Kerr  Effects  in  the  Infra-red  Spectrum. 
65 
should  be  stated,  however,  that,  because  of  unavoidable  un- 
steadiness, points  determined  by  deflexions  of  less  than 
10  mms.  must  be  regarded  as  uncertain,  and  this  applies 
particularly  to  those  points  near  the  limits  of  the  spectral 
region  covered.  In  the  curves,  the  rotations  given,  measured 
in  minutes  of  arc,  are  as  determined  for  the  two  surfaces — 
unless  otherwise  specified;  and  for  reversal  of  the  magnet, 
and  hence  to  be  put  on  the  ordinary  basis  for  the  measure- 
ment of  the  Kerr  effect,  should  be  halved  so  as  to  obtain  the 
Fig.  4. —  Variation  of  Kerr  Effect  with  wave-length.     Constant  field. 
Double-refracting  polarizing  apparatus. 
-80 
-70 
&-60 
1 
K-50 
5 
-40 
-30 
-20 
WM£LENGTH 
rotation  for  a  single  surface.  On  the  same  basis  as  the 
results  on  the  Faraday  rotation,  they  are  really  fourfold, 
since  the  rotations  in  the  former  case  were  stated  for  only  a 
single  throw  of  the  magnet.  As  is  customary,  the  sense  of 
the  rotation  is  taken  as  negative,  being  opposite  in  direction 
to  current  in  the  magnet-coils. 
To  check  the  results  with  as  large  a  change  as  possible  in 
the  apparatus,  the  curves  were  all  repeated  with  the  glass- 
plate  polarizer  and  analyser  replaced  by  a  double-image  prism 
Phil.  Mag.  S.  6.  Vol.  11.  No.  61.  Jan,  15)06.  l  F 
, 
X 
© 
X 
f 
\ 
\ 
Sr£ 
EL 
J 
D 
\ 
\ 
\ 
\ 
©\ 
\ 
X 
*L      > 
\ 
s 
A 
A 
N 
^x. 
\ 
^ 
N 
V 
A 
> 
A 
