496     Prof.  Wilson  and  Mr.  Gold  :  Electrical  Conductivity  of 
Capacity. 
Self- 
induction. 
Alternations 
per  second. 
d2. 
1      1 
d2     dx' 
33000  cms. 
136500  cms. 
7-14  xlO4 
0-4 
24 
6400    , 
1-62  xlO5    ! 
04 
24 
1600    „ 
., 
3-24  xlO5    i 
034 
2-6 
150    „ 
., 
10-8   xlO5 
030 
3-0 
1.30    „ 
4100  cms. 
62  xlO0    ! 
0-30 
3-0 
It  will  be    seen  that  the  value  of  -=- T  varies   as   the 
d2      a1 
number  of  alternations  per  second  is  changed.  Unfor- 
tunately it  was  not  found  possible  to  obtain  the  variation  of 
d2  with  the  frequency  very  exactly:  and  all  that  can  be 
said   is    that   changing    the    frequency    from    7*14  XlO4   to 
6*2  XlO6  per   second  does  not  change  y —  —  by  more  than 
25  per  cent,  of  its  mean  value. 
If  we  suppose  that T  varies  as  nx,  where  n  is    the 
d2      a  i 
frequency,    then    the    results  just    given    show   that    x   lies 
between  +O05  and  —0*05. 
For  a  pure  capacity  #=0,  and  for  a  pure  self-induction 
#=  —  1;  so  that  it  appears  that  the  flame  behaves  nearly 
like  a  pure  capacity.  That  it  does  not  include  much  con- 
ductivity in  parallel  with  the  capacity  is  shown  by  the 
fact  that  it  could  be  nearly  balanced  by  three  condensers. 
In  the  section  below,  on  the  theory  of  conductivity,  it  is 
shown  that  these  conclusions  from  the  experiments  might 
have  been  anticipated. 
To  obtain  the  apparent  specific  inductive  capacity  of  a  salt 
vapour,  it  is   only  necessary  to   multiply  the  numbers   for 
1  ~ ~T  given  above  by  2d1=6'66  and  add  on  unity.     The 
apparent  specific  inductive  capacities  so  obtained  vary  from 
about  100  for  the  strongest  solution  of  K2C03  sprayed  to 
about  4  for  the  solution  of  LiCl  *. 
Some  experiments  were  made  to  find  out  how  the  con- 
ductivity varied  with  the  distance  between  the  electrodes. 
*  These  values  of  the  apparent  specific  inductive  capacity  have  of 
course  no  relation  to  the  true  specific  inductive  capacity,  which  must 
be  nearly  unit  v. 
