The small white flowers of sapodilla are borne on short pedicels in the leaf axils. Piatos and Knight (1975) found sapodilla to be self-incompatible, indicating that cross-pollination might be necessary. While production of some cultivars such as 'Prolific' is high in areas such as Florida, fruit yield among cultivars is variable, a characteristic which may be due to self-incompatibility. Gonzalez and Feliciano (1953) examined flowering and fruit set in 'Ponderosa' sapodilla and found that the trees could not be self-pollinated, either naturally or artificially. The best method of pollination was determined to be cross-pollination by hand, which resulted in a 39.6% fruit set. Open pollination resulted in only 5% fruit set in that study. Similar results were observed in 'Kalipatti' sapodilla by Relekar et al. (1991). Mulla and Desle (1990) found that the highest percent fruit set was observed in trees cross-pollinated with pollen of the same cultivar and ranged from 20 to 34%. There was no sign of parthenocarpy and self-pollination resulted in only negligible fruit set in some of the cultivars examined.

Reddi (1989) conducted a series of studies in which he showed that thrips ( Thrips hawaiiensis Morgan and Haplothrips tenuipennis Bagnall) are the principle pollinators of sapodilla in India. The thrips apparently take shelter in the flowers and live on the pollen grains, nectar, and stigmatic exudations. They collect pollen grains while feeding on these components and transfer then to other flowers when the food reserves are exhausted. Bees have been observed in flowering sapodilla plantings (Sambamurty and Ramalingam 1954), although examination of the insects indicated that they were not carrying pollen.

Reddi's study showed several traits which indicated that wind is not an important pollinating agent, nor are large insects. Sambamurty and Ramalingham (1954), however, suggest that wind is an important factor in sapodilla pollination. Reddi also showed that sapodilla flowers are not self-pollinating, although pollen transfer is generally limited to a single tree unless trees are closely spaced. This information would provide evidence that sapodilla is not self-incompatible as previously suggested. Pollen size and viability are quite variable between cultivars, possibly causing the varied results in fruit set (Minhas and Sandhu 1985).

Much like avocado, sapodilla produces many more flowers than developed fruits. The great variability in fruit set may be due to differences in flower abortion or fruit drop, although differences in this phenomenon between cultivars has not been specifically examined in sapodilla. The major period of fruit drop occurs in the first five weeks following fruit set and as little as 1.6% of the flowers produced by a tree may develop into fruit (Relekar et al. 1991). Fruit set is highly variable, even within a cultivar. Gonzalez and Feliciano (1953) suggested that tree vigor may be related to flower production and fruit set.

Source :http://www.hort.purdue.edu/newcrop/proceedings1996/V3-439.html