Three quarters of the seven million cases of infectious tuberculosis diagnosed in the world are in developing countries. While the disease is curable, half a million people die of it each year. Problems of diagnosis and treatment relate more to access to services and money for drugs than to anything else. CIETmexico provides training in tuberculosis diagnosis to medical graduates performing community service (pasantes) and also to voluntary health promoters from the communities. A project in the late 1980s evaluated the different methods of case identification in remote rural communities in Guerrero. The diagnoses made by medically trained personnel were compared with those of local volunteers with only brief training. Both these methods compared favourably with the exising system of passive surveillance based on voluntary attendance at health centres. Quality control was studied in 94 health centres that send sputum samples to state laboratories; no false positives or false negatives were found among samples analyzed by voluntary promoters or pasantes. The cost, the number of positive diagnoses and efficiency of each method was measured and compared. It was shown that, where voluntary promoters take part in the surveillance, cases were diagnosed earlier in life and a greater number of cases found. The cost of diagnosing a case of TB by voluntary promoters proved to be three times less than the pasantes. A further survey of students in the faculties of Medicine and Chemical Sciences illustrated that their knowledge of and attitude towards the diagnosis of tuberculosis by baciloscopy was strongly conditioned by their undergraduate training.