72  Filtration  of  Drinking  Water.  {AFebJruYr'y??mn, 
million,  the  filter  should  not  require  scraping  more  than  thirteen  to 
fifteen  times  per  year,  but  in  times  of  freshet  I  have  seen  the  experi- 
mental filters  shut  down  and  scraped  once  a  week.  This  condition 
is  exceptional,  however.  After  scraping,  the  bacterial  counts  are 
high,  so  the  water  is  run  to  waste  for  a  period.  My  observation  on 
this  point  was  that  forty-eight  hours  usually  elapsed  before  the  counts 
returned  to  the  normal.  The  depth  of  water  covering  the  filter  is 
about  four  feet ;  this  head  forces  the  water  through  the  sand.  As  the 
deposits  accumulate  there  is  diminished  flow,  gradually  decreasing 
until  the  pressure  is  insufficient  to  deliver  the  3,000,000-gallon  rate, 
then  the  filter  is  shut  down,  drained  and  scraped. 
The  operation  of  scraping  the  filter  consists  of  removing  about 
3^  inch  in  depth  of  sand  from  the  surface,  together  with  the 
deposit  of  mud,  etc.,  after  which  the  surface  is  raked  even  and  filtra- 
tion proceeds  as  before.  At  the  end  of  a  year  the  total  amount  of 
scrapings  is  washed,  loss  made  up,  and  returned  to  the  filter  in  one 
operation. 
In  order  to  be  sure  that  there  are  neither  holes  nor  ways  through 
the  filter,  an  ingenious  procedure  is  adopted  as  follows :  Large 
quantities  of  a  culture  of  Bacillus  Prodigiosus  are  applied  to  the 
surface  of  the  filter  at  regular  intervals  of  half  an  hour  each  for  a 
period  of  twenty-four  hours.  Millions  of  these  organisms  are  intro- 
duced at  each  application.  Test  samples  of  the  effluent  are  taken 
every  fifteen  minutes  during  the  time  of  the  trial.  Plates  are  made 
of  each  sample  and  counted.  The  Bacillus  Prodigiosus  grows  best 
upon  agar-agar  producing  a  bright  red  colony  distinguished  at  a 
glance  from  the  ordinary  water  bacteria.  If  there  are  any  ways  or 
openings  in  the  filter  the  bacilli  are  sure  to  be  found  in  the  effluent. 
When  the  test  was  applied  to  this  city's  experimental  filters  we 
found  but  two  plates  showing  red  colonies  out  of  several  hundred 
trials,  and  it  is  just  possible  that  these  were  accidental  contamina- 
tions, proving  conclusively  the  excellent  construction  of  those  beds. 
In  order  to  have  the  results  of  all  bacteriological  investigations 
comparable,  it  is  absolutely  necessary  that  the  methods  of  manipu- 
lations be  uniform.  One  most  important  stride  in  this  direction  was 
made  by  Fuller  when  he  demonstrated  that  the  medium  most  suit- 
able for  cultivation  of  bacteria  for  water  work  is  that  composed  of 
gelatin  10  per  cent.,  peptone  1  per  cent.,  salt  y2  per  cent.,  dissolved 
in  meat  infusion  representing  the  soluble  portion  of  500  grammes  of 
