• Diurnal and seasonal cycles of PE rate over three different saturated soil textures were assessed using a unique lysimeter. • The textures showed different PE sums, with the largest differences during daytime in spring and summer. • PE differences over textures were reproduced well by the energy balance method. • Simplifying the estimate of ground heat flux results in large deviations from measured PE. Calculating actual bare soil evaporation (E a ) based on potential evaporation (PE) is a widely followed approach in many disciplines such as hydrogeology, hydrology and agricultural sciences. The influence of soil texture on PE is rarely considered in these approaches. In this work, seasonal and diurnal cycles of PE over saturated bare soils were assessed using a unique lysimeter experiment in the Guanzhong Basin, China. The assessment was made for three different soil textures including fine sand (PE fine ), coarse sand (PE coarse ) and gravel (PE gravel ). Meteorological variables, soil heat flux and soil temperatures were measured at a high temporal resolution for more than 14 consecutive months. Potential evaporation rates over saturated bare soil showed clear differences between fine sand, coarse sand and gravel on an annual basis. PE fine was higher than PE coarse and PE gravel (7.3 % and 11.0 % respectively). The differences between measured PE rates over different surfaces were especially pronounced during daytime in spring and summer, but showed minor differences in autumn and winter. These results are quantitatively explained with detailed calculations of the surface energy balance and it is found that differences in available energy over the soil textures, related to different albedos, as well as different porosities and thermal properties for the materials (which influence soil temperatures) explain the differences. This shows that PE is different for different soil textures, which is neglected in most hydrological studies. Nevertheless, the full-form Penman-Monteith equation can reproduce the PE differences over soil textures quite well in autumn and winter, but a simplified approach to calculate the ground heat flux does not allow to reproduce PE differences between textures.