Big-bracted dogwoods (Cornus sp.) are well-known plants in North America and eastern Asia where they occur as wild, generally spring-flowering understory trees. They are also popular ornamental landscape plants, and many economically important cultivars are propagated and sold across North America, Europe, and Asia. Starting in the late 1960s, Elwin Orton of Rutgers University in New Jersey (USA) utilized three geographically disjunct species of dogwoods, C. florida (eastern North America), C. nuttallii (western North America), and C. kousa (East Asia), in an extensive interspecific hybridization program. He was successful in developing the first-ever interspecific F₁ hybrids of these species, several of which have become staple items in the ornamental nursery trade due to their enhanced ornamental qualities and resistance to diseases. The original F₁ plants are still alive at Rutgers University. While they have been available for decades in horticultural commerce, the interspecific hybrid crosses were never formally described and their scientific hybrid names were never published. For the C. kousa × C. florida hybrids, the name Cornus ‘rutgersensis’ has been used on occasion in the horticultural trade, but without proper citation and description. Here, it is formally named Cornus × rutgersensis Mattera, T. Molnar, & Struwe, hybr. nov. For the C. kousa × C. nuttallii hybrids, no previous name has been used, and it is hereby named Cornus × elwinortonii Mattera, T. Molnar, & Struwe, hybr. nov. The need for providing scientific names for commonly used horticultural hybrids is discussed. Holotype material for both hybrid names was collected from the original F₁ hybrids for full documentation, typification, and description. The comparative intermediate development of leaves, inflorescence structures, and fruit types of the hybrids and their parents is discussed and illustrated. Etymology, phenology, and cultivation aspects of these hybrids and their cultivars including backcrosses to C. kousa are also presented.

Cornaceae , East Asia, horticulture, hybridization, nomenclature, North America

The circumboreal genus Cornus L. (Cornaceae, Cornales; APG III 2009) contains about 60 species divided into ten subgenera (Fan and Xiang 2001). Species in this genus express a wide variety of morphologies, from low herbaceous ground covers, such as the boreal-temperate species C. suecica L., to multi-stemmed shrubs, such as C. sericea L. It also includes small to large trees, such as C. kousa Buerger ex Miq. and C. nuttallii Audubon ex Torr. & A.Gray, the latter of which can grow up to 24 m tall. Some taxonomists have divided the genus up into six genera, but molecular studies have shown that Cornus in the current circumscription is monophyletic (Xiang et al. 2006).

Several species of Cornus have large, showy petaloid bracts located under tight head-like, multi-flowered inflorescences. These species form the monophyletic big-bracted (BB) clade sensu Xiang et al. 2006, and are mostly spring-flowering trees of North American and East Asian forests. The members of this clade are classified into three different subgenera: Cynoxylon, Discocrania, and Syncarpea (Xiang et al. 2006). The most commonly known big-bracted species in North America are C. florida L. and C. nuttallii of subgenus Cynoxylon and C. kousa of subgenus Syncarpea. Seed and clonally propagated big-bracted dogwoods are popular ornamental landscape trees in subtropical to temperate regions around the world. Their most conspicuous characteristics are their large, white or red petaloid floral bracts, showy red fruits, and brightly colored fall foliage (Li et al. 2009). Cornus kousa can be easily distinguished from the other two species by its round, fleshy multiple fused fruits formed from a whole flower head (as opposed to single, separate drupes from each flower arranged in clusters). It can also be identified by its acute or acuminate floral bracts, whereas the others have bracts that are rounded or retuse (Harrison 2009).

Typical horticultural uses of the big-bracted dogwoods include container, specimen, or shade plantings in suburban landscapes, display gardens, and parks (Gilman and Watson 1993a, b, Mohlenrock 2006). In the eastern and southeastern USA, Cornus florida is a common component of native deciduous forests, gardens, and home landscapes. It is among the first trees to bloom with conspicuous flowers in the spring in North America, with a range of cultivars available that express dwarf to vigorous growth habits and white, pink or red floral bracts. Cornus kousa is also a common component of ornamental landscapes in the eastern USA. It blooms about a month later than C. florida (after the leaves have developed), has a more vase-shaped growth habit, and most have white floral bracts, although a few forms with light pink bracts exist (Cappiello and Shadow 2005, Dirr 2009, Rhoades et al. 2011). The use of C. nuttallii in landscaping is much more limited than the former two species, due to limited winter hardiness in the eastern USA and it is cultivated mostly in the Pacific Northwest (USA), where it is native. Dogwood sales in the USA account for over 11% of the total deciduous flowering tree market, amounting to nearly 31 million USD in 2009 (Fulcher et al. 2012, NASS 2007).

