Diagnosis is always difficult when a rare condition gives rise to a common symptom. The Zollinger-Ellison syndrome causes symptoms which are often indistinguishable from those of peptic ulceration, yet it affects only about one in 1000 ulcer patients. The first major advance in the management of this condition was the demonstration by Gregory et al. that the syndrome was due to hypersecretion of gastrin [1], and the subsequent development of a radioimmunoassay for gastrin. Fasting serum gastrin assay (performed after anti-secretory drugs have been stopped for at least 48 hours) has since been used as the mainstay of diagnosis, but although serum concentrations in excess of 1000 pg/ml are diagnostic, concentrations are usually lower and in about 30 per cent of cases they overlap the range of gastrin concentrations found in uncomplicated duodenal ulceration [2]. The excessive production of gastrin results in a 4-6-fold increase in parietal cell mass [3] and there is typically a high basal acid secretion (> 15 mmol HCl/hour) which is greater than 60 per cent of the maximal acid output after administration of exogenous pentagastrin. In some cases, however, there may be a misleading response to pentagastrin, as reported by Collins et al. who describe their experience of 16 patients with Zollinger-Ellison syndrome [4]. In such cases the secretin provocation test may be very helpful: intravenous administration of exogenous secretin usually results in a marked increase in serum gastrin and gastric acidity in patients with Zollinger-Ellison syndrome while having little if any effect on duodenal ulcer patients. The effect of intravenous secretin on serum gastrin on its own may be unreliable, as shown by Collins et al., but an acid output of > 18 mmol/hour after intravenous secretin is highly suggestive of the diagnosis [5]. The diagnosis should be considered in any patient with peptic ulceration who has persistent diarrhoea: this was a major feature in 38 per cent of patients in the series reported here. It should also be excluded in patients with multiple or unusually distal ulcers, ulcers that are resistant to conventional medical therapy and in patients with recurrent ulcers after surgery. The median duration of symptoms of 5.5 years before diagnosis reported by Collins et al. typifies the difficulties in diagnosis and, in view of this delay and the fact that about 60 per cent of gastrinomas are malignant, it is not surprising that surgery has rarely proved curative. As a result the standard policy for many years has been to recommend total gastrectomy, paying little attention to finding and removing the primary tumour. This has changed radically since the introduction of potent antisecretory drugs.