Mammalian hibernation consists of energy-saving torpor bouts (periods of controlled reduction in body temperature [Tb]) interspersed with brief arousals to normothermic Tb. Frequency and duration of torpor bouts and arousals can affect winter survival and are thought to be influenced by an optimization balancing the energetic benefits of prolonged torpor against the physiological and ecological costs (e.g., accumulation of metabolic wastes). Female little brown bats (Myotis lucifugus) spend their fat reserves more slowly than males during winter, presumably so they can emerge from hibernation in good condition to initiate pregnancy. We used temperature telemetry over three winters to test a prediction of the optimization hypothesis that female M. lucifugus would use longer torpor bouts and/or shorter arousals than males. Females did conserve energy relative to males by adjusting the magnitude and duration of arousals but not the duration of torpor bouts. Although torpor bout duration did not vary by sex for adults, it did vary by age. Adults initially used longer torpor bouts than young-of-the-year, but this difference declined as cave temperature warmed in spring. Males and females in better condition spent more energy during hibernation, again via increased arousal duration rather than decreased torpor bout duration. Longer arousals by males could increase reproductive fitness if males mate with torpid females throughout winter. Our results highlight demographic differences in winter behavior for small hibernators facing extreme energy limitation in cold climates and illustrate the influence that reproductive costs have on hibernation energetics.