Residential water demand is studied with microdata and with allowance for (1) seasonal differences in price elasticity, (2) a dynamic adjustment process, (3) a marginal price specification, (4) cross‐price effects between peak and off‐peak demand, and (5) the inclusion of a detailed set of household demographic variables including accurate measures of age distribution and household income. The findings are that the peak (summer) price elasticity of demand is more than twice the off‐peak elasticity; cross price effects are important at the 5% level of significance; variables measuring household income, property value, related property features, and age distribution are simultaneously significant; and finally, peak period adjustment rates are found to be less than off‐peak rates.