In humans and other mammals, the release from growth-inhibiting conditions, such as glucocorticoid excess, leads to supranormal linear growth. The prevailing explanation for this catch-up growth involves a central nervous system mechanism that compares actual body size to an age-appropriate set-point and adjusts growth rate accordingly via a circulating factor. Although such a neuroendocrine "sizostat" was hypothesized more than 30 yr ago, its existence has never been confirmed experimentally. Here we show that suppression of growth within a single growth plate by locally administered glucocorticoid is followed by local catch-up growth that is restricted to the affected growth plate. Thus, the catch-up growth cannot be explained by neuroendocrine mechanism but, rather, must arise from a mechanism intrinsic to the growth plate. To explain this finding, we propose that the normal senescent decline in growth plate function depends not on age per se, but on the cumulative number of stem cell divisions, and that glucocorticoid administration, by suppressing stem cell proliferation, delays senescence, resulting in catch-up growth after the growth-inhibiting agent is removed.