298         Prof.  H.  A.  Bumstead  on  the  Heating  Effects 
were  carefully  made,  little  difficulty  was  experienced  in 
maintaining  the  vacuum  for  considerable  periods ;  the  rise  in 
pressure  due  to  leakage  was  usually  less  than  0*002  mm.  per 
day. 
In  the  preliminary  experiments  for  testing  the  working  con- 
ditions of  the  radiometer,  the  ebonite  wheel  and  its  metal  strips 
were  replaced  by  a  light  ebonite  frame  which  carried  two  strips 
of  platinum  foil  (one  opposite  each  vane),  through  either  of 
which  a  known  current  could  be  sent.  By  this  means  one 
could  readily  find  the  pressure  of  maximum  sensitiveness,  and 
compare  different  quartz  fibres  and  different  forms  of  the 
suspended  system.  The  best  pressure  appeared  to  be  between 
0*03  and  0'08  mm.,  and  within  this  region  the  variation  of 
sensitiveness  with  pressure  was  slow  ;  in  the  subsequent  work 
a  pressure  between  these  two  limits  was  usually  employed. 
From  a  knowledge  of  the  resistance  of  the  platinum  strips  and 
of  the  current  employed,  it  appeared  that  a  deflexion  of  one 
millimetre  (scale  distance,  196  cm.)  corresponded  to  an 
emission  of  about  0'0I  erg  per  second  from  each  square 
centimetre  of  platinum  surface.  The  deflexions  were  also 
found  to  be  proportional  to  the  energy  generated  in  the  strips. 
Of  course  the  radiometer  action  depends  primarily  on  the 
temperature  of  the  surface,  and  with  a  surface  of  different 
emissivity,  the  deflexion  for  a  given  emission  of  energy  will 
be  different. 
The  Rontgen  bulb  finally  employed  was  a  very  large  one 
made  by  Miiller  and  obtained  from  Isenthal  &  Co.  The 
diameter  of  its  spherical  portion  was  17  cm.  and  the  electrodes 
were  big  enough  to  bear  a  very  heavy  discharge.  The  anti- 
cathode  was  sealed  into  the  bottom  of  the  water-tube  so  that 
it  was  in  direct  contact  with  the  water,  and  I  have  frequently 
caused  the  water  in  the  tube  to  boil  without  seriously  heating 
the  rest  of  the  tube.  It  was  provided  with  an  automatic 
vacuum  adjuster  which  worked  well ;  the  focus  was  sharp 
and  the  Rontgen  rays  obtained  were  very  powerful  and  fairly 
steady  in  intensity  and  "  hardness/''  The  bulb  was  driven  by 
an  18-inch  Apps  coil  and  a  rotating  mercury-jet  interrupter. 
The  coil  was  in  an  adjoining  room  and  about  6  metres  from  the 
radiometer ;  its  orientation  was  adjusted  so  that  it  produced 
very  little  effect  upon  the  magnetic  u  torsion  head  *"  of  the 
instrument — less  than  4  mm.  with  the  largest  currents  used: 
this  was  always  in  one  direction  and  could  be  applied  as  a 
correction.  The  secondary  leads  were  gutta-percha  covered 
and  supported  by  silk  ribbons  ;  where  they  passed  from  one 
room  to  the  other,  through  the  wooden  frame  over  a  door,  they 
were  enclosed  in  long  glass  tubes.    As  the  rays  used  were  not 
