Understanding plant reproduction requires knowledge of genetic contributions through pollen and seeds. Since direct genetic assessments of fitness through pollen are often intractable, reproductive ecologists use components of male fitness such as pollen production and pollen removal as surrogates tot paternity. However, we know little of the strength of the relationship between these components and actual paternity. Here, I report on a study undertaken to examine the relationship of pollen production and removal with paternity in Fragaria virginiana, a wild strawberry. A morphological marker was used to track paternity in experimental arrays exposed to native pollinators. Relative pellen production proved to be a poor predictor of relative paternity in most arrays, and over all arrays there was no significant correlation between relative paternity and relative pellen production. In contrast, relative pollen removed correlated significantly and positively with proportion of seeds sired, suggesting that a plant's contribution to the pool of removed pellen is a good predictor of its male reproductive success. Deviations from expected paternity based on relative pollen removal suggest a systematic overestimation of the siring success of plants with low pollen removal. And, in at least one specific case, low pellen removal may be explained by delayed anther dehiscence, which could lower the effectiveness of the removed pellen.