530        The  Electrical  Conductivity  of  Metallic  Oxides. 
increases  rapidly  with  the  temperature.  Moreover,  in  the 
present  paper  it  has  been  shown  that  the  conductivity  of  the 
oxide  increases  in  the  same  manner  as  the  rate  of  emission  of 
corpuscles  from  its  surface. 
The  fact  that  mixtures  of  certain  oxides  conduct  better 
than  either  of  their  constituents,  is  capable  of  explanation  on 
the  corpuscular  theory  of  metallic  conduction.  For  there  is 
ample  proof  that  the  ease  with  which  a  corpuscle  can  escape 
from  an  atom  depends  on  its  surroundings.  In  an  atom 
placed  by  itself  and  left  to  its  own  resources,  a  corpuscle  has 
nothing  to  help  it  to  escape  except  its  own  velocity;  but  if  it 
is  surrounded  by  other  atoms,  the  other  corpuscles  exert  forces 
on  it  and  so  help  it  to  escape.  A  corpuscle,  for  instance, 
might  escape  from  an  atom  in  the  solid  or  liquid  state  when 
it  could  not  escape  in  the  gaseous  state.  Liquid  mercury  is 
a  good  conductor  of  electricity,  and  if  there  are  the  same 
number  of  atoms  in  mercury  vapour  as  in  the  liquid  we  should 
expect  a  much  higher  conductivity  in  the  vapour  on  account 
of  the  larger  free  path  of  the  corpuscles  ;  but  neither  Maxwell 
nor  Hittorf  could  find  any  trace  of  conductivity  through  hot 
mercury  vapour.  Professor  Thomson  *  found  that  at  high 
temperatures  mercury  vapour  was  a  better  insulator  than  air 
under  similar  conditions  ;  while  Strutt  f  has  shown  that  up  to 
a  full  red  heat  the  conductivity  of  saturated  mercury  vapour 
is  only  about  one  ten -millionth  part  of  that  of  the  liquid.  It 
seems  not  improbable,  therefore,  that  there  might,  in  certain 
cases,  be  many  more  free  corpuscles  in  a  mixture  of  two 
oxides  than  in  either  oxide  separately.  Even  neglecting  this 
influence  of  the  surroundings  on  the  freedom  of  the  corpuscles, 
wTe  should  expect  to  obtain  a  mixture  of  better  conductivity 
than  either  of  its  constituents  by  mixing  two  metallically 
conducting  bodies,  one  containing  a  large  number  of  free 
corpuscles,  but  in  which  the  mobility  of  the  corpuscles  was 
small,  and  the  other  in  which  the  mobility  of  the  corpuscles 
was  very  great,  but  containing  only  a  few.  On  mixing  a 
little  of  the  first  substance  with  the  second  a  good  conducting 
mixture  would  be  obtained. 
I  do  not,  therefore,  think  that  ^sernst's  reasons  for  consi- 
dering the  conductivity  of  heated  metallic  oxides  to  be  electro- 
lytic are  conclusive.  The  experiments  described  in  this  paper 
seem  to  show  that  by  far  the  greater  part  of  the  current  is 
carried  by  negatively  charged  corpuscles,  as  in  metals.  In 
the   case  of  those  oxides  or  mixtures   of  oxides  in  which  a 
*  J.  J.  Thomson,  Phil.  Mao-.  [5]  xxix.  p.  364. 
t  Hon.  R.  J.  Strutt,  Phil.  Mag.  [6]  iv.  p.  596  (1902). 
