Nowadays, most water meters are mechanical and intended to be installed on pipes completely filled with water. But the pipelines of a water supply network may contain air, which poses a metrological problem: if this air flows through the domestic intakes, it can propel the moving part of meters, resulting in an overestimation of water consumption. Of how much? There is a surprising lack of field data on this topic. So, the case of one house is reported: it is located at the top of a steep and sparsely occupied street, with water typically supplied for a few hours per day. The house's meter (multi-jet) was estimating a huge and erratic consumption: several times more than what would be normally expected on average, and with some daily peaks exceeding the built storage capacity (underground cistern plus roof tank). After one year of monitoring, including the installation of a few devices, it is concluded that: (1) the house's meter was affected by air in the water supply network (most likely for different reasons, of which three are discussed); (2) a small air-release valve installed just upstream from the meter did not solve the problem; (3) another mechanical meter (single-jet) installed just downstream was also affected by air (although to a lesser extent), and (4) reliable estimates of water consumption were finally obtained with an ultrasonic meter installed at the domestic intake (and with a mechanical meter installed at the roof tank's outlet). Thus, the case reported emphasizes the need to study more how air in pipelines affects mechanical water meters, and to sometimes consider alternatives for measuring domestic water consumption.