177 



to the leak produced by the natural ionization in the cham- 

 ber, to the radiation which comes directly from the Ra and 

 has not all been cut off by the interposed screens, such as 

 L,L, and to secondary y radiation from surrounding bodies, 

 which may be able to enter the chamber. This latter quan- 

 tity was found to be by no means negligible, and so far as 

 I know has not been allowed for in many experiments by 

 previous observers. To obtain with great accuracy the zero 

 reading necessary for subtraction, is a course exceedingly 

 difficult. 



In the present experiment a close approximation to its 

 value was found by inserting a large Pb plug in the coni- 

 cal hole and placing plates of Pb above this, so as to make 

 in all a thickness of about 20 cm. This, as shown by pre- 

 vious observers, should be sufficient to reduce even the hard 

 ■y radiation, which streamed through the conical hole, to 

 almost a negligible amount, while it should not affect to any 

 great extent the secondary radiation coming from surround- 

 ing bodies. 



The results of experiments with different materials used 

 as radiators are given in table ii., and in fig. viii. are shown 

 by full-line curves the thickness of radiator multiplied by 

 its density, plotted horizontally against the logarithm of the 

 corresponding currents vertically. The value of the current 

 with no radiation in position at p^y has in each case been 

 reduced to the common value 100, and the zero for the ordi- 

 nates has in some of the curves been altered so as to pre- 

 vent overlapping. It will be seen that Pb, Bi, and Hg give 

 results almost identical : while Sn, Zn, Al, and C all giwQ 

 results very similar to each other, but differing from those 

 of the first group in that a more sudden drop of the curve 

 is shown for the initial thicknesses of radiator. From a 

 value of r/A =7 onwards the curves for all the substances 

 become practically parallel. 



The Pb screens L,L, fig. vii., were now removed, and 

 radiators were used in the form of large, flat plates, the 

 first of which was placed horizontally just below the floor 

 of the ionization chamber, merely separated from it by a 

 small thickness of ebonite for insulation. 



Successive thicknesses of radiator were added immediately 

 below the first. By this means it was thought possible to 

 allow a large proportion of all the emergent secondary y 

 radiation to add on its effect in the chamber. The results 



