The effective delayed neutron fraction from fission was determined for an unreflected uranium (93.2 wt% 235U) metal sphere from the ratio of time-correlated counts in a randomly pulsed neutron measurement to those in a Rossi-α measurement. In the randomly pulsed neutron measurements, a 252Cf source was placed in the sphere which contained a fission counter that, because of its location, did not count neutrons directly from the source. Neutrons from spontaneous fission of 252Cf initiated fission chains in the sphere, and the fission counter detected events from the interaction of neutrons from these fission chains with the uranium of the fission counter. A Type I time analyzer was triggered each time a 252Cf nucleus fissioned and recorded the time distribution of neutrons from the fission chains initiated by neutrons from californium at t = 0. The delayed neutron fraction by this method (60.2 ± 0.8 × 10−4) is ∼11% lower than that from other measurements or calculations that are all in agreement. This low value may be due to an improper theoretical formulation for the correction of point kinetics for spatial effects. The value of this correction factor estimated by another theoretical formulation is 30% larger. An 11% larger correction for spatial effects would produce agreement between this measurement and previously measured results.