Abstract
Meat of the procedures proposed in the past to correct the coincidence and dead time losses in droplet counters FSSP* or ASSP** are based on empirical formulae. A more rigorous analysis of the statistical behavior of these counters yields a better estimate of the high droplet concentration measurements and moreover, it becomes possible to explain the fluctuations of the operating characteristics of the probes (rejection criteria). A new method is then proposed to detect heterogeneities at scales smaller than the sampling scale and to estimate the concentration peak values in the sample, with the current parameters measured by the FSSP.