Seedlings in the 2019 high-density nursery photographed 22 Oct. 2020. The nursery was planted 30 May 2019 using greenhouse-grown seedlings from seed planted in Nov. 2018. The four weakest rows in the center contain F1 hybrids from highbush × tetraploid V. stamineum crosses. The more vigorous rows on the left contain seedlings obtained by intercrossing BC1 seedlings that had V. stamineum as the recurrent parent, and those on the right contain F2 seedlings from F1 × F1 crosses.
Spurlike branches on an upright cane of a BC1 seedling (highbush the recurrent parent) transplanted to the field nursery in May 2017. Photographed 4 Nov. 2020. The scale is in centimeters.
BC1 seedlings with inflorescence structure intermediate between that of highbush and V. stamineum. Peduncles elongate well before the flowers open. Photographed 22 Oct. 2020.
Clusters of berries on an F1 hybrid in the greenhouse showing long pedicels and peduncles. The flowers had been pollinated with pollen from a highbush cultivar. The plant had been defoliated by mites during berry development.
Loose clusters of berries on an F1 hybrid in the greenhouse after the flowers had been pollinated with pollen from a highbush cultivar.
Internal berry color of freshly sliced mature berries of tetraploid selection FL 15-669 (a highbush cultivar × tetraploid V. stamineum F1 hybrid) and highbush cultivar FL06-96.
Large berries from a diploid plant of V. stamineum (clone FL15-695) harvested from a forest in Alachua County, FL, July 2015. Pedicels remain attached to the berries when they abscise from the peduncle. The scale is in centimeters.
Click on author name to view affiliation information
Vaccinium stamineum (deerberry) is a highly variable diploid species in section Polycodium. Deerberry is native on excessively drained sandy soils from southeastern Ontario, south through the Florida peninsula to Lake Okeechobee, west to eastern Texas and southeastern Kansas. The V. stamineum used in this study were tall plants (2–4 m) native in north Florida, with a plant architecture similar to rabbiteye blueberry (V. virgatum). Starting in 2013 with crosses between tetraploid highbush cultivars (section Cyanococcus) and colchicine-doubled V. stamineum, hundreds of F1 and thousands of later-generation seedlings were grown and evaluated in high-density field nurseries at Citra in North Florida. The populations studied included F1, F2, backcrosses to each parent species, and BC1 × BC1 seedlings. The goal of the study was to assess the feasibility of introgressing into highbush blueberry cultivars desirable traits from V. stamineum (drought tolerance, red-flesh berries, new flavor components, open flowers with short corolla cups and exserted anthers and stigmas) without introducing horticulturally problematic characteristics (bitter skin, berries that shatter when ripe, difficult vegetative propagation). Vigor averaged very low in F1 seedlings, higher in F2 seedlings and in seedlings from backcrosses to V. stamineum, and highest in seedlings from backcrosses to highbush. Most crosses yielded numerous plump seeds, but crosses to produce F1 hybrids yielded fewer than 10% as many seeds as highbush × highbush crosses. Most vegetative, flower, and fruit traits that differentiate highbush from V. stamineum were intermediate in F1 seedlings. Backcross seedlings more closely resembled the recurrent parent. Variability in morphological characters was high in every generation, giving much opportunity for selection. Some seedlings from backcrosses to highbush (≈5%) appeared to have the vigor, berry quality, and yield potential required in commercial cultivars. Producing highbush cultivars that strongly express a particular V. stamineum trait might best be accomplished by growing large, segregating F2 populations from which parents for backcrosses can be selected.
Cultivated highbush blueberries are clones that were developed by crosses involving Vaccinium section Cyanococcus species. Section Cyanococcus includes northern and southern highbush cultivars (complex hybrids based on V. corymbosum and related Cyanococcus species), lowbush (V. angustifolium), rabbiteye (V. virgatum), and various undomesticated diploid, tetraploid, and hexaploid species native in eastern North America.
The genus Vaccinium has wide distribution, with numerous native species in the tropical highlands of both the Old and New World. Vander Kloet (1997) divided the genus into 35 sections. Several sections, including Cyanococcus, include diploid, tetraploid, and hexaploid species. Within sections, homoploid crosses are usually successful and produce vigorous, fertile hybrids. There is a strong triploid block, and intrasectional crosses between diploid and tetraploid species normally fail or give only a few seedlings. These may be triploid or tetraploid, with the tetraploids being the result of unreduced gametes from the diploid parent.
A few intersectional Vaccinium hybrids have been discovered or created. The most widely studied natural intersectional hybrid {[V. intermedium = V. myrtillus (section Myrtillus)] × V. vitis-idaea (section Vitis-idaea)} was first reported from northern Germany in 1826 (Ritchie, 1955). V. intermedium and both parent species are diploid. The hybrid, which has arisen at various times and places in Europe, is vigorous, sometimes even heterotic, and partially fertile (Hiirsalmi, 1977; Rousi, 1966). By contrast, most experimentally produced diploid intersectional Vaccinium hybrids have been weak and sterile.
