Temperature has profound effects on ectotherms such as plants, invertebrates, and lower vertebrates (Hochachka and Somero 1984; Cossins and Bowler 1987). The impact of temperature is often depicted graphically as a “thermal performance curve,” which plots performance as a function of body temperature (Huey and Stevenson 1979). Typically, performance increases gradually with temperature up to some maximal or optimal level but then declines precipitously as temperature approaches damaging levels and upper lethal limits. The impact of temperature on Darwinian fitness of ectotherms can be depicted similarly. However, one must first choose an appropriate measure of fitness (Stearns 1982). Several measures are available (Tuljupurkar 1990; Roff 1992; Stearns 1992; Carey 1993; Kozlowski 1993; Charlesworth 1994), but r (“intrinsic rate of increase”; see “Material and Methods”) and Ro (“net reproductive rate”; see “Material and Methods”) are by far the two most commonly used ones. The intrinsic rate of increase is the rate of population increase in a closed population, assuming constant age-specific schedules of death and reproduction and a stable age distribution, whereas the net reproductive rate is the average number of female offspring born to a female over her lifetime, again assuming constant agespecific schedules of death and reproduction (Carey 1993). Both measures estimate population growth rates, but r is scaled to time, whereas Ro is scaled per generation and is independent of time. Surprisingly, however, whether these