It is widely held that most of the oxidation in thermally sprayed coatings occurs on the surface of the droplet after it has flattened. The evidence in this paper suggests that, for the conditions studied here, oxidation of the top surface of flattened droplets is not the dominant oxidation mechanism. In this study, a mild steel wire (AISI 1025) was sprayed using a high-velocity oxy-fuel (HVOF) torch onto copper and aluminum substrates. Ion milling and Auger spectroscopy were used to examine the distribution of oxides within individual splats. Conventional metallographic analysis was also used to study oxide distributions within coatings that were sprayed under the same conditions. An analytical model for oxidation of the exposed surface of a splat is presented. Based on literature data, the model assumes that diffusion of iron through a solid FeO layer is the rate limiting factor in forming the oxide on the top surface of a splat. An FeO layer only a few thousandths of a micron thick is predicted to form on the splat surface as it cools. However, the experimental evidence shows that the oxide layers are typically 100x thicker than the predicted value. These thick, oxide layers are not always observed on the top surface of a splat. Indeed, in some instances the oxide layer is on the bottom, and the metal is on the top. The observed oxide distributions are more consistently explained if most of the oxide formed before the droplets impact the substrate.