Family and twin studies have shown clearly that Type I (insulin-dependent) diabetes mellitus has a genetic basis. However, only within the past decade has it been possible to systematically attempt to identify the genes that increase susceptibility to this disorder using linkage and association analysis of genetic markers distributed across the genome. More than 20 putative diabetes-predisposing genes have been localised in addition to HLA region susceptibility genes detected more than 25 years ago. Unfortunately, the effects of most diabetes-predisposing genes are weak, with the exception of HLA region susceptibility genes (which contribute about half of the genetic risk). The overall effects of diabetes-predisposing genes could be weak because the susceptibility gene occurs in only a small proportion of diabetic patients or the susceptibility gene (although it might be common) produces only a modest increase in risk, probably by acting in concert with other such genes to cause disease. The weak effects of these genes have made them difficult to locate, difficult to confirm in independent studies and difficult to isolate by genetic procedures. This paper summarizes the major challenges that have faced geneticists in mapping Type I diabetes genes, and reviews the progress achieved to date. The rewards of the genetic studies will be twofold: increased understanding of the causes of Type I diabetes, facilitating creation of preventative therapies, and enabling clinicians in the future to use genetic information to predict which children are predisposed to diabetes in order to target them for preventative therapies.