At the tetraploid level, intersectional Vaccinium hybrids can be fertile. The most widely studied has been the cross between the highbush blueberry cultivar Rancocas and tetraploid V. uliginosum (section Vaccinium; Rousi, 1963). The hybrids were fertile enough to allow the production of thousands of seedlings in backcrosses to tetraploid highbush. Some, but not all, tetraploid F1 hybrids between highbush cultivars and colchicine-produced V. arboreum (section Batodendron) were both male and female fertile (Lyrene, 2013). Tsuda et al. (2013) obtained fertile hybrids by crossing colchicine-produced tetraploids of the Japanese native blueberry (V. bracteatum) with the highbush blueberry cultivar Spartan. Ehlenfeldt and Polashock (2014) reported fertile hybrids between highbush cultivars and tetraploid V. padifolium (section Hemimyrtillus). A self-fertile F1 hybrid between V. corymbodendron (section Pyxothamnus) and diploid V. vitis-idaea (section Vitis-idaea) was reported by Ehlenfeldt et al. (2018).
Vaccinium stamineum (deerberry) is the only species in Vaccinium section Polycodium, but it is highly polymorphic. Deerberries are endemic in North America and are found from southwestern Ontario to central Florida and west to eastern Texas (Vander Kloet, 1988). As noted by Vander Kloet (1988), Ballinger et al. (1979) found that deerberries resemble cranberries (V. macrocarpon section Oxycoccus) more than blueberries (section Cyanococcus), not only in fresh-market and culinary qualities, but also in the types of anthocyanins and flavanols the berries contain. Deerberry plants in Florida are capable of maturing full crops of fruit in the forest, even under xeric conditions (Sharpe and Sherman, 1971). Ballington (1995) carefully studied large populations of V. stamineum in North Carolina and South Carolina, and concluded that deerberries had a high potential for domestication as a new crop.
Intersectional hybrids involving sections Cyanococcus and Polycodium have been reported previously. Coville produced viable diploid hybrids between V. stamineum and the diploid Cyanoccus species V. myrtilloides (Darrow and Camp, 1945). Uttal (1987) reported a natural hybrid between V. pallidum and V. stamineum. Several F2 hybrids from V. darrowii × V. stamineum yielded tetraploid seedlings when crossed to tetraploid highbush in North Carolina (Ballington et al., 1997).
Tetraploid hybrids between highbush blueberry cultivars and colchicine-induced tetraploid V. stamineum have been reported previously (Lyrene, 2016, 2018). This article describes further progress in our efforts to recombine the genes of highbush cultivars and V. stamineum. The results of additional crosses involving F1, F2, and backcross parents are given, along with a description of the phenotype of hybrid and backcross populations after several years in the field. Prospects for developing cultivars with novel characteristics, such as red flesh, adaptation to droughty soils, flowers that are accessible to a wider range of pollinating insects, and exotic new flavors, are discussed in light of what has been seen in hybrid and backcross populations.
All crosses were made in a bee-proof greenhouse between 2013 and 2020 using potted plants. The crosses were made to study the fertility and crossing behavior of various parental combinations and to produce seedlings for evaluation in the field. BC1 seedlings in this study were of two types, depending on whether the recurrent parent was highbush or tetraploid V. stamineum. Unless otherwise specified, BC1 hereafter refers to seedlings in which highbush was the recurrent parent, because they were most numerous and most extensively studied. Except in cases when the stigmas were far exserted from the corolla (e.g., in V. stamineum, some F1 hybrids, and BC1 seedlings when V. stamineum was the recurrent parent), the flowers on the seed parent were emasculated before anthesis; flowers not used in crosses were removed as they opened. Pollen was transferred to the stigmas by thumbnail. The F1 plants used in crosses were selected from high-density field nurseries after examination throughout the year to make sure they were valid hybrids. Their identity as valid intersectional hybrids was confirmed by flower and inflorescence morphology when they flowered in the greenhouse (data not shown). As would be expected from intersectional hybrids, the hybrids were easily distinguishable from seedlings of the parental species after they flowered and fruited.
Unilateral cross compatibility has been reported in Vaccinium (Hiirsalmi, 1977; Wenslaff and Lyrene, 2003; Zeldin and McCown, 1997). Before 2019, all of our F1 hybrids were produced using V. stamineum as the pollen parent. In 2019 and 2020, some crosses were made using tetraploid V. stamineum as the seed parent, and highbush cultivars or F1 hybrids as pollen parents (Table 1). The reciprocal crosses were attempted to determine whether they would give seed and whether the seedling phenotype would differ, depending on the direction in which the cross was made.
Most crosses in which V. stamineum was the seed parent were made using second-generation tetraploid V. stamineum seedlings. These were produced by intercrossing tetraploid V. stamineum plants produced using colchicine. The advantages of using second-generation tetraploids were that they could easily be produced in large numbers and they avoided sectorial chimeras (some branches diploid, others tetraploid), which had been a problem with first-generation colchicine tetraploids (Lyrene, 2016). The V. stamineum plants used in the 2019 and 2020 crosses were confirmed as tetraploids based on pollen tetrad size. Both the flowers and the pollen tetrads from tetraploid plants were substantially larger than those from diploids (e.g., pollen tetrad diameters, typically 60 vs. 40 nm).
