The validity of the poisson assumption for analysis of most-probable-number results

Abstract

A test of the validity of the Poisson assumption for sample replicates in dilution series of finite length is proposed and its properties are examined by using Monte Carlo simulation. The test is based on an examination of the number of intervals between complete sterility and complete infection in a series. The test is applied to a data set of routine influent coliform samples at the Chicago water supply intake. By this test, the data set is rejected as being drawn from a Poisson replication. Tables for direct application to a 3-dilution, 5-tube decimal series are presented, and their application is illustrated.