Fourteen recirculating cooling water systems were surveyed during the summer, 1981, to see what factors might influence the prevalence of Legionella pneumophila. The effect on the organism of three anti-microbials was studied, each in two systems, by intermittent treatment at two week intervals. L. pneumophila was isolated from six of the 14 cooling systems at the beginning of the trial but by the end was present in ten. An association was found between the presence of the organism and the concentration of dissolved solids, and chlorides and the pH. There also appeared to be associations with exclusion of light and higher water temperatures. Repeated tests on eight untreated systems showed that two were consistently infected, three became and remained infected, one was infected on a single occasion and two were never infected with L. pneumophila. Treatment of a contaminated system, either with a 10 p.p.m mixture of a quaternary ammonium compound and tributyltinoxide or slow release chlorine briquettes (maximum recorded free chlorine level 1.2 p.p.m.), did not eliminated legionellae. Treatment of two infected towers with a chlorinated phenol (100 p.p.m.) eliminated legionellae for at least three days, but after 14 days the organism was again found.