Formulas for deriving probabilities of dying from exposure rates, person-years lived between 2 ages, and stationary population numbers, must lead to consistent results. No difficulty occurs when using integral calculus. In practice, however, events are observed over discrete intervals; in this case inconsistent results may be obtained for various life table variables when the survivorship function is nonlinear over the age interval x to x plus n. The purpose of this note is to establish the necessary conditions in order to obtain consistent results when computing probabilities of dying, person-years lived, and stationary populations, over discrete age-intervals. (author's)