126 
Pedigree  Seed  Corn. 
Needless  to  say  the  constant  external  conditions  above 
postulated  are  never  actually  maintained,  and  therefore  the 
conditions  of  different  years  are  favourable  to  different  races. 
Notwithstanding  this  a consideration  of  such  purely  hypo- 
thetical cases  as  the  above  leads  to  the  conclusion  that,  whilst 
in  some  cases  quite  rapid  changes  in  the  racial  character  of 
aggregates  occur,  in  other  cases  fairly  uniform  varieties  may 
have  arisen  in  course  of  time  from  mixed  types. 
For  instance,  Chevallier  barley  as  we  have  it  now  is 
probably  all  descended  from  the  “few  ears”  (to  use  Chevallier’s 
own  words)  from  which  the  variety  was  originally  raised  in 
1823.  We  know  nothing  about  the  ancestry  of  these  “ few  ears,” 
but  it  seems  extremely  probable  that  the  present  uniformity 
of  all  the  differently  named  Chevalliers  now  in  cultivation  is 
largely  a consequence  of  natural  selection.  There  have  been 
from  time  to  time  many  reselections,  but  the  progeny  of  these 
are  all  very  much  alike  in  racial  characters.  It  is  practically 
impossible  to  distinguish  these  different  selections  when  they 
are  growing  together,  and  the  prolonged  series  of  Danish 
experiments  to  be  referred  to  later  failed  to  show  any  appre- 
ciable difference  between  several  of  them  in  either  yield  or 
quality  of  grain. 
This  is  not,  of  course,  to  say  that  repeated  “ pedigreeing  ” 
of  a good  variety  is  not  desirable.  It  is  indeed  obviously 
necessary  if  the  racial  character  of  the  grain  is  to  be  main- 
tained against  the  operation  of  all  the  factors  which  make  for 
divergence  from  the  original  type. 
It  has  been  fairly  well  proved  by  Johannsen  that  main- 
tenance of  established  characters  is  all  that  can  be  effected  by 
reselection  within  an  absolutely  pure  race  of  self -fertilising 
cereals,  and  that  increased  productiveness  is  not  obtainable 
by  selecting  the  most  productive  individuals.  This  is,  of 
course,  a statement  in  a special  case  of  the  application  of 
the  doctrine  which  has  now  been  very  generally  accepted, 
that  characters  acquired  by  the  individual  from  effects  of 
environment  only  are  not  inherited.  This  doctrine  as  a 
general  formula  is  still  in  dispute,  but  such  knowledge  as  we 
have  is  in  favour  of  its  application  in  the  case  of  self -fertilising 
cereals. 
The  best  piece  of  evidence  in  this  connection  is  perhaps 
afforded  by  the  history  of  Major  Hallett’s  Pedigree  Races. 
In  the  case  of  his  pedigree  barley  the  most  prolific  plant 
has  been  selected  year  after  year  for  a great  number  of  years, 
and  forms  the  starting  point  of  a “pure  race.”  The  very 
desirable  effect  has  been  produced  of  maintaining  absolute 
uniformity,  and  there  is  no  evidence  so  far  as  I am  aware  that 
the  progeny  of  the  later  selections  differs  from  that  of  the 
