The main purpose is to examine if radiation, considered as the dominant mechanism of fuel bed preheating, can account for experimental results obtained from laboratory fires conducted in no-wind conditions. For this, two one-dimensional radiative models for fire spread in a porous forest fuel bed are derived from albini's model (1986), The sensitivity of these two models to uncertain input parameters, particularly the emissive powers of the ignition interface and of the flame, are shown. The predictions of spread rates from the two models are then compared against the observed values for different fuel loads and slopes for two kinds of pine litter. In most cases, the observed effects are poorly predicted when fixed emissive powers are used. Thus, in a second step, the necessary values of emissive powers to be used for predicting the observed effects are researched. Although reasonable values of emissive powers are apparently required in most cases, it is suggested that the heat loss due to free convection in the energy balance should be replaced by convective heat transfer depending on the gas temperature profile, in order to give better estimates of the observed effects.