An experiment was conducted in 2020 to test the possibility that selfing might have occurred while highbush pollen was being applied to the stigmas of unemasculated V. stamineum flowers. The experiment was conducted because a surprisingly large number of seeds had been obtained in 2019 by pollinating tetraploid V. stamineum stigmas with highbush or F1 pollen. This raised the question of whether some second-generation tetraploid V. stamineum plants might be self-compatible, producing selfed seedlings rather than hybrids after highbush pollen was placed on the stigmas of unemasculated flowers. V. stamineum stigmas extend far beyond the anthers and the corolla, and highbush pollen was applied easily without emasculation. Although care was taken to avoid selfing while applying highbush pollen, some self-pollination could have occurred if the seed parents were self-compatible. Diploid V. stamineum has been considered highly self-incompatible (Cane et al., 1985), but tetraploids have never been tested.
In the 2020 experiment, branches on each of seven tetraploid V. stamineum plants were divided into two groups. Flowers on some branches were pollinated with an abundance of self pollen, and those on others with an abundance of highbush pollen. None of the flowers was emasculated. Most of the flowers were pollinated multiple times over a period of a week or more.
Five crosses of a new type were attempted in 2020. Flowers of tetraploid V. stamineum were pollinated with pollen from tetraploid Vaccinium arboreum (section Batodendron). Like V. stamineum, V. arboreum (sparkleberry) is a diploid blueberry native in Florida and throughout the southeastern United States. Plants of the two species sometimes occur together in the forest, but V. stamineum normally flowers earlier and occurs on drier sites. Pollen used in these crosses was a composite from three colchicine-produced tetraploid V. arboreum plants that had been planted in a garden between 2005 and 2010 (Lyrene, 2011). The plants had previously been determined to be tetraploid by pollen size and stomate size, and by the production of large numbers of seedlings when crossed as pollen parent to tetraploid highbush cultivars (Lyrene, 2011). The branches from which the pollen was taken in 2020 were verified as tetraploid based on pollen size a few days before pollen was collected for use in crosses. V. stamineum flowers were pollinated over a period of 6 weeks starting in early Mar. 2020. Fresh V. arboreum pollen was collected from the garden for use each week.
The number of flowers pollinated and the number of plump seeds obtained from each cross in 2019 and 2020 were recorded (Table 1). The first 10 berries that ripened from each cross were opened individually and the plump seeds in each berry were counted.
Various plant and berry characteristics were observed in the greenhouse on plants used as seed parents. Observed characteristics included flower and inflorescence morphology, plant vigor, berry size, berry color (wax, skin, and flesh), picking scar, stemminess, firmness, flavor, and tendency for berries to shatter when ripe.
The seeds from each cross were planted on the surface of peat in 4-L pots in a greenhouse in November of the year the cross was made. For crosses that produced large numbers of seed, only enough seeds were planted to provide at least 120 seedlings. During January and February (2–3 months after sowing the seed), young seedlings were transplanted to trays of peat. For crosses that produced numerous seedlings, no more than 120 seedlings were saved per cross. The seedlings were grown in the greenhouse for ≈3 months.
In May of each year, from 2015 through 2020, seedlings from the previous year’s crosses were transplanted to a high-density field evaluation nursery at the University of Florida Plant Science Unit in Citra, FL (Table 2). The nurseries were prepared by mixing a 3-cm layer of pine bark chips into the top 20 cm of the sandy soil. Seedlings were spaced ≈20 cm apart in rows that were ≈12 m long, with 45 cm between rows. The nurseries were irrigated with overhead irrigation and fertilized to maximize the growth of highbush blueberry seedlings planted in the same nursery. About 10,000 highbush and 400 V. stamineum seedlings were planted along with the hybrids in each of the nurseries.
Each nursery was maintained for 4.3 years, allowing the seedlings in the older nurseries to be observed through four fruiting seasons. The nurseries were observed weekly throughout the year, and more frequently during the months of flowering and fruiting. The goal was to assess vigor, fruitfulness, and other characteristics of each type of seedling. During November, December, and January of each year, selected plants were dug, potted, and moved to the greenhouse for use as parents. Nighttime temperature in the greenhouse was kept above 5 °C, but was otherwise allowed to vary with weather conditions, and fell to less than 10 °C on many nights in December, January, and February. It was discovered that many F1 plants that had no visible flower buds in the field nursery in early November and appeared to be too small and weak to flower the next spring, would flower if dug and potted in early November, chilled at 5 °C for 6 weeks in a dark cooler, and then moved to the greenhouse in late December. On most plants, flowers, sometimes numerous, were evident on the new leafy shoots that appeared from January through March.
Despite the close spacing of plants in the evaluation nurseries, vigor of F1 hybrids and BC1 hybrids in which V. stamineum was the recurrent parent was so low that plants had enough space to grow and develop during 4 years in the nursery. F1 populations were examined at the end of the first growing season to remove plants that did not appear to be the intended hybrids. This was necessary because highbush plants that appeared with the hybrids in some populations were so vigorous comparatively that they crowded out nearby F1 plants after the second year in the nursery if not removed. In BC1 populations in which highbush was the recurrent parent, and in BC1 × BC1 populations, there was much variation in vigor, and after the first growing season in the field, competition among plants began to eliminate the weaker plants or prevent them from flowering. After the second year in the field nurseries, observations of fruit characteristics could only be made from the strongest, most competitive plants.
