The objective of this research is to improve the comprehension of the hydrological behaviour of natural catchments. The main originality of this work is to associate different types of measurement in order to obtain a better vision of hydrological processes responsible for streamflow generation. First, the hydrological behaviour is studied at the catchment scale by the application of environmental tracing. A three-component mixing model based on the silica and calcium concentrations of water allows one to distinguish the contributions of direct precipitation, soil water and groundwater during flood generation. Despite the different hydrological responses observed between the four subcatchments studied, a common behaviour is apparent. Soil contribution increases with a rise in the basin humidity. The subsurface water dominates the generation of major floods, which occur in wet conditions. In order to discover the processes responsible for the important soil water contributions, a large-scale time-domain reflectometry experiment (64 probes) was conducted. On the whole, this experiment indicates that the water flow in soil is spatially quite heterogeneous and depends on local properties. Macropore flows were clearly identified during a rainfall simulator experiment. Preferential flows may be responsible for the important contribution of soil water and the heterogeneity of the soil moisture. In order to test this hypothesis, a dye-tracing experiment was done. This new investigation confirms that an important part of soil water reaches the stream by preferential flows. So as to synthesize all these observations, a conceptual model is proposed. This model respects both the hydrochemical responses highlighted by the environmental tracing experiment and the observations done at the local scale. This conceptual model suggests that the important contribution of soil water is due to the extent of the hydrographic network and the role of preferential flows.