Return to METRIA 10

New Interspecific Hornbeam Hybrids (Carpinus)
Bred for Improved Tolerances to Environmental Stresses

Susan J. Wiegrefe and Lisa C. Berg

The Morton Arboretum, Lisle, IL 60532

A paper from the Proceedings of the 10th Metropolitan Tree Improvement Alliance Conference held in St. Louis, MO, September 30 and October 1, 1998, co-sponsored by the Landscape Plant Development Center and the Society of Municipal Arborists.


Although as many as 45 species of Carpinus exist, they are under-represented in North American plant collections and planted landscapes. A review of the horticultural literature and data about the site conditions from which Carpinus have been collected in their native habitats indicate that they possess adaptations which would enable them to withstand the rigors of urban growing conditions. In 1996, controlled crosses were made among the species from which reproductive specimens are available in an effort to determine interspecific crossabilities and to generate progenies for further use in developing trees with superior tolerances to environmental stresses. Twenty trees growing at four arboreta and representing nine species were used in 49 controlled crosses. Preliminary results indicate that seven previously undocumented interspecific hybrids were generated: C. caroliniana x C. betulus, C. caroliniana x C. cordata, C. caroliniana x C. coreana, C. caroliniana x C. laxiflora, C. caroliniana x C. orientalis, C. caroliniana x C. tschonoskii, and C. cordata x C. japonica. Inferences about crossability were made based on numbers of fruit filled and support for hybridity is derived from intermediate leaf morphologies.

Keywords: Carpinus, interspecific hybridization, crossability, environmental stress tolerance, tree breeding.


Members of the genus Carpinus possess such understated elegance that they are often overlooked in the process of selecting landscape plants. Closer inspection reveals attributes of considerable value for urban trees. This is especially true when viewed through the eyes of a plant breeder who contemplates the recombination of characters possible through (a couple of generations of) interspecific hybridizations.

There are as many as 45 species of Carpinus found throughout temperate portions of the northern hemisphere (Bean 1970). Nearly half of these species are not in cultivation and a number of others have only recently been introduced to the western world from Asia (Rushforth 1986). In spite of our limited knowledge of this group, there is a growing body of evidence that the species that comprise this genus possess a wide variety of adaptations which would be useful in coping with urban stresses. Without exception they also possess ornamental traits of considerable variety.

A previous report by Santamour (1978) on the crossability among a subset of these species indicates that interspecific crossability within the two sections of the genus is generally good.

In 1996, while working for the Landscape Plant Development Center, the senior author began a project of making interspecific crosses among the available species of Carpinus. The goals were to determine cross compatibilities and produce hybrids which could be evaluated themselves and also ultimately used to create a second generation of trees. From these F2 progenies, individuals would be selected for possessing the desired combination of traits from the original species.

Plant Materials

The species that were used in this project, their nativity, and positive traits are as follows:

Carpinus betulus is native to continental Europe and Asia Minor, where it is often a dominant species in the forest canopy. It is long lived and strong wooded, tolerant of a wide range of soil conditions from coarse sand to clay, as well as acid or alkaline soil pH (Dirr 1990). The species possesses tolerance of nitrous oxides (Hilliard 1980), and is frequently used in urban settings in Europe. Many selections have already been made for desirable growth forms and other aesthetic effects, such as 'Columnaris', 'Pendula', 'Globosa', 'Purpurea', 'Variegata', and 'Quercifolia'.

C. caroliniana, the only North American species in the genus, grows on stream banks, and as an understory species on mesic to wet sites. Besides its small stature and shade tolerance it brings cold hardiness, tolerance of poorly drained soil, and the potential for brilliant red autumn foliage coloration to the genetic mix.

C. coreana is endemic to South Korea and is another diminutive species. This specimen, growing at the Morris Arboretum, was ten foot tall at ten years of age. The species grows on the seashore and is tolerant of salt spray, exposed conditions, and reflected heat (Yinger 1987, Meyer 1995). Its branchlets tend to be pendulous and its small, shiny leaves assume a burgundy hue in fall. Although its winter hardiness is insufficient for regions that experience a more continental climate, through hybridization we hope to be able to utilize its positive traits.

