In order to evaluate the response at both the gametophytic and sporophytic level of a selection based on the pollen competitive ability and to compare its effects with those obtainable from a conventional sporophytic procedure, three recurrent selection plans were developed in maize starting from the same F2 population. Two gametophytic recurrent selection procedures at high (GSH) and low (GSL) selection intensity were performed by utilizing, to advance the populations, kernels taken from the base (GSH) or apex (GSL) of ears obtained from pair-crosses of randomly chosen plants. The third scheme was a sporophytic full-sib recurrent selection procedure (SS); the only selection criterion was the machine-harvestable grain yield of the families. In a sixyear period of selection, six cycles of both GSH and GSL and three cycles of SS were performed. The source and the selected populations (16 entries) were tested for pollen performance and for sporophytic traits. The selection cycles advanced through GSH showed a progressive increase, as compared to GSL, in pollen tube length measured at 4 h of in vitro culture. The SS cycles response was intermediate at 4 h whereas at 2 h it exceeded both GSH and GSL. A slight decrease in pollen diameter was evidenced in populations advanced with GSL procedure. The SS selection caused a marked increase for grain yield, lateness, leaves per plant and plant height. No response was shown by gametophytic selection for grain yield. The GSH procedure, however, led to an increase in kernel weight and to a decrease in kernel moisture, leaf number and plant height, as compared to GSL. Though gametophytic selection showed limited effects on sporophytic traits, it can be considered an efficient tool to supplement conventional sporophytic selection.