Most of the crosses made in the greenhouse in 2019 and 2020 produced numerous seeds (Table 1). Exceptions were highbush × tetraploid V. stamineum crosses and crosses in which one parent was a Backcross-1 of the type (highbush × V. stamineum) × V. stamineum. Similar crosses had given low seed numbers in previous years (Lyrene, 2016, 2018; and unpublished).
In the 2020 experiment, which involved seven tetraploid V. stamineum plants used as female parents, both self-pollinated flowers and flowers that received highbush pollen produced berries and seeds (Table 3). On most V. stamineum females, the number of seeds obtained did not differ significantly for the selfed and crossed branches. This suggests that most of the seedlings from the highbush pollen were true hybrids, not seedlings from self-pollination. Had most of the seedlings from crossed flowers been the result of unintended self-pollination, more seed would have been expected from flowers that were intentionally self-pollinated with an abundance of pollen than from flowers where efforts were made to avoid self-pollination and highbush pollen was applied in abundance. Both the selfed and the intersectional pollinations gave fewer seeds than crosses between tetraploid V. stamineum genotypes (Table 1), suggesting that partial self-incompatibility reduced seed production from self-pollinated flowers, and intersectional incompatibility reduced seed production from flowers pollinated with highbush pollen.
After one growing season in the field, most seedlings from the V. stamineum × highbush and V. stamineum × F1 crosses made in 2019 had phenotypes consistent with their hybrid origins. Seedlings for which F1 hybrids had been the pollen parent were more vigorous than those with highbush as the pollen parent. This would not be expected if seedlings from both types of pollinations had resulted from selfing.
Except for having larger flowers, second-generation tetraploid V. stamineum seedlings were similar to diploids in vigor and macroscopic morphology. Both were as vigorous as seedlings from highbush × highbush crosses growing in the same nursery.
Crosses in which a diploid section Cyanococcus species (V. darrowii, V. elliotti, and V. fuscatum were tried) was the seed parent and diploid V. stamineum was the pollen parent gave numerous hybrid seedlings, but these were invariably very weak. Some of these seedlings probably would have survived to flower had they been potted and placed in a greenhouse, but this was not attempted.
With rare exceptions, F1 hybrids obtained by crossing tetraploid highbush cultivars with tetraploid V. stamineum were far weaker than either parent (Fig. 1). F1 hybrids varied greatly in vigor depending on the particular highbush and V. stamineum plants that were crossed. Sibling seedlings within some crosses were highly variable in vigor, but in other crosses all the seedlings were very weak. About half the F1 seedlings from the most vigorous half of the crosses could probably have been brought to flowering if potted and grown in a greenhouse. When the strongest F1 plants were dug, potted, and placed in the greenhouse, most of them increased in vigor but remained less vigorous than highbush seedlings. The weakness of the F1 hybrids in the high-density field nurseries was probably a result, in part, to excess soil moisture. Overhead irrigation scheduling was adjusted to maximize the growth of highbush seedlings planted in other parts of the same nursery. These were far more vigorous and required much more water than the hybrids. Year after year, F1 seedlings were much more susceptible to leaf mites (Eotetranychus clitus) in the greenhouse than either parent growing in the same greenhouse. To prevent defoliation, it was necessary to control mites. Of the original four tetraploid V. stamineum plants used repeatedly in crosses, FL13-1145 consistently produced the most vigorous F1 seedlings, although nothing about its appearance distinguished it as special.
Citation: HortScience horts 56, 4; 10.21273/HORTSCI15619-20
F2 seedlings, produced by intercrossing unrelated F1 hybrids, were much more vigorous than F1 seedlings (Fig. 1). In part, this was probably because only the strongest F1 plants were used as parents, but gametic and zygotic selection were also probably important, because only the most functional gametes from the F1 parents were likely to make zygotes and only the most functional zygotes were likely to make viable seed.
An additional increase in vigor was seen in BC1 seedlings if highbush was the recurrent parent. Like F1 seedlings, BC1 seedlings were highly variable in vigor, both within and between crosses. Only about 10% of the thousands of BC1 seedlings grown in the field for 3 years or more appeared to be as vigorous as typical highbush seedlings. Poor leaf health, resulting from diseases and physiological disorders, was one reason for the weakness of many seedlings. Seedlings from BC1 × BC1 crosses were no more vigorous than seedlings of the parent generation, and vigor continued to be highly dependent on which particular plants served as parents.
BC1 seedlings from crosses in which V. stamineum was the recurrent parent were less vigorous than those with highbush as the recurrent parent, but they were extremely variable in vigor. About 30% of these seedlings, after three growing seasons in the field, had medium vigor and healthy leaves. The V. stamineum backcross seedlings were more vigorous and looked more like V. stamineum than the most vigorous F1 plants. In growth habit and vegetative appearance, the BC1 seedlings were very distinct from V. stamineum. In the field, some of the plants flowered profusely but produced little fruit. This may have been the result of low male or female fertility and the fact that they flowered late, when little pollen was available from other tetraploid blueberries growing nearby. Several of the BC1-to-V. stamineum seedlings produced defective flowers in which the styles were missing or poorly developed.