The Rutgers University dogwood breeding program began in 1965 under the direction of horticultural plant breeder Dr. Elwin Orton. The early goals of the program were to develop novel cultivars of Cornus florida and C. kousa with improved aesthetic qualities, including pink and red floral bracts, unique growth habits, and superior disease resistance. Several years after the program started, attention was turned toward developing interspecific hybrids between these two species as well as between C. kousa and C. nuttallii, to help reach these goals (Elwin Orton personal communication). Because of differences in flowering times between the species, which can span more than a month, Orton used two approaches to make the hybrid crosses. First, he collected, dried, and stored pollen from earlier flowering plants to apply to the stigmas of those that bloomed later in the field and greenhouse. Second, he manipulated bloom times through the careful use of cold chambers and warm greenhouses to artificially break dormancy and match flowering times of container-grown plants to those in the field (E. Orton personal communication). Orton was ultimately successful in his interspecific hybridization attempts and is credited as being one of the first to create C. florida × C. kousa and C. kousa × C. nuttallii F₁ hybrids (Dirr 2009). To date, eleven interspecific cultivars, comprising eight from C. florida × C. kousa crosses and three from C. kousa × C. nuttallii hybrids, have been named, released, and patented through the Rutgers University dogwood breeding program (Table 1). The two classes of interspecific hybrids display intermediate morphological and phenological characteristics between the parental species (Cappiello and Shadow 2005, Dirr 2009, E. Orton personal communication, Orton 1990a, 1990b, 1990c, 1990d, 1990e, 1991, 2014, Orton and Gant 1993a, 1993b, 2006a, 2006b, 2004, 2007, 2011). Many also show increased vigor (rates of growth) compared to their parent species, as well as improved stress tolerance.

According to the International Code of Nomenclature for algae, fungi, and plants (abbreviated hereafter as ICN), a hybrid between two plant species can be given two types of scientific names to classify them within the taxonomic system of plant biodiversity (McNeill et al. 2012: Art. H1). Either the hybrid is listed with the name of the two parents separated by a multiplication (×) sign, such as in the oak hybrid Quercus alba × Quercus bicolor, or they may be given a unique name with the species epithet preceded by a multiplication (×) sign, such as Quercus × jackiana for the same hybrid (Haines 2011). For hybrids in horticulture and commerce, the second option is preferred since it provides a simpler name that is easier for horticulturalists and the public to learn, catalogue, use on labels, and remember. It also provides a scientific name that fits into existing databases already in use for commercial plants. For the two flowering dogwood hybrids discussed here, no formal scientific names have been proposed, although ‘Cornus × rutgersiensis’ and ‘Cornus × rutgersensis’ (sometimes without the multiplication sign, ×) have been used in popular and horticultural literature for many years to indicate Cornus florida × Cornus kousa hybrids (e.g., Gayraud 2013, Cubey et al. 2014; Shearer and Ranney 2013, Wikipedia 2014). Those names are currently invalid since, according to the ICN, all proposed scientific names, including hybrid names, require that they be formally published and described and be represented by a type specimen. A type specimen is the specimen to which the name is permanently attached and which is publicly available for consultation (McNeill et al. 2012).

The name Cornus × rutgersensis is proposed for the hybrid C. kousa × C. florida. A new name is also proposed, Cornus × elwinortonii, honoring our colleague Dr. Elwin Orton, for the hybrid he created between C. kousa and C. nuttallii. Full morphological descriptions, typification, illustrations, horticultural information with cultivar names, disease response, and a discussion on the formation of intermediate morphological traits with regard to leaf size, inflorescence structure, and fruits are provided for each of these new names. In doing this, we provide both formal names and summarize information of general botanical interest of these popular garden plants for botanists and horticulturalists.

Morphological intermediacy in hybrids. Interspecific hybrids are commonly intermediate in their morphology between their parents (e.g., Tovar-Sanchez and Oyama 2004). However, in hybrids between the big-bracted dogwoods, there is the added complication of the parental species having either a multiple, berry-like fruit or single-seeded drupes, and remarkably different inflorescence buds, bract morphology and phenological development. Despite such large differences, the hybrids clearly express intermediate phenotypes and provide good examples of ‘halfway’ morphologies created through hybridization. Intermediate traits include leaf size, inflorescence structure, and fruit type, which are three important ornamental characteristics of big-bracted dogwoods. The shape of the bract shape is also intermediate in these hybrids, leading to increased variation in bract shape. Also, the intermediate flowering times allow for a lengthening of the display of the ornamental bracts across the big-bracted clade. Both hybrids discussed in this paper also display novel characteristics not seen in previous dogwood cultivars. For example, ‘KN30-4’ Venus® displays larger floral bracts than in other hybrid cultivars or in the species of Cornus known to us.

