Heating serum at 56 degrees is used to inactivate complement in several immunological assays. During heating, both heat-labile and heat-stable anticomplementary activity (ACA) develop. While heat-labile ACA can be completely inactivated, heat-stable ACA increases progressively with continued heating. Heat-stable ACA develops in deaggregated IgG and in normal, but not in hypogammaglobulinaemic, human and porcine serum heated at 56 degrees suggesting that this ACA is due to formation of immunoglobulin aggregates. These aggregates would produce false-positive tests for immune complexes and could inhibit a variety of cell-mediated reactions in assays which incorporate heat-inactivated serum. Other temperatures were tested to determine whether endogenous haemolytic activity could be destroyed without forming immunoglobulin aggregates. At 53 degrees both endogenous haemolytic activity and heat-labile ACA were inactivated and formation of heat-stable ACA in normal serum was minimal. ACA, however, could be induced in deaggregated IgG at 53 degrees. Moreover, the degree of heat-induced aggregation of IgG in vitro at either temperature was directly proportional to IgG concentrations and inversely related to albumin concentrations. Thus, pathological sera with these protein alterations might form more aggregates during heating than normal sera. These data suggest the following: (1) heat inactivation of complement at 53 degrees for 90 min is preferable to the traditional 56 degrees; (2) in any assay where immunoglobulin aggregates might interfere, normal serum may be an inadequate control and correlations will need to be made between serum IgG and albumin concentrations and the results obtained in these assays.