In the 2019 greenhouse, four V. stamineum backcross seedlings that were pollinated with highbush pollen produced berries of the expected phenotype (more like deerberries than blueberries) and made a considerable amount of seed (Table 1). Seedlings from these four crosses were substantially more vigorous than most highbush × V. stamineum F1 hybrids, and their leaves had a healthier green color.
The deerberry plants used as parents resembled highbush blueberries in plant architecture (crown forming, with 2–10 major canes arising from a small base, reaching a height of 2–3 m after several years in field nurseries). Except for low vigor, the bush structure of the F1 hybrids resembled the parents. Seedlings from backcrosses to highbush generally resembled highbush in architecture. After 3 years in the field, a few F2 and backcross seedlings (especially those with V. stamineum as the recurrent parent) produced upright canes bearing large numbers of short, stiff side branches oriented perpendicular to the upright canes. These seemed analogous to fruiting spurs in apricots, almonds, apples, and pears. Branches of this kind are rare in highbush and not common in deerberry. The apical 65 cm of one vigorous, upright cane on a BC1 plant examined in November bore 50 spurlike branches (Fig. 2). These spurs were stiff, woody, and 2 to 6 cm long. Each had four to six nodes, with a bud at each node. The previous May, several canes on the plant shown in Fig. 2 had produced clusters of ripe berries extending basipetally from the apex for 40 to 60 cm on the major canes, giving the appearance of cattail (Typha) spikes.
Citation: HortScience horts 56, 4; 10.21273/HORTSCI15619-20
Most tetraploid V. stamineum plants were highly fertile, whether they had been produced by colchicine treatment or by intercrossing colchicine-derived tetraploids. An exception was FL13-872, which was widely used as a pollen parent because it was one of the first tetraploids available, in which only about half of the microspores appeared to be normal under the microscope at ×250 magnification. F1 hybrids varied widely in fertility (Lyrene, 2016). The most vigorous seedlings obtained by backcrossing F1 hybrids to highbush were typically very fruitful and produced berries with numerous seeds when open-pollinated in the field. An open-pollinated polycross involving 25 BC1 plants in 15-L pots placed outside of the greenhouse, in which the only pollen source was other BC1 hybrids, produced an abundance of berries, and these contained numerous well-developed seeds. The increased fertility in the BC1 population compared with the F1 was probably a result, in part, of the use of the most fertile F1 plants as parents and, in part, to gametic and zygotic selection for increased fertility.
Highbush blueberry cultivars are clones, propagated by stem cuttings or in vitro. V. stamineum of the type used in this study (2–4 m tall with a few stout canes) is nearly impossible to propagate from stem cuttings except using highly juvenile cuttings. Juvenile cuttings are not available from plants that have been selected for superior fruit. Juvenile shoots suitable for rooting in a mist propagation bed can be produced in vitro from mature plants of V. stamineum (Lubell and Brand, 2018; A. Garcia, personal communication), but efficient methods for producing field-ready plants from V. stamineum microcuttings need to be developed. Nonjuvenile softwood cuttings of most F1 hybrids rooted readily under mist, but cuttings of others did not (Lyrene, 2016). Cuttings from some hybrids fail to root because they defoliate quickly when placed under mist, but cuttings from other hybrids maintain healthy leaves for several months under mist but do not form roots. Some seedlings from backcrosses to highbush are also slow to root from stem cuttings.
The highbush cultivars used in the crosses ripen from 10 Apr. through 15 May when grown in the field in North Florida. Berries on the diploid V. stamineum plants from which the tetraploids were derived ripen in the forest in North Florida from late June through late July. BC1 seedlings with highbush as the recurrent parent typically ripened 3 to 4 weeks later than highbush cultivars when the two were grown together in field plots at the research station at Citra, FL, but there was much variation among the BC1 seedlings. BC1 hybrids that have a low-chill requirement but late flowering and ripening might be used to extend the blueberry harvest season in low-chill production areas. In Florida, late flowering decreases losses from freezes but increases problems with flower thrips, which often destroy late flowers and prevent pollination.
Highbush flowers are borne on short peduncles that do not persist after the berries are harvested. Flowers and berries on some cultivars are tightly clustered. Tight berry clusters can become a problem as berry size is increased through breeding. In V. stamineum, both the pedicels and peduncles are much longer, with much more space between the berries. In F1 and BC1 hybrids, peduncle and pedicel length range widely between the parental types, but the racemes of most hybrids are more open than in highbush (Figs. 3–5).
Citation: HortScience horts 56, 4; 10.21273/HORTSCI15619-20
Citation: HortScience horts 56, 4; 10.21273/HORTSCI15619-20
Citation: HortScience horts 56, 4; 10.21273/HORTSCI15619-20
Flowers of blueberry species in section Cyanococcus are adapted to buzz pollination by native bees such as Bombus and Habropoda. Achieving adequate pollination using honeybees has been a problem for growers in commercial blueberry plantations in Florida and Georgia. Flowers of most highbush cultivars have a long corolla tube, which has a constricted aperture and anthers located well inside the tube. By contrast, in V. stamineum, the corolla cup is short, with a wide opening, and both the stigma and anthers extend well beyond the end of the corolla cup (Cane et al., 1985). The pollen and stigmas are readily accessible to honeybees and other nonsonicating pollinators. Flower structure in F1 and F2 hybrids was surprisingly variable. All F1 plants had flower morphology somewhere between that of the parents; but, in some, flowers were more like those of V. stamineum and, in others, more like those of highbush. Flowers on both F1 and BC1 plants tended to be smaller than on highbush. Flowers of all F1 plants had anther awns, which are absent in highbush and conspicuous in V. stamineum, but these were invariably smaller than in V. stamineum. Flower structure remained quite variable in BC1 (to highbush) hybrids, and plants could be selected in which the pollen is readily accessible to honeybees. In many BC1 hybrids, the pollen discharge tubes extend all the way to the stigma, and both are near the aperture of the corolla tube. In highbush flowers, the anthers are situated deep inside the corolla tube.
