72       Faraday  and  Kerr  Effects  in  the  Infra-red  Spectrum. 
exhibit  its  magnetic  properties.  If  more  careful  tests  on  it, 
as  well  as  on  the  non-magnetic  alloys  of  nickel,  can  show  that 
the  Kerr  effect  is  not  entirely  dependent  on  magnetic 
properties,  a  considerable  advance  will  be  made  towards  the 
explanation  of  the  phenomena. 
Summary. 
1.  The  electromagnetic  rotatory  dispersion  of  carbon 
bisulphide  has  been  measured  by  infra-red  methods  over  a 
range  of  spectrum  extending  from  the  sodium  lines  to 
X=4'3/>t,  and  found  to  be  correctly  represented  by  a  formula 
which  takes  account  of  the  absorption-band  beyond  8  /j,, 
showing  that  an  infra-red  absorption-band  may  affect  the 
rotatory  dispersion,  much  as  it  does  ordinary  dispersion,  oyer 
a  considerable  range  of  spectrum. 
2.  The  magnetic  metals  and  magnetite  show,  after  wave- 
length  1  //,  in  the  infra-red,  a  decrease  of  the  Kerr  rotation, 
with  increase  of  wave-length.  The  complete  rotatory  disper- 
sion-curves, made  by  supplementing  the  results  for  the  infra- 
red by  existing  observations  for  the  visible  spectrum,  show  a 
marked  resemblance  to  a  typical  dispersion-curve  in  the 
region  of  an  absorption-band,  indicating  the  existence  in 
metals  of  something  analogous  to  a  region  of  resonance- 
absorption,  extending  over  the  visible  spectrum. 
3.  The  particular  cases  of  nickel  and  magnetite  are  notable, 
for  the  rotation  appears  to  vanish  for  a  particular  wave- 
length in  each  case,  and  then  change  in  sign.  The  Kerr 
rotation  for  the  magnetic  alloy,  Heusler's  metal,  if  it  exists 
at  all,  is  less  than  one  tenth  of  that  for  iron  or  steel,  although 
the  magnetic  properties  of  the  metal  are  quite  comparable. 
4.  Although  the  results  do  not  allow  of  definite  conclusions 
as  to  whether  the  hypothesis  of  molecular  currents,  or  of  the 
Hall  effect,  should  be  accepted  in  explaining  magnetic  rota- 
tion, the  indications  are  that  the  latter  theory  holds  for 
carbon  bisulphide — i.  e.  it  presents  a  case  analogous  to  that 
of  sodium  vapour  ;  while  the  curves  of  the  magnetic  metals 
require  the  former  explanation,  although  the  reversal  of 
nickel  might  perhaps  be  considered  as  evidence  for  the  Hall 
effect. 
In  conclusion,  I  wish  to  express  my  sincere  thanks  to 
Prof.  B.  W.  Snow  for  his  kindness  in  furnishing  apparatus 
and  supplies  ;  to  Prof.  A.  Trowbridge  for  assistance  in  pro- 
curing certain  specimens  of  metals  ;  and  especially  to  Prof. 
C.  E.  Mendenhall,  at  whose  suggestion  the  work  was  under- 
taken, and  whose  advice  throughout  its  progress  has  been  of 
the  greatest  service. 
Physical  Laboratory,  University  of  Wisconsin. 
July.  190-5. 
