Using market prices for crude-oil futures options and the prices of their underlying futures contracts, we estimate the volatility skew in two ways. As a benchmark for our theoretical model, on each date we first estimate a cross-sectional polynomial structure for each maturity to demonstrate the strength and weaknesses of a purely-mechanical model. We then apply to the empirical data a Merton-style jump-diffusion model, with a rich structure of cross-sectional constraints on the parameters. Both models are tested with respect to their mark-to-market accuracy over time, as well as their efficacy in hedging intertemporal option price changes. The postulated Merton-style model is shown to yield useful parameters from which market prices can be computed, option prices can be marked-to-market and (imperfectly) hedged, as well as an informationally-rich structure covering the time period of the turbulent past six months.