Ann Chanon and Pablo Jourdan
Interspecific hybrids among Aesculus species, where possible, offer opportunities for improvement of a variety of horticultural attributes for key members of this genus. The authors have begun a program to develop such hybrids, focusing on A. parviflora, the bottlebrush buckeye, and A. pavia, the red buckeye. Preliminary findings on the floral biology and fruit set in A. parviflora are reported here. The results indicate that crosses with A. parviflora as the pistillate parent may be difficult, although possible.
Aesculus parviflora, the bottlebush buckeye, was first described in 1788 by Walter. It is native to South Carolina, Georgia, and northern Florida in populations that are isolated from one another (Harden, 1957). These native populations are most often found in association with rivers where the seeds are dispersed by water (Wyatt, 1985).
Despite its southern origin, this shrubby, stoloniferous plant is hardy to Zone 4. It ranges in height from 8-15 ft. and blooms for two to three weeks in mid- to late summer (Clark, 1982). White flowers are borne on panicles 8-18" in length. Flowers are either staminate or perfect, and both produce long slender filaments that extend from the flower, giving the panicles the 'bottle brush' appearance.
Aesculus parviflora is an excellent landscape plant; it does well in a sunny or a shaded location and can tolerate either alkaline or acid soil. The leaves unfurl in early spring, and they maintain healthy, dark-green foliage throughout the year. It does not suffer from the many foliage problems (e.g., scorch) associated with other Aesculus species (Dirr, 1990).
The plants have only moderate to poor seed set, and the reasons for this are uncertain. Such low seed set hampers efficient propagation of the plant in the trade. Much research to date has focused on the difficulties of asexual propagation from shoots and stolons (Fordham, 1987).
Aesculus pavia L., the red buckeye, is also native to the southeastern United States. It was the first of the North American Aesculus species to be described by Plukenet in 1696 (Taylor, 1982). Within Aesculus, four members are recognized in the section Pavia: A. flava, A. glabra, A. pavia, and A. sylvatica. These species grow in close proximity and have overlapping ranges.
Natural interspecific hybrids of A. pavia are quite common within this section, and flower color is only one of the traits affected (Taylor, 1982). In addition, hybrids between A. pavia and A. hippocastanum, a member of a different section, are known and have been named A. x carnea (Upcott, 1936).
A. pavia is a large shrub to small tree, depending on pruning, and is hardy to Zone 4. It ranges in height from 10-25 ft. and about the same in spread. Panicles bloom in May and range in size from three to eight inches; flowers can range from yellow to a dull red to a bright crimson, depending on the specimen.
The plant is relatively rare in cultivation, which may be due in part to variability in flower quality and to a fairly slow growth rate. Red buckeye is best propagated by seed, but grafting is possible. Seeds gathered in early fall should be planted immediately since they deteriorate with storage (Dirr and Heuser, 1987).
Harden (1957) noted that Aesculus species hybridize easily, but no natural hybrids between bottlebrush and other buckeyes have been observed. All Aesculus share the same number of chromosomes 2n = 20 with the exception of A. x carnea, which is polyploid with 2n = 40 (Hoar, 1927). It is unclear whether this lack of interspecific hybrids with bottlebrush buckeye is the result of incompatibility barriers or just due to the separation in bloom time, since most Aesculus flower in early spring, but the bottlebrush does so in summer. Some attempts have been made to produce artificial hybrids between A. parviflora and other Aesculus, but these have not yet been successful (R. Marquard, personal communication).
This research is examining the barriers that so far seem to prevent the formation of hyrids between A. pavia and A. parviflora. To this end, this study is examining in some de-tail the floral biology of these species, focusing first on A. parviflora, looking at inflorescence structure, pollen longevity, and seed set after controlled intra- and inter-specific pollination. Some preliminary findings are reported here.
Fresh pollen was collected from nine different A. pavia trees, four from the Dawes Arboretum, two from The Ohio State University's Columbus campus, two from the Holden Arboretum, and one from a private residence in Lake County. Pollen was collected by rubbing freshly dehisced anthers across a nylon screen onto a collecting dish. Between 25 and 50 mg of pollen were placed into size 00 gelatin capsules and stored over a drying agent in a -20°C freezer.
Pollen viability was assessed prior to use for hand pollination and was estimated using an in vitro pollen germination method. A sample of pollen was taken from the freezer, equilibrated at room temperature, and applied to the germination medium. The medium contained 10% (w/v) sucrose and was made following the Brewbaker and Kwack (1963) procedure. The plates were incubated upside down for 24 hours.
Following incubation, a drop of cotton blue stain was placed on the areas where the pollen had been applied and covered with a cover slip. A pollen grain was counted as germinated if the pollen tube was at least twice the diameter of the pollen grain.
Panicles on A. parviflora became visible after bud expansion and were observed weekly until flowering began. Panicles were tagged, and the number of staminate flowers removed was recorded daily. The complete flowers were emasculated when the buds were approximately six millimeters in length, and the anthers were just visible. The panicles were then covered with pollination bags to exclude insects and other sources of contamination.
When the styles reached approximately 1 cm in length, pollinations began. Flowers were either pollinated once or pollinated daily until the stigmas turned black. Each day, fresh pollen was taken to the field where it was allowed to warm up prior to application. The stigmas were coated with pollen by moving the stigma up and down the side of the gel cap until the entire surface was covered. Fruit set was assumed to have occurred if the ovary swelled and turned from white to green. About 10 weeks following pollination, the fruits were harvested.
