24  Some  Secondarj/  Actions  of  Manures  upon  the  Soil. 
a plot  receiving  nitrate  of  soda  to  compare  with  one  receiving 
ammonium  salts,  while  the  other  treatment  was  identical  on 
both  plots. 
Table  III. — Mechanical  Analyses  of  Rothamsted  Soils  with 
and  without  Sodium  Nitrate: 
Percentages  after  ignition. 
Mean  of  5 plots 
No 
sodium  nitrate 
With 
sodium  nitrate 
Fine  gi’avel  ........ 
2-2 
21 
Coarse  sand  ........ 
61 
6-7 
Fine  sand  ........ 
18-8 
18-8 
Silt 
29-5 
29-9 
Fine  silt  ........ 
14-0 
13-9 
Clay 
17-9 
1.5-0 
It  will  be  seen  that  the  percentage  of  clay  is  distinctly  less 
on  the  plots  which  had  received  nitrate  of  soda,  and  though  the 
difference  may  not  appear  to  be  great  it  is  without  doubt  a 
real  one,  because  it  was  found  to  exist  in  each  pair  of  soils 
used  in  the  comparison  (for  the  actual  figures  see  Trans.  Chem. 
Soc.  1904,  5<5,  964),  and  it  is  contrary  to  what  would  have  been 
expected  from  the  behaviour  of  the  soils.  Of  the  reality  of 
the  differences  we  have  moreover  another  indication  in  the 
fact  that  when  the  tile  drains,  which  are  laid  beneath  the  whole 
length  of  the  narrow  strips  constituting  the  plots  on  the 
Broadbalk  wheat  field,  begin  to  run,  the  water  flowing  from  the 
drains  beneath  the  nitrate  plots  is  always  faintly  turbid  and 
carries  a very  light  cloud  of  fine  mud,  whereas  the  water  from 
the  plots  receiving  ammonium  salts  is  always  crystal  clear. 
Evidently  the  washing  out  of  the  finest  clay  particles  which 
we  see  going  on  in  the  drainage  water  has  been  so  continuous 
that  the  quantity  remaining  in  the  soil  has  been  definitely  re- 
duced by  the  one  to  five  per  cent,  shown  in  the  different  analyses. 
It  is  also  evidence  in  the  same  direction  to  find  that  in 
the  earlier  years  of  the  experiment  the  drain  beneath  the 
unmanured  plot  ran  more  frequently  than  that  below  the 
nitrate  of  soda  plot,  whereas  of  late  years»  since  so  much  of 
the  finest  stuff  has  been  washed  out,  the  drains  have  been 
running  more  frequently  beneath  the  nitrate  plots. 
The  turbid  aspect  of  the  drainage  water  from  the  two  sets 
of  plots  suggested  another  experiment,  which  provided  the 
clue  to  the  different  texture  of  the  two  plots.  Small  equal 
portions  of  soil  from  the  plots  were  weighed  out,  and  each  was 
shaken  up  with  a large  bulk  of  pure  water  ; the  resulting 
muddy  liquids  were  then  put  to  stand  separately  in  similar 
tall  jars.  A certain  time  is  occupied  before  the  suspended  soil 
