The adjuvant activity of Bordetella pertussis was investigated, both at the cellular and humoral levels, when the bacterial adjuvant was given at various times after the primary antigenic stimulus of both 2 x 10(7) (suboptimal dose) and 4 x 10(8) (optimal dose) of sheep erythrocytes. In all experiments, both adjuvant and sheep erythrocytes were administered by the intraperitoneal route. Adjuvant activity was measured on the basis of the early and late phases of the primary response and on the degree of priming for the secondary immune reaction. A maximal adjuvant activity was found in mice which had received B. pertussis vaccine simultaneously with the antigen. Adjuvant effectiveness became less as the time interval between the injection of antigen and adjuvant increased. Adjuvancy also depended on the amount of antigen used as the primary antigenic stimulus. With 4 x 10(8) sheep erythrocytes, significantly increased priming for the secondary response was produced only when B. pertussis cells were administered within a period of 24 hr. When the bacterial adjuvant was administered either 48 or 72 hr after the primary antigenic stimulus, adjuvancy was found to be limited to the late phase of the primary response and to the prolonged development of antibody-forming cells during the secondary immune reaction. In contrast, significantly enhanced priming for the secondary response was detectable when the adjuvant was administered as late as 48 hr after primary immunization with 2 x 10(7) sheep erythrocytes. When the bacterial adjuvant was administered either 6, 24, 48, or 72 hr after the primary immunization with 2 x 10(7) sheep erythrocytes, the early phase of the primary 19S and 7S hemolysin response was found to be suppressed, and adjuvancy became detectable only thereafter.