C. laxiflora is native to the mixed deciduous forests of Japan and Korea where it grows to the moderate size of 45 feet. Specimens growing in full sun at the Morris Arboretum in Philadelphia, however, grew to 12 feet in 10 years (Fig. 1), indicating the possibility of slower growth and smaller ultimate size in more exposed settings. The species exhibits a very graceful, uncongested branching habit and beautiful, apricot to orange fall color.

C. orientalis is native to southeastern Europe and east to the Caucasus. It grows on woodland edges and often on dry, stony hillsides (Bobrov 1970). Key traits of interest are: short stature (it grows slowly to 25 feet), drought and heat tolerances, wide branching angles, and the characteristic persistence of its leaves throughout the winter.

C. tschonoskii is native to Japan and continental northeast Asia. It is an early successional species (Tanouchi and Yamamoto 1995) which tolerates exposed conditions as well as exhibiting a graceful growth habit.

C. turczaninovii is also from Japan, Korea, and northern China where it grows in open woodland and scrub. This is how Bean (1970) sums up its attributes: "Makes a small, neat, bushy tree and colors rich brown and orange in autumn." It is also presumed that its exposed native habitat has enforced natural selection for tolerance to drought and exposure, and a fair degree of cold hardiness.

That was the last of the species from the section referred to as Eucarpinus. They are all characterized by their tight-fitting, smooth bark, sometimes fluted (or muscled) branches, and their infructescences in which the nutlets are exposed under bracts that flare out above them (Rushforth 1986). The second section, Distegocarpus, contains fewer species and they are characterized by rougher bark, less fluting, and nutlets which have the bracts tucked in around them in the infructescence. The species used from this section are:

C. cordata, another native to Japan and northeastern Asia. It has attractive fruits, striking diamond-shaped lenticels, and a greater tendency towards developing a dominant central leader than some of the other species.

C. japonica is endemic to Japan. It is a small tree growing to 30 feet and some form a vase shape and flat top (Dirr 1975; personal observation) that many landscapers find desirous for providing good sight clearance and shade with minimal height. Bean (1970), however, states that it forms a pyramidal crown. This apparent contradiction may be explained by different germplasm being introduced into Europe and North America or by individual observers having very few specimens to draw their conclusions from, as this is still a relatively rare plant in cultivation. Its leaves are striking, with a shiny surface, deep corrugations, and jagged serrations and can assume a deep red fall color. The infructescences are also ornamental and may be tinged pink before fading to light tan.


In order to have sufficient number of blooming specimens, the plant materials of four arboreta were used: the Arnold Arboretum (Jamaica Plains, MA), the Morris Arboretum (Philadelphia, PA), the Holden Arboretum (Kirtland, OH), and the Morton Arboretum (Lisle, IL). Pollen was collected from forced branches from 14 accessions of 9 species.

Seven individuals from 4 species were used as seed parents. Forty-nine crosses including "no pollen" controls were performed. A minimum of 4 inflorescences were pollinated in each cross, though ones which suffered insect injury or were missing at the time of seed harvest are not included in the statistics which follow. Pollen germinability was tested on the first day each cross was made following Marquard (1992).

Pollinations were carried out using techniques passed on to the senior author by Robert Marquard (1986). Staminate inflorescences were removed from the branch tips used in the crosses and synthetic sausage casings with one end stapled shut were used as isolation bags. Trees were bagged well before receptivity occurred on all trees used as seed parents with the exception of C. cordata. When the pistillate inflorescences were determined to be receptive through visual inspection, a hole was made in the bag with a paper punch and pollen introduced by blowing pollen out of a plastic disposable pipette (Fig. 2). The hole was then taped shut using transparent tape. Pollination was repeated two to three days later whenever possible to ensure pollination of slower-developing inflorescences and to bracket receptivity.

Fruits were collected when the bracts were beginning to dry and the nutlets were separated from the bracts. Floatation of open-pollinated nutlets in water or ethanol did not correspond well with cut-test results as an indicator of filled seeds. Therefore, a glancing slice was made in all fruits to determine the number of filled seed.

