One of the major problems in the use of heat flow data in quantitative manner has been the variability of closely space measurements. It is suggested that this variability is directly related to hydrothermal circulation in an 'effectively permeable' oceanic crust. As a consequence, only where this crust is sealed from the seawater by an extensive and thick layer of impermeable sediment can reliable estimates be made of the flux at depth. Heat flow data from such areas are analyzed. It is shown that the scatter in the data is low and that in most areas the mean heat flow through a province of known age is close to that predicted by plate models which account for the increase in depth with increasing age of the ocean floor. From this analysis it is argued that these heat flow means are a reliable estimate of the flux at depth. In order to be able to use heat flow measurements to investigate the age of the crust, the thickness of the lithospheric plate, upper mantle convection, and local anomalies, it is necessary to substantiate this analysis and to find an adequate explanation of the low mean heat flow observed at the center of the sediment bulge in the equatorial Pacific.