192  Mr.  G.  K.  Winter  on  the  Corona  seen  during 
through  the  Savart's  polariscope  to  the  eye*.    Fig.  1,  Plate  II. 
will  perhaps  render  this  description  clearer. 
Although  quite  convinced  myself  of  the  fact  of  the  radial  po- 
larization of  the  corona,  I  was  anxious  this  time  to  place  it 
beyond  doubt  by  taking  actual   measurements  of  it  in  such  a 
position  that,  if  the  polarized  light  proceeded  from  the  unobscured 
portion  of  the  earth  reflected  into  the  atmosphere  and  again 
back  to  the  eye,  it  could  not  be  measured.     I  therefore  chose  the 
southern  limb  for  my  observations,  and  carefully  got  my  bands 
radial  to  the  sun,  and  consequently  making  but  a  small  angle 
with  the  horizon,  before  totality,  keeping  the  field  as  nearly 
as  I  could  in  the  same  position-angle  with  respect  to  the  sun 
by  means  of  the  right-ascension  tangent-rod  during   totality. 
Immediately  totality  commenced,  the  white-centred  bands  ap- 
peared.    I  turned  the  axis  of  the  frame  with  the  glass  plates 
until  the   bands   disappeared.     The   angle  the   plates  had  to 
be   turned  through  was  35°.     I  then  turned  the  declination 
tangent-screw  slightly,   so  as  to  get  a  portion  of  the  corona 
a  small  distance  from  the  limb  (I  think  about  10')  into  the 
field.     The  plates  had  then  to  be  turned  through  an  angle  of 
45°  before  the  bands  disappeared.     Three  other  measurements 
were  taken  in  about  the  same  position,  the  result  showing  that 
the  polarization  increased  considerably  with  distance  from  the 
limb. 
It  is  evident  that  if  the  polarized  light  were  really  due  to  the 
reflection  from  the  unobscured  portion  of  the  earth,  it  would  be 
polarized  in  a  plane  nearly  at  right  angles  to  the  plane  of  my 
bands,  and  consequently  its  polarization  could  not  be  neutralized 
by  the  plates  of  glass  in  the  position  in  which  they  were  used. 
When  the  plates  were  inclined,  so  as  to  neutralize  the  corona 
polarization,  I  saw  faint  black-centred  bands  on  the  portion  of 
the  moon's  disk  in  the  field.  I  did  not  observe  any  when  the 
plates  were  at  right  angles  to  the  axis  of  the  telescope;  but  I 
think  I  should  have  noticed  them  if  they  had  existed;  so  that 
although  there  was  a  sensible  amount  of  light  on  the  moon's 
disk,  sufficient  to  show  bands  when  polarized  by  the  glass  plates, 
I  do  not  think  it  was  perceptibly  polarized  itself. 
With  regard  to  the  evidence  of  the  spectroscope,  the  existence 
of  the  bright  line  1474  of  KirchhofFs  scale  (first  observed,  I  be- 
lieve, by  Mr.  Pogson  at  the  eclipse  of  August  1868)  seems  fully 
established.  This  linef,  which  is  also  seen  in  aurorse,  would 
lead  us  to  suppose  that  a  gas  existing  in  the  higher  regions  of 
*  The  bands  of  the  polariscope  were  white-centred  when  in  the  plane  of 
polarization.  0 
t  Kirchhoff  and  Angstrom,  I  believe,  suppose  this  line  to  belong  to  the 
spectrumof  iron. 
