Geochronological results from central Europe indicate that the duration of Pennsylvanian time is only 19 m.y., compared to the Harland et al. and Palmer estimates of 34 m.y. Prior calculations of Pennsylvanian cycle periods from the midcontinent of North America suggesting a fit with Milankovitch orbital parameters may well be in error; as a consequence, other mechanisms for possible eustatic sea-level changes represented in those cycles are needed. Calculation of cycle periods of 100 ka or less lack precision in stratigraphic intervals representing ages characterized by error margins of millions of years. Thus, cycle periods may be less reliable as an indicator of global process than previously considered, particularly in rocks of Paleozoic and early and middle Mesozoic age.