304 WATER HYGIENE 
the examination of a sample of entirely unknown quantity. The quan- 
titative value of a series of tests 1s lost, unless all or, at least, a large 
portion of the smallest dilutions tested have given negative results. 
In reporting a single test, 1t is preferable merely to record results as 
observed, indicating the amounts tested and the result in each, rather 
than to attempt expression of the result in numbers of B. cols per cubic 
centimeter. In summarizing the results of a series of tests, however, 
it is desirable for the sake of simplicity to express the results in terms 
of the numbers of B. cold per cubic centimeter or per 100 c.c. To con- 
vert the result of fermentation tests to this form, the result of each test 
is recorded as indicating a number of B. colt per cubic centimeter equal 
to the reciprocal of the smallest decimal or multiple fraction of a cubic 
centimeter giving a positive result. For example the result: 10 c.c. 
plus; 1 c.c. plus; 0.1 ¢.c. minus would be recorded as indicating one 
B. cole per zubic centimeter. An exception should be made in the case 
where a negative result is obtained in an amount larger than the smallest 
portion giving a positive result; for example, in a result such as 10 c.c. 
plus; 1 ¢.c. minus; 0.1 ¢.c. plus. In such a case, the result should be 
recorded as indicating a number of B. cols per cubic centimeter equal to 
the reciprocal of the dilution next larger than the smallest one giving a 
positive test, this being a more probable result. 
Where tests are made in amounts larger than 1 c.c. giving an average 
of results less than one B. cols per cubic centimeter, it 1s convenient to 
express results in terms of the number of B. col per 100 c.c. 
The following table illustrates the method of recording and aver- 
aging results of B. colz tests 
Result of Tests in Amounts Designated Indicated No of B coh 
~y] 
10 e.c. 1 cc. Olce Ol c.c. perce per 100 c.c 
plus minus minus minus 01 10 
plus plus minus minus 10 100 
plus plus plus minus 10 0 1000 
plus plus plus plus 100 0 10,000 
plus plus minus plus 10 6 1,000. 
1211 | 12,110 
24 0 | 2,499 
The above method of expressing results is not mathematically alto- 
gether correct. The average number of B. colz per cubic centimeter, as 
thus estimated is not precisely the most probable number calculated by 
