Radiation  from  Ordinary  Materials.  219 
The  experiments  consisted  in  measuring  the  ionization 
through  the  box,  then  placing  a  plate  of  the  metal  investigated 
on  the  window  and  noting  the  increase  in  the  ionization 
caused  by  the  rays  from  it,  which  penetrated  the  window. 
The  plate  was  then  raised  a  known  distance  from  the  window, 
and  the  effect  of  the  layer  of  air  on  the  intensity  of  those 
rays  was  then  noted  ;  comparison  of  the  observations  gave 
the  penetration  of  the  rays.  The  ionization  was  measured 
by  the  compensation  method  (§  4). 
Only  three  metals  were  measured  in  this  manner — lead, 
copper,  and  platinum.  As  an  example  of  the  observations 
recorded  some  figures  are  given,  which  are  the  mean  of  a 
large  number  of  readings. 
Normal  ionization  in  the  box .  .   -821  x  10 
—3 
With  lead  plate  on  window    .  .    -881  X 10 
Lead  plate  *5  cm.  from  window.  -872  x  10~3 
Lead  plate  L0  cm '865  xlO-3 
Lead  plate  2  cm -852 X  10~3 
Lead  plate  3  cm -844 X  10~" 
diff.6-0xl0~5,  ratio  LOO 
•5-1 
•85. 
4-4 
•74 
31 
•52 
■38 
We  can  ascertain  to  what  values  of  a  these  readings 
correspond  in  the  following  manner.  In  §  7  the  ionization 
in  a  layer  of  thickness  x  next  to  the  plate  has  been  calculated . 
The  difference  between  this  and  the  whole  ionization  is  the 
portion  existing  beyond  this  layer  ;  this  latter  ionization  is 
caused  by  the  rays  which  have  passed  through  the  thickness  «r. 
A  reference  to  the  analysis  will  show  that  a  layer  of  thickness 
d  of  a  material  density  p  is  equivalent  in  all  respects  to  a 
thickness  of  air  x  if  pd  —  x.  Hence  if  fit  is  the  ratio  of  the 
ionization  caused  by  rays,  which  pass  through  a  thickness  of 
material  equivalent  to  a  layer  of  air  x,  to  the  whole  ionization 
which  would  be  caused  in  the  absence  of  that  layer,  at  =  1  —  — , 
Vx    •  •  P 
where  —  is  found  from  Table  I. 
1\ 
The  ratios  given  in  the  last  column  of  the  figures  for  lead 
represent  fit  for  the  values  *5,  1,  2,  3  of  x.  The  corresponding 
numbers,  if  a  =  12,  are  '80,  '12,  '52,  *36  ;  if  a=13,  they 
are  '87,  '74,  -54,  '39.  We  thus  see  that  a  for  lead  lies 
between  12  and  13,  and  confirm  the  results  given  by  fche 
curves. 