Deerberries in the forest are larger than those of any other wild North American Vaccinium species (Fig. 6). Mean weight of 100 individually weighed berries selected randomly from each of three heavily loaded V. stamineum plants in the forest in Alachua County, FL, on 8 July 2020 were 1.24, 1.37, and 1.15 g, respectively, with sems of 0.057, 0.035, and 0.023. Mean berry weight for 23 wild tetraploid highbush blueberry plants (V. fuscatum) from Alachua County, FL, ranged from 0.14 to 0.54 g, with an overall mean for the 23 plants of 0.25 g (Lyrene and Sherman, 1980). Berries of the highbush cultivars grown in Florida are large because of 6 to 10 generations of selection for large berries. Berries on the most vigorous and fertile BC1 plants averaged about as large as berries from random seedlings from highbush × highbush cultivar crosses, indicating that V. stamineum introgression had not reduced berry size.
Citation: HortScience horts 56, 4; 10.21273/HORTSCI15619-20
Ripe berries from highbush cultivars typically have black skin, green to white flesh, and variable amounts of surface wax that gives the berries of some varieties a powdery blue appearance. Mature berries from the wild V. stamineum plants used in this study had skin that was black after the surface wax was removed. Because of variations in berry-surface waxes, mature V. stamineum berries in the forest ranged from shiny black to powdery blue depending on the plant.
Unlike the section Cyanococcus species and the cultivars bred from them, V. stamineum is polymorphic for flesh color. About half the plants in the forest from which our parents came had a flesh color of white–green, like that of highbush, and about half had flesh color ranging from light pink through bright red to dark purple. These colors were consistent over years on individual plants, some of which have been observed yearly for 8 years or more. The tetraploid V. stamineum plants used in this study were obtained by colchicine treatment of seeds or seedlings with a flesh color that was unknown at the time of colchicine treatment. By chance, two of the first four tetraploid plants used in crosses had pink to red flesh in fully ripe berries.
Some F1 hybrids had pink to red flesh (Fig. 6; Lyrene, 2016). The F1 plants used in backcrosses were selected for vigor and fertility, without regard to flesh color. The berry flesh of some BC1 plants in the field was somewhat pink, but the pink was less intense than in the pigmented F1 hybrids and only appeared in berries that were fully mature. Mature berries from numerous BC1 plants were compared with those from numerous highbush seedlings growing in the same nursery. Pink flesh was much more common among seedlings in the BC1 populations. None of the BC1 plants had enough pigment to be considered commercially purple fleshed. A project to obtain highbush-type cultivars with deep-red flesh should start with the selection of V. stamineum parents having the most intense purple flesh, because the character is extremely variable in V. stamineum.
When extracting seeds from crosses, it was noticed that a 50:50 (by volume) mix of water and fresh berries homogenized in a food blender produced a puree that was normally much brighter red with F1 and BC1 berries than with highbush berries. When berries from wild diploid V. stamineum plants were processed in the same way, the puree was bright red, even if the berries had white–green flesh. These bright colors were stable for more than a week at room temperature. The difference between highbush and V. stamineum puree pigment intensity may be a result of the thicker skin of V. stamineum berries or a greater pigment concentration in the skin.
When highbush blueberries ripen, an abscission zone forms between the berry and the pedicel, and harvested berries are free of pedicels (i.e., they are not stemmy). In V. stamineum, the abscission zone is at the opposite end of the pedicel, and the berries come off the bush with the pedicel attached (Fig. 7). The F1 hybrids tended to be intermediate, with much variation from plant to plant. Most BC1 plants resembled highbush in stem separation, and the berries were mostly not stemmy. In wild highbush and in some of the earlier released cultivars, the skin tended to tear when the berry was separated from the pedicel. These cultivars were said to have a bad scar. In V. stamineum, the pedicel is long and very thin compared with highbush. When the pedicels are forcibly removed from the berries, only a small scar is formed. The scar on most BC1 plants was small and dry.
Citation: HortScience horts 56, 4; 10.21273/HORTSCI15619-20
The diploid V. stamineum population from which the parents came was polymorphic for berry shattering. Some plants that seemed to be loaded with berries at the start of the ripening season never held a large number of ripe berries, because the berries fell as they ripened. On only 20% to 30% of the plants were most of the ripe berries retained long enough on the bush so the plant appeared, at some time, to be loaded with ripe fruit. Berry shattering was highly variable among both the first- and second-generation tetraploid V. stamineum plants. Most F1 and F2 seedlings and nearly all backcross seedlings held their berries well, but shattering was a problem with a few.
