818  Dr.  0.  G.  Barkla  on 
From  these  we  see  that  the  ionization  produced  by  the 
secondary  radiation  from  the  intense  primary  beam  was  re- 
duced by  68'3  per  cent,  of  its  original  value,  while  that  pro- 
duced by  the  secondary  radiation  set  up  by  the  weak  primary 
beam  was  diminished  by  70  per  cent. 
The  difference  between  these  wTas  within  the  limits  of 
possible  error  of  experiment.  We  thus  see  that  the  character 
of  the  secondary  radiation  from  calcium  did  not  depend  to 
any  appreciable  extent  on  the  intensity  of  the  primary. 
Similar  experiments  were  made  when  iron  was  used  as  the 
radiating  substance.  The  sheet  of  aluminium  when  placed  in 
the  path  of  the  secondary  beam  diministied  its  ionizing  effect 
in  successive  experiments  by  906  per  cent.,,  90*9  per  cent., 
90*9  per  cent.,  and  90*5  per  cent.  In  the  first  experiment 
the  radiation  emitted  by  the  tube  was  very  weak.  Before 
the  second  the  tube  was  heated,  consequently  it  worked  much 
more  easily  and  produced  a  very  much  greater  ionization. 
For  the  third  it  was  removed  to  something  like  double  the 
distance  from  the  iron  radiator,  and  for  the  fourth  it  was 
brought  back.  The  variation  in  absorption  as  shown  by  the 
above  numbers  must  have  been  exceedingly  small,  if  it  existed 
at  all. 
Experiments  have  been  made  at  different  times  to  deter- 
mine if  the  intensity  of  secondary  radiation  was  proportional 
to  that  of  the  primary.  In  all  cases  so  far  investigated,  the 
proportionality  has  been  demonstrated  within  two  or  three 
per  cent. 
Variation  in  Penetrating  Power  of  the  Primary  Radiation. 
It  was  of  interest  to  determine  in  what  way  and  to  what 
extent  the  secondary  radiation  from  a  particular  substance 
depended  on  the  character  of  the  primary  radiation.  From 
substances  of  low  atomic  weight  it  had  been  seen  that  the 
character  was  dependent  solely,  or  almost  solely,  on  that  of 
the  primary  and  to  an  inappreciable  extent  on  the  nature  of 
the  radiator. 
It  was  found  that  as  the  difference  in  character  between 
the  secondary  and  primary  rays  became  more  marked  by 
increasing  the  atomic  weight  of  the  radiator,  the  effect  on  the 
secondary  of  a  change  in  character  of  the  primary  diminished. 
From  those  substances  which  emitted  a  radiation  varying 
in  intensity  in  different  directions  when  the  primary  beam 
was  polarized,  there  were  considerable  variations  in  ab- 
sorbability, the  secondary  becoming  more  penetrating  as  the 
primary  became  more  penetrating.  The  radiation  from  those 
substances,  however,  which  produced  considerable   transfor- 
