We study the arterial and venous circulation of the normal leg by strain gauge plethysmography and venous occlusion (thigh tourniquet). We propose the application of a simplified linear physical model of the venous circulation. It helps to analyse the plethysmographic data recorded during and after the congestion. It ignores the arterial inflow and consider the post-occlusive venous volume decay in function of time as being monoexponential. The venous compliance (C) is measured when the volume has reached a steady-state level during the congestion (known pressure). The time-constant (T) characterizes the volume decay in function of time when the occlusion is released. The tourniquet is successively inflated with two levels of pressure (30 and 60 mm Hg) in order to check if the system is actually linear as predicted by the model. The venous outflow is not strictly monoexponential and the model is only suitable to describe the beginning of the curve. The compliance does not behave linearly, the values measured at 30 mm Hg, being higher than at 60 mm Hg ($ 26%). The time-constant T is slightly influenced by the level of pressures. The calculated resistance is therefore lower at low pressure. We also study the arterial inflow before and after the venous congestion (3 min, 60 mm Hg). We observe a post-venous occlusion hyperaemia (mean rest flow: 5.2%/min, mean hyperemic flow: 12.1%/min) followed by a drop of the inflow (mean minimal flow: 3.4%/min). We evaluate the quantitative influence of neglecting the arterial inflow on the computing of the venous properties. The simplification appears acceptable.