While ecological causes of sociality (or group living) have been identified, proximate mechanisms remain less clear. Recently, close connections between sociality, glucocorticoid hormones (cart) and fitness have been hypothesized. In particular, cort levels would reflect a balance between fitness benefits and costs of group living, and therefore baseline coil levels would vary with sociality in a way opposite to the covariation between sociality and fitness. However, since reproductive effort may become a major determinant of stress responses (i.e., the cart-adaptation hypothesis), cort levels might also be expected to vary with sociality in a way similar to the covariation between sociality and fitness. We tested these expectations during three years in a natural population of the communally rearing degu, Octodon degus. During each year we quantified group membership, measured fecal cortisol metabolites (a proxy of baseline cort levels under natural conditions), and estimated direct fitness. We recorded that direct fitness decreases with group size in these animals. Secondly, neither group size nor the number of females (two proxies of sociality) influenced mean (or coefficient of variation, CV) baseline cortisol levels of adult females. In contrast, cortisol increased with per capita number of offspring produced and offspring surviving to breeding age during two out of three years examined. Together, our results imply that variation in glucocorticoid hormones is more linked to reproductive challenge than to the costs of group living. Most generally, our study provided independent support to the cort-adaptation hypothesis, according to which reproductive effort is a major determinant, yet temporally variable, influence on cart-fitness covariation. (C) 2011 Elsevier Inc. All rights reserved.