produced  by  Rontgen  Rays  in  Different  Metals.       309 
More  direct  experimental  evidence  upon  this  point  was 
obtained  by  reversing  the  position  of  the  ebonite  disk  with 
reference  to  the  case,  and  putting  the  vanes  on  the  side  of  the 
strips  next  the  window  instead  of  on  the  further  side.  The 
effects  observed  were  the  same  as  in  the  other  position. 
3.  The  possibility  of  unequal  values  of  the  emission  co- 
efficient for  the  aluminium-covered  surfaces  of  the  two  metals 
is  negatived  by  the  experiments  with  light  in  which  the  two 
metals  produced  equal  effects. 
I.  If  much  heat  were  lost  by  conduction  through  the 
ebonite  support,  or  the  copper  wire  by  which  the  strips  were 
earthed  (diameter,  1/4  mm.),  the  central  part  of  the  zinc 
strip  would  be  cooler  than  the  lead  on  account  of  its  greater 
conductivity  and  thickness.  If  this  were  the  cause  of  the 
higher  temperature  of  the  lead,  it  would  act  also  when  the  strips 
are  warmed  by  light,  which  is  not  the  case.  Experiments  with 
light  were  made  with  the  full  aperture  of  the  circular  window, 
so  that  the  ebonite  was  illuminated  as  well  as  the  strips  ;  and 
also  with  a  cardboard  screen  in  front  of  the  window  with  a 
rectangular  hole  which  allowed  the  light  to  fall  on  the  strips 
alone,  and  not  on  the  ebonite.  No  difference  was  observed  in 
the  two  cases. 
5.  The  observed  results  cannot  be  due  to  electrostatic  effects 
produced  by  the  negative  corpuscles  of  the  secondary  rays. 
The  strips  were  earthed,  and  the  two  vanes  were  connected 
by  a  conductor ;  under  these  circumstances,  no  electrostatic 
effect  could  be  produced  except  an  instability  due  to  both 
vanes  having  a  different  potential  from  the  two  strips.  Such 
an  instability  was  frequently  observed  after  the  vanes  had 
become  electrified  by  bumping  against  the  strips,  but  no 
other  electrical  effect.  Moreover,  the  coincidence  of  the 
rates  of  motion  of  the  radiometer  when  illuminated  by  Rontgen 
rays  and  by  ordinary  light,  shows  that  the  deflexions  pro- 
duced by  the  rays  were ,  caused  by  changes  of  temperature 
and  not  by  electrical  forces. 
6.  The  last  consideration  also  excludes  the  (somewhat 
remote)  possibility  that  the  effect  is  produced  by  direct 
impact  of  the  secondary  rays. 
7.  It  is  concewable  that  a  slight  difference  in  the  distance 
of  the  lead  and  zinc  strips  from  the  vanes  might  cause  a 
considerable  difference  in  sensitiveness  of  the  two  vanes,  and 
thus  account  for  the  preponderance  of  the  lead.  If  this  were 
the  case,  however,  the  effect  should  have  appeared  in  the 
experiments  with  ordinary  light.  I  also  tested  the  effect 
upon  the  sensitiveness  of  varying  the  vane  distance  by  taking- 
deflexions  with  the  zero  near  the  top  and  near  the  bottom  oi' 
