Biodegradable implants made from cholesterol and lecithin (C:L) were used to deliver a recombinant antigen (recombinant Dichelobacter nodosus pili) and adjuvant (Quil A) to sheep. Implants (5.5-×1.8-mm) were placed subcutaneously and compared to a conventional vaccination regime (2 injections, 4 weeks apart) for antibody responses and tissue compatibility. Release profiles of antigen and adjuvant were also studied in vitro and in vivo. The presence of Quil A in vaccine implants had a marked effect on the rate at which antigen was released with 29 and 44% being released in the first 24 h from implants containing pili alone and pili with Quil A, respectively. Sheep produced significant levels of antibody when immunized with implants, however the response was short-lived and of significantly lower intensity than the response stimulated by two injections of antigen with Quil A (P<0.01). A second implant system was developed where implants coated with C:L to delay antigen release, were used in combination with uncoated implants to deliver a priming dose and boosting dose of antigen. Antibody titres stimulated by the double implant system were equivalent to those stimulated by a conventional regime of two injections (four weeks apart) for the first six weeks of the experiment. After this time, antibody levels in the groups which received implants dropped significantly. In vitro studies revealed that some of the implant coatings had caused a delay in the release of antigen (the rate of release peaked at 72 h), however this was not long enough to provide a significant boosting effect. In all cases, implants were well tolerated by sheep and caused less local reaction than injected vaccines.