The relative merits of either barium enema or colonoscopy for investigating lower gastrointestinal tract symptoms is still unclear. We studied the value of double contrast barium enema (DCBE) as the initial evaluation modality. We reviewed our 10-year experience of double contrast enemas as read by consultant radiologists. The study also aimed to identify which lesions are usually missed. We reviewed clinical data for all patients who underwent DCBE within the 6 months prior to surgical resection of colorectal cancer between April 1989 and April 1999. Patient demographics and tumour characteristics were analysed for their effects on the likelihood of the lesions being missed at DCBE. There were 706 patients included in the study, 54.2% were male and the mean age was 63.7 years (SEM=0.5 years). The site along the colon and rectum of tumours missed by DCBE corresponded with the frequency of tumour occurrence at each site. The overall rate of missed lesions was 4.1% (29 of 706 patients); these patients were found on subsequent endoscopy to harbour cancer. Tumours less than 3 cm in length and with lesser extent of circumferential involvement were more likely missed at DCBE (p=0.05 and p=0.01, respectively). Age, sex, and tumour grade and stage were not significant predictors of the likelihood of missed lesions. Of the 29 patients with missed lesions, 77.2% had a serum concentration of carcinoembryonic antigen (CEA) above the normal range (3.5 µg/l). The mean follow-up was 65.3 months (SEM=1.8 months). The overall survival for this series was 60.1%. The inaccuracy of the initial DCBE was not found to cause statistically significant differences in the stage of the tumour at diagnosis nor the overall survival of the patients in our series. Smaller cancers without circumferential involvement may be missed when DCBE is performed to evaluate lower gastrointestinal symptoms. Further evaluation by colonoscopy must be recommended when symptoms persist, especially in the context of a raised CEA level.