As animals age, they are expected to invest successively more energy in reproduction as they have fewer subsequent chances to reproduce (the “restraint” hypothesis). Conversely, the oldest animals may show restraint in reproduction because even a small increase in energy expended during reproduction may lead to death. Alternatively, both young and very old animals may lack the ability to maintain high levels of energy expenditure (the “constraint” hypothesis), leading to reduced reproductive success. Many studies have observed an increase in reproductive success with age followed by a reduction at the end of life, but fewer studies have examined the proximate mechanisms, which provide a context for understanding ultimate causes. I examined over 30 behavioural and physiological metrics of aging in two species of free-living long-lived seabirds (thick-billed murres Uria lomvia and black-legged kittiwakes Rissa tridactyla) and a short-lived passerine (tree swallows Tachycineta bicolor). For all species, reproductive success was high at intermediate ages. In support of the “restraint” hypothesis, when birds were stressed glucocorticoid hormones, which direct energy away from reproduction and towards survival, were higher in young birds (swallows) and both young and very old birds (kittiwakes and murres). When birds were handicapped older birds expended more energy. When challenged exogenously, there was no change in hormone levels with age, implying that they were “choosing” to be restrained. Resting metabolic rate (RMR) declined linearly with age in both seabird species. T3, which I show is indicative of RMR in birds, also declined with age, demonstrating that the reduction in metabolism was strategic and not due to changing body composition. In contrast, daily energy expenditure in both seabird species during breeding was constant with age while antioxidant capacity became elevated during middle age, and further increased with age. Several measures of behavioural performance did not vary with age. I conclude that hormonal cues lead to greater investment in adult’s energy stores over its offspring’s energy reserves (restraint hypothesis) at the start of life. At the end of life, both hypotheses were supported; energy expenditure was constrained by senescence, leading to increased restraint in investing additionally in offspring.