This study was carried out in order to evaluate the variations in the physicochemical characteristics of the water in an experimental distribution system. The primary objectives of the study were : - to compare the disinfectant efficiency of chlorine and monochloramine - to establish quantitative correlations among bacterial density, concentration of residual disinfectant, and concentration of biodegradable organic matter. The finished waters were obtained from a water treatment pilot plant characterized by : prechlorination (average treatment rate : 1.4 mgCl2 l-1 and residual alter sand filtration : 0.08 mgCl2 l-1), coagulation-flocculation-sedimentation (FeCl3 treatment rate : 30 to 60 mg l-1 adjusted to the raw water turbidity below 0.3 NTU), sand filtration (filtration rate : 6 h-1) and post-disinfection with chloramine (average treatment rate : 1.8 mgCl2 l-1) or chloramine (average treatment rate : 1.66 mgCl2 l-1). The concentrations of post-disinfectant used were chosen in order to maintain chlorine at 0.2 to 0.5 mgCl2 l-1 and monochloramine at 1 mgC2 l-1 after the first 24 hours residence time in the experimental distribution system. The experimental distribution system is composed of three parallel loops connected in series (31 m length, 100 mm diameter, cement lined cast iron, water velocity : 1m s-1). The configuration and operation of the system permitted a residence time of 24 hours in each loop (that is 72 hours for the whole system). Appropriate sample tap locations facilitated removal of bulk water samples. Special sampling parts also permitted sampling of cement coupons for determination of attached biofilm. The measured parameters were : residual oxidant (DPD method), DOC, BDOC (28 days of incubation at 28 °C with a bacterial inoculum), cell density in the bulk water phase (CFU ml-1) and in the biofilm (CFU cm-2) after 15 days of incubation at 20-22 °C. Total cells were enumerated using the epifluorescence direct count technique. For each experiment, all the sampling sequences were carried out on each of three days, after quasi steady-state was achieved in the system (4 to 6 weeks after starting each experiment). The data were analysed in order to characterize the treated and distributed waters; the results discussed here are based on the averages of the measured parameters tram the water samples and biofilm samples taken after the system achieved quasi steady-state. Characteristics of the treated waters The treated waters were characterized by important variations at the DOC, BDOC and cell density. For example, the concentrations of DOC showed a seasonal variation ranging from 0.8 to 1.3 mg Cl-1 in winter to 1.6 to 2.6 mg Cl-1 in summer. The treated waters contained approximately the same concentrations of residual disinfectant, averaging 1.6 mgCl2 l-1 for chlorine and 1.5 mgCl2 l-1 for monochloramine. However, a significant difference cell density was found between the two post-disinfectants. Cell densities by the epifluorescence direct count technic were 1.6 x 103 ml-1 (0.3 % of CFU ml-1) with chlorine and 6.3 x 104 ml-1 (0.03 % of CFU ml-1) with monochloramine. The difference on cell density between the post-chlorination and the post-chloramination treatments has been observed systematically, and may be explained either as cellular lysis with chlorine or an interference when using epifluorescence counting for chlorinated waters. Characteristics of the distributed waters Whatever the season, depletion of oxidant (chlorine or monochloramine), and elimination of dissolved organic matter (DOC, BDOC) occured during the first hours of circulation of water in the distribution system. The net result was an increase in bacterial cell density. During the first hours of circulation of the waters in the distribution system, depletion of the disinfectant occured. Depletion was more rapid for chlorine (-0.05 mgCl2 l-1 h-1) titan for monochloramine (-0.02 mgCl2 l-1 h-1), winch is considered more stable titan chlorine. Even in the presence of a residual disinfectant in the distribution system, microorganisms are present in the water phase (104 to 106 cells ml-1 by epifluorescence direct count; 1 % CFU ml-1 after 15 days of incubation at 20-22 °C) and in the biofilm (105 to 106 cells cm-2 by epifluorescence direct count; 1 % CFU ml-1 after 15 days of incubation at 20-22 °C). The bacterial density increased white the disinfectant residual decreased. The apparent growth rate of the attached biomass (µfix) in loop 2 of the chlorinated distribution system (equivalent to 48 hours detention), was close to the µfix calculated for loop 2 of the chloraminated distribution system : the values were 0.0043 h-1 and 0.005 h-1 respectively. In addition, the change in the organic matter (expressed as DOC) occured in two steps : - a slight increase in DOC during the 24 first hours of residence time (loop 1), when increased residual disinfectant were present. - a bacterial consumption of DOC after 24 hours of residence time (loops 2 and 3), even in the presence of small concentrations of disinfectants. In loop 2 (48 hours residence time of the water in the system; chlorine : 0.01 mgCl2 l-1, chloramine : 0.3 mgCl2 l-1 h-1), the rates of DOC elimination averaged 13 mgCl2 l-1 h-1 and 0.42, mgCl2 l-1 h-1, respectively in the chlorinated and chloraminated distribution systems. This decrease in DOC concentrations was related to the increase in bacterial density. Relationships between cell density, oxidant and organic matter Linear relationships between the concentration of residual oxidant (chlorine or monochloramine) and the cell density in the water phase or in the biofilm show that : - whichever oxidant was used, the pipe loop sections without residual disinfectants were characterized by about 5 x 106 attached cells per cm2 (4 to 10% were able to grow on agar medium in 15 days at 20-22 °C) and by 4 x 105 planktonic cells per ml (1 % CFU ml-1); - consistently, in the sections of the system with a residual disinfectant, the bacteria (CFU and epilluorescence counting) in the water phase were more sensitive to the residual disinfectant (chlorine or monochloramine) than the bacteria attached to the pipe walls (biofilm). However, there was a difference in effectiveness between the two disinfectants; chlorine was more efficient in controlling planktonic bacteria and biofilm bacteria than was monochloramine. For example, to achieve complete inactivation of the planktonic bacteria (CFU ml-1) a constant chlorine residual of 0.5 mgCl2 l-1 was required throughout the whole distribution system, compared to 2.5 times more chloramine to achieve the same efficiency. Finally, with equivalent concentrations of residual disinfectant, the microbiological quality of the chlorinated distribution waters was better than that of the chloraminated distribution waters. From loop to loop, linear relationships between ∆DOC and cell density pointed out that the presence of biodegradable organic matter can explain the bacterial proliferation in the distribution system. For example, a concentration of BDOC as low as 0.1 mgC l-1 resulted in an increase in the cell density : an additional accumulation of 7.5 x 105 attached cells cm-2 and 4 x 104 planktonic cells ml-1 was observed in the experimental distribution system at quasi steady-state. Consequently, the control of the biodegradable organic carbon remains one of the prime objectives in order to achieve biologically stable distribution waters.