The Bending of Electric Waves round the Earth. 757 



secondary radiation, apparently generated by lead and capable 

 of penetrating readily all substances, except lead, even after 

 reflexion from wood &c. 



(6) The initial variations in the absorption coefficient of 

 the radium 7-rays have been shown to depend much on the 

 nature of the absorber and on the disposition employed. 

 Zinc, tin, and aluminium, for a certain disposition, absorbed 

 the rays quite exponentially but with somewhat high values 

 of X. In another disposition zinc absorbed exponentially 

 with normal value of X from the thickness sufficient to 

 eliminate /3-rays up to 6 cm. 



(7) The value of the absorption coefficient, over the higher 

 ranges, of thickness for which absorption is always strictly 

 exponential, can be varied within fairly wide limits, being 

 for example always diminished when the rays first traverse 

 a denser substance ("hardening"). The abnormal ratio 

 ^UrxA-Ea for lead (1*465) previously obtained is so accounted 

 for. In general, the ratio A-urx/^Ea may be taken to be from 

 1-2 to 1-3. 



Physical Chemistry Laboratory, 

 Glasgow University. 



LXXVIII. On the Bending of Electric Waves round the 

 Earth. III. By J. W. Nicholson, M.A., B.Sc* 



IN the second note on this subject "\, M. Poincare's second 

 investigation was considered, and it was indicated that 

 when a radial oscillator is placed close to the earth, regarded 

 as a perfect conductor, the effect produced by diffraction at 

 any point in the geometrical shadow can only differ from 

 an exponential type with large negative argument by a series 

 of inverse powers of ka, where ka is a magnitude roughly of 

 order 10 -6 . In the meantime, the writer has continued the 

 examination of these " residual" terms depending on har- 

 monics of low order, and has proved that they do in fact 

 have a zero sum for points on the surface of the sphere. 

 They play their part in the remarkable compensation of the 

 terms of the harmonic series, so that the total effect at a 

 point on the surface is, in fact, exponential. M. Poincare's 

 exponential factor is thus valid for surface points, although 

 his mode of investigation in its present form does not apply 

 to points not on the surface, even though fairly close to it. 

 The proof of the evanescence of these terms in inverse 



* Communicated by the Anthor. 

 t Phil. Mag. March 1910. 



