Recombination  of  Ions  in  Air  and  other  Gases.         481 
likely  to  lose  an  electron  as  if  it  were  not  associated  with  other 
atoms.  Perhaps,  therefore,  the  molecule  as  a  whole  loses  two 
or  three  electrons,  and  its  electric  field  is  the  more  intense. 
Recombination  of  this  kind  must  also  be  easier,  the  shorter 
the  free  path. 
It  will  be  clear  from  the  foregoing  that  certain  effects 
occur  which  are  characteristic  of  a  process  of  initial  recom- 
bination, a  process  which  is  a  priori  not  improbable.  The 
question  now  arises  as  to  whether  any  other  cause  could 
produce  the  same  effects. 
When  we  consider  the  great  increase  of  current  in  a 
complex  gas  which  is  caused  by  an  increase  in  the  electric 
force  applied,  we  cannot  but  ask  whether  any  of  it  is  due  to 
the  production  of  other  ions  by  those  actually  made  by  the  a 
particle.  Could  the  electric  force  aid  the  freed  electron  to 
gather  speed  sufficient  to  ionize  by  collision  ?  A  process  of 
this  kind  would  be  similar  in  its  results  to  initial  recom- 
bination, in  that  it  would  be  independent  of  strength  of  current 
and  depth  of  ionization  chamber.  It  seems  probable,  how- 
ever, that  its  effectiveness  would  rather  be  increased  than 
diminished  by  lowering  the  pressure  ;  and  also  it  would  be 
difficult  to  account  for  the  existence  of  a  saturation  value  of 
the  current.  Nevertheless,  it  does  not  seem  safe  as  yet  to  say 
that  no  such  process  occurs.  Probably  further  light  could 
be  thrown  on  the  subject  by  an  investigation  into  the  total 
number  of  ions  produced  in  different  gases  under  varying 
conditions.  Some  initial  experiments  of  this  kind  will  be 
described  presently. 
Rutherford  has  recently  shown  that  the  a  particle  of  Ra  C 
has  only  lost  40  per  cent,  of  its  velocity  when  it  censes  to 
ionize.  If  this  fact  is  considered  in  conjunction  wTith  our 
investigations  into  the  form  of  the  ionization  curves  for 
gaseous  mixtures,  it  is  at  once  clear  that  the  a  particle  stops 
ionizing  in  every  gas  when  its  speed  has  fallen  to  precisely 
the  same  value.  For,  if  not,  the  ionization  curve  for  a 
mixture  would  show  a  superposition  of  simple  curves,  of 
which  effect  there  is  no  trace.  This  and  other  considerations 
seem  to  show,  as  we  have  already  said  ( Philosophical 
Magazine,  September  1905),  that  the  a  particle  performs 
the  same  number  of  acts  of  ionization  in  every  gas.  If,  then, 
we  find  the  total  saturation  current  to  be  different  in  different 
gases,  we  must  come  to  the  conclusion  that  either  the  ions  in 
the  gases  of  higher  conductivity  produce  others  by  the  help 
of  the  electric  field,  or  that  in  the  gases  of  lower  conductivity 
some  of  the  ions  made  by  the  a  particle  do  not  get  tree,  ev<  n 
under  conditions   of  saturation,  from   their   parent  atoms,  or 
