Aims/hypothesis Previous work has demonstrated that beta cell amount (whether measured as beta cell mass, beta cell volume or insulin-positive area) is decreased in type 2 diabetes; however, recent findings suggest that mechanisms other than death may contribute to beta cell failure in this disease. To better characterise beta cell mass and function in type 2 diabetes, we performed morphological, ultra-structural and functional studies using histological samples and isolated islets. Methods Pancreases from ten non-diabetic (ND) and ten matched type 2 diabetic organ donors were studied by insulin, glucagon and chromogranin A immunocytochemistry and electron microscopy (EM). Glucose-stimulated insulin secretion was assessed using isolated islets and studies were performed using independent ND islet preparations after 24 h exposure to 22.2 mmol/l glucose. Results Immunocytochemistry showed that the fractional islet insulin-positive area was lower in type 2 diabetic islets (54.9± 6.3% vs 72.1±8.7%, p <0.01), whereas glucagon (23.3 ± 5.4% vs 20.2 ± 5.3%) and chromogranin A (86.4±6.1% vs 89.0±5.5%) staining was similar between the two groups. EM showed that the proportion of beta cells in type 2 diabetic islets was only marginally decreased; marked beta cell degranulation was found in diabetic beta cells; these findings were all reproduced after exposing isolated ND islets to high glucose. Glucose-stimulated insulin secretion was 40-50% lower from type 2 diabetic islets ( p <0.01), which again was mimicked by culturing non-diabetic islets in high glucose. Conclusions/interpretation These results suggest that, at least in subgroups of type 2 diabetic patients, the loss of beta cells as assessed so far might be overestimated, possibly due to changes in beta cell phenotype other than death, also contributing to beta cell failure in type 2 diabetes.