Papers. 31 



a spherical wave-front, the two similar vessels, being syinetrically placed 

 with respect to the source, should be equally effected by the y rays, 

 though the results of the equal effects may not be the same. If, on the 

 other hand, the y rays are any type of corpuscular radiation (in the New- 

 tonian sense) made of a finite number of pai-ticles. the effect in the ioniza- 

 tion -vessels would be unequal over short periods of time. To compare 

 the number of ions produced in the two vessels, the electrodes were con- 

 nected to an electrometer, one vessel being positively the other negatively 

 charged. The positive and negative currents from the two cans were 

 balanced as closely as possible for long periods of time, and so there was 

 no large steady drift, of the electrometer. The quartz fibre electrometer 

 (Proc. Camb. Phil. Soc. vol. 15, p. 106, 1909) showed fluctuations in this 

 balance. In the second apparatus a box-shaped ionization-can with a 

 central plane electrode was used. The positive ions formed in one half 

 of the can were received on one side of the flat electrode, the negative ions 

 from the other half of the can on the other side of the electrode. By 

 making the can airtight and thoroughly drying the contained air, complete 

 "saturation" was produced with a field of only 8 volts per cm. Large 

 fluctuations were observed when the ionization currents from the two 

 halves were balanced, the source of y rays being placed outside the can 

 in the plane of the central electrode. This experimental result would be 

 explained if (1) the y rays from radium are projected particles, or (2) if 

 the number of ions produced in air by a constant source of rays is subject to 

 fluctuations. 



We are. continuing the experiments with a view to determining what 

 part each of these factors plays in producing the fluctuations observed. 



The radium used in these experiment* was lent by the Royal Society of 

 London. 



