662 Mr. J. A. Crowther on the Secondary 



But, by our previous experiment, 



that is 



P*GoV*+C*=p % GY a +.C t 



or &l&o=P%Jp%- 



The required ratio of the coefficients of secondary radiation, 

 of the two gases, is thus given directly, as the ratio of two 

 pressures, which can be read off on the scale of the pressure- 

 gauge. 



It was considered possible that the constants V and C 

 might depend somewhat upon the hardness of the primary 

 rays. This, however, was not found to be the case. Experi- 

 ments were made with the bulb giving very soft rays with 

 an equivalent spark-gap of about 1 cm., and also with the 

 bulb at various hardnesses up to an equivalent spark-gap of 

 5 cms., which was as hard as the coil could be made to drive 

 the bulb steadily ; but when the chambers were filled with 

 gases such as air, carbon dioxide, hydrogen, or sulphur 

 dioxide, no alteration in the balance could be observed whether 

 the bulb was hard or soft. We must conclude therefore that, 

 for the purposes of this experiment, these constants may be 

 regarded as independent of the hardness of the bulb. 



One further possible source of error remains to be con- 

 sidered, namely, the absorption of the rays, both primary and 

 secondary, by the gas in the gas-chambers. This effect * 

 was known to be quite small, except in the case of a few 

 vapours. It is also proportional to the pressare of the gas. 

 Hence the corrected term might be written in the form 



where X is some constant depending upon the size and shape 

 of the gas-chamber. By making comparisons at different 

 pressures, it would be possible to calculate X, or if the pressure 

 were sufficiently low, the correcting factor depending upon 

 the square of the pressure would become negligibly small. 

 The method actually employed was to plot the values of the 

 ratio Gr/Gr , obtained at various pressures, and by producing 

 the curve thus obtained (it was always practically a straight 

 line) to zero, to find the value this ratio would have at an 

 infinitely small pressure, when, as appears from the above 

 expression, the correction should disappear. 



Asa matter of fact the correction involved was never very 

 large. With such relatively light gases as air, carbon dioxide, 



* .Rutherford, PHI. Mag. [6] xliii. p. 241 (1897). 



