Changes in oxidized (HbO), reduced (Hbr), and total hemoglobin (Hbt) contents were monitored by near infrared spectroscopy in human forearm skeletal muscle during oscillatory variations of the effective venous pressure. Laser diode pulses (wavelength, 775, 810, 865, 904; duration, 60 ns) were directed to the muscle by means of an optic fiber bundle and detected with a sensing fiber positioned at 1.5 cm from the emitting bundle. Sinusoidal pressure waves (frequency, 1 and 2 cycles/min; amplitude, 10-15 mm Hg) generated by a piston pump were transmitted to a sphygmomanometer cuff placed on the arm, the mean cuff pressure (Pc) being 20, 40 and 60 mm Hg. Variations of HbO, Hbr and Hbt were computed from the optical signals and processed by Fourier analysis to characterize their amplitude and phase relative to the cuff mean pressure oscillations (Posc). Oscillations of HbO, Hbr, Hbt were observed for all mean cuff pressures, the amplitude of the variations being decreased with increasing Pc. For Pc equals 20 mm Hg, the oscillations of HbO and Hbt were nearly in phase with Posc whereas the oscillation of Hbr were out of phase with HbO and Posc. Increasing Pc resulted in an increase of the phase difference between HbO and Posc, Hbr remaining out of phase with Posc. These trends could be predicted with a lumped model of the forearm vasculature, suggesting that the technique could be used to asses mechanical characteristics of vascular beds.