[Illustration: A Display ready for Firing, Dresden, September 1st, 1899. Firework Portraits of the King and Queen of Saxony on the right, Bismarck second from the left.] He then deals with tourbillions and table wheels. The latter consisted of a circular table with a central pivot, round which is free to revolve a bar which forms the axle of a wheel, the hub of which runs on the edge of the table. When the wheel is turned, the hub running on the edge of the table moves it forward in a circular path round the pivot. This principle is applied to similar and more elaborate devices. The name tourbillions, as before mentioned, is by other writers differently applied. The section dealing with cut-out and transparent devices is of little interest. These devices were an attempt to give variety from the monotonous repetition of turning cases and gerbs. To-day the use of colour cases, lances, a much-enlarged range of fountain and similar compositions, including aluminium and other brilliant fires, has obviated the employment of effects which cannot be rightly considered as pyrotechnic. The moving and stationary pieces considered in this and preceding chapters give a good general idea of the firing methods in compound fireworks. As we have already noted, the difference of designs and effects at the present time is infinite, so that it would be impossible in a work of the present size to give anything approaching a complete survey of what has been accomplished. But it is hoped that enough has been said to give the reader some idea of the methods adopted and the lines upon which the modern pyrotechnist works. CHAPTER VII FIREWORK COMPOSITIONS It may have been remarked in the foregoing chapters that, although the ingredients composing the firework mixtures are given, generally the proportions are not. The reason for this is two-fold: primarily, as we have seen in the chapter on rockets, the proportion of the ingredients of a firework varies in accordance with its size. So that to give the proportions of the compositions of any one type of firework would often require as many formulæ as there are sizes. Secondly, and perhaps more importantly, the quality and purity of chemicals as supplied in bulk vary so enormously that a constant series of experiments has to be conducted in order to ascertain what modifications and adjustments are necessary in the formulæ to give the required standard of performance.