Four aspects seemed variable in F1 populations: sugar, acidity, bitterness, and the presence of aromatic or spicey flavors not found in section Cyanococcus species. Deerberries in the source population had high Brix. Ten plants growing in the forest whose average Brix was determined using a hand refractometer on 16 July 2020 (stated as the average of three single-berry samples per plant), and ranged from 15.0 to 22.0 °Brix. Typical values for cultivated highbush range from 13 to 16 °Brix (DeGraw et al., 2016). Acidity was not measured but, by taste, ranged from low to medium compared with the average for highbush cultivars. Berries of V. stamineum are sometimes stated in the literature to be inedible or barely edible because of their astringency or bitterness. In our deerberry source population, berries never seemed astringent, but ranged from slightly bitter to strongly bitter. Bitterness seemed to be concentrated in the skin and was less noticeable in the flesh. Bitterness was not detected in the berries of any BC1 plants in which highbush was the recurrent parent. The berries of some F2 and BC1 plants had pleasant aromatic or spicey flavors not found in section Cyanococcus species. Deerberries from the forest, when dried at room temperature and stored for 3 months without refrigeration, were sweet, somewhat bitter, and had a flavor component suggestive of pine.
V. stamineum seeds are slightly larger than those of highbush, but the berries are large and juicy and do not seem more gritty than highbush berries. Deerberries have thick skin. This gives the berries a pleasant “pop” or crunch when eaten, but the skin residues that persist during chewing sometimes elicit complaints. Thick skin did not seem problematic in the BC1 plants.
No formal studies were done, but deerberries held in paper bags at room temperature for 10 d seemed to stay firm without shriveling or becoming moldy just as well as the best highbush cultivars. Berries collected from the ground beneath the plants in the forest remain firm for a week or more after they fall.
Obtaining plants that meet the collective requirements of growers, marketers, and consumers starting with weak intersectional hybrids will not be easy. It was not hard to get hybrid seedlings by crossing highbush cultivars with tetraploid V. stamineum plants. Many (but not all) of the tetraploid F1 hybrids had surprisingly high fertility and fruitfulness, and, in some cases, high berry quality, but nearly all of them were weak, and many were susceptible to disease, insect, and physiological problems. In September, October, and December, when highbush cultivars often make their best growth in Florida, V. stamineum plants in the forest are often losing their leaves. In October, in high-density nurseries, the foliage on many (but not all) F2 and BC1 seedlings appeared less healthy than on highbush seedlings. In the greenhouse, defoliation by leaf mites was much more severe on F1 blueberry–deerberry hybrids than on the parent taxa.
Fall flowering was a problem with some V. stamineum–highbush advanced-generation hybrids. Fall flowering in North Florida usually results in loss of berries resulting from winter freezes, and berries that ripen in early winter can have high acidity and low sugar contents. Many hybrids that lost flowers and berries from winter freezes still produced numerous berries in the spring because of the flower buds that remained dormant until late winter.
A problem in the introgression program has been the occasional appearance of highbush plants and plants of unknown origin among the F1 seedling populations. Because large numbers of flowers must be pollinated to generate relatively few F1 hybrids, even a low level of self-pollination can produce a substantial percentage of nonhybrid plants. Even with careful emasculation and removal of all unpollinated flowers from female parents in the greenhouse, some nonhybrid plants appeared in some populations of hybrid seedlings. In the high-density field evaluation nursery, these plants were almost always far more vigorous than the F1 hybrids. Because of the close spacing of the plants, a single highbush plant could overshadow and kill several true hybrids. It was necessary to identify and remove nonhybrid plants from the nursery before the second growing season. On rare occasions, a plant appeared among the F1 seedlings that resembled highbush in most aspects, but seemed to have one or more traits that could only have come from V. stamineum. These anomalous seedlings are a puzzle, but southern highbush cultivars have a complex genetic background, and it is not always safe to conclude that a particular seedling with surprising characteristics could not be a highbush seedling.
The phenotypic variation observed indicates that commercial cultivars could probably be selected from seedlings in BC1 or BC2 populations following highbush × V. stamineum crosses. In addition to the characters already discussed, several useful traits might be obtained in highbush cultivars using genes from V. stamineum. Improved drought tolerance is likely to be found in the hybrids. Deerberries grow naturally on excessively drained, coarse-textured upland sands; highbush are native on moist flatwoods soils where peat accumulates.
Cultivars with a flower structure that makes the pollen readily accessible to honeybees could be selected. The ability to flower on new spring growth flushes rather than from overwintering flower buds formed the previous autumn might be useful in some environments. Highbush blueberries that lose their leaves prematurely in the fall as a result of insects, leaf diseases, hurricanes, or other causes, form few flower buds, and few flowers open in the spring. During winter, highbush flower buds can be damaged by freezes, blueberry bud mites, or gall midges. In Florida forests, V. stamineum normally begins to defoliate early in the fall as a result of insects, diseases, drought, low soil nutrients, and genetic factors. No flower buds are visible on V. stamineum in January, but the plants can still flower heavily 3 months later and produce heavy berry crops on the spring growth flushes. Many of the F1 hybrids appeared to have inherited the ability to form flowers on spring growth flushes even when no flower buds could be seen in midwinter.
