Studies of marine animals such as fish and plankton require knowledge of their density (number of animals per unit volume). Standard acoustical techniques, such as echo integration, give accurate estimates of density, but only with a carefully calibrated system. We have found the statistical nature of the echo useful in making rough estimates of density without a calibrated system. In particular, we are studying the probability density function (PDF) of the maximum (peak) value of the echo achieved in a time gate. One useful property of this PDF is that for the two extremes in which the echoes from the individual scatterers overlap and do not overlap, the PDFs take on drastically different shapes. The difference is so great that one can visually judge from the PDF whether or not the echoes were overlapping. For a given ping duration and sonar beamwidth, there is a critical density above which the echoes overlap (or conversely below which the echoes do not). We use this property to estimate densities of animals. Examples are given in this paper of order of magnitude estimates of the density of fish in Trout Lake, Wisconsin and Lake Michigan, and a lower bound estimate of 1‐cm‐size plankton density in the Gulf Stream Boundary near Cape Hatteras, North Carolina. The technique is shown to be simple and robust. It was quickly implemented without calibration of the system. The technique is good for fast preliminary results and is a good check on the accurate, more involved techniques.