The staining of gram-positive and gram-negative cultures with acridine orange in metabolically active and inactive states was investigated using a Bactoscan, direct epifluorescent filter technique (DEFT), and standard plate count as the reference method. The evaluation of the bacterial cultures in the Bactoscan revealed a linear relationship between Bactoscan counts (pulses) and the quantity of pure culture suspension used. But the proper detection of bacteria with the fluorescence optic methods was dependent on the type of microorganism and the physiological state of the cells. The Bactoscan and DEFT underestimated the bacterial counts of gram-negative cultures as compared with standard plate counting. When stained with acridine orange, metabolically active bacteria showed more orange fluorescence and a lower percentage of green fluorescent cells as compared with inactive bacteria. Bactoscan pulse height analysis (PHA) diagrams, graphs of the detected pulses and their intensity, showed low pulses of inactive bacteria. Many of these weak pulses were eliminated from counting because of their faint fluorescent staining. In contrast, PHA diagrams of metabolically active microorganisms showed bright staining and, therefore, high pulses. A complete count of these bacteria was possible. These investigations point out that discrepancies between the fluorescence optical counting methods and the standard plate count depend strongly on the staining of the cultures with acridine orange and, therefore, on the type of microorganism and the metabolic state of the cells measured.