For the isolation and cultivation of Legionella pneumophila from tap water in hospitals, we compared different media and selection techniques. A second part of the study compared the L. pneumophila yields from different water samples at identical sites. A total of 210 water samples (500 ml each) were collected from two selected sites in each of 21 hospitals. Warm water samples were collected after flow times of 0, 5, 10, and 15 min; in addition, one cold water sample was collected. Filtration was used to concentrate all samples. Following filtration, 0.1 and 1 ml each of untreated samples, heat-treated samples (3 min, 59 degrees C), and acid-treated samples (pH 2.2, 15 min) were spread onto the selective media MWY (SR 118; Oxoid) and BMPA alpha (SR 111; Oxoid), and samples from 12 hospitals were also spread onto GVPC medium (SR 152; Oxoid). A total of 72 (34%) of the 210 samples from 12 hospitals were positive. With respect to the positive Legionella cultures, there was no significant difference between the selective media MWY, BMPA alpha, and GVPC. With the BMPA alpha supplement, more samples were positive following heat treatment (P < 0.05) or acid treatment (P < 0.05) than without any further treatment. For the maximum yield of Legionella colonies with minimum additional microbial flora, acid treatment was the most effective, and by all methods, the GVPC supplement was the most selective. For routine water tests in hospitals for differentiating between systemic and local contamination, acid treatment of the concentrated samples, the use of different selective media, and the correct selection of sampling sites are recommended.