Fe availability is critical for optimal lymphocyte proliferation; however, the minimum required levels are unknown. Such information is valuable when assessing in vitro immune responses in Fe-deficient subjects, because serum (Fe) added to the culture medium may replete lymphocytes. To address this issue, splenic lymphocytes obtained from seventeen 3-month-old C57BL/6 mice were incubated without and with 1 mg/l concanavalin A or 50 μg/l anti-CD3 antibody in media that contained between 0·113 and 9·74 μmol Fe/l. Fe was provided by either fetal calf serum (FCS, 0–100 ml/l), newborn calf serum (NBCS, 0–100 ml/l), or NBCS (10 ml/l) plus ferric ammonium citrate. As expected, the rate of DNA synthesis increased with Fe levels (P<0·01). Maximum DNA synthesis was obtained with 2·26 μmol Fe/l (50 ml FCS/l) for concanavalin A and 0·895 μmol/l (20 ml FCS/l) for anti-CD3-treated cells. In serum-free media (0·113 μmol Fe/l), the proliferative responses to concanavalin A were below the background, while they rose 5·5-fold in anti-CD3-treated cells (P<0·05). In apotransferrin-supplemented media (0·13 μmol Fe/l), the proliferative responses to concanavalin A and anti-CD3 antibody were 18·6 and 71 %, respectively, of that obtained with 4·66 μmol Fe/l (100 ml FCS/l). Interleukin 2 secretion also followed the same trend as lymphocyte proliferation. Since differences between both mitogens persisted after FCS was substituted with NBCS, we can rule out an effect on ribonucleotide reductase activity, or by other serum growth factors. We speculate an Fe effect at an early step of T-cell activation. Data suggest that the minimum Fe concentration required for lymphocyte proliferation varies with the mitogen.