The seeds were placed in a moist mixture of 50% peat and 50% perlite in a zip-lock plastic bag and placed at room temperature for one month followed by cold stratification (32-36 F) for 5 months. The seeds and medium were then spread over a larger volume of the stratification medium, drenched with 1% Benlate, and placed in a warm greenhouse (70° F minimum) for germination. One seedlot, C. cordata x C. cordata germinated during stratification.

Visual assessment of leaf morphology of the putative interspecific hybrids was done in the summer of 1998. DNA analyses using RAPD markers are planned.

Crossability Data

Of the seven species that were used in interspecific crosses onto C. betulus cultivars only one, C. tschonoskii, resulted in filled seeds being produced and germinated (Table 1). The feasibility of that cross had already been established by Santamour (1978). The single filled seed in the cross between C. betulus ‘Quercifolia’ and C. cordata was unexpected due to the cross being intersectional as well as interspecific and previous indications of sexual incompatibility between the sections (Santamour 1978). It was especially surprising since intraspecific crosses on that individual produced zero and two seeds from two entire inflorescences which can produce as many as twenty nutlets (personal observation). The low fruit set may be explained by generally low fertility of the individual or poor timing of the pollinations relative to style receptivity. Further testing will be needed to determine the crossability between C. betulus and C. cordata.

Seven interspecific crosses produced filled seeds when C. caroliniana was used as the seed parent. In all cases where the same pollen source was used to pollinate the two C. caroliniana individuals at the Holden Arboretum, filled seed were produced on both plants. This indicates that crossability between the species involved is likely to be the general rule, i. e., the phenomenon is not specific to the individuals used. The number of filled seeds produced (per inflorescence harvested) using C. tschonoskii, C. laxiflora, and C. betulus pollen was of the same order of magnitude as with the intraspecific crosses. Although the number of seeds produced with the C. coreana, C. cordata, and C. turczaninovii pollens was similar to those produced in the controls where no pollen was applied, as will be shown in the section discussing evidence of hybridity, this does not rule out the possibility of crossability between the species.

Within section Distegocarpus, the number of filled seeds produced using C. japonica pollen on C. cordata was almost as great as when pollen of another C. cordata was used (Table 1). No filled seeds were produced when C. cordata pollen was used to pollinate C. japonica. Without having pollen from a second Japanese Hornbeam to use for the intraspecific cross for comparison (i. e., a positive control), we cannot say whether improper timing of pollination or low individual tree fertility may have accounted for the lack of filled seed.

Evidence of Hybridity

Although DNA testing has yet to be done on the seedlings generated in these interspecific crosses, morphological evidence indicates they are indeed hybrids due to the observance of some of the characteristics of the putative pollen parent in the progeny. The evidence we observed is listed below. In accordance with convention, the seed parent is listed first and the pollen parent is listed second. In all cases comparisons were made to plants of a similar age, not to the adult plants which were actually used in the cross.

C. betulus x C. tschonoskii seedlings exhibit a level of pubescence intermediate to the two parent species.

C. caroliniana x C. betulus The serrations on the leaf margins were finer than in straight C. caroliniana. Seedlings from C. betulus ‘Quercifolia’ crosses exhibited some mildly lobulate leaves.

C. caroliniana x C. cordata. On the second year branches of the C. cordata and the putative hybrid there were raised horizontal lenticels not found on C. caroliniana.

C. caroliniana x C. coreana. The leaves of these seedlings possess a secondary vein number intermediate to the two parents, and a rounded outline of the primary teeth and a shorter length relative to its width, both characteristics from the pollen parent.

C. caroliniana x C. laxiflora. (Individual not located at the time of the observations and photographs)

C. caroliniana x C. orientalis. Relative to pure C. caroliniana Seedlings, leaves from the hybrids possessed an higher number of secondary veins and larger primary teeth.

C. caroliniana x C. tschonoskii. Hybrid seedlings possessed much greater pubescence than did pure C. caroliniana.

C. cordata x C. japonica. Compared to pure C. cordata the leaves are narrower relative to their length, have an higher secondary vein number, a reduction in secondary tooth size, and an intermediate leaf base shape (Fig. 3).

