Acute infection is accompanied by a characteristic reduction in circulating eosinophils. This study examined the generally held assumption that the eosinopenia of infection is a manifestation of adrenal stimulation. Trichinosis, Escherichia coli pyelonephritis, and early subcutaneous pneumococcal abscess were used as experimental infections of limited severity. Trichinosis is associated with eosinophilia, but pyelonephritis and pneumococcal infection produce eosinopenia. An assay for serum corticosterone was developed that is sufficiently sensitive to be performed with the small volumes of blood obtained sequentially from individual mice. The corticosterone response to trichinosis fits the sterotyped reaction previously reported for several other bacterial, viral, and rickettsial infections. The peak concentrations of corticosterone in serum from mice with trichinosis was approximately twice normal and occurred at the onset of clinical illness. Serum corticosterone levels gradually declined to the normal range over the next several days. E. coli pyelonephritis produced a similar adrenal response, although the peak serum corticosterone caused by pyelonephritis was less than the serum corticosterone occurring during the first peak of eosinophilia during trichinosis. Infection of a subcutaneous air pouch with penumococci produced eosinopenia within 6 h after inoculation, but there was no rise in serum corticosterone during the first 12 h of the pneumococcal infection. In addition, the eosinopenic response produced by a 12-hpneumococcal abscess occurred mice adrenalectomized 1-4 days before infection with pneumococci. The eosinopenia of acute infection cannot be ascribed to adrenal stimulation.