ALL TOGETHER NOW: THE DEVELOPMENT AND USE OF HERMAPHRODITE BREEDING LINES IN ACTINIDIA DELICIOSA

There are compelling horticultural reasons for converting kiwifruit from dioecy to hermaphroditism, such as no male vines in the orchard, just fruit-bearing vines, and perhaps less reliance on bee pollination. But consumers evaluate fruit, not the particular reproductive system which produces the fruit, and hermaphroditic kiwifruit selections must be improved in fruit quality as well as agronomic traits. Development of hermaphroditic kiwifruit began in 1977. Actinidia deliciosa is not always strictly dioecious, and sometimes males planted as pollinisers in orchards carry bisexual as well as staminate flowers. Crosses between some of these fruiting males and the dominant female cultivar ‘Hayward’ resulted in rare hermaphrodite variants. The best of these, with fertile pollen in all flowers followed by big fruit, became the basis of a series of crosses from 1987 on, aimed at testing the inheritance and stability of the hermaphroditic trait and investigating the genetics of sex determination. There were 4 types of crosses – hermaphrodite × fruiting male, hermaphrodite selfed, hermaphrodite × hermaphrodite and female × hermaphrodite. The female × hermaphrodite and hermaphrodite × hermaphrodite crosses were superior, with higher average fruit weights and only female and hermaphrodite progenies, with average fruit weight in hermaphrodites equivalent to that in females. All these families were constructed within one kiwifruit population, derived from the first seed introduction to New Zealand. Next, selected hermaphrodites, with >90% viable pollen and >100 g fruit, were crossed with unrelated females from other accessions, selected for superior fruit quality and productivity. Production of only female and hermaphrodite progenies continued. Hermaphroditism is now being introgressed into elite lines. There are also compelling breeding reasons for converting kiwifruit from dioecy to hermaphroditism: selection can then occur on both ovule and pollen parents, so selection response is R_µ = i h² σ_P rather than R_µ = ½ i h² σ_P.