In this study we examine the implications of excitation saturation on fluorescence recovery after photobleaching (FRAP) experiments. In particular we present both experimental and theoretical evidence that fluorescein, one of the most frequently used fluorophores in FRAP, does not always comply with the basic assumptions that are made in many FRAP models: an invariant bleaching illumination intensity distribution (BID) in combination with first-order photobleaching kinetics. High light intensity levels, which are typical for the photobleaching phase of FRAP experiments, can cause excitation saturation of fluorescein in the excited triplet state. We show by experiments and computer simulations that under such saturating conditions the higher-order diffraction maxima of the BID substantially contribute to the photobleaching process and can no longer be neglected. As a result, the bleached regions are larger than expected theoretically from the FRAP models. Although this effect is not always directly evident from the FRAP experiments, neglecting it may shift the calculated diffusion coefficient by as much as over one order of magnitude. We present a discussion on the implications of this saturation effect on various types of FRAP models.