The immunocompetence handicap hypothesis proposes that the immunosuppressive effect of testosterone enforces honesty of sexual signalling via a physiological trade–off between signal intensity and immuno–competence. However, evidence that testosterone is immunosuppressive is scant, particularly in birds. I studied the correlation between immunocompetence and testosterone in superb fairy–wrens (Malurus cyaneus), a species with intense intersexual selection. Males are seasonally dichromatic and testosterone increases during the moult from dull brown eclipse plumage into bright nuptial plumage. I determined the primary antibody response to immunization with sheep red blood cells (SRBCs) in (i) control and testosterone–implanted males in captivity, and (ii) a cross–section of free–living males with basal and elevated testosterone (in eclipse plumage, moulting and in nuptial plumage). Experimental treatment with testosterone decreased the likelihood of an antibody response to SRBCs in captive birds. In contrast, free–living males which had acquired the nuptial plumage and had naturally elevated testosterone were more likely to respond to SRBCs than males in eclipse plumage with basal testosterone levels. The association between higher immunocompetence and higher immunosuppressive testosterone could arise if both are positively correlated with male phenotypic quality. In addition, the association could result if males compensate for potential immunosuppression by enhancing their humoral immune responses, particularly since high testosterone is linked to other demanding activities such as moulting and courtship displays.