In the present study we observed that lipopolysaccharide (LPS) administration provoked a characteristic reduction in body weight gain, food consumption, saccharin (but not water) consumption and nocturnal locomotor activity. It has been previously suggested that the ability of LPS to suppress the consumption of, and preference for, a palatable solution such as saccharin without altering water consumption, may represent an anhedonic response. The results of the present study demonstrate that chronic treatment with the tricyclic antidepressant (TCA) desipramine (7.5 mg/kg; i.p.) prevented LPS-induced anorexia, loss of body weight, the antidipsogenic effect and hypoactivity. In contrast, chronic treatment with the antidepressants paroxetine (7.5 mg/kg; i.p.) and venlafaxine (10 mg/kg; i.p.) failed to alter any of the LPS-induced behavioural responses. Furthermore, chronic treatment with desipramine (and to a lesser extent paroxetine) reduced the consumption of, and preference for, saccharin suggesting that these antidepressant treatments induce an "anhedonic" response in their own right. In conclusion, chronic desipramine treatment attenuated LPS-induced depressive-like behavioural symptoms in the rat. However, chronic treatment with paroxetine and venlafaxine did not significantly alter LPS-induced behavioural responses. The results of the present study support the hypothesis that TCA's may exert part of their anti-depressive efficacy through their effects on the immune system. However, this property does not appear to be shared by newer antidepressants which possess a better side effect profile than the TCA's. The suppressive effect of TCA's on proinflammatory cytokine secretion is discussed as a mechanism by which these agents alter LPS-induced behavioural responses.