A retrospective study has been made of all adult patients admitted to Manchester Royal Infirmary with exocrine pancreatic disease between 1968 and 1974, in order to define the factors which influence the variable mortality and morbidity rates in published accounts of patients with acute pancreatitis. The most plausible explanation is that some series with low mortality rates include a variable number of patients with relapsing acute pancreatitis and acute exacerbations of chronic pancreatitis. Both these pathological entities have a negligible mortality and morbidity rate compared with single attacks of acute pancreatitis. The difficulties encountered by the clinician in determining the prognosis of acute pancreatitis at the time of admission to hospital are discussed. Attention is drawn to the differing role of alcohol as an aetiological factor in relapsing chronic pancreatitis and acute pancreatitis.