The concentrations of Legionella pneumophila in cooling towers may vary considerably over short periods of time, producing significant fluctuations throughout the year. Despite genetic variability, in small geographical areas the same indistinguishable pulsed-field gel electrophoresis patterns may be shared among different cooling towers and persist over time.The involvement of cooling towers in some community outbreaks of legionellosis has been demonstrated (3,6,8,9,11), and it has been suggested that there may be an association between high Legionella counts in cooling towers and the occurrence of an outbreak of legionellosis (12). Although not all environmental isolates are associated with human disease, the present guidelines aimed at preventing legionellosis have proposed ranges of risk involving different environmental insults based on Legionella counts (4). However, the fluctuations in Legionella counts over short periods of time have led some authors to question these risk ranges (1).Since the recovery of a high inoculum of Legionella in a cooling tower is not necessarily related to an outbreak or sporadic cases of legionellosis, genetic studies are necessary to establish the link between environmental and clinical isolates. The "gold standard" technique for genotyping Legionella is the analysis of chromosomal restriction patterns by pulsed-field gel electrophoresis (PFGE). Isolation of Legionella strains from a cooling tower which are genetically indistinguishable from clinical isolates is considered to be clear evidence of the implication of the tower in an outbreak (1, 9). However, little is known regarding the genetic variability of Legionella within the same tower and in cooling towers from areas in the vicinity. Likewise, to our knowledge, no study has determined the persistence of molecular patterns in isolates in cooling towers. All these circumstances have epidemiological and legal consequences when assigning responsibility for an outbreak of legionellosis to a particular installation.The main objectives of this study were to establish the fluctuations in Legionella counts over time and to determine the genetic variability and persistence of PFGE patterns in cooling towers.Fifteen cooling towers (A to O) were selected within a radius of 3 km. Four of them (J, K, M, and N) shared the same main line for water distribution. The other cooling towers had a different domestic distribution system but were supplied by the same municipal water network. All the cooling towers followed the controls and hygiene measures required by Spanish regulations (RD 865/2003), and we were able to verify that the levels of disinfectant (sodium hypochlorite [0.2 to 0.5 ppm residual] in conjunction with the discontinuous dosage of nonoxidant biocide [dibromonitrilopropionamide]) were in accordance with legislation at the time of sampling. Cooling water samples were cultured for Legionella fortnightly over a 1-year period. The samples were concentrated by filtration, and four plates of GVPC agar (Oxoid, Wesel, Germany) we...