Embryos of the fork-tailed storm petrel survived exposure to 26, 30, and 34 C. However, during artificial incubation at 26 C, metabolic rate did not increase, and presumably development did not progress. When incubated at 30 and 34 C, eggs showed a gradual increase in metabolism commensurate with development, but success of pipping and hatching was much greater at 34 C than at 30 C. The reduction of metabolic Qio to 1.55 at temperatures above 34 C suggests that the optimum incubation temperature is near 34 C. Embryos tolerated periodic chilling-24-h exposure to 10 C every fourth day, which simulates environmental conditions when parents neglect the egg during foraging trips. During chilling to 10 C, metabolic rate was only 5% of that at 34 C, and development was presumably arrested. The energy expended by embryos during absence of the parents should thus be negligible. How- ever, the total incubation period is increased, and development is slow even when the egg is warm. Low conductance of the eggshell to water vapor, 2.1 mg H20 (day- torr)-1, is an adaptation which minimizes water loss during incubation. Natural incubation of the one-egg clutch is prolonged because of periodic chilling, low temperature of incubation in the subterranean burrow, and slow rate of development. Tolerance of a broad range of temperatures during development and a slow rate of development represent a suite of physiological adaptations which facilitate breeding success of a species with patterns of parental behavior that impose unusual physio- logical demands on the developmental process.