Fertility of Triploid Highbush Blueberry

Eight highbush blueberry (V. corymbosum L.) triploids (2n = 3x = 36) were crossed with diploids (2n = 2X = 24), tetraploids (2n = 4x = 48), and hexaploids (2n = 6x = 72). No plants were recovered from 4021 3x × 2x crosses. One triploid was relatively fertile in 3x × 4x and 3x × 6x crosses, which is most likely attributable to 2n gamete production in the triploid. The lack of fertility of triploids, which do not produce 2n gametes, in crosses with diploids and tetraploids suggests that the production of gametes with numerically balanced (n = 12 or 24) chromosome numbers is extremely low. In addition, the inability to recover progeny from 3x × 2x crosses also suggests that aneuploid gametophytes and/or zygotes, including trisomics, are inviable in blueberry. Pollen stainability was also highly reduced in triploids. Frequency distributions of anaphase I pole chromosomal constitutions of three triploids were significantly different from one another. Two of the three distributions were shifted toward the basic chromosome number of 12, with one triploid having 25% poles with 12 chromosomes. However, the sterility of 3x × 2x and 2x × 3x crosses indicates that lagging chromosomes during meiotic anaphases are probably not excluded from gametes, resulting in unbalanced gametes in blueberry. Triploids can be used as a bridge to facilitate gene transfer from the diploid and tetraploid levels to the hexaploid level in blueberry. Aneuploids have been useful in genetic studies involved with gene localization and for gaining greater insight into the basic nature of the genome, with trisomics having been particularly useful (Khush, 1973; Khush et al., 1984). The most common and dependable source of trisomics in many plant species have been triploids. The frequency of trisomics recovered from trip- loid by diploid crosses depends on the species and, in many cases, the direction in which the cross is made (Levan, 1942). In some species, however, triploid by diploid crosses are either completely unsuccessful, as in aster (Avers, 1954), or produce only a few diploid progeny, as in watermelon (Kihara, 1951). Aneuploid series between the tetraploid and hexaploid levels have been developed in blueberry (Vorsa, 1988; Vorsa et al., 1986). However, the usefulness of these aneuploids for local- izing genes and linkage groups is limited since these aneuploids are not phenotypically distinguishable from one another or from euploids (Vorsa et al., 1986). Polyploidy is considered to have a buffering effect on the imbalance caused by aneuploidy (Khush, 1973). Thus, the four-genome minimum present in these aneu- ploids may obscure any aneuploidy effects. A trisomic series derived from triploids in blueberry could be useful in gaining a better understanding of the genetics of this crop. Triploids in blueberry have only recently been realized (Dweikat and Lyrene, 1988; Megalos and Ballington, 1988; Vorsa, 1990). A cross between a complex tetraploid (2n = 4x = 48) hybrid derived from largely V. corymbosum L. ancestry and a wild diploid (2n = 2x = 24) V. corymbosum clone resulted in eight triploid (2n = 3x = 36) and three tetraploid hybrids (Vorsa, 1990). Three additional triploids derived from two unrelated crosses of tetraploid by diploid V. corymbosum were also available (Megalos and Ballington, 1988). The purpose of our study was to determine the potential value