SummaryThe normal or jigsaw‐puzzle method of planning cannot be applied to the production of an engine factory. One has to fall back upon a kind of statistical automatism, like in the planning of road traffic as opposed to the planning of railway traffic.In order to do this kind of planning efficiently, it is necessary to know the statistical relationships between the degree of occupation (number of working hours divided by number of available hours) of the machines, the waiting time and the velocity of flow through the factory.The statistical analysis of some of these quantities in an Amsterdam plant, manufacturing medium size diesel engines, showed that Erlang's formulas of waiting time in telephone traffic (with exponential distribution of holding times) are applicable.These formulas are used to prove that the highest degree of occupation is not the best one from an economic point of view. Formulas and graphs are given for finding the optimum degree of occupation in engine factories and other works where the same conditions appl