We report on a 32—yr study of a population of Spanish Imperial Eagles, Aquila adalberti, which increased during the first 16 yr of study but remained stable during the last 16 yr. We analyzed changes in the mean and variance of fecundity in relation to population density to test predictions of two hypotheses of density—dependent fecundity. According to the interference" hypothesis, as density increases, frequency of agonistic encounters increases, resulting in a relatively uniform decrease in habitat quality. Consequently, mean fecundity decreases, and no relationship is expected between density and variance in fecundity. For the habitat heterogeneity hypothesis, however, as density increases, a greater proportion of individuals are forced to occupy lower quality habitats. Thus, mean fecundity decreases and fecundity variance must increase. Additionally, for this hypothesis, fecundities in good sites are expected to be equal in both low— and high—density situations. An inverse relationship between fecundity and population size was found in this eagle population. Annual variance in productivity showed significant increases over the study period. This trend was inversely related to mean productivity. Variance in productivity was related to the year of pair establishment, being higher in recently occupied territories. Mean and variance of the longer term territories remained constant during the study period. These results are in accordance with the habitat heterogeneity hypothesis.