Whole-tree isoprene emission rates of a chambered white oak were measured throughout the growing season to develop a seasonal emissions model and to compare with emissions estimated using current leaf algorithms and forest canopy models. Emissions increased from mid-May to a maximum in mid-July, and the highest rates occurred late in June and throughout July. Rates decreased from early August to mid-October. A model developed to characterize the whole-tree emissions accounted for 80% of the variability of observed emissions over the growing season. Peak isoprene emissions and photosynthesis occurred at the same time, but peak emissions continued longer, suggesting that peak photosynthetic rates were not necessary for peak isoprene emissions. Measured light intensities 2 m and 3.3 m down into the tree canopy corresponded to intensities estimated by using current canopy models; however, intensities at 1 m were 14% lower than predicted by modeling. Measured median leaf-to-air temperature differences were 2.0 ¿C at the canopy top and 0.5 ¿C or less in the canopy. Median values of leaf temperatures estimated using a leaf energy balance procedure were slightly lower than air temperatures at all canopy levels, although differences were not more than 0.9 ¿C. The most recently developed biogenic emissions model, which assumes that leaf and air temperatures are the same, predicted July whole-tree emission rates fairly closely, although high emission rates were slightly underpredicted using the model. Leaf temperature adjustments used in previous canopy models were applied to this model, and in this case, predicted rates underestimated measured rates when measured rates exceeded 80 μg C g-1 h-1. ¿ 1997 American Geophysical Union