Family studies suggest a strong genetic component in the aetiology of non-insulin dependent diabetes (NIDDM), with evidence for a major gene of co-dominant or dominant effect. A gene-dosage effect, whereby diabetes develops earlier in people with two susceptibility genes than in those with one susceptibility gene is likely. The search for the diabetes gene has led to the cloning and characterization of many genes involved in controlling glucose homeostasis. These include the insulin, insulin receptor, glucose transporter, amylin and glucokinase genes. Molecular techniques have permitted rapid screening of these genes in NIDDM patients and controls. There is now a rather contradictory genetic literature for NIDDM, with weak disease associations reported and refuted for most candidate genes. However, pedigree analyses and DNA sequencing of available candidate genes and their regulatory regions have failed to implicate any of these in the common form of diabetes, NIDDM. Methodical application of random clones in well-defined NIDDM families may be the strategy of choice in finding the NIDDM genes, given the wide range of genes potentially involved in the glucose and lipoprotein metabolic disturbances seen in NIDDM.