125, 
of the results of our experience, so the following tentative 
standards are given for tropical countries generally, and India 
in particular. But it must be distinctly understood that 
they are not hard-and-fast rules. 
Let us first dispose of filtered samples. All samples of 
water that passes through filters should be free from organisms 
generally. From a study of filters in India, it is perfectly 
possible to get a water which should contain no lactose fer- 
menting fecal organism in 100 c.c. and only about 50 to 
roo colonies on agar. Consequently any divergence from 
this rule should be carefully investigated. The commonest 
causes of failure of filters in this country are :— 
(1) A very highly polluted water being put into the 
filters direct without previous purification ; 
(2) Unsuitable nature of the sand ; 
(3) Too frequent scraping ; 
(4) Wrong design of the filters themselves, and bad 
manipulation ; and 
(5) Presence of frogs and fish in the clean water reservoir. 
"When any sample of filtered water is unsatisfactory, 
that is to say, shows fecal organism of any kind, present in 
large numbers, careful investigation should be made as to 
which of these causes is the one operating in any given 
installation. 
RAW WATERS—LAKE WATERS. 
A good lake water should contain less than 100 colonies 
per c.c. (on agar at 37°C.). 
No lactose fermenters in 50 C.c. 
No lactose fermenting organism of class 1. 
Glucose fermenting organisms may be present in I or 5 C.c. 
A fairy or usable lake water should not contain more 
than 400 organisms (on agar at 37°C.). 
Lactose fermenters should not be present in less than I c.c. 
Glucose fermenters may be present in ‘I c.c. 
No organism of class 1 should be present in ‘20 c.c. 
Lactis e#rogenes should very largely ‘predominate. 
If a sample is taken from the bottom of a lake it should 
contain mostly bacillus cloace. 
