﻿558 
  Williams 
  — 
  Twist 
  in 
  Steel 
  and 
  Nickel 
  Rods. 
  

  

  Method 
  of 
  Taking 
  Observations. 
  

  

  The 
  rods 
  used 
  in 
  this 
  work 
  were 
  suspended 
  by 
  the 
  upper 
  

   end 
  in 
  a 
  vertical 
  solenoid* 
  used 
  in 
  previous 
  investigations. 
  

   A 
  concave 
  mirror 
  was 
  attached 
  to 
  the 
  lower 
  end, 
  which 
  threw 
  

   the 
  image 
  of 
  an 
  incandescent 
  filament 
  upon 
  a 
  slit 
  behind 
  

   which 
  a 
  photographic 
  film 
  moved 
  and 
  on 
  which 
  were 
  traced 
  

   the 
  deflections 
  of 
  a 
  spot 
  of 
  light: 
  

  

  For 
  graphs 
  1, 
  2, 
  3 
  and 
  4, 
  film 
  T 
  1? 
  fig. 
  1, 
  graphs 
  1, 
  2, 
  3 
  and 
  

   4, 
  film 
  T 
  3 
  and 
  graphs 
  5, 
  6, 
  7 
  and. 
  8, 
  film 
  T 
  4 
  , 
  fig. 
  2, 
  the 
  rods 
  

   were 
  demagnetized 
  before 
  each 
  graph 
  was 
  made 
  ; 
  that 
  is, 
  

   everything 
  being 
  in 
  adjustment 
  a 
  decreasing 
  alternating 
  cur- 
  

   rent 
  was 
  sent 
  through 
  the 
  solenoid 
  and 
  then 
  graph 
  1 
  was 
  run 
  

   off. 
  After 
  the 
  maximum 
  field 
  strength 
  had 
  been 
  attained 
  for 
  

   graph 
  1 
  the 
  field 
  was 
  suddenly 
  broken 
  and 
  the 
  rod 
  again 
  de- 
  

   magnetized 
  by 
  the 
  decreasing 
  alternating 
  current 
  field. 
  

   Graph 
  2 
  was 
  then 
  run 
  off 
  but 
  with 
  the 
  magnetic 
  field 
  in 
  the 
  

   opposite 
  direction 
  to 
  that 
  of 
  graph 
  1. 
  This 
  was 
  repeated 
  for 
  

   graphs 
  3 
  and 
  4. 
  In 
  all 
  cases, 
  odd-numbered 
  graphs 
  are 
  for 
  

   fields 
  directed 
  upwards 
  in 
  the 
  solenoid 
  while 
  even-numbered 
  

   graphs 
  have 
  their 
  fields 
  downward. 
  Graphs 
  1 
  and 
  3 
  are 
  dupli- 
  

   cates 
  under 
  the 
  same 
  conditions, 
  which 
  also 
  hold 
  for 
  graphs 
  

   2 
  and 
  4, 
  5 
  and 
  7, 
  etc. 
  In 
  all 
  of 
  my 
  previous 
  work 
  I 
  have 
  

   demagnetized 
  in 
  this 
  way 
  before 
  each 
  graph 
  was 
  taken 
  so 
  that 
  

   the 
  rods 
  would 
  have 
  the 
  same 
  treatment 
  before 
  taking 
  the 
  

   next 
  graph. 
  

  

  The 
  parallel 
  straight 
  lines 
  in 
  the 
  photographs 
  indicate 
  defi- 
  

   nite 
  currents 
  flowing 
  in 
  the 
  solenoid. 
  The 
  values 
  of 
  the 
  field 
  

   strengths 
  for 
  these 
  currents 
  may 
  be 
  obtained 
  from 
  figs. 
  3 
  and 
  4. 
  

  

  I 
  have 
  gone 
  to 
  some 
  length 
  in 
  speaking 
  of 
  the 
  magnetic 
  

   treatment 
  which 
  these 
  rods 
  received 
  before 
  each 
  graph 
  was 
  

   made, 
  as 
  an 
  inspection 
  of 
  the 
  curves 
  in 
  figs. 
  3 
  and 
  4 
  shows 
  that 
  

   the 
  previous 
  history 
  makes 
  a 
  great 
  difference 
  in 
  the 
  results 
  if 
  

   one 
  wishes 
  to 
  duplicate 
  them. 
  Graphs 
  1 
  and 
  3 
  or 
  2 
  and 
  4 
  in 
  

   any 
  of 
  the 
  figures 
  show 
  that 
  by 
  similar 
  treatment 
  remarkable 
  

   duplications 
  may 
  be 
  obtained 
  after 
  a 
  cyclic 
  state 
  has 
  been 
  

   established. 
  

  

  To 
  show 
  what 
  the 
  effect 
  would 
  be 
  if 
  the 
  rods 
  were 
  not 
  de- 
  

   magnetized, 
  film 
  T 
  2 
  , 
  T 
  5 
  and 
  graphs 
  5, 
  6, 
  7 
  and 
  8 
  of 
  film 
  T 
  3 
  

   have 
  been 
  added. 
  The 
  field 
  was 
  reversed 
  for 
  each 
  graph 
  just 
  

   as 
  in 
  the 
  case 
  for 
  the 
  other 
  films, 
  but 
  the 
  demagnetization 
  was 
  

   omitted. 
  Even 
  for 
  graph 
  1 
  a 
  field 
  was 
  previously 
  imposed 
  on 
  

   the 
  rod 
  opposite 
  to 
  that 
  which 
  it 
  was 
  to 
  have 
  for 
  graph 
  1. 
  In 
  

   each 
  case 
  the 
  field 
  was 
  broken 
  suddenly, 
  and, 
  without 
  demag- 
  

   netization 
  the 
  circuit 
  was 
  reversed 
  for 
  the 
  succeeding 
  graph. 
  

   Just 
  as 
  in 
  the 
  case 
  of 
  the 
  graphs 
  showing 
  demagnetization 
  so 
  

  

  *Phys. 
  Rev., 
  xxxiv, 
  258, 
  1912. 
  

  

  