In addition to adaptation to excessively drained mineral soils, open flower structure, and red flesh, V. stamineum could be the source of various nutraceutical compounds, such as iridoids. Iridoids are secondary metabolites, present in various plant families, that may exhibit a range of pharmacological activities, such as cardiovascular, hypoglycemic, anti-inflammatory, hepatoprotective, antimutagenic, antispasmodic, antitumoral, antiviral, immunomodulatory, and purgative effects (Viljoen et al., 2012). Plants that produce iridoids have a history of use in traditional medicine in treating inflammatory conditions (Viljoen et al., 2012). Leisner et al. (2017) measured iridoid content in the berries of 84 Vaccinium accessions, including many Cyanococcus cultivars and several wild species from other sections. Of the 71 Cyanococcus cultivars sampled, ‘Ozarkblue’ had the greatest iridoid content measured as monotropein (113 ng·mg–1 dry fruit weight). Of the non-Cyanococcus species sampled, V. stamineum had the greatest content of monotropein (2371 ng·mg–1 berry dry weight), and V. arboreum had 1692 ng·mg–1. If iridoids can contribute to human health, both V. stamineum and V. arboreum could serve as gene sources for increasing iridoids in highbush cultivars.
In attempting to move single traits, such as red flesh, wide-open flowers, or exotic flavors, from V. stamineum into highbush cultivars, growing large F2 populations might be the key to success. F2 seedlings are relatively easy to produce in large numbers, and the plants are much more vigorous than F1 seedlings. F2 seedlings are vigorous and fertile enough to be evaluated for plant and fruit characteristics in high-density field nurseries. F2 seedlings are extremely variable for many of the traits that distinguish highbush and V. stamineum. This high variability indicates that at least some homoeologous chromosome pairing is occurring in the F1 hybrids (Ramsey and Ramsey, 2014). To minimize inbreeding and maximize diversity, the F1 parents should be chosen so that every F2 seedling has four distinct genotypes as grandparents. Superior F2 seedlings can be used in backcrosses to produce cultivars.
Seedlings in the 2019 high-density nursery photographed 22 Oct. 2020. The nursery was planted 30 May 2019 using greenhouse-grown seedlings from seed planted in Nov. 2018. The four weakest rows in the center contain F1 hybrids from highbush × tetraploid V. stamineum crosses. The more vigorous rows on the left contain seedlings obtained by intercrossing BC1 seedlings that had V. stamineum as the recurrent parent, and those on the right contain F2 seedlings from F1 × F1 crosses.
Spurlike branches on an upright cane of a BC1 seedling (highbush the recurrent parent) transplanted to the field nursery in May 2017. Photographed 4 Nov. 2020. The scale is in centimeters.
BC1 seedlings with inflorescence structure intermediate between that of highbush and V. stamineum. Peduncles elongate well before the flowers open. Photographed 22 Oct. 2020.
Clusters of berries on an F1 hybrid in the greenhouse showing long pedicels and peduncles. The flowers had been pollinated with pollen from a highbush cultivar. The plant had been defoliated by mites during berry development.
Loose clusters of berries on an F1 hybrid in the greenhouse after the flowers had been pollinated with pollen from a highbush cultivar.
Internal berry color of freshly sliced mature berries of tetraploid selection FL 15-669 (a highbush cultivar × tetraploid V. stamineum F1 hybrid) and highbush cultivar FL06-96.
Large berries from a diploid plant of V. stamineum (clone FL15-695) harvested from a forest in Alachua County, FL, July 2015. Pedicels remain attached to the berries when they abscise from the peduncle. The scale is in centimeters.
Contributor Notes
P.M.L. is the corresponding author. E-mail: Lyrene@ufl.edu.
Seedlings in the 2019 high-density nursery photographed 22 Oct. 2020. The nursery was planted 30 May 2019 using greenhouse-grown seedlings from seed planted in Nov. 2018. The four weakest rows in the center contain F1 hybrids from highbush × tetraploid V. stamineum crosses. The more vigorous rows on the left contain seedlings obtained by intercrossing BC1 seedlings that had V. stamineum as the recurrent parent, and those on the right contain F2 seedlings from F1 × F1 crosses.
Spurlike branches on an upright cane of a BC1 seedling (highbush the recurrent parent) transplanted to the field nursery in May 2017. Photographed 4 Nov. 2020. The scale is in centimeters.
BC1 seedlings with inflorescence structure intermediate between that of highbush and V. stamineum. Peduncles elongate well before the flowers open. Photographed 22 Oct. 2020.
Clusters of berries on an F1 hybrid in the greenhouse showing long pedicels and peduncles. The flowers had been pollinated with pollen from a highbush cultivar. The plant had been defoliated by mites during berry development.
Loose clusters of berries on an F1 hybrid in the greenhouse after the flowers had been pollinated with pollen from a highbush cultivar.
Internal berry color of freshly sliced mature berries of tetraploid selection FL 15-669 (a highbush cultivar × tetraploid V. stamineum F1 hybrid) and highbush cultivar FL06-96.
Large berries from a diploid plant of V. stamineum (clone FL15-695) harvested from a forest in Alachua County, FL, July 2015. Pedicels remain attached to the berries when they abscise from the peduncle. The scale is in centimeters.
