62  Mr.  L.  R.  Ingersoll  on  the  Faraday  and 
Part  II. 
The  Kerr  Effect  *. 
Dm  Bois  f  has  shown  that,  for  the  visible  spectrum,  the 
rotation  in  the  Kerr  effect  in  general  increases  with  longer 
wave-length  in  the  light  used.  This  is  notably  true  in  the 
case  of  iron  or  steel,  while  cobalt  and  nickel  exhibit  weakly 
denned  minima  in  the  green  and  yellow  respectively,  and 
magnetite  appears  to  reach  a  maximum  in  the  yellow. 
Inasmuch  as  this  is  entirely  different  from  the  magnetic 
dispersion  shown  by  most  substances — e.  g.  carbon  bisulphide 
— -and  is  not  satisfactorily  accounted  for  by  theory,  it  is  a 
matter  of  considerable  interest  to  determine  whether  this 
anomalous  nature  persists  throughout  the  whole  spectrum,  or 
is  merely  characteristic  of  the  limited  region  studied,  which 
includes  less  than  *3  /-t. 
This  necessitates  infra-red  measurements,  and  accordingly, 
when  it  was  found  in  connexion  with  the  work  on  the 
Faraday  effect  that  rotations  of  the  order  of  a  degree  could 
be  measured,  the  apparatus  was  modified  to  suit  the  case  of 
reflexion,  and  the  effect  found  to  be  measurable.  The 
results  so  far  obtained  show  fairly  good  agreement  among 
themselves,  indicating  as  great  an  accuracy  as  might  reason- 
ably be  expected,  in  view  of  the  difficulties  of  the  problem. 
Since,  however,  they  are  of  quite  an  unexpected  character 
and  point  to  certain  rather  important  conclusions;  it  would 
seem  most  desirable  to  verify  them,  with  apparatus  modified 
to  secure  greater  sensibility,  and  with  more  perfect  surfaces, 
and  thus  to  explain  or  remove,  if  possible,  inconsistencies 
which  must  now  detract  from  their  certainty.  To  re-model 
the  apparatus,  however,  and  to  make  a  more  complete  study 
of  the  problem,  including  the  important  case  of  rotation  by 
transmission  through  thin  films  of  magnetic  metals,  may 
require  considerable  time,  so  the  first  stage  of  the  work  is 
here  presented,  and  is  not  intended  to  be  regarded  as  final 
and  complete,  but  rather  in  the  light  of  a  preliminary  report. 
Method. 
The  rotation  was  produced  by  reflexion  at  the  surfaces  of 
two  small  mirrors  of  the  metal  to  be  tested,  mounted  near 
the  ends  of  the  pole-pieces  of  the  magnet.  To  double  the 
effect  for  greater  facility  of  measurement,  and  at  the  same 
*  Read  in  part  before  meeting-  of  American  Physical  Society,  Chicago, 
April  1905. 
t  Wied.  Jinn,  xxxix.  p.  25  (1890) ;  Phil.  Mag.  [5]  xxix.  p.  253  (1890). 
