Statistics of enumerating total coliforms in water samples by membrane filter procedures

Abstract

A commonly used assumption is that the microbial counts of a given water sample are randomly distributed according to Poisson statistics and, in particular, that replicate analyses of the same sample are so distributed. Using Fisher's index of dispersion, as well as a newly developed test, analysis of prior data on membrane filter total coliform organisms in a variety of water samples indicated that, in many cases, the assumption of Poisson statistics was incorrect, but that the data were consistent with a negative binomial distribution. This finding has significance in the interpretation of total coliform counts for regulatory as well as research purposes, and further research to delineate the responsible mechanisms is required. © 1986.