The kinetics of ATP reagents not affected by product inhibition or other forms of inactivation of luciferase during the measurement time has been clarified. Under these conditions the decay rate of the light emission expressed as percentage per minute is a measure of luciferase activity and can be given as the rate constant k (min-1), directly reflecting the degradation of ATP in the luciferase reaction. Three types of reagents with different analytical characteristics and different application possibilities have been identified. Stable light-emitting reagents are suitable for measurements of ATP down to 1000 amol. This is the only type of reagent suitable for monitoring ATP-converting reactions, i.e., assays of enzymes or metabolites, assays of oxidative phosphorylation, photophosphorylation, and so on. A higher luciferase activity resulting in a slow decay of the light emission by approximately 10% per minute (k = 0.1 min-1) gives a reagent suitable for measurements down to 10-100 amol. The slow decay of light emission allows use of manual luminometers without reagent dispensers. A further increase of the luciferase activity resulting in a decay rate of approximately 235% per min (k = 2.35 min-1) and only 10% of the light emission remaining after 1 min is suitable for measurements down to 1 amol corresponding to half a bacterial cell. With this type of flash reagent the total light emission can be calculated from two measurements of the light intensity on the decay part of the light emission curve. This new measure is not affected by moderate variations in luciferase activity, but only by changes in quantum yield and self-absorption of the light in the sample. Flash-type reagents require the use of reagent dispensers. The stringent requirements for ATP-free cuvettes, pipette tips, and contamination-free laboratory techniques make it unlikely that flash reagents would be useful in nonlaboratory surroundings. A potential application for this type of reagent is sterility testing. In general, it is concluded that one should select the ideal ATP reagent carefully for each application. Obviously the reagents used in a particular application do not have to match the decay rates given earlier exactly. However, various applications of the ATP technology and the properties of manual and automatic luminometers fall quite nicely into categories corresponding to the properties of the three reagents described. The rapidly growing interest in ATP technology has already resulted in the development of a greater variety of luminometers, from hand-held instruments to high-throughput systems. The continuation of efforts in both reagent and instrument development will undoubtedly result in many new applications.