high,  the  u.\yh;icnu>niol)ni  Ijands  lyiut,r  between  the  solar  lines  1)  and 
b  were  separate  and  of  suitable  depth  of  shade.  IJy  the  aid  of  a 
haeniocrit  a  .solution  of  haemoglobin,  containing  o-2  per  cent,  b)' 
volume  of  healthy  human  red  cells,  was  next  prepared.  A  glass  cell 
one  or  two  centimetres  high  was  then  fdled  with  this  solution,  while 
the  diluted  urine  was  placed  in  a  second  cell,  the  height  of  which 
could  be  varied.  The  haemoglobin  bands  of  the  second  cell  were 
then,  by  varying  the  height  of  the  column  employed,  made  to  match 
those  ol  the  first,  for  which  purpose  a  Zeiss  comparison  spectroscope 
was  employed. 
fi)'  this  method  it  is  possible  to  estimate  percentages  of  haemo¬ 
globin  in  cases  m  which  colouring  matter  is  present  in  the  urine  in 
amount  sufficent  to  render  haeinoglobinometer  determination  difficult 
or  impossible.  This  method  may,  of  course,  be  used  when  the 
pieceding  method  is  also  available.  The  results  obtained  by  the  two 
methods  correspond  closely.  The  first  method  is  more  rajhdly  carried 
out,  and  is,  therefore,  when  available,  preferable  to  the  second. 
Me/hod  3.  ^  I  he  estimation  w'as  made  by  boiling  the  solution  of 
haemoglobin,  fen  c.cm.  of  the  centrifugali.sed  urine,  which  did  not  as 
a  rule  require  filtering,  were  rendered  fairly  acid  with  acetic  acid, 
boiled  and  after  cooling  transferred  to  a  graduated  centrifugal  tube.' 
Centrifugahsation  was  continued  until  the  bulk  of  the  precipitate 
became  constant.  Its  volume  was  then  measured.  Previously  by 
means  of  the  haemocrit,  a  number  of  specimens  of  urine  (with  the 
addition  of  5  per  cent,  to  25  per  cent,  of  water)  were  iirepared 
containing  known  percentages,  by  volume,  of  healthy  red  blood  cells 
(in  the  mqist  condition),  which  were  subsequently  laked.  To  these 
solutions  an  equal  volume  of  urine  was  added,  and  the  solutions 
lendered  very  slightly  acid  with  acetic  acid  and  then  boiled.  The 
bulk  of  the  centrifugalised  precipitates  was  next  determined  as  above 
desejibed,  and  from  this  a  table  constructed,  which  permitted  the 
piecipitates  obtained  in  haemoglobinuria  to  be  c.xpressed  in 
i:ieicentagc  form.  As  this  method  depends  upon  the  precipitation 
of  globulin,  and  is  not  affected  to  any  considerable  extent  by  the 
change  undergone  by  haemoglobin  in  the  urine,  it  may  therefore  be 
eiiqiloyed  to  determine  with  fair  approximation  the  total  amount  of 
haemoglobin  originally  present  in  the  urine.  The  method  ’is, 
however,  only  approximate,  since  the  volume  of  the  precipitate  varies 
