Dormancy is a condition in which an herbaceous perennial does not sprout for one or more growing seasons. To test whether dormancy is an adaptive response to environmental stress, we defoliated and shaded individuals of two rare geophytic orchids, Cypripedium calceolus and Cephalanthera longifolia, in five Estonian populations early in the growing season in 2002 and 2003. We also censused plants at the same time, and conducted one more census in 2004. Mark–recapture models were used to estimate the probabilities of dormancy (d, the complement to resighting, p), and apparent survival (φ). Apparent survival varied little by treatment, with Cypripedium and Cephalanthera surviving at 0.986 ± 0.014 and 0.974 ± 0.021 (mean ± se), respectively. In contrast, treatment impacted dormancy dramatically. For both Cephalanthera and Cypripedium, defoliated (def.) plants were most dormant (0.320 ± 0.055 and 0.095 ± 0.036, respectively). However, while both control (cont.) and shaded (sh.) plants were roughly equally least dormant in Cypripedium (dcont. = 0.048 ± 0.020 vs. dsh. = 0.045 ± 0.021), the least dormant Cephalanthera had been shaded (0.182 ± 0.040 vs. dcont. = 0.206 ± 0.050). We conclude that dormancy may allow the plant to buffer stress in the short term without increasing mortality risk.