26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 7I 



ejected with the gas, which accumulates with each successive shot. 

 This soHd matter must offer considerable frictional resistance to 

 motion along the U-pipe, and, since the mass of gas is only of the 

 order of a gram, must necessarily act to stop the flow in a very short 

 time. This interval of time was great enough, however, so that this 

 second method afforded a satisfactory check upon the first method. 



A possible modification of the above two methods would have been 

 to provide some sort of trap-door arrangement whereby the gases, 

 after having been reduced in speed in a container as just described, 

 would have been prevented from returning upward into the 3-inch 

 pipe P by this trap, which would be sprung at the instant of firing. 

 In this way gaseous rebound would be entirely eliminated. It was 

 found, however, that results with the two methods already described 

 could be checked sufficiently to make this modification unnecessary. 



The tanks were exhausted by a rotary oil pump, No. i, of the 

 American Rotary Pump Company, supported by a water jet pump. 

 In this way the pressure in the cylindrical tank could be reduced to 

 1.5 mm. of mercury in 25 minutes and to the same pressure in the 

 circular tank, in 10 minutes. The pressures employed in the experi- 

 ments ranged from 7.5 mm. to 0.5 mm. 



METHODS OF DETECTING AND MEASURING GASEOUS REBOUND 



With the two tanks used in the experiments, it was obviously 

 impossible to eliminate gaseous rebound entirely, from the fact that, 

 even if the velocity of the gases is reduced to zero, there still remains 

 the effect of introducing suddenly a certain quantity of gas into the 

 tank. It became Tiecessary, then, to devise some means of detecting, 

 and, if possible, of measuring, the extent of the rebound. 



Three devices were employed, one for detecting a force of 

 rebound, and two for measuring the magnitude of the impulse per 

 unit area produced by the rebounding gas. These latter devices, 

 from the fact that quantitative measurements were possible with 

 them, will be called " impulse-meters." 



TISSUE PAPER DETECTOR 



The detector for indicating the force of the rebound consisted of 

 a strip of delicate tissue-paper, I, plate 6, figure 2, and text figure 

 5 (a), 0.02 mm. thick, with its ends glued to an iron wire, W, as 

 shown in figure 5 (a). This iron wire was fastened to the yoke Y, 

 plate 7, figute i, and held the tissue paper, with its plane horizontal, 

 between the chamber and the wall of the 3-inch pipe, P. In many of 



