The compositional variation of plagioclase and the partitioning of major elements between plagioclase and melt have been experimentally measured as a function of the cooling rate. Crystals were grown from a basaltic melt at a pressure of 500MPa under (i) variable cooling rates of 0.5, 2.1, 3, 9.4, and 15°C/min from 1250°C down to 1000°C, (ii) quenching temperatures of 1025, 1050, 1075, 1090, and 1100°C at the fixed cooling rate of 0.5°C/min, and (iii) isothermal temperatures of 1000, 1025, 1050, 1075, 1090, and 1100°C. Our results show that euhedral, faceted plagioclases form during isothermal and slower cooling experiments exhibiting idiomorphic tabular shapes. In contrast, dendritic shapes are observed from faster cooled charges. As the cooling rate is increased, concentrations of Al+Ca+Fe+Mg increase and Si+Na+K decrease in plagioclase favoring higher An and lower Ab+Or contents. Significant variations of ^pl–liqKd are also observed by the comparison between isothermal and cooled charges; notably, ^pl–liqKd_Ab–An, ^pl–liqKd_Ca–Na and ^pl–liqKd_Fe–Mg progressively change with increasing cooling rate. Therefore, crystal–melt exchange reactions have the potential to reveal the departure from equilibrium for plagioclase-bearing cooling magmas. Finally, thermometers, barometers, and hygrometers derived through the plagioclase–liquid equilibria have been tested at these non-equilibrium experimental conditions. Since such models are based on assumption of equilibrium, any form of disequilibrium will yield errors. Results show that errors on estimates of temperature, pressure, and melt-water content increase systematically with increasing cooling rate (i.e. disequilibrium condition) depicting monotonic trends towards drastic overestimates. These trends are perfectly correlated with those of ^pl–liqKd_Ca–Na, ^pl–liqKd_Ab–An, and ^pl–liqKd_Fe–Mg, thus demonstrating their ability to test (dis)equilibrium conditions.