Reproduction is energetically expensive for both sexes, but the magnitude of expenditure and its relationship to reproductive success differ fundamentally between males and females. Males allocate relatively little to gamete production and, thus, can reproduce successfully with only minor energy investment. In contrast, females of many species experience high fecundity-independent costs of reproduction (such as migration to nesting sites), so they need to amass substantial energy reserves before initiating reproductive activity. Thus, we expect that the relationship between energy reserves and the intensity of reproductive behavior involves a threshold effect in females, but a gradual (or no) effect in males. We tested this prediction using captive vipers (Vipera aspis), dividing both males and females into groups of high versus low body condition. Snakes from each group were placed together and observed for reproductive behavior; sex-steroid levels were also measured. As predicted, females in below-average body condition had very low estradiol levels and did not show sexual receptivity, whereas males of all body condition indices had significant testosterone levels and displayed active courtship. Testosterone levels and courtship intensity increased gradually (i.e., no step function) with body condition in males, but high estradiol levels and sexual receptivity were seen only in females with body reserves above a critical threshold. The energetics of reproduction has been a central theme in the ecological study of life history variation (Hirshfield and Tinkle, 1975;Morris, 1987Morris, , 1992Winkler and Wallin, 1987;Stearns, 1992; Monaghan and Nager, 1997), and more recently this topic has become an important focus in behavioral endocrinology (Cherel, Mauget, Lacroix, and Gilles, 1994;Bronson, 1998). It is obvious that organisms cannot reproduce in the total absence of energy reserves (Frish, 1978;Frish and McArthur, 1974), but the relationship between the amount of energy reserves and reproductive output (or effort) is complex and varies among species. In some species there is a linear relationship between energy reserves and reproductive output. In others there appears to be no discernible relationship (e.g., if clutch and offspring sizes are fixed) or a complex nonlinear relationship between the two variables. For example, many organisms are "capital breeders" that rely on previously gathered energy stored in the form of body reserves (i.e., fat bodies, proteins) rather than on current energy intake to fuel the energetically costly process of reproduction (Drent and Daan, 1980;Jö nsson, 1997; Bonnet, Bradshaw, and Shine, 1998). In such species, life history theory predicts that a threshold level of energy reserves is necessary for a reproductive cycle to begin, rather than a linear relationship in which an increment in energy availability adds an increment in reproductive output (Schaffer, 1974;Bull and Shine, 1979;Stearns, 1992). Several empirical studies provide evidence for nonlinear relationships between...