Z'2' SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 7 1 



There is a large difference, however, with InfalHble powder, with 

 all three nozzles. P^or the medium nozzle a comparison c?f experi- 

 ments 4 to 12, inclusive, with 35 and 36 shows that the increase 

 amounts to 22 per cent of the velocity in air. 



2. The medium nozzle gives, in general, greater velocities than the 

 short or the long nozzle with the same length of chamber, 1, and 

 approximately the same charges of powder. In all cases, the short 

 nozzle gives less velocity than the medium or the long nozzle, which 

 is to be expected. 



3. The results show no appreciable dependence of the velocities 

 upon the pressure in the tank between 7.5 mm. and 0.5 mm., and it is 

 safe to conclude that the velocities are practically the same from 

 atmospheric pressure down to zero pressure, except as regards 

 the slight increase of velocity with decreasing pressure already 

 mentioned. 



4. A comparison of the results when the chamber moved under 

 the influence of the spring with those in which the chamber was 

 merely lifted, show that the agreement of results obtained by the two 

 methods is good, provided the displacement in the direct lift experi- 

 ment is small (compare experiments 40 and 41 with 26). If, on the 

 other hand, the displacement in the direct lift experiment is large, 

 this method gives considerably less velocities than the spring method 

 A comparison of experiments 35 and 36 with 34, 37 and 43 makes it 

 evident that all the velocities obtained by experiments in zvhich the 

 lift exceeded 4 cm. are from 300 to 600 ft./ sec. too small. This is a 

 very important conclusio'n, for it means that the highest velocities in 

 vacuo, recorded in table III, are doubtless considerably less than 

 those which zvere actually attained. 



5. A comparison of the results obtained by means of the circular 

 tank with those obtained by means of the cylindrical tank shows that 

 the velocities range about 100 ft./sec. higher for the circular tank — 

 a difference that is so small as to be well within the accidental varia- 

 tions of the experiments. 



Concerning the behavior of the cylindrical tank under different 

 conditions, a comparison of experiments shows that the velocities 

 are much the same for all cases. Hence it is safe to conclude that 

 the rebound, at least for small charges, is not excessive even if an 

 empty tank is used, providing it is sufficiently large. 



A check of some interest, on the effectiveness of the cylindrical 

 tank, with the retarder, J, J', in position inside, was the sound of 

 the shot, which resembled a sharp blow of a hammer on the lower 