Panicles were observed from a number of individual plants at locations in and around Columbus as well as the Holden Arboretum. Of the 400 panicles observed, the average number of flowers per panicle was 319. On average, 43% of those panicles produced no complete flowers. Those panicles that had complete flowers produced an average of eight complete flowers per panicle but had a range that was from one to 74 complete flowers. The complete flowers occurred in all parts of the panicle and were most often found in the upper-most third of the panicle. This was unlike other species of Aesculus, where most of the complete flowers are located at the base of the panicle (Harden, 1956).
Pollen from the six A. pavia, the A. x carnea, and two A. parviflora sources was germinated on artificial medium to evaluate their viability as pollen sources. Both A. parviflora samples germinated at greater than 90% efficiency. The A. x carnea sample germinated at approximately 50% efficiency. Germination for the various A. pavia pollen samples germinated in a range from 56-77%. All of the pollen sources were judged fit for use in the pollination experiments.
Five A. parviflora plants were used for the pollination experiments. Two of those plants produced very little to no seed, but this may have been the result of the 1998 summer drought conditions. At 10 weeks post-pollination, fruits were harvested, and the percent of fruit set was calculated (Table 1).
| Table 1. Pollination and Fruit Set on Aesculus parviflora Using Various Pollen Sources and Pollination Methods. | |||
|---|---|---|---|
| Pollen Source | Number of Pollinations | Number of Fruit Set | Relative Percent Fruit Set |
| Unpollinated | 150 | 0 | 0.0 |
| Interspecific Pollinations | |||
| A. x carnea | 275 | 4 | 1.5 |
| A. pavia | 951 | 69 | 7.3 |
| Intraspecific Pollinations | |||
| A. parviflora* | 353 | 75 | 21.2 |
| Open pollination | 144 | 36 | 25.0 |
| Self pollination | 230 | 31 | 13.5 |
| * Combined data of multiple paired pollinations. | |||
Fruit set was affected by the pollen source. The unpollinated controls set no fruit, which suggests that pollination is required for fruit development and that parthenocarpy and apomixis were unlikely. This agrees with previous research by Hardin (1955), where no parthenocarpy and apomixis had been reported. In addition, the absence of fruit in such plants suggests the risk of pollen contamination may be low, and accidental pollinations would be rare.
Pollinations using A. pavia pollen produced, on average, 7.3% fruit set, but the range was wide, from a low of 2.3% to a high of 26.1% fruit set depending on the A. pavia individual (genotype).
Pollinations with A. x carnea pollen resulted in only a 1.5% fruit set; whether this low set is the result of differences in chromosome numbers or just low pollen fertility is not clear at this time. The intraspecific pollination (among different A. parviflora individuals) had a fruit set of 21.2%.
Self pollinations (within the same plant) had a 25.0% fruit set, suggesting perhaps that self-incompatibility is not a problem in this species; however, additional tests are needed before firm conclusions can be made. Both of these controlled pollinations resulted in higher fruit set than the 13.5% observed from open pollinated flowers (bulked pollen from multiple A. parviflora sources). These results indicate:
Further work is needed to optimize the interspecific hybridization system for A. parviflora and A. pavia. Factors to be considered include:
The authors wish to express their sincere gratitude to the staffs of the Holden Arboretum and Dawes Arboretum for permission to work there and to Dale Killian at the Bexley Parks and Recreation Department at Jeffrey Mansion for permission to collect pollen samples.
Special thanks to Robert Marquard and others for providing helpful suggestions and encouragement throughout this investigation and for allowing the use of laboratory facilities.
Brewbaker, James and Beyoung Kwack. 1963. The essential role of calcium ion in pollen germination and pollen tube growth. American Journal of Botany. 50(9):859-865.
Clark, Ross. 1982. Bottlebrush buckeye: mysteries of a southern delight. Morton Arboretum Quarterly. 18(1):1-5.
Dirr, Michael and Charles Heuser. 1987. The reference manual: woody plant propagation. Varsity Press, Inc., Athens, GA.
Dirr, Michael. 1990. Manual of woody landscape plants: their identification, ornamental characteristics, culture, propagation, and uses. 4th Ed. Stipes Publishing Co., Champaign, IL.
Fordham, Alfred. 1987. Bottlebrush buckeye (Aesculus parviflora) and its propagation. International Plant Propagators Society. 37:345-347.
Hardin, James. 1957. Studies in the Hippocastanaceae. IV. Hybridization in Aesculus. Rhodora. 59:185-203.
Hardin, James. 1957. A revision of the American Hippocastanaceae. Brittonia. 9(3/4):145-171.
Hardin, James. 1956. Studies in the Hippocastanaceae. II. Inflorescence structure and distribution of perfect flowers. American Journal of Botany. 43:418-424.
Hardin, James. 1955. Studies in the Hippocastanaceae. I. Variations within the mature fruit of Aesculus. Rhodora. 57:37-42.
Hoar, Carl. 1927. Chromosome studies in Aesculus. Botanical Gazzette. 84:156-170.
Taylor, N. P. 1982. Aesculus pavia. Curtis's Botanical Magazine. London. 184:81-84.
Upcott, Margaret. 1936. The parents and progeny of Aesculus carnea. Journal of Genetics. 33:135-149.
Wyatt, Robert. 1985. Aesculus parviflora in South Carolina: phytogeographical implications. Bulletin of the Torrey Botanical Club. 112(2):194-195.