In typhoid perforation patients, Salmonella typhi was isolated from blood in 4%, ileal contents in 23%, peritoneal pus in 13% and from mesenteric lymph nodes in 71%. While isolation of S. typhi was made from patients with less than 4 days of chloramphenicol therapy, cultures were negative from these sites after 5 days of therapy; however, S. typhi appeared to remain viable in the lymph nodes even after such therapy. All isolates of S. typhi were sensitive to chloramphenicol. Significant SAT titers (0 > or = 1/240) were obtained in only 7/21 (33%) of patients. The perforated group had lower geometric mean titers (0-1/138; H-1/46), when compared to matched patients with uncomplicated typhoid fever (0-1/476; H-1/148). This difference was significant (0- p < 0.005; H- p < 0.0025). The two groups (uncomplicated and perforated) showed no significant difference in total serum IgG, IgM and IgA or isohemagglutinin levels, indicating that the apparent hyporeactivity was not due to a generalized humoral immunodeficiency. Mesenteric lymph node histology showed hyporeactivity in both the T cell and B cell zones. These findings are discussed with the suggestion that S. typhi-specific host immunological hyporeactivity could be an explanation for these observations and a basis for the pathogenesis of perforation. Aerobic cultures of the peritoneal pus gave 39 isolates from 25 patients; the predominant isolates were Escherichia coli (24) and Klebsiella pneumoniae (12). On no occasion was S. typhi the predominant isolate. Gentamicin and kanamycin were the only two antibiotics which were consistently effective in vitro against the aerobic isolates from peritoneal pus.