On the Statistical Interpretation of some Bacteriological Methods employed in Water Analysis

Abstract

We were recently consulted by an officer serving on the Western Front as to the significance attaching to ordinary bacteriological methods of gauging the potability of waters. He wished to know what was the probability that a given water supply did not contain more than a certain proportion of bacteria in the unit volume, it having been found that particular samples tested showed no growth while, perhaps, larger samples had done so, or that so many out of a series of samples of the same size had given positive results. Having obtained what seemed to us a reasonable solution of the particular problem proposed, we thought that the results might interest other officers and bacteriologists who have to do similar work. A survey of the criteria actually used by bacteriologists when they form an opinion as to the purity of waters seems to us to emphasise the need for some discussion.