Planting cotton (Gossypium hirsutum L.) according to an empirically determined earliest planting date increases the probability that fully-emerged seedlings will be exposed to suboptimal temperatures which significantly affect subsequent development and yield. Most cotton chilling-stress studies have dealt with the effects of temperatures < 10-degrees-C on germinating seeds and emerging seedlings. In this study, the responses of 10-d-old seedling roots and shoots of three cultivars of cotton to 10, 15, 20, 25, 30, and 35-degrees-C were determined, using an adaptation of a seed-vigor testing procedure that employs paper-towel scrolls and environmental chambers. Inhibition of root and shoot growth (i.e., root and shoot elongation and fresh weight accumulation) at cool temperatures was less in 'Coker 315' than in 'Deltapine 61' or 'Paymaster 145'. The capacities of Coker 315 and Deltapine 61 for recovery from a 5-d exposure to suboptimal temperatures during a subsequent 48 h at 30-degrees-C were greater than that of Paymaster 145. Chilling (10-25-degrees-C) increased root, but not shoot, relative water content in each cultivar. The levels of growth inhibition caused by nonoptimal temperatures, and the capacities for regaining pretreatment water relations and resuming growth after chilling stress, differed according to treatment temperature and cultivar. Both decreasing chilling response and increasing poststress recovery capacity offer possibilities for crop improvement in cotton.