The time dependence of the 2.5-MeV neutron emission from the Joint European Torus (JET) is reliably measured using fission chambers. The absolute calibration of these chambers is required to an accuracy of 10% or better for a range of intensities that may cover six or more decades. At JET, this calibration is now achieved by use of activation techniques, the most convenient of which involves fissionable materials (thorium and uranium) and delayed neutron counting. Because delayed neutron counting is unfamiliar in the fusion community, particular care is taken to obtain confirmation of the results based on this method by comparison with measurements made using the conventional activation procedure (involving indium, nickel, and zinc as target materials). As the activation measurements can be influenced appreciably by the weak emission of 14-MeV neutrons, this contribution is measured separately using high threshold energy activation reactions (in copper and silicon). Neutron transport calculations are employed to relate the measured local fluences of both 2,5- and 14-MeV neutrons to the total yields from the plasma. Absolute calibration accuracies of 6 and 8% are claimed for 2,5- and 14-MeV neutron yields, respectively; the accuracy of the 14-MeV to 2,5-MeV yield ratios is 6%.