A series of experiments to determine the survival characteristics of enviro
nmental and faecal coliforms in a 1.3 km long experimental pipe distributio
n system is described. In the first experiments, tertiary treated effluent
(10(3)-10(4) coliforms ml(-1)) was inoculated directly into the distributio
n system. Coliform organisms were not detected in any samples taken downstr
eam of the inoculation point. By comparison, laboratory jar tests showed lo
w level survival for coliforms at the same chloramine residuals (0.3 mg l(-
1)) for up to 48 h. In the second series of experiments, a by-pass pipe in
the experimental distribution system was isolated, drained, and filled with
tertiary treated effluent (10(3)-10(4) coliforms ml(-1)) to simulate the c
onditions in a dead-end. Coliform numbers were monitored and found to decre
ase rapidly, but they were still detectable at low levels after 7 weeks. Th
e water in the by-pass section was then released into the main pipe-rig and
sampled downstream. No coliforms were detected in water samples or in samp
les swabbed from the pipe walls. Finally, the flow in the main pipe-rig and
in the by-pass was increased to dislodge any deposits and biofilm into the
water. The absence of coliforms in any water samples taken during or after
the flow increases is contrary to the widely accepted understanding that c
oliforms survive and grow in biofilms on pipe walls.