Am.  Jour.  PharnO 
February,  1896.  J 
Character  of  the  Urine. 
93 
For  a  reliable  sugar  test  in  urine  it  is  not  adapted,  as  this  secre- 
tion is  liable  to  reduce  it  when  glucose  is  absent.  In  chrysophanic 
urines  this  result  takes  place  readily,  the  supernatant  urine,  after  the 
precipitation  of  mercury,  possessing  a  red-brown  color. 
It  resembles,  in  this  respect,  diabetic  urine  ;  due  to  the  large 
excess  of  alkali  in  Sachse's  solution,  the  sugar  is  sometimes  cara- 
melized, and  the  mixture  will  have  a  brown  color. 
The  same  results  may  be  expected  in  Knapp's  alkaline  mercuric 
cyanide  solution. 
(6)  The  picric  acid  and  potash  method  of  Braun  and  Johnson  is 
also  liable  to  produce  a  dark  color,  as  with  glucose,  when  applied 
to  chrysophanic  urine,  due  to  the  action  of  the  alkali  upon  the 
acid. 
In  this  test,  however,  no  dark  red  color  is  produced,  unless  the 
other  reducing  constituents  of  urine  are  factors  also.  The  yellow 
picric  acid  evidently  retards  the  formation  of  the  deep  mahogany 
color  produced  in  the  case  of  glucose.  But  if  the  urine  be  rich  in 
chrysophanic  acid,  the  alkaline  mixture  will  also  turn  quite  dark  on 
boiling. 
Several  other  tests  are  open  to  the  same  objection,  namely,  the 
action  of  the  alkali  upon  chrysophanic  acid,  when  applied  to  such 
urine. 
This  may  be  the  case,  for  instance,  in  Penzoldt's  diazobenzolsul- 
phonic  acid  reaction  in  the  presence  of  strong  alkali,  the  purple 
color  produced  by  the  latter  in  chrysophanic  urine  being  liable  to 
be  mistaken  for  the  red-blue  tint  produced  in  diabetic  urine. 
Source  of  Error  in  Diabetic  Urine. — We  have,  so  far,  considered 
the  liability  of  chrysophanic  urine  to  be  a  source  of  error  in  exam- 
ining healthy  urine  for  small  amounts  of  glucose. 
But  it  can  also  interfere  in  diabetic  urine,  and,  in  some  cases, 
retard  the  accuracy  of  certain  tests. 
In  such  cases  it  must  be  supposed  that  rhubarb  or  similar  medi- 
cines have  been  administered  to  a  diabetic  patient,  or  one  slightly 
suffering  from  glycosuria. 
Some  of  the  recently  introduced  tests  for  glucose,  also  acting  in 
alkaline  media,  depend  upon  decolorization  of  the  test  liquid,  the 
glucose  present  acting  as  reducing  agent.  One  of  these  tests  is 
Crissler's  reaction  with  safranine.  This  is,  for  ordinary  purposes,  an 
excellent  reaction,  where  an  accurate  quantitative  estimation  of 
