Kerr  Effects  in  the  Infra-red  Spectrum.  43 
L.  Grunmach  *.  None  of  these  experimenters,  however, 
made  any  attempt  to  obtain  radiations  of  uniform  and  known 
wave-length  for  their  work ;  and  G.  Moreau  f  seems  to  be 
the  only  one  who  has  heretofore  done  this  and  so  obtained 
measurements  of  magnetic  rotatory  dispersion  in  the  infra- 
red. He  determined  the  rotation  in  a  tube  of  carbon 
bisulphide,  placed  in  a  powerful  magnetizing  solenoid,  for 
wave-lengths  ranging  from  079  ^  to  1*42  /x.  A  beam  of 
sunlight  was  passed  in  turn  through  a  polarizer,  carbon- 
bisulphide  tube,  and  analyser,  and  its  intensity  after  disper- 
sion by  a  prism  of  flint  glass  was  measured  by  a  thermopile 
placed  in  the  spectrum.  The  method  of  observation,  in 
brief,  consisted  in  determining  the  azimuth  of  the  plane  of 
polarization  of  the  light  emerging  from  the  carbon  bisulphide 
by  two  observations  of  the  intensity  of  radiation  for  two 
positions  of  the  analyser  90°  different  in  azimuth.  By 
repeating  this  when  the  current  was  passing  in  the  solenoid, 
the  rotation  due  to  the  magnetic  field  could  readily  be 
calculated.  Measurements  by  ordinary  polarimetric  means 
allowed  the  calculation  of  the  field-strengths  by  the  use  of 
known  data,  and  the  effect  of  this  field  at  the  galvanometer 
was  annulled  by  placing  near  by  a  small  secondary  spool  in 
shunt  with  the  main  solenoid.  He  found  rotations  amount- 
ing in  general  to  several  degrees,  being  at  \  =  079  /x  equal 
to  52  per  cent,  of  the  rotation  for  sodium  light,  and  at 
X=1'42  //,,  32  percent.,  with  an  estimated  maximum  error 
averaging  about  10  per  cent. 
In  the  present  work  the  magnetic  rotation  has  been 
measured  for  thirty  wave-lengths  between  X  =  0*58  /j,  and 
4\3  \x\  carbon  bisulphide,  the  substance  generally  used  in 
magnetic  polarimetry,  being  chosen  for  the  test/  because  of 
its  high  rotatory  power,  combined  with  transparency  in  the 
infra-red.  A  powerful  electromagnet  was  used,  instead  of 
a  solenoid,  to  furnish  the  magnetic  field  ;  for  the  gain  in 
field-strength  allowed  the  use  of  a  shorter  carbon-bisulphide 
tube  without  sensible  loss  of  rotation,  and  with  the  great 
added  advantage  of  less  absorption  of  radiation  in  the  infra- 
red. To  simplify  the  problem,  all  other  conditions  upon 
which  the  rotation  depends,  such  as  field-strength,  length  of 
carbon-bisulphide  tube,  and  temperature, — save  the  wave- 
length of  the  light  used, — were  kept  as  nearly  as  possible  the 
same  throughout  all  the  measurements,  and  hence  need  not 
be  accurately  known,  since,  being  constant,  they  could  in  no 
way  affect  the  form  of  the  dispersion-curve,  the  determination 
of  which  was  the  desired  end  o\'  the  experiment. 
*  Wied.  Ann.  \i\.  p.  85  (1881). 
t  Ann.  Chim.  Phys.  [7]  i.  i>.  227  (1894). 
