Abstract The most unpredictable and uncontrollable wildfires are those that burn in the crowns of live vegetation. The fuels that feed these crown fires are mostly live, green foliage. Unfortunately, little is known about how live fuels combust. To understand how live fuels burn, piloted ignition experiments were performed with lodgepole pine and Douglas-fir. The thermal behavior (thick versus thin) of both live and dead needles was explored. Both live and dry needles were shown to behave as thermally intermediate solids in this apparatus. Additionally, samples were collected throughout the growing season to take advantage of the natural variation in moisture content and chemical composition. This data set was then compared to several correlations found in the literature to determine whether live fuel ignition can be predicted by moisture content alone and to test the applicability of these models to the wildland fire problem. Many of the correlations from the literature for ignition time with moisture content fail to capture the trends with live fuels. A linear regression of the ignition time with moisture content only predicts 74–80% of the variability suggesting that there is another mechanism controlling ignition time of live fuels. Based on the hypothesized difference in water storage between live and wet dead fuels, the chemical composition of the live needles was included in an empirical model for ignition time. Including chemical composition improved the prediction accuracy for Douglas-fir needles only. Because the thermal properties of live foliage are largely unknown, it is possible that the predictions from more physically-based models would show improvement with more accurate values of density, thermal conductivity, and specific heat.