The Baltic Sea became connected to the open sea about 8500 years B.P., and that marked the beginning of the still continuing Litorina Sea stage. There have since been significant long-term variations of the salinity from a maximum of some 10--15? to the present of 7--8?. The salinity variations have been known since the beginning of the twentieth century, and various mechanisms have been proposed to explain these. However, so far, no one has actually tried to quantify and thereby rank the relative importance of the different proposed mechanisms. In this paper we used a simple oceanographic model to quantify the impact of changes in cross-sectional areas of the inlets to the Baltic. We found that such changes cannot explain the full variation of the salinity. We propose that the remaining salinity variations (0.5--5?) are explained by changes in net freshwater input in the range 15--60%. The calculated variations in the freshwater supply are in large coherent with known variations of the northern European climate. We show that other possible factors as increased mixing, frontal displacements, higher deep water salinity, and larger depths in Kattegat and larger oscillating flows across the sills due to increased sea level variability are less likely to have made substantial contributions to the salinity variations in the Baltic Sea.