RESUMO
Twenty-four freshwater sampling sites (11 river water, 6 spring water, and 7 groundwater) were selected from 4 sampling areas located in the northern and central parts of Taiwan. A total of 125 water samples were collected during a 5-month sampling period, and the numbers of total heterotrophic bacteria, total coliforms, fecal coliforms, enterococci, Aeromonas hydrophila, and Salmonella spp. were enumerated. Besides the traditional membrane filtration method, total coliforms and Escherichia coli were also simultaneously enumerated using the Colilert (Quanti-Tray/2000) method. On average, 94% and 80% of the water samples assessed with the Colilert method had equal or higher total coliform and E. coli counts, respectively, as compared with the membrane filtration method. Furthermore, when m-FC agar was used to enumerate fecal coliforms, 18% of the samples failed to yield the typical bluish colonies, while E. coli were counted in the same samples using the Colilert method. The data indicate that the m-FC agar culture method is inadequate for the enumeration of fecal coliforms in subtropical water samples. Significant correlations were observed between the total number of bacteria and various indicator bacteria in river water samples, but no such correlations were found for groundwater and spring water. This finding suggested that the river water was polluted by anthropogenic sources. The counts of total coliforms, E. coli, and other indicator bacteria were significantly correlated in all river water samples, while in groundwater and spring water, significant (p<0.01) correlation was only observed with enterococcal counts. The presence of total coliforms/E. coli generally implies the presence of fecal pollution possibly including pathogenic enteric bacteria. However, no Salmonella spp. were detected in any of the 107 water samples analyzed. The results of this study suggest that the use of these commonly employed microbial indicators for assessing subtropical water quality, especially in a pristine body of water (ie, mountain spring and groundwater), is highly questionable.