348 
Nitrogen  as  Nitric  Acid,  in  the  Soils 
We  now  turn  to  the  determinations  of  nitric  acid  in  the  soils 
of  the  various  plots  in  Broadbalk  Field.  We  must,  however,  at 
starting  beg  the  reader  to  remember  the  considerable  difficulty 
of  the  task  we  have  attempted.  The  mean  composition  of  the 
soil  of  a plot  352  yards  long  has  to  be  calculated  from  samples 
taken  from  six  holes  equally  distributed  in  its  length ; that  the 
mixture  prepared  from  such  samples  should  sometimes  fail  to 
represent  accurately  the  average  composition  of  the  soil  is 
surely  to  be  expected.  In  two  instances  in  Broadbalk  Field, 
and  one  in  Hoos  Field,  the  surface  soil  from  individual  holes 
has  contained  so  high  a proportion  of  nitric  acid  and  chlorine, 
as  to  lead  to  the  supposition  that  the  sample  had  been  taken 
from  a spot  contaminated  by  the  droppings  of  horses  when 
ploughing ; or  where  possibly  a bag  of  manure  had  been  set 
down  and  emptied  during  sowing.  In  these  cases  the  results 
given  have  been  obtained  from  a mixture  of  five  holes,  or,  in 
the  case  of  Hoos  Field,  of  three  holes  only.  There  is  also 
always  some  difficulty  in  preserving  a large  number  of  soil- 
samples,  in  this  case  160,  entirely  free  from  accidental  contami- 
nation before  they  are  put  into  bottles ; and  in  one  instance, 
that  of  the  third  depth  of  the  unmanured  Plot  16,  it  has  been 
necessary  to  exclude  the  results  of  the  analysis  from  the  table,  the 
nitric  acid  found  being  far  larger  than  seemed  possibly  consistent 
with  the  truth.  A small  amount  of  error  will,  indeed,  distinctly 
affect  the  result ; the  addition  of  one  grain  of  nitric  nitrogen 
to  1144  lbs.  of  soil  would,  in  fact,  make  a difference  of  1 lb.  per 
acre  in  the  quantity  calculated  for  a depth  of  27  inches.  It  is 
clear,  therefore,  that  it  will  be  unsafe  to  found  conclusions  on 
small  differences  between  different  plots ; we  must  rather  aim  to 
seize  the  main  features  of  the  results. 
In  calculating  the  results  of  the  analyses  into  pounds  per 
acre,  it  has  been  assumed  that  the  weight  of  soil  per  acre  in 
the  first  9 inches  is  2,552,203  lbs. ; in  the  second  9 inches, 
2j706,573  lbs. ; and  in  the  third  9 inches,  2,750,601  lbs.,  these 
weights  being  the  mean  of  all  the  samples  taken  on  the 
present  occasion,  excluding  only  the  surface  soil  of  Plot  2.  The 
analysis  of  this  soil  (receiving  farmyard-manure)  is  reckoned  on 
2,456,509  lbs.,  the  weight  calculated  from  its  own  samples. 
It  is  of  interest  to  note,  that  even  among  the  artificially 
manured  plots  there  is  an  obvious  tendency  to  give  lower  weights 
of  surface  soil  the  heavier  the  crops,  that  is,  the  greater  the  root 
development,  and  the  amount  of  crop-residue.  Hence  the 
calculation  of  the  analyses  relating  to  the  soils  of  such  plots  on 
the  rather  higher  average  weights,  will  so  far  give  results  some- 
what in  excess  ; whilst  those  relating  to  the  poorer  plots  will  be 
proportionally  low. 
