obtaining as high a velocity of ejection as possible. Some of these chambers were made of soft tool-steel, and others of tough nickel steel. Fig. 3 shows a chamber taken apart. A charge of dense smokeless powder, P, Fig. 1, was fired electrically. The resulting gases, after blowing out the small pasteboard wadding, B, entered a tapered nozzle, expanded, and left the nozzle with high velocity. It will thus be seen that I obtained the work of expansion of the gases, much as is done in the De Laval steam turbine.
The efficiencies and velocities which I have obtained in this way are truly remarkable. My highest efficiency, or rather "duty", is over 63 per cent. This is the highest ever obtained from a heat engine; the best reciprocating steam engine giving 21 per cent, and the Diesel (internal combustion) engine, about 40 per cent. My average velocities of ejection are, in most cases, over 7,000 ft/sec, and the highest is slightly under 8,000 ft/sec. This is far in excess of any velocity hitherto obtained except by minute quantities of matter in electrical discharge tubes. The highest velocities and efficiencies were obtained with a large chamber.
After most of the above experiments had been performed, a question arose as to the possible effect of the air in the nozzle and immediately beyond. It seemed likely that some fraction of the above velocity (measured, of course, ballistically, by the recoil of the chamber) might have been due to reaction against the air. This would evidently mean that, as as high altitudes were reached, and the density of the air