An Investigation of Errors in Direct Counts of Aquatic Bacteria by Epifluorescence Microscopy, with Reference to a New Method for Dyeing Membrane Filters

A number of methods for observing freshwater bacteria by epifluorescence (incident light fluorescence) microscopy are examined. The suitability of each method for quantitative studies using black membrane filters is assessed. In spite of inadequacies it was considered that the use of acridine‐based fluorochromes provided the best available estimate of the bacterial population. The errors which may arise when these stains are used were examined and it was noted that small changes in methodology could cause significant differences in the results obtained. The largest errors were associated with changes in volume of sample filtered and the pore size of the membrane used. A procedure for sample treatment is suggested and a new method for dyeing membrane filters is given which allows the use of 0·22‐μm pore size membranes of the cellulose ester and polycarbonate type.