^MJSSm*}        Methods  of  Water  Analysis.  107 
same  certain  basis  as  the  detection  of  poisons.  With  the  knowledge 
that  typhoid  fever  is  usually  caused  by  the  drinking  water  and  after 
the  discovery  by  Koch  that  cholera  is  of  similar  origin,  it  was  ex- 
pected that  the  typhoid  bacilli  and  the  cholera  spirilla  could  be 
detected  in  the  suspected  water.  Unfortunately,  disappointment 
followed  all  attempts  in  this  direction.  It  soon  became  evident  that 
while  a  certain  water  has  been  the  cause  of  either  a  cholera  or 
typhoid  epidemic,  as  established  by  all  evidence  at  hand,  neither  the 
cholera  spirillum  or  the  typhoid  bacillus  could  be  detected  in  such 
waters.  The  cause  for  this  failure  was  found  in  the  great  predomi- 
nance of  water  bacteria  which  overgrow  and  obscure  the  few  specific 
parasites,  rendering  their  discovery  impossible.  The  effort  may  be 
compared  to  looking  for  a  needle  in  a  haystack.  While  not  entirely 
abandoned,  the  search  for  specific  microorganisms  has  not  been 
made  the  object  of  routine  examinations ;  and  until  some  satisfactory 
method  is  devised  by  which  the  saprophytic  bacteria  may  be  entirely 
eliminated  and  the  number  of  the  specific  microorganisms  increased 
so  as  to  have  them  present  in  very  small  quantities  of  the  water,  the 
bacteriologist  must  depend  upon  other  data  upon  which  a  conclusion 
as  to  the  quality  of  the  water  may  be  reasonably  based.  It  was 
thought  for  a  time  that  the  number  of  bacteria  in  the  water  could 
serve  as  an  index  of  pollution,  and  a  number  of  standards  of  bacte- 
rial purity  have  been  suggested  by  various  authorities.  Thus,  Koch 
considers  100  bacteria  per  cubic  centimetre  as  the  safe  limit  for 
drinking  water;  Miquel  raises  the  standard  to  1,000;  Crookshank 
agrees  with  this  standard,  while  Mace  and  Migula  claim  that  250  to 
500  bacteria  is  the  highest  limit  for  a  good  drinking  water.  These 
or  any  other  arbitrary  standards  based  on  mere  number  of  bacteria  > 
are  as  fallacious  as  the  "  standards  "  proposed  from  time  to  time  for 
ammonias,  nitrites,  nitrates,  etc.  In  the  first  place,  the  number  of 
bacteria  in  water  will  vary  greatly  with  the  medium,  the  reaction  of 
the  medium,  the  length  of  time  the  colonies, are  allowed  to  develop, 
dilution,  etc.,  as  may  be  seen  by  the  following  data: 
(1)  Time  of  Plating. — It  makes  considerable  difference  whether 
the  water  is  plated  immediately  upon  collection  or  is  allowed  to  stand 
for  some  time  before  plating.  At  room  temperature,  the  bacteria 
multiply  enormously,  so  that  if  the  plating  is  done  several  hours 
after  collection  of  the  sample,  an  originally  pure  water  may  be  con- 
demned on  the  bacterial  count.    On  the  other  hand,  if  packed  in 
