In deciding whether the precaution of removing the bollard should have been taken, the trial judge relied on a statement in a safety audit report that one bollard “is unlikely to slow cyclists” and rejected the evidence of the Council’s engineer that the bollard was not removed because it would tend to slow down cyclists: [40], [42], [46]. Whereas two bollards may have a greater deterrent effect on cyclists, it does not follow that one bollard will have no significant effect on cyclists. If one bollard is a hazard, it must also be an obstacle which may deter speeding cyclists: [53].

Even on the assumption that two bollards had social utility, but one bollard did not, it did not follow that one bollard constituted an unreasonable “hazard” and required removal in the exercise of taking reasonable care for the safety of cyclists using the pathway. The plaintiff did not establish that the Council had failed to take reasonable care for her safety in failing to remove the bollard at some point between 2016 when the first bollard was removed, and the accident in 2020: [56]-[57].