Thirty-three trees were identified from commercial orchards in Queensland during the 1990–91 and 1991–92 seasons to investigate the potential of improving the agronomic and fruit quality aspects of Kensington Pride mango. These selections were grafted and planted in a replicated field trial in the Burdekin district for comparative evaluation. Cropping characteristics were recorded annually until the trees reached 10 years of age. Small yield variations were observed between the selections, with no significant differences between the top 19�selections. No single selection had outstanding yields, however 2 selections had significantly lower yields than the rest. These findings indicate the difficulty in significantly improving Kensington Pride mangoes through selection from existing commercial germplasm. Sixteen of the best performing selections from the field trial were used as a top group to investigate the cropping characteristics of the cultivar. The average annual increase in yield and fluctuations from this average were modelled. The first significant crop was 15.7 kg per tree in year 4 and modelled yields increased by 23.3 kg per tree per year thereafter to year 10. There were significant annual fluctuations from the modelled trend with yields varying from between –44.7 to +35.7% of the expected yield. As fruit number per tree increased, average fruit weight decreased by 0.14g for every extra fruit.