In commercial fruit tree orchards, consistently high yields are necessary for a durable economy. The Swedish pear cultivar 'Carola' has been noted for low setting in some orchards, possibly due to insufficient pollination. In this study, fruit set was evaluated in a research orchard where `Carola' had been planted together with four potential pollinators. Total yield and number of fruits was noted during three and four years, respectively. In 2003, seeds were germinated from the harvested `Carola' fruits, and the paternity of three seedlings from 50 trees was determined with RAPD analysis. 'Clapp's Favourite' had sired 39.6% of the seedlings, closely followed by `Seigneur d'Esperen' (30.7%) and 'Clara Frijs' (26.7%) whereas 'Skanskt sockerparon' only sired 1.1% of the seedlings. The remaining 2.3% appeared to have been derived by selfing. Pollen-limited seed set was indicated at surprisingly short distances; accumulated number of fruits on the `Carola' trees was significantly higher when separated by only 2 m from one of the two most efficient pollinators, 'Clapp's Favourite' or 'Seigneur d'Esperen‘, compared to trees 4—l0 in away in the same row. Number of viable seeds per fruit was also higher in fruits from trees immediately adjacent to the pollinators, suggesting an effect of improved pollination success. The importance of very short inter-cultivar distances for efficient pollen transfer became even more clear when comparisons involved the true pollination distances as determined by RAPD; the accumulated yields decreased linearly from 55 kg at a 2 in distance to only 17 kg at 13 m.