160  Nitrates  and  Ammonia  in  Water.  {^'l^'Jm^ 
ard  solution  added  and  calculate  to  entire  quantity  of  solution  if 
an  aliquot  portion  was  taken;  the  number  of  cubic  centimetres  of 
standard  solution  (required  for  the  10  c.c.  of  water  started  with) 
X  0*14  will  give  the  parts  of  nitrogen  as  nitrates  per  million  parts 
of  water. 
AMMONIA  DETERMINATIONS. 
The  importance  attached  particularly  to  the  presence  of  albumi- 
noid ammonia,  exceedingly  small  quantities  of  which  suffice  to  con- 
demn a  sample  of  water,  and  the  difficulty  of  getting  the  reagents 
perfectly  free  from  ammonia  makes  this  the  most  difficult  determi- 
nation in  a  water  analysis.  For  some  years  past  the  writer  has 
carried  out  these  determinations  in  a  manner  which  insures  the 
greatest  possible  accuracy  with  the  least  amount  of  extra  manipu- 
lation. 
The  alkaline  potassium  permanganate  solution  is  made  by  simply 
dissolving  the  chemicals  in  distilled  water;  it  is  useless  to  attempt 
freeing  this  solution  from  ammonia  as  it  cannot  be  kept  so. 
In  making  an  analysis,  the  retort  and  condenser  are  rinsed  with 
water,  and  the  retort  charged  with  30  c.c.  of  the  alkaline  perman- 
ganate solution,  and  200  c.c.  clear  river  or  distilled  water ;  to  avoid 
dangerous  bumping  of  the  boiling  liquid  some  copper  turnings  are 
placed  in  the  retort.  The  distillate  is  collected  in  portions  of  10  c.c, 
and  these  set  aside  in  regular  order,  and  distillation  continued  until 
about  60  c.c.  are  left  in  the  retort.  (The  operation  so  far  is  for  the 
purpose  of  freeing  apparatus  and  reagents  from  ammonia,  and 
enables  one  in  most  cases  to  secure  sufficient  ammonia-free  water 
to  make  up  the  solutions  for  comparisons.  The  first  six  portions 
of  the  distillate  will,  as  a  rule,  contain  all  of  the  ammonia ;  to  each 
of  these  is  added  0  4  c.c.  Nessler's  reagent ;  should  the  last  one 
still  indicate  the  presence  of  ammonia,  more  of  the  distillates  in 
their  regular  order  will  have  to  be  tested).  100  c.c.  of  the  water  to 
be  tested  are  now  added  through  the  tubulure,  and  100  c.c.  distil- 
late collected  in  portions  of  10  c.c.  and  set  aside  in  regular  order; 
in  nesslerizing  these  it  must  be  remembered  that  if  a  precipitate  is 
produced  it  will  be  impossible  to  make  a  colorimetric  comparison; 
it  is  therefore  a  good  plan  to  first  nesslerize  the  fourth  portion,  then 
the  third  and  second.  Should  any  of  these  yield  a  very  deep  color 
it  will  enable  one  to  dilute  the  first  distillate  with  one  or  more  of 