None of the seeds produced in the no-pollen controls germinated so that tests will not be able to be done to determine whether a low level of apomixis is possible in these taxa or if some pollen contamination occurred during the study.

The low germination rate in this study was disappointing, but the fact that intraspecific seedlots fared as badly as those of interspecific crosses indicates that post-zygotic genetic incongruity does not appear to be the cause. The use of a cut test to determine the number of filled seeds may have contributed to embryo mortality. Consequently, seeds generated in crosses made in 1998 season were not cut, even though an inflated number of filled seeds may have been presumed as a result.


Although additional verification is needed before we can definitively say the seedlings produced in this study are hybrids, our preliminary data indicates we have generated seven new interspecific hybrids: C. caroliniana x C. betulus, C. caroliniana x C. cordata,

C. caroliniana x C. coreana, C. caroliniana x C. laxiflora, C. caroliniana x C. orientalis, C. caroliniana x C. tschonoskii, and C. cordata x C. japonica. Two other crosses, C. caroliniana x C. turczaninovii and C. betulus ‘Quercifolia’ x C. cordata may be possible as well.

These results indicate that there may be widespread crossability within and perhaps between the sections of Carpinus. As other species are collected from around the world and additional seed sources of the species already in cultivation are obtained, especially from harsher, less accessible regions, additional sources of tolerances to biotic and abiotic stresses can be anticipated. With additional breeding and selection work using these genetic resources, the day may not be far off when Hornbeam is the first tree that comes to mind when looking for a beautiful and tough small tree.

Literature Cited

Bean, W. J. 1970. Trees and Shrubs Hardy in the British Isles, Vol. 1, 8th Ed. Rev. M. Bean and J. Murray Publ., Ltd., London, England. 845 pp.

E. G. Bobrov. 1970. Carpinus L., pp 202-207 In N. Landau (Trans.), Flora of the U.S.S.R. Vol. 5. Keter Press, Jerusalem, Israel.

Dirr, M. 1978. The Hornbeams - Choice Plants for American Gardens. Amer. Nurseryman. 148(10): 10-11, 46, 48, 50, 52.

______. 1990. Manual of Woody Landscape Plants, 4th Ed. Stipes Publ. Co., Champaign, IL. 1007 pp.

Hilliard, N. E. 1980. Environmental Factors and Plant Selection for Urban Landscapes. M. S. Thesis, State University of New York, Syracuse, NY. April 1980. 174 pp.

Marquard, R. D. 1992. Pollen Tube Growth in Carya and Temporal Influence of Pollen Deposition on Fertilization Success in Pecan. J. Amer. Soc. Hort. Sci 117(2): 328-331.

______. 1996. Director of Research, The Holden Arboretum, Kirtland, OH. (Personal communication.)

Meyer, P. W. 1995. Executive Director, The Morris Arboretum, Philadelphia, PA. (Personal communication.)

Rushforth, K. 1986a. Hornbeams and Hop Hornbeams. The Plantsman 7(3): 173-191.

Santamour, S. F. 1978. Interspecific hybridization in Carpinus. Metro. Tree Impr. Alliance (METRIA) Proc. 1: 73-79.

Tanouchi, H. and S. Yamamoto. 1995. Structure and Regeneration of Canopy Species in an Old-growth Evergreen Broad-leaved Forest in Aya District, Southwestern Japan. Vegetatio 117(1): 51-60.

Yinger, B. R. 1987. Plant Profiles: Carpinus coreana. Publ. Gard. J. Am. Assoc. Bot. Gard. Arbor. 2(1):15.


This work was made possible by a grant from the Electrical Power Research Institute to Dr. Harold Pellett, Department of Horticulture, University of Minnesota, St. Paul, MN and the Landscape Plant Development Center. I (SJW) would like to thank Dr. Pellett and the LPDC for the opportunity to work on this project while at the U of MN and for additional funding of Carpinus breeding research at the Morton Arboretum through the LPDC competitive grants program. I’d also like to thank the four LPDC member institutions who gave me access to their plant collections for this study, and to the staff at each arboretum who assisted me in my research and offered me their friendship as well.

Return to METRIA 10

Web Crafter: Anne S. Napier ~ Email:

Format updated July 24, 2009