Monte Carlo simulation of photon migration in tissue was used to assess the sampling depth, measuring depth and photon pathlength in laser Doppler flowmetry. The median sampling depth and photon pathlength in skin, liver and brain tissue were calculated for different probe geometries. The shallowest median sampling depth found was 68 microns for a 120 microns diameter single fibre probe applied to a one-layered skin tissue model. By using separate transmitting and receiving fibres, the median sampling depth, which amounted to 146 microns for a 250 microns fibre centre separation, can be successively increased to 233 microns when the fibres' centres are separated by 700 microns. Total photon pathlength and thereby the number of multiple Doppler shifts increase with fibre separation, thus favouring the choice of a probe with a small fibre separation when linearity is more important than a large sampling depth. Owing mainly to differences in the tissue g-value and scattering coefficient, the median sampling depth is shallower for liver and deeper for brain, in comparison with skin tissue. For skin tissue, the influence on the sampling depth of a homogeneously distributed blood volume was found to be limited to about 1 per cent per percentage increase in tissue blood content, and may, therefore, be disregarded in most practical situations. Simulations show that the median measuring depth is strongly dependent on the perfusion profile.