Secondary Rdntc/en Radiation from Carbon. 771 



V. Method of Experimenting. 



The radiator was set at an angle of 30° with the beam. 

 S was then placed with its axis making 150° with the primary 

 beam, and the ratio of the secondary to the primary deflexion 

 obtained. S was then set with its axis at right angles to the 

 primary and on the same side of the radiator, and a similar 

 ratio found. Dividing the first ratio by the second, we get 



the value of T — . Several corrections are necessarv. and of 



these two will vary as the hardness of the primary beam 

 varies, namely, the correction due to 



(1) The polarization of the primary beam. 



(2) The secondary radiation from the air and tertiary 

 radiation from surrounding objects. 



When a bulb is very soft we have a maximum amount of 

 polarization, but with the cathode stream inclined at 45° to 

 the vertical it is very unlikely that the theoretical value of 



•— will be greater than 1*795 — an increase of only 2^ per cent. 



of its value with an unpolarized beam. When the beam is 

 very hard it exhibits no polarization. 



The correction due to the air and other stray radiations and 

 the normal leak of the electroscope were rendered negligible 

 by the use of a thick radiator which gave rapid deflexions. 

 Since the radiator is set so that the normal to it bisects the 

 two directions along which we successively measure the 

 intensity of the secondary radiation, then the radiation 

 from any small portion of carbon in the thick radiator will 

 traverse exactly the same thickness of carbon ; and so the 

 radiation of any particular absorbability will be absorbed 

 by exactly the same amount by the material of the radiator, 

 and the deflexions obtained will give the same ratio as if 

 none of the secondary radiation was absorbed in the carbon 

 or air. The total air and other stray radiations were never 

 more than 4 per cent, of the total deflexion obtained ; and as 

 a large portion of this was cut down by the carbon radiator 

 itself, and their ratio in the two directions approximately 

 the same as for the carbon radiator, they were therefore 

 neglected. 



This was also justified by the following experiment. A 

 very thin sheet of pure copper was mounted on an aluminium 

 frame and then set with its plane making an angle of 25° to 

 the primary beam. The copper was arranged so that the 

 rays fell on the central portion of it. None whatever fell on 



