A common method used to determine reactivity is to measure the reactor period and then refer to the inhour equation which relates asymptotic period to reactivity. The implicit assumption in such a method is that the neutron population is varying exponentially with time. The conditions necessary for this assumption to be valid have been obtained by a quantitative examination of the time behavior of the neutron population. The results obtained show that under common experimental conditions, several minutes may be required following a positive step change in reactivity in order to achieve accuracy by means of a period determination. In addition, in the case of the reactor with a constant extraneous source, a significant increase in this waiting time can result even if the reactor is initially only very slightly subcritical. Whereas in principle both positive and negative reactivities may be obtained from period observations, it is pointed out that in practice serious objections exist for the case of negative reactivities.