Summary: In this article it will be proved that the commonly used formula for dead time corrections :[xxx] (Ν' is the true and Ν the measured count rate, r the dead time) is correct for a G.M.-counter tube and also for high count-rates. However τ seem to be a function of N. Three often used methods for measuring the dead time τ are discussed (the two samples method ; the method of a series of samples in which the relative intensities are very well known; the method of a shortlived radioisotope), and also for each of these three methods the best way to calculate the dead time from the experiments. The two samples method, combined with a very simple formula of SCHULZE, seems to be the most attractive solution. The theory has been illustrated by a number of experiments according to the three different methods.