For some reason, the bacteria that inhabit drinking water systems are difficult to detect by culturing them. More than 99% of those shown to be present by staining them with dyes and counting them with the aid of a microscope just don’t grow into countable colonies when spread onto a nutrient medium in a petri dish. They are called “non-fermenters or “viable but nonculturable organisms.” Part of the reason for the error in plate counts may be that many are trapped in a clump of biofilm containing hundreds of individuals, which grows out as a single colony. Or, it may be that their growth requirements (nutrients, pH, temperature, atmosphere, salinity, etc.) are simply too exotic and different from what is provided by the “standard” test conditions. However, the main reason is believed to be the phenomenon of bacterial dormancy: each specific strain of bacteria seems to require a rest of some weeks or months after growing actively for a while. At any moment, the bacterial population of a municipal distribution system will be comprised mostly of three to seven specific types, each following its own schedule of growth, decline, dormancy, and regrowth, and almost all of them are dormant most of the time. That means that, regardless of the specific plate count procedure used, the true bacterial numbers are at least 100 times, perhaps as much as 1000 times higher. Thus, plate counts are not very accurate or informative. It is folly to try to compare individual plate count results; only controlled bacteriological testing programs, in which the same trained person takes the samples in the same way, at the same place, at the same time of day, etc., yield scientifically valid data that can be compared. Even then, the variation is so great that the averages must differ by more than a “log” (logarithm) or power of ten to be considered statistically different. For example, an average influent plate count of, say, 350/mL and an average ‘filter effluent plate count of 3450/mL are not significantly different numbers, statistically speaking.

Tests for coliform organisms are done on 100 mL volumes, either by filtering 100 mL through a membrane and then culturing the membrane, or by adding special chemicals to a bottle containing 100 mL of sample. The special nutrient medium and temperature are inhibitory to most non-coliform organisms, but many unwanted HPC organisms may grow anyway, making the true coliforms difficult or impossible to detect. When that happens, the lab analyst may reject the sample with a “TNTC” notation (Too Numerous To Count) and request a new one. In the U.S., any repeat coliform tests are to be analyzed using a “Presence-Absence” test procedure, in which chemicals called “MMO” and “MUG” are added to 100 mL of the sample. If there are any total coliform organisms in the sample, their metabolism will turn the MMO yellow by the next day. If there are any fecal coliform organisms (E. coli) in the sample, they will turn the MMO yellow and also metabolize the MUG to something that fluoresces chartreuse when viewed with UV illumination.