The striking opportunities for competition in the pollination stage of the life cycle of flowering plants have been remarked on for many years, yet it is still not clearly known how much genetic variation for the rates of growth of pollen tubes exists in natural populations. Here we review the evidence for such variation, and discuss some of the possible mechanisms by which this thype of variation could be maintained. If faster pollen tube growth rates tend to be correlated with higher fitness of the propeny sired, then most of the variation would be expected to be due to detrimental mutations, and therefore subject to purifying selection. If, however, variation is maintained in populations, it is likely that genetic factors for fast pollen tube growth would reduce fitness at other stages of the life cycle, resulting in negative genetic correlations between pollen and progeny quality. Future studies of pollen competition and its effects on progeny fitness components should be designed to avoid confounding effects of such factors as inbreeding depression and self-incompatibility types, enabling these possibilities to be distinguished.