826  Dr.  C.  G.  Barkla  on 
secondary  radiation  connect  the  absorption  and  atomic 
weights  o£  elements  in  the  first  and  part  of  the  second  long- 
chemical  periods,  and  that  the  latter  part  of  such  a  curve  has 
been  obtained  with  the  latter  part  of  the  third  long  period. 
These  are  the  periods  shown  by  Mc  Clelland ,  by  experiments 
on  the  secondary  rays  from  substances  subject  to  ft  and  y 
rays  from  radium.  His  curves  also  show  the  second  short 
period.  The  results  obtained  in  these  experiments  from 
substances  of  atomic  weights  less  than  32  have  not  been 
plotted,  because  they  were  made  under  different  conditions, 
the  radiating  layer  of  gas  absorbing  only  a  very  small 
fraction  of  the  incident  primary  radiation.  These  substances 
were  found  to  emit  rays  differing  little  in  character  from  the 
primary. 
It  will  be  noticed  that  some  substances  emitted  a  radiation 
whose  ionizing  effects  were  diminished  more  by  an  absorbing 
plate  than  were  those  of  the  direct  primary  beam.  The  rays, 
however,  were  produced  by  the  penetrating  portion  of  the 
primary  beam,  so  the  transformation  was  one  to  greater 
absorbability.  This  was  seen  by  placing  aluminium  screens 
in  the  primary  beam  before  it  fell  on  the  radiating  substance, 
for  the  diminution  of  the  secondary  ionization  was  con- 
siderably less  than  when  the  same  screen  was  placed  in  the 
path  of  the  secondary  beam.  The  reason  for  this  is  obvious 
when  we  consider  the  different  penetrating  powers  of  the 
constituents  of  the  primary  beam. 
[Later  experiments  have  shown  that  variations  in  the 
primary  beam  have  such  an  enormous  effect  on  the  character 
of  the  radiations  from  the  substances  referred  to  as  being 
inconstant,  that  the  relative  positions  of  antimony  and  iodine 
may  have  to  be  reversed.  The  values  first  obtained  with  an 
approximately  constant  primary  beam  are,  however,  un- 
altered in  this  paper,  as  the  true  positions  are  not  known 
with  certainty.] 
Theory. 
The  theory  which  has  been  -  shown,  to  account  for  the 
phenomena  of  secondary  radiation  from  certain  gases  and 
light  solids  may  be  extended  to  explain  the  results  of  ex- 
periments on  metals. 
In  light  atoms  the  corpuscles  or  electrons  have  sufficient 
freedom  to  move  almost  entirely  independently  of  each  other 
under  the  influence  of  the  primary  pulses,  consequently*  to 
emit  a  radiation  whose  penetrating  power  is  the  same  as  that 
of  the  primary,  and  whose  intensity  depends  on  the  direction 
*  Transactions  of  Royal  Dublin  Society.  May  17,  1905. 
