Long-term population effects of changes in atmospheric CO2 will be largely determined by reproductive effort. Our research objectives were to quantify variability in seed production and rate of maturation among individual Pinus taeda L. (Pinaceae) trees growing in elevated CO2 (ambient plus 200 μL L−1) since 1996. Estimating tree fecundity in nature is frustrated by the difficulty of counting seeds from individual trees and the need for long-term data. We have used a hierarchical Bayes approach to model individual tree fecundity, accounting for the complexity of experimentation in a natural setting over multiple years. The study presented here demonstrates large variability in natural fecundity rates and contributes to our understanding of how both interannual variation and population heterogeneity influence elevated CO2 effects. We found that trees growing under elevated CO2 matured earlier and produced more seeds and cones per unit basal area than ambient grown trees. By 2004, trees grown in high CO2 had produced an average 300 more seeds per tree than ambient grown trees. Although there was a trend toward decreasing mean CO2 effect (difference in fecundity between elevated and ambient treatments) over time, the hierarchical analysis indicates that this decrease comes from the emergence of a few highly fecund ambient grown trees by 2002, rather than acclimation or downregulation among the fumigated trees. The most important effect of increased CO2 in forest ecosystems may be the increase in fecundity reported here. Although biomass responses can sometimes be large, the increase in fecundity can have long-term impacts on forest dynamics that transcend the current generation.