An apparent excess risk of all-cancer mortality among 473 of the largest US cities was found in relation to surface drinking water supplies. The increased risk for 100% surface water versus 100% ground water use was slight, about 2%, but statistically significant. This finding agrees with reports from several earlier studies in smaller geographical regions of the US, Great Britain and Canada. A relationship was further supported by the replication of this association within the larger of the 11 independent regions studied. Our data suggest that the association with surface water may be specific to cancer mortality. The increased risk would be expected to be greater than 2% if analyses were restricted to cancers of sites previously related to the use of surface drinking water.