Advances in recombinant DNA technology have made possible the advent of a new generation of safer, better-defined subunit vaccines. Because vaccines based on these weakly immunogenic antigens require an adjuvant for efficacy, we undertook the development of a safe and efficacious adjuvant suitable for widespread human administration. Vaccines formulated with aluminum salts (alum), the only adjuvant thus far utilized with vaccines approved in the United States for human administration, were necessarily adopted as a benchmark for the minimum acceptable activity of a new adjuvant. Our goal was to develop an adjuvant that significantly exceeded aluminum hydroxide in potency, while retaining equally low toxicity. By the early 1990s a wide variety of approaches to adjuvant development had been described (Allison and Byars, 1990; Edelman, 1980; Gregoriadis and Panagiotidi, 1989; Warren et al., 1986). Two major mechanisms of adjuvant activity have been repeatedly cited in this literature: the depot effect, whereby long-term release of antigen results in increased immune response; and coadministration of immunostimulators, which specifically activate portions of the immune system in, as yet, incompletely defined fashions. The prototypic strong adjuvant, complete Freund’s adjuvant (CFA), combined these functions by releasing a mixture of immunostimulatory mycobacterial cell wall components along with antigen from a water/mineral oil/Arlacel A emulsion depot over an extended period of time (Freund, 1956). CFA remains the reference standard for potent adjuvant activity; however, it is now considered too toxic in many cases for use even in laboratory animals. In addition to the aluminum salts, several adjuvants based on the depot effect alone have been studied. These include incomplete Freund’s adjuvant (IFA), which lacks the potent, but toxic, cell wall components, and Adjuvant 65 (water/peanut oil/mannide monooleate), a yet further detoxified water/oil formulation (Hilleman et al., 1972a,b). Despite extensive study, neither formulation was approved for human administration. We chose to avoid water/oil emulsions. A more recent version of the depot approach, controlled release of antigen from synthetic polymer microspheres, remains a promising area of study (Cohen et al., 1991; O’Hagan et al., 1991) but appears to have an unacceptably long development time for our purposes.