mum values of the Z scores (standardized values), and by a large coefficient of variation. These skewed variables are total number of fruit, total yield, blossom density, fruit set, fruit density, yield density, cluster size, and fruit size. The skewness of the pawpaw population in the Wye collection towards floriferousness, higher fruit set larger fruit, and greater yields reflects the origins of our accessions from the historic collections of cultivars. It is particularly significant that blossom and yield traits are so highly skewed, as pawpaw is notorious for low yields, a major obstacle to commercial production.
Many, if not all, of the problems outlined in the earlier section have potential genetic solutions within the gene pools of Asimina triloba and the other Asimina species. Those other species are relevant because the transfer of traits from them to A. triloba has been proven realistic and practical. Kral (1960) reports 6 different hybrid combinations of the Florida dwarf pawpaw species (A. tetramera excluded) as occurring naturally, and as being common where human disturbance has brought into contact different species; he reports that hybrid individuals appeared fertile. Zimmerman (1938, 1941) reports obtaining fertile hybrids of A. triloba (female) with A. obovata, A. incana and A. longifolia (syn, A. angustifolia). Swartz crossed A. pygmaea with A. triloba pollen, producing seed which failed to germinate (Peterson, 1986). In spring 1989, the author easily crossed A. triloba with pollen from A. obovata, A. reticulata and A. parviflora, and crossed A. parviflora with pollen from A. triloba; the success rate was about 50 percent, followed by normal fruit and seed development (germination has not been ascertained as of the time of this writing). Thus, all species of Asimina have been shown to cross easily with A. triloba except A. tetramera which has not been tested. The potential contributions of all 8 species to pawpaw improvement are summarized in Table 5.
Another possibility for improvement in pawpaw exists in wider crosses between Asimina and other genera in Annonaceae, notably Annona. A succinct and provocative discussion of the breeding possibilities in Annonaceae was published by Clift (1977), although others have also mentioned the possibility of using the Annona genus in the improvement of A. triloba (Amer. Genet. Assn., 1917; Zimmerman, 1940). The traits Clift identified within Annona as potentially valuable were fruit quality (especially flavor and aroma), adaptedness to dry soils, tolerance of saturated soils, bright skin colors, bright flesh colors, thick skin, compact growth habits, and hardiness. The degree of difficulty to be encountered in achieving Asimina x Annona crosses is unknown. Taxonomically, Annona belongs to the same tribe as Asimina, Unoneae, but to the other subtribe, Annonineae (Hutchinson, 1964). In addition to different climatic adaptations, the two genera differ most noticeably in the structure of their fruits: Annona fruits are compound, whereas Asimina fruits are simple. Both genera are very similar in pollination biology: flowers are strongly protogynous with
591