whose effects eclipsed that of steel. It has also been used where extra effect is wanted, that is, more tail in rockets and tourbillions. It is, however, not much used in the former case to-day, as the presence of steel in a composition which is to be charged on a steel spindle introduces a decided element of risk into the operation. The introduction of steel and iron was the first use of metals in firework making, probably the next metal to be introduced was antimony, either black (sulphide) or regulus. Jones (1765) was already using what he calls crude antimony; this was probably the black sulphide. Before the introduction of genuine colour, and while the chemicals which had been adopted for pyrotechny were still very limited in number, attempts were made to obtain either a semblance of colour or some variety in stars and garnitures by the addition of such substances as powdered glass, brass, sawdust, beech raspings, which appear to have functioned as do the iron or steel in the compositions already discussed, except that there would be no coruscation even with the brass. These additions would merely show as red-hot particles in the jet of fire. Kentish gives two gerb compositions, one of which contains coke grains, and the other porcelain grains, which would apparently produce cognate results; the use, however, of both these ingredients is now almost if not quite obsolete.