We studied the transport and deposition of quasi-monodisperse nanoparticles (with mobility diameter d(m) ranging from 1 nm to 52 nm) inside an extrathoracic airway model (which showed all relevant features of a computed tomography based model) by numerical simulation and, for part of the size range (14.5 nm <= d(m) <= 52 nm), by experimental measurement. For each particle size, three physiologically relevant steady inspiratory flow rates (Q=10, 20 and 30 l/min) were considered. The validation of the numerical results consisted of (i) a comparison of numerical flow results to available Particle Image Velocimetry and Large-Eddy Simulation data and (ii) a comparison of numerical particle deposition simulations to new nanoparticle deposition measurements as well as previously published experimental and numerical deposition data. Computations of the laminar-turbulent flow were carried out using an open-source steady-state solver in conjunction with the k-omega SST turbulence model. The particle motion and their subsequent deposition was computed by the so-called Eulerian mass transport model. Quasi-monodisperse carbon nanopartides (14.5 nm <= d(m) <= 52 nm) were generated by spark ignition and their deposition was measured in a cast of the extrathoracic airway model. Both simulations and experiments showed that particle deposition increases with decreasing flow rate and decreasing particle size. A direct comparison between our numerical and all available experimental deposition results in fact showed a good overall agreement in the entire flow and size range. In the particle size range that could be studied experimentally in the same extrathoracic airway model geometry (d(m) >= 14.5 nm), absolute differences were of the order of a few percent, in the face of very low deposition fractions in this nanoparticle size range (with all experimental deposition fractions less than 5%). However, when normalizing to the largest deposition for the smallest flow rate (Q = 10 l/min), the dependency on flow rate was very similar between experiments and simulations. An adjustment of the turbulent Schmidt number (0.9-0.5) led to a slightly 'better agreement with experimental values. Finally, the simulations also showed that within the extrathoracic model, particular deposition patterns appeared that were similar for different particle sizes, however with more marked deposition hot-spots for the smallest particle size. (C) 2012 Elsevier Ltd. All rights reserved.