A detailed comparison of exptl. obtained, spatially and temporarily resolved NO concn. profiles in an SI engine with different modeling approaches was performed. The primary goal was to evaluate potential exptl. problems assocd. with the use of laser-induced fluorescence (LIF) for quant. NO measurements under engine conditions. Quant. measurements from engine runs with propane fuel using two different approaches for LIF detection of NO in IC engines were assessed in terms of sensitivity to temp., fluorescence quenching, and interference by mol. oxygen and flame intermediates. It was found that results obtained with a scheme using excitation near 248 nm and detection at 230 +- 10 nm allow a direct comparison with modeling results because spatial variations of temp. and quenching corrections are negligible. Using 225-nm excitation, a significant distortion of the profiles close to the flame front was revealed due to an addnl. contribution of O2-LIF signal. Calcd. NO profiles were obtained from a one-dimensional computation with detailed chem. of 56 species using the MixFla code as well as using a simple model including only the Zeldovich reactions and a variable prompt-NO contribution. Computations with the full kinetics code were performed for fixed initial conditions, while the simple model allowed implementation of the time history of the temp. in the unburned gases as well as of pressure to simulate engine conditions. While the peak NO concn. of the full calcn. and the measurements agree well, the spatial distribution is very different. The exptl. detd. flat NO distribution in the center of the flame is not captured by the model. The simple model on the other hand yielded a spatial profile similar to the measurements. However, the abs. NO no. densities were obtained by scaling the results to exptl. data for the middle of the combustion chamber and adopting the prompt-NO contribution from the detailed kinetics calcn.