As the study of type 1 diabetes moves towards preventive therapy, the role of adjuvants needs to be addressed. Incomplete Freund's adjuvant (IFA) is thought of as "immunologically inert" as, unlike complete FA (CFA), it has no components designed to provoke an immune response. We investigated the effect of IFA as an immunomodulator on the disease process leading to type 1 diabetes in the non-obese diabetic (NOD) mouse. 24 NOD mice were injected intradermally (i.d.) at 8 and 12 weeks of age with a 1:1 mixture of IFA and saline; 24 controls received saline alone. Splenocytes were tested against antigens thought to be involved in the disease process, namely insulin, a GAD peptide, a beta-casein peptide, a Glut-2 peptide and concanavalin A (ConA) as a non-specific antigen. In the IFA experiment diabetes incidence was 13% compared to 38% in the controls (p < 0.05). In vitro, splenocytes from IFA treated animals showed non-specific immunosuppression with ConA (p < 0.01), whereas the response to 1-casein and Glut-2 was raised in IFA treated animals with respect to controls. ELISA using supernatants from IFA treated animals, showed a typical Th2 cytokine pattern, whereas controls showed a Th1 pattern. In conclusion, IFA alone can reduce diabetes incidence in the NOD mouse apparently by modulating the immune response towards beta-cell related specific antigens. As IFA has been adopted as an adjuvant in preventive trials in the NOD mouse, this might have implications for the interpretation of previous and future results.