First, we have stlldiect the relations between 'the dynamics of calcium (Ca) and strontium (Sr) in different ecosystems on acid and calcareous soils, in order to evalllate if Sr could be used as a tracer of Ca. Variations of Ca and Sr concentrations were strongly and linearly correlated at each level of the ecosystems. Ho"Yvever, the slope of the correlations varied in each compartment of each ecosystem. In general, raot Sr/Ca ratio was lower to leaf Sr/Ca ratio, because of a preferential upward transfer of Ca versus Sr. Hence leaf litterfall was enriched in Ca in comparison to raot litter. This process, together with a preferential adsorption of Ca on the exchange sites of organic horizons, determined a preferential accumulation of Ca versus Sr in the soil surface horizons. In acid soils, the Sr/Ca ratio of spruce raots was lower than the Sr/Ca ratio of the soil exchange eomplex, whieh suggested a preferential uptake of Ca. This study confirmed the general analogy between Ca and Sr, but underlined the limitations of its use in ecosystem studies. Then, we have used the variations of Sr isotopie ratio to evaluate the dynamies of minerai elements in fore st ecosystems under boreal, tropical and temperate climate. In French Guyana, in two soil units differing in soil texture, we have measured the Sr isotopie ratio of three fore st species differing by their rooting depth and pattern. In the two units, the isotopie signature of bioavailable Sr inereased strongly with soil depth. But the isotopie signatures of al! speeies were similar to that of bioavailable Sr in the uppest sail horizon, which demonstrated that deep roots did not provide nutrients to the tropical ra in forest. In the North of Syveden, 'vve have measured the isotopie ratio of trees and sa ils in a pine and spruce mixed stand grovving on a peat and a podzol. Root isotopie signatures '\;vere higher at depth. Pine organs isotopie ratios were lower than that of spruce. Moreover spruce 1'Oot in the podzol E horizon had much higher isotopie ratio th an pine roots. A model of Sr circulation was llsed and showed that the interspeeific differences in Sr is6topie ratio were rdated to the intensity) of Sr cycling in litterfaH, and to a specifie weathering in the rhizosphere of spruce. Measurements a lso showed a deeper water and Sr uptake of pine roots. In the Vosges Mountains, we have studied the time develoDment of water and mineraI stress in a spruce plantation. Betvveen 1950 and 1980, before fore;t decline, tree ring èl l3C decreased eontinuously, suggesting an ,inerease of }vater stress. After 1980, 6 13C values stabilis~cJ and direct, measurements showed a clecrease in transpiration and water stress in relation to needle loss. During the last ten years, a set of budget showed a deerease in soil base saturation at that site. Between 1950 and 1990, tree ring Sr isotope ratio decreased. A model was used and simulated a similar decrease wh en the leaching of exchangeable Sr was increased, or when roots were allowed to take up nutrients in more superficial soil layers. These two mechanisms are very likely to have caused the mineraI defieiencies of the stand and may have increased its sensitivity to water stress. These results show the advantage of combining isotopie measurements and dynamie modelling 111order to understand the dynamics of mineraI elements in forest ecosystems.