Abstract-A popular model for ureilites assumes that during anatexis in an asteroidal mantle, pressure-buffered equilibrium smelting (partial reduction coincident with partial melting) engendered their conspicuous mafic-silicate-core mg diversity (75-96 mol%). Several mass-balance problems arise from this hypothesis. Smelting inevitably consumes a large proportion of any plausible initial carbon while generating significant proportions of Fe metal and copious proportions of CO gas. The most serious problem concerns the yield of CO gas. If equilibrium smelting produced the ureilites' entire 21 mol% range in olivine-core mg, the proportion of gas within the asteroidal mantle (assuming plausibly low pressure <∼80 bar) should have reached ≥85 vol%. Based on the remarkably stepwise cooling history inferred from ureilite texture and mineralogy, a runaway, CO-leaky process that can loosely be termed smelting appears to have occurred, probably triggered by a major impact. The runaway scenario appears likely because, by Le Châtelier's principle, CO leakage would tend to accelerate the smelting process. Also, the copious volumes of gas produced by smelting would have led to explosive, mass-leaky eruptions into the vacuum surrounding the asteroid. Loss of mass would mean diminution of interior pressure, which would induce further smelting, leading to further loss of mass (basalt), and so on. Such a disruptive runaway process may have engendered the ureilites' distinctive reduced olivine rims. But the only smelting, according to this scenario, was a short-lived disequilibrium process that reduced only the olivine rims, not the cores; and the ureilites were cooling, not melting, during the abortive "smelting" episode.