34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. "J i 



DISCUSSION OF POSSIBLE EXPLANATIONS 



1. The fact that the velocities are higher in vacuo than in air seems 

 expHcable only by there being conditions of ignition different in vacuo 

 from those in air ; although this may also have been due to the air in 

 the nozzle interfering with the stream-lines of the gas, thus produc- 

 ing a jet not strictly unidirectional. It should be remarked that the 

 highest velocity in vacuo recorded, experiment 25, may have been 

 due to unusually good circumstances of ignition ; but may also have 

 been due, in part, to being performed in the circular tank. 



2. The fact that the medium nozzle gives in general velocities 

 higher than the long nozzle shows that very likely after traveling 

 the distance from the throat equal approximately to the length of 

 the medium nozzle, the gas is moving so rapidly that it fails to 

 expand fast enough to fill the cross-section of the nozzle. A dis- 

 continuity in flow is produced at the place where the gas leaves the 

 wall of the nozzle, and this produces eddying and a consequent loss 

 of imdirectional velocity. The efficiency could doubtless be increased 

 by constructing the nozzle in the form of a straight portion, corre- 

 sponding to a cone of 8° taper, for the length of the medium nozzle, 

 with the section beyond this point in the form of a curve concave to 

 the axis of the nozzle. 



CONCLUSIONS FROM EXPERIMENTS 



1. The experiments in air and in vacuo prove what was suggested 

 by the photographs of the flash in air, namely, that the phenomenon 

 is really a jet of gas having an extremely high velocity, and is not 

 merely an effect of reaction against the air. 



2. The velocity attainable depends to a certain . extent upon the 

 manner of loading, upon the circumstances of ignition, and upon the 

 form of the nozzle. Hence, in practice, care should be taken to 

 design the cartridge and the nozzle for the density of air at which 

 they are to be used, and to test them in an atmosphere of this 

 particular density. 



It is with pleasure that the writer acknowledges the use, as 

 honorary fellow in physics, of the laboratory facilities, and especially 

 the rotary pump, at the Physics Laboratory at Clark University 

 where these experiments were performed. 



SIGNIFICANCE OF THE ABOVE EXPERIMENTS AS REGARDS 

 CONSTRUCTING A PRACTICAL APPARATUS 

 It will be well to dwell at some length upon the significance of the 

 above experiments. In the first place, the lifting power of both 



