The challenge hypothesis has been very successful in explaining patterns of testosterone secretion in response to social stimuli in avian species. However, there have been few studies in nonavian vertebrates. We tested the challenge hypothesis in male northern fence lizards (Sceloporus undulatus hyacinthinus). These males are highly territorial and nonparental. Consequently, the challenge hypothesis predicts that plasma testosterone concentrations will be insensitive to aggressive interactions. Testosterone concentrations indeed were not significantly affected by either a short (3–15 min) simulated territorial intrusion (“challenge”) in June or a longer (50–60 min) intrusion in July. Levels of corticosterone were elevated in challenged males in the long, but not the short, intrusion. Challenged males displayed significantly more intense territorial behaviors than did unchallenged control males. The intensity of territorial behavior changed significantly across the active season and was positively related to testosterone concentrations. Thus, while testosterone concentrations do not appear to be involved in rapid changes in aggressive behavior in the fence lizard, they probably are important in larger-scale behavioral changes throughout the season.