The influence of particular enzyme activities on the flux of metabolites in a pathway can be estimated by 'modulating' enzymes (i.e. changing turnover or concentration) and measuring the response in various parts of the system. By controlling the nuclear ration of two genetically different nuclear types in heterokaryons, the enzyme concentrations at four different steps in the arginine pathway were decreased over a range. This range was extended by the use of bradytrophs, mutant strains specifying enzymes with greatly diminished enzyme activities. Strains altered simultaneously at more than one step were also constructed by genetic recombination. By measuring the outputs of the pathway and the steady-state concentrations of intermediate pools, the fluxes in different parts of the pathway were calculated. This allowed the construction of flux/enzyme relationships, the slope of which is a measure of the sensitivity of a flux to the change in enzyme activity at that step. All fluxes were found to be considerably buffered for quite substantial decreases in the activities of all enzymes. Mass action plays an important part in this phenomenon, as do inhibition and repression. Because of the existence of expansion fluxes in growing systems, we find quantitatively different fluxes in different parts of the single pathway. For the same reason some enzyme modulations given decreased fluxes in one part and increased fluxes in another. The understanding of control in the pathway thus involves consideration of many mechanisms operating simultaneously and the estimation of changes in the whole system. The concept of a 'rate-limiting step' is found to be inadequate and is replaced by a quantitative measure, the Sensitivity Coefficient, which takes account of all the interactions. It is shown that control of the flux is shared among all the enzymes of the pathway. The results are discussed in terms of the theory of flux control.