Although it is generally accepted that extraterrestrial material is the source of metals in the upper atmosphere, the relative abundances of mesospheric metal atoms and ions present us with a conundrum. Lidar observations have consistently shown that the abundances of neutral metals in the atmosphere and the abundances of these metals in the meteoric material that falls to Earth are significantly disproportionate. The column density of neutral sodium is perhaps 2 orders of magnitude larger than that of calcium, while the abundances in meteorites are approximately equal. By contrast, ion mass spectroscopy has shown that the abundances of the meteoric ions match reasonably well those in the meteorites. We present here a model that attempts to address these discrepancies. At the heart of the model is the concept of differential ablation, which suggests that more volatile metals sublimate earlier in the descent of a cosmic dust particle than do the less volatile components. We model three different meteoric metals: sodium, magnesium, and calcium. Results suggest that sodium ablates to a greater extent than does calcium and that it ablates at a substantially higher altitude. Deposition at lower altitudes leads to more rapid conversion of the atomic calcium into complexes through three-body reactions. Thus the depletion of calcium arises from both a decrease in deposition and an increase in the rate of removal of that which is deposited. We examine the behavior of the model in several respects, comparing predicted results with measurements and finding reasonable agreement. We argue that the success of this model indicates that differential ablation is a key factor in the determination of the relative abundances of meteoric metals in the mesosphere. ¿ 1998 American Geophysical Union