Patients (98) with chronic bronchitis were given killed influenza vaccines i.m. to study the systemic (serum) and local respiratory (sputum) antibody response to vaccination; 49 received a whole virus vaccine, and 49 a split-product vaccine. Adverse reactions could not be correlated to the content of chick cell agglutinating units, total protein, chick cell protein, endotoxin or induction of fever in rabbits. No difference in the local or systemic side effects was noted for the 2 bivalent vaccines. A 4-fold rise in serum hemagglutination inhibition antibody [HIA] occurred in 76% of all vaccines for the influenza A/Eng/72 and B/Hong Kong/72 antigens. When dividing the groups according to their prevaccination antibody titer, there was a significant difference in the immune response between persons with little or no detectable antibody before vaccination and persons with antibody present at greater than the lowest test dilution. Thus, the immune response of a vaccinated group as a whole depends on what percentage of patients in the group have no detectable antibody before vaccination. Sputum HIA determinations were comparable to sputum neutralizing antibody concentrations in a small group of 8 patients. Sputum HIA did not correlate with the concentration of sputum protein or immunoglobulin types A or G. A 4-fold rise in sputum HIA against the influenza A antigen occurred in 31% of vaccinees with no detectable antibody before vaccination but was noted in only 6% of vaccinees with prevaccination sputum antibody present in greater than the lowest test dilution. A simple method of nonenzymatic homogenizaiton of viscous sputum samples is described that permits large-scale studies of local antibody in patients with chronic bronchitis.