﻿926* 
  Prof. 
  Bragg 
  and 
  Mr. 
  Madsen 
  : 
  An 
  Experimental 
  

  

  ,->.— 
  

  

  The 
  representative 
  points 
  do 
  not 
  lie 
  closely 
  on 
  a 
  smooth 
  

   curve. 
  To 
  some 
  extent, 
  no 
  doubt, 
  this 
  is 
  due 
  to 
  experimental 
  

   error, 
  for 
  the 
  measured 
  quantities 
  are 
  very 
  small. 
  But 
  we 
  

   arc 
  inclined 
  to 
  think 
  that 
  the 
  departures 
  from 
  regularity 
  are 
  

   to 
  some 
  extent 
  real. 
  We 
  hope 
  to 
  settle 
  this 
  point 
  partly 
  by 
  

   improving 
  our 
  apparatus, 
  ancl 
  so 
  attaining 
  greater 
  accuracy 
  

   of 
  measurement, 
  and 
  partly 
  by 
  obtaining 
  better 
  knowledge 
  of 
  

   the 
  corrections 
  to 
  be 
  applied 
  to 
  the 
  observations. 
  For 
  our 
  

   present 
  purpose, 
  it 
  is 
  enough 
  to 
  draw 
  a 
  smooth 
  curve 
  such 
  as 
  

   AA, 
  and 
  to 
  ignore 
  the 
  variations 
  from 
  it. 
  

  

  The 
  figure 
  -hews 
  a 
  rapid 
  decrease 
  in 
  the 
  emergence 
  radia- 
  

   tion- 
  a- 
  we 
  proceed 
  from 
  C 
  to 
  Te. 
  There 
  is 
  then 
  a 
  small 
  rise 
  

   to 
  W 
  and 
  Fb. 
  This 
  is 
  easily 
  explained, 
  and 
  exemplifies 
  a 
  

   very 
  important 
  feature 
  <>t' 
  the 
  problem. 
  We 
  must 
  remember^ 
  

   no! 
  only 
  that 
  X 
  p 
  is 
  fairly 
  constant 
  for 
  the 
  heavy 
  atoms, 
  but 
  

   also 
  that 
  the 
  lead 
  screen 
  did 
  not 
  completely 
  remove 
  the 
  soft 
  

   rays 
  : 
  in 
  other 
  words, 
  that 
  the 
  pencil 
  of 
  7 
  rays 
  was 
  not 
  quite 
  

   homogeneous. 
  

  

  This 
  will 
  bo 
  dearly 
  understood 
  by 
  comparing 
  the 
  curve 
  

   AA 
  with 
  the 
  curve 
  BB, 
  which 
  represents 
  the 
  results 
  we 
  

   obtained 
  when 
  the 
  lead 
  plug 
  was 
  removed 
  and 
  tin 
  1 
  7 
  rays 
  had 
  

   passed 
  through 
  only 
  the 
  wall 
  of 
  the 
  vessel 
  containing 
  the 
  

   radium 
  before 
  entering 
  the 
  chamber. 
  There 
  was 
  therefore 
  a 
  

   quantity 
  of 
  -oft 
  radiation 
  in 
  the 
  stream 
  of 
  y 
  rays 
  ; 
  the 
  effect 
  

   was 
  to 
  increase 
  considerably 
  the 
  emergence 
  radiation 
  of 
  Pb 
  

   and 
  \V 
  relatively 
  to 
  that 
  of 
  the 
  other 
  substances. 
  This 
  was. 
  

   as 
  it 
  should 
  have 
  been. 
  We 
  know 
  that 
  hard 
  7 
  rays 
  pay 
  no 
  

   attention 
  to 
  atomic 
  structure, 
  but 
  that 
  soft 
  rays 
  distinguish 
  

   between 
  atom- 
  of 
  different 
  weight. 
  In 
  the 
  table 
  given 
  by 
  

   Wigger 
  (Jahrbuch 
  der 
  Radioaktivitdt, 
  11)05, 
  p. 
  432) 
  the 
  

   value- 
  of 
  X 
  p 
  for 
  thin 
  sheets 
  of 
  Pb 
  and 
  Zn 
  are 
  '068 
  and 
  *039 
  

   respectively 
  : 
  but 
  when 
  the 
  rays 
  have 
  been 
  hardened 
  by 
  

   passing 
  through 
  2*8 
  cm. 
  of 
  lead, 
  each 
  coefficient 
  has 
  become 
  

   nearly 
  '02. 
  Thus, 
  the 
  hard 
  rays 
  treat 
  lead, 
  and 
  zinc 
  alike* 
  

   but 
  softer 
  ray- 
  are 
  more 
  absorbed 
  by 
  the 
  former 
  than 
  the 
  

   latter, 
  weight 
  for 
  weight. 
  "When 
  a 
  stream 
  of 
  7 
  rays 
  is 
  

   passed 
  through 
  a 
  lead 
  plate, 
  the 
  soft 
  rays 
  are 
  rapidly 
  con- 
  

   verted 
  into 
  ft 
  rays. 
  Consequently, 
  a 
  thin 
  lead 
  plate 
  produces 
  

   a 
  large 
  quantity 
  of 
  emergence 
  radiation, 
  due 
  principally 
  to 
  

   the 
  slow 
  ft 
  rays 
  produced 
  by 
  the 
  soft 
  7 
  rays. 
  The 
  effect 
  is 
  

   further 
  discussed 
  later 
  on 
  in 
  the 
  paper. 
  But 
  this 
  is 
  rapidly 
  

   used 
  up 
  ; 
  consequently, 
  the 
  radiation 
  measured 
  on 
  the 
  far 
  

   side 
  of 
  a 
  lead 
  plate 
  is 
  relatively 
  large 
  when 
  the 
  plate 
  is 
  thin, 
  

   and 
  falls 
  off 
  more 
  quickly 
  at 
  first 
  than 
  it 
  does 
  subsequently, 
  

   the 
  plate 
  being 
  gradually 
  thickened. 
  The 
  logarithmic 
  curve 
  

   of 
  Pb 
  has 
  a 
  rapid 
  initial 
  fall, 
  as 
  several 
  observers 
  have 
  shown. 
  

  

  