NO. 2 METHOD OF REACHING EXTREME ALTITUDES 33 



cap of the 12-mch pipe. The mipact was most clearly discernible 

 when the hand was on the lowest part of the tank. The sound, in the 

 case of the circular tank, did not appear to come from any particular 

 part. When the tank was grasped during firing, a throb of the entire 

 tank was noticed. 



6. Concerning the proportion of the measured reaction that is 

 due to gaseous rebound, the tissue paper detector, as has already 

 been explained, does not give any information. All that this detector 

 really shows is that the force exerted by the initial upward rush of 

 gas past the chamber is not excessive. The fact that the tissue paper 

 is sometimes torn and sometimes not under identical conditions of 

 firing, shows that either this force differs more or less in various 

 parts of the tank {i. e., the upward rush of gas is not perfectly 

 homogeneous) ; or that the tissue paper is weakened by each succes- 

 sive shot. This last explanation is the more probable ; for fine 

 particles of the wadding rush upward with the gas, as is proved by 

 fine markings on the smoked glass, and also from the fact that, after 

 a. number of shots, the tissue paper is found to be perforated with very 

 small holes. 



The gaseous rebound could not be measured accurately with the 

 direct-lift impulse-meter. Thus of all the experiments in which this 

 meter could be used, 15 to 26 inclusive, only two, 16 and 22, gave 

 readable displacements, the failure to obtain readable displacements 

 in the other cases being doubtless due to friction, as already men- 

 tioned. It will be noticed that the impulse is under one per cent. 



The spring impulse-meter used in the last five experiments gave 

 reliable results because of the very slight friction during operation. 

 This impulse-meter shows that, if the momentum of the chamber 

 were to be corrected for gaseous rebound, this correction would he 

 much less than one per cent of the momentum of the chamber. But 

 as has been stated above, the impulse of the rebound at the chamber 

 must be less than that at the impulse-meter, from the fact that gases 

 may pass readily behind the chamber, and exert a downward pres- 

 sure, and also because of friction in the 3-inch pipe. The effect of 

 gaseous rebound is therefore negligible, and no account of it has 

 been taken in calculating the velocities and efficiencies. 



It now becomes possible to find, from the experimental results, 

 the highest velocity in vacuo upon which dependence may be placed. 

 This is evidently the result of experiment 45 and is 2.34 km./sec. 

 or 7,680 ft. /sec. It is Avell worth noticing, however, that experiment 

 50 would have given, without doubt, a velocity even higher, had fric- 

 tion properly been taken into account. 



