The production of time-correlated neutrons from six 241 AmO 2 Be(α, n) neutron sources has been measured using a high level neutron coincidence counter (HLNCC-II). Five of the sources were contained in standard Amersham X.2 capsules while the sixth was in the larger X.3 format. The results quantitatively support the hypothesis that the predominant source of time-correlated neutrons arises from the Be(n, 2n) process initiated by the primary 241AmO 2/Be(α, n) neutrons. On this basis, analysis of the experimental data yields an (n, 2n) (α, n) ratio of ≈ 1.8% for the X.2 sources and ≈ 2.1% for the X.3 source. This work suggests that passive neutron coincidence counters could be used to good effect in the measurement and testing of integral (n, 2n) cross section data. Tentative conclusions are that the EFF-2 evaluation of the Be(n, 2n) double differential cross section is to be preferred over the ENDF/B-VI evaluation and that the integrated-over-angle Be(n, 2n) cross sections appearing in these evaluations may be approximately 20% low.