BACKGROUND Experimental deposition was studied using three different mouth-throat models: (1) the standard United States Pharmacopeia induction port (IP), (2) the idealized human mouth and throat replica developed by the University of Alberta (UofA replica), and (3) the conductive rubber mouth-throat cast from a human subject developed by Lovelace Respiratory Research Institute (LRRI cast). METHODS Both solid and liquid monodispersed fluorescent particles in the size range of 2-30 μm in diameter were delivered into the devices at flow rates of 15, 30, and 60 L min(-1). For solid particles, the study was conducted with and without grease coating inside the devices to investigate the effects of particle bounce. CONCLUSIONS Large amounts of rebounded particles were found for the IP and UofA replica without the coating treatment, while particle bounce was only observed at the large particle size for the LRRI cast. The UofA replica and LRRI cast agreed well for solid particles with coating treatments and liquid particles. The deposition results from this study were also compared to data of in vivo deposition studies from the literature. The deposition efficiencies in the UofA replica and LRRI cast were within the range of in vivo data, which showed a large scatter.