Conductivity  of  Metallic  Oxides.  513 
those  of  6~l  we  ought  to  obtain  a  straight  line.  In  diagram  II. 
the  ordinates  are  the  values  of  log10  c  — -J  log10  0,  the  abscissae 
being  the  corresponding  values  of  0~1. 
It  will  be  seen  that  seven  points  representing  the  largest 
values  of  6~l  (i.  e.  the  lowest  temperatures)  lie  very  nearly  on 
a  straight  line.  The  fact  that  the  points  above  1300°  C.  do 
not  lie  on  this  line  is  probably  due  to  disturbing  causes  which 
affect  the  resistance  to  be  measured.  At  these  high  tempera- 
tures, the  lime  would  probably  have  some  chemical  action  on 
the  platinum  electrodes,  causing,  possibly,  a  decrease  in  the 
contact  resistances.  It  was,  for  instance,  always  found  that 
the  lime  was  fixed  to  the  electrodes  after  it  had  been  strongly 
heated,  but  not  when  lower  temperatures  were  used.  This 
explanation  would  account  for  the  fact  that  the  value  found 
for  the  conductivity  at  high  temperatures  was  greater  than 
it  should  be  on  the  present  theory.  Diagram  II.  may 
therefore  be  taken  as  showing  that  the  increase  of  con- 
ductivity of  lime  with  temperature  follows  the  same  law  as 
the  emission  of  corpuscles  from  its  surface. 
When  the  above  series  of  experiments  had  been  completed, 
tests  were  made  to  see  if  any  polarization  could  be  detected 
which  would  show  that  some  part  of  the  conduction  through 
lime  was  electrolytic.  The  lime  was  connected  to  a  battery 
for  a  few  seconds  and  a  current  allowed  to  pass  through  it. 
The  battery  was  then  cut  out  and  a  delicate  d'Arsonval 
galvanometer  substituted.  This  indicated  a  small  polarization 
current  in  the  opposite  direction  to  the  original  current  and 
dying  away  logarithmically  with  the  time.  The  polarization 
current  when  the  lime  was  at  800°  C.  was  very  small,  but  it 
increased  as  the  temperature  was  raised,  the  increase  being 
no  doubt  due  to  the  decrease  in  the  resistance  of  the  lime.  A 
battery  of  100  volts  was  next  connected  up  in  series  with  the 
slab  of  lime  and  galvanometer,  the  latter  being  shunted  for 
measuring  larger  currents.  The  current  through  the  galvano- 
meter gradually  decreased,  and,  if  the  E.M.F.  was  left  on  for 
a  few  seconds  and  the  key  connected  to  the  lime  reversed,  a 
larger  deflexion  in  the  opposite  direction  was  obtained  which 
also  decreased  with  time.  But  if  the  current  was  allowed  to 
pass  for  several  minutes  in  one  direction  through  the  lime, 
the  galvanometer  deflexion  decreased  to  about  a  tenth  of  its 
original  value,  and,  on  reversing  the  key,  the  deflexion 
obtained  in  the  opposite  direction  was  only  slightly  greater 
than  the  small  value  just  before  reversing.  This  increased 
with  time  slowly  at  first,  but  soon  very  rapidly,  and  then 
more  slowly  again  as  a  maximum  deflexion  was  reached. 
the     current    then    decreased    with    time,    at    a    gradually 
