46  Mr.  L.  R.  Ingersoll  on  the  Faraday  and 
ratio  the  rotation  is  at  once  calculated.  This  method  as  so 
far  described  has  about  the  same  possibilities  of  sensibility  as 
that  of  Moreau, — with  the  advantage,  however,  that  the 
rotation  is  obtained  almost  directly,  instead  of  as  the  differ- 
ence between  two  large  angles  ;  but  its  ability  to  measure 
small  rotations  may  be  greatly  increased  by  the  following 
simple  scheme.  Suppose  it  is  possible,  after  measuring  the 
intensity  I,  to  increase  the  bolometer  sensibility  in  a  known 
ratio,  say  twenty  times,  and  at  the  same  time  to  shift  the  zero 
of  the  galvanometer  so  that  the  reading,  which  the  radiation  I 
in  connexion  with  the  greatly  increased  sensibility  would 
otherwise  throw  far  off  the  scale,  may  still  be  brought 
within  the  field  of  vision.  In  this  way,  the  change  of  inten- 
sity dl,  which  is  in  general  only  a  small  fraction  of  I,  may 
be  effectively  magnified  and  made  to  produce  a  galvanometer 
deflexion  of  the  same  order  as  the  latter,  with  a  corresponding 
increase  in  the  possible  accuracy  of  measurement,  for  the 
change  of  sensibility  may  be  readily  effected  by  cutting  out 
resistance  in  series  with  the  galvanometer,  and  the  shift  of 
the  zero  by  simply  adjusting  the  balance  of  the  bolometer. 
Description  of  Apparatus. 
Before  describing  the  apparatus  in  detail,  certain  special 
requirements  of  the  problem  should  be  noticed.  Since  rota- 
tion appears  as  a  change  of  perhaps  only  a  few  per  cent,  in 
the  energy  at  any  wave-length,  it  is  evident  that  for  accurate 
measurement  a  very  intense  spectrum  is  demanded.  This 
must  be  formed  moreover  of  light  which  has  been  polarized, 
transmitted  down  the  axis  of  the  magnet,  analysed,  and  which 
has  suffered  a  number  of  other  losses  by  reflexion  and  absorp- 
tion ;  hence  the  necessity  for  a  very  brilliant  source.  It  is 
clear,  too,  that  this  source  must  be  exceptionally  constant, 
for  changes  of  intensity  might  be  falsely  interpreted  as 
rotations. 
These  requirements  have  been  met  by  the  use  of  a  Nernst 
glower,  with  special  protective  covering,  as  a  source.  The 
formation  of  a  parallel  beam,  as  is  generally  used  in  polari- 
metry,  which  should  be  of  sufficient  intensity,  and  of  small 
enough  cross-section  to  suit  the  necessarily  small  apertures  in 
the  poles  of  the  magnet,  was  found  impracticable :  a  conical 
beam  was  therefore  used,  which  converged  to  a  focus  at  the 
only  points  where  small  aperture  was  demanded, — between 
the  poles  of  the  magnet  and  on  the  slit  of  the  spectrometer. 
The  latter  instrument  had  short-focus  mirrors,  and  was 
designed  to  give  as  intense  a  spectrum  as  possible  without 
