the Diffusion of Gases. 249 



100 oxygen should be replaced by 105.41 air ; and 



100 air should replace 94.87 oxygen. 



When confined in a straight diffusion tube, there is uniformly 

 an expansion ; but it is unnecessary to recount experiments per- 

 formed with the straight tube, as the divisions are not minute. 



Experiment 1 Thermometer 64. Barometer 29.82 inches. 

 Diffusion-instrument with bulb, divided into two-hundredths of 

 a cubic inch ; also standard bulb and tube, close at top, to afford 

 corrections for changes in temperature and pressure, as before ex- 

 plained. Both diffusion-instrument and standard were filled with 

 pure oxygen from chlorate of potash, and placed in glasses over 

 water, covered by a bell-jar, of which the inside was moistened. 

 A few minutes were purposely allowed to elapse before the quan- 

 tity of gas in either instrument was noted, as the quantity oscil- 

 lated for a little. The diffusion-instrument contained 795 mea- 

 sures oxygen, and the standard 828, at the outset. In two hours 

 the expansion in diffusion-instrument, corrected from the stand- 

 ard, was 6 measures ; in four hours and a half, 1 3 measures ; in 

 fifteen hours, 29 measures ; in twenty hours, 34 measures ; in 

 twenty-nine hours, 41 measures ; in thirty-eight hours, the ex- 

 pansion was at a maximum, namely, 43 measures. In explana- 

 tion of the long duration of this and the following experiments, 

 it may be stated, that the plug was fully half an inch in thick- 

 ness. 



795 measures oxygen and vapour have therefore been replaced 

 by 838 measures air and vapour. 



0.9487 diffusion volume of oxygen by experiment. 



This is the exact theoretic number, a coincidence, however, 

 which we must view as accidental. 



Experiment 2. In a careful repetition of this experiment with 

 another specimen of oxygen gas, the results approached very 



