The effect of sera from patients with metastatic breast cancer on the response of normal lymphocytes to phytohaemagglutinin (PHA) was studied. A preliminary experiment indicated that dialysis of sera against tissue culture medium RPMI-1640 revealed differences between patient and control sera which were not apparent using undialysed sera. The ability of a dialysed serum sample to support the PHA response of normal lymphocytes, as measured by 3H-thymidine incorporation, was called the serum support of lymphocyte stimulation (SSLS). Sera having an SSLS value less than 30% of the control mean in the same experiment were called inhibitory sera. Dialysed sera from a proportion of patients with metastatic breast cancer on various treatments were found to inhibit the response of normal lymphocytes to PHA compared with control sera. This effect was not due to cytotoxicity of the inhibitory sera. The SSLS of inhibitory sera remained lower than that of most controls when experimental conditions such as concentration of reagents and duration of incubation were varied. The ability of undialysed sera to support a mixed lymphocyte reaction was significantly correlated with the SSLS of the same sera when dialysed. The presence of inhibitory sera was apparently related mainly to treatment. In the earlier (Phase 1) experiments, treatment with chemotherapy or oestrogens was associated with inhibitory sera, whereas treatment with androgens or corticosteroids was associated with sera which supported PHA stimulation well. In later (Phase 2) experiments, there was a higher proportion of persons with inhibitory sera among patients receiving one or both of the newer hormonal agents, Tamoxifen and Aminoglutethimide, than among controls or untreated patients. In the phase 1 experiment, advanced disease and disease of long duration appeared to be associated with inhibitory sera, but these associations were not confirmed in Phase 2. Higher patient age in both phases and greater tumour differentiation in Phase 1 were associated with inhibitory sera, but in Phase 1, the presence of inhibitory sera was unrelated to prognosis. Prognosis and tumour differentiation were not studied in Phase 2. Lymphocytes from breast, cancer patients responded to PHA significantly less well than did control lymphocytes. However, there was no correlation between the PHA response of patients' lymphocytes and the SSLS of sera from the same patients. A low molecular weight (10,000 dalton) inhibitory substance appeared to be present in a pool of two very inhibitory patients' sera in much greater concentration than in control sera. However, no similar substance could be demonstrated in other less inhibitory sera. Measurement of several serum proteins revealed a borderline significant negative linear relationship between the serum concentration of ai-antitrypsin and the logarithm of the SSLS. Variability in the concentration of ai-antitrypsin could account for less than 10% of the variability in log SSLS. Thus, apparently more than one substance is involved in determining SSLS. / Science, Faculty of / Microbiology and Immunology, Department of / Graduate