Despite much experimentation precipitation of anhydrite from solution in conditions similar to those occurring in sedimentary environments has not yet been reproduced in the laboratory. To resolve this long-standing contradiction we have monitored the precipitation and stability behaviors of calcium sulfate during experiments lasting up to two years. Calcium sulfate was precipitated from solution between 40 and 120 °C at three different salinities and the formed solid phase was sampled at different time intervals (from 2 min up to 2 yr). We found that below 80 °C gypsum is the sole primary phase and in the range of 80 to 120 °C gypsum and bassanite are the primary phases. The stability of the latter increased with increasing salinity. As expected, we did not observe primary anhydrite precipitation, but over time phase transition occurred and anhydrite eventually appeared at temperatures > 80 °C. We show that intrinsic thermodynamic and kinetic properties severely constrain the precipitation of anhydrite (compared to gypsum and bassanite), and consequently, a considerable amount of time (e.g. > 2 yr at 60 °C) is needed for anhydrite to form. Even so, at a geological time-scale, anhydrite can be considered as a pseudo-primary phase thus resolving the long-standing paradox of our inability to directly precipitate anhydrite in the laboratory at temperatures below 120 °C and the abundant presence of anhydrite in evaporitic environments. Our results also show that at low water activity, bassanite becomes an important phase, which could be relevant to explain its presence on the surface of Mars.