Experimentally induced models of insulin-dependent diabetes are often not completely devoid of endogenous insulin. This is an important consideration when claims of an insulin substitute are being investigated. To test the efficacy of antidiabetic plants using STZ-induced or alloxan-induced diabetes in rodents, a convenient procedure is to commence plant therapy within a few days of STZ/alloxan administration before hyperglycemia becomes severe.Efficacy can then be judged by a slower progression and less severe hyperglycemia. If the study is continued until a parallel placebo (untreated) group develops ketoacidosis and requires insulin, this suggests that antidiabetic activity in an insulin-dependent state. However, it is possible that the therapeutic intervention has prevented complete -cell destruction.This can be seen if insulin concentrations are measured and animals survive when the intervention is discontinued. Because some natural regeneration of islets can occur from islet remnants, long-term survival cannot be exclusively attributed to the therapeutic intervention. An alternative protocol to test efficacy in an insulin-dependent state is to introduce plant therapy to spontaneously or experimentally induced models which have already developed severe hyperglycemia and are controlled by exogenous insulin injections. When plant treatment is introduced and the dosage titrated up, evidence of reduced hyperglycemia or a reduction of insulin dosage without deterioration of glycemic control can be used as indices of efficacy