The use of pollen ratios to estimate the fluctuations of the late-glacial and holocene tree-limits and the timber line is discussed. An attempt is made with fossil data from some sites from the eastern part of the Pyrenees. Comparisons with the records of Pinus stomata indicating local occurence of Pinus show a good correlation between the two complementary approaches and emphasize the value of the pollen ratio. However, in all cases, the components of the ratio have to be discussed. For reconstructing the timber line, local floristic studies are necessary so that the most appropriate pollen indicators can be chosen.