In long-lived clonal plants, the overall size of a clone is often used to estimate clone age. The size of a clone, however, might be largely determined by physical or biotic interactions, obscuring the relationship between clone size and age. Here, we use the accumulation of mutations at 14 microsatellite loci to estimate clone age in trembling aspen (Populus tremuloides) from southwestern Canada. We show that the observed patterns of genetic divergence are consistent with a model of increasing ramet population size, allowing us to use pairwise genetic divergence as an estimator of clone age. In the populations studied, clone size did not exhibit a significant relationship with microsatellite divergence, indicating that clone size is not a good proxy for clone age. In P. tremuloides, the per-locus per-year neutral somatic mutation rate across 14 microsatellite loci was estimated to lie between 6 x 10(-7) (lower bound) and 4 x 10(-5) (upper bound).