Mice injected intravenously with sheep erythrocytes (sRBC) demonstrate a transient increase in splenic cAMP levels (4-fold), which peak at 2 min after injection and return to basal levels by 20 min. In addition to the change in cAMP, an increase in splenic cGMP levels (1.5-fold) occurs beginning 5-10 min after sRBC injection, and persists for up to 7 days. During this period cAMP levels remain at or below control levels in the spleen. There is no change in 3':5' cyclic nucleotide levels in the liver and a small increase (1.2- to 1.3-fold) in the thymus at the time when splenic cyclic nucleotide levels are elevated. The changes in splenic cyclic nucleotide levels appear to be dependent on the presence of thymus-derived (T) lymphocytes, since little change occurs in cAMP and changes in cGMP are absent in athymic nude mice. In addition, cAMP levels were increased by pretreatment of normal mice with cortisone acetate, which selects for mature T lymphocytes. Agents that block autonomic nervous system functions have no effect on the early sRBC-induced changes in cyclic nucleotide levels.Indomethacin, an inhibitor of prostaglandin synthesis, reduces the change in cAMP level by 50% and blocks the change in cGMP levels completely. Secondary stimulation with sRBC results in a larger increase in cAMP levels than that seen with a primary injection of sRBC, indicating the presence of specific antigen-sensitive memory cells. Changes in splenic cyclic nucleotide levels cannot be detected at early times after the injection of soluble protein antigens such as bovine serum albumin and keyhole limpet hemocyanin. Salmonella H antigen stimulates a 2-fold increase in cAMP levels, the increase occurring more slowly than with sRBC. The in vivo changes incyclic nucleotide levels are correlated with known changes in cyclic nucleotide levels which have been documented in vitro in both T-cells and T-cell-dependent-B-cell (bone-marrow derived) antibody responses.