Generally, C. × rutgersensis and its parents display similar tree shape and form, but the hybrid displays increased vigor and growth (Fig. 4). In C. × elwinortonii, tree shape and growth habit appear similar to the parents, C. kousa and C. nuttallii, but the hybrid is significantly more vigorous than C. kousa and shows increased growth in younger trees. Cornus × elwinortonii can be significantly larger in stature (to 8 m; Fig. 2) than most trees of C. kousa (to 6 m; Gilman and Watson 1993b), but hybrid tends to be significantly shorter than C. nuttallii (to 12 m, occasionally to 22.9 m; Gucker 2005).

The leaves of C. × rutgersensis are intermediate between the two parents, being longer and wider than C. kousa and shorter and narrower than C. florida. A similar phenomenon was recorded in the Quercus crassifolia Bonpl. × Q. crassipes Bonpl. hybrid complex (Fagaceae; Tovar-Sanchez and Oyama 2004) and in crosses between the herbs Brassica oleracea L. and Sinapis alba L. (Brassicaceae; Hansen and Earle 1996). The leaves of C. nuttallii and C. kousa are narrower (5-7 cm) than their offspring, C. × elwinortonii (5-8 cm), and the leaves of the hybrid can also have a crinkled appearance, which is not characteristic of either parent. Such novel hybrid characteristics are not unusual and have also been reported in the Quercus crassifolia × Q. crassipes complex (Tovar-Sanchez and Oyama 2004) and in Carica papaya L. × Vasconcellea cauliflora (Jacq.) A.DC. (reported as C. cauliflora Jacq.; Caricaceae; Magdalita et al. 1996).

Inflorescence bud morphology and development shows dramatic differences between the parents of C. × rutgersensis and is also correlated with large differences in floral bract display (and anthesis). The floral bracts of Cornus florida are displayed before vegetative bud-break in early spring, whereas in C. kousa the floral bracts are displayed after the foliage is fully developed. In C. florida, the inflorescence bud consists of two pairs of floral bracts (inner and outer) tightly clinging to a well-developed inflorescence head. Underdeveloped vegetative bracts are present but do not cover the inflorescence. Cornus kousa has two pairs of floral bracts that tightly cling to the underdeveloped inflorescence. In addition, they are tightly covered by two pairs of vegetative bracts. The hybrid displays an intermediate flower bud in which floral bracts cling to the inflorescence and vegetative bracts cling loosely to the flower head (Fig. 3c and 5). As expected, intermediate inflorescence bud development leads to intermediate floral bract display and flowering time. The flowering period of Cornus × rutgersensis ranges from the end of flowering in C. florida to the beginning of flowering in C. kousa. Correlation of morphological variation in floral bud shape to intermediate flowering time has also been reported in hybrids between Fraxinus excelsior L. and F. angustifolia Vahl (Oleaceae; Gerard et al. 2006). Notably, in C. × rutgersensis, the floral bracts tend to only weakly cover the inflorescence during overwintering, resulting in 10–45% of the flowers being naked (exposed).

Differences in the inflorescence buds also exist for C. × elwinortonii and its parents, C. nuttallii and C. kousa. The underdeveloped inflorescence head in C. kousa is tightly covered by two pairs of floral and vegetative bracts. Cornus nuttallii has a completely exposed inflorescence head, where the small floral and vegetative bracts do not cover the developing flower buds. Cornus nuttallii, native to the Pacific Northwest (USA), is exposed to milder winter temperatures than C. kousa from eastern Asia. Cornus × elwinortonii displays an intermediate bud ranging from completely exposed to completely covered. In Cornus × rutgersensis, there is strong variation in the degree of coverage by the bracts, with 10-45% naked to nearly completely covered floral buds.

Distinct differences between inflorescence architectures can also be observed between parents and their hybrids. In Cornus kousa, all flowers in the inflorescence are fused, creating a densely merged ball of flowers, while in C. florida the flowers are not fused, creating a more open structure. The flowers in their hybrid, C. × rutgersensis, are densely packed and at first appear to be fused together; however, they are separate even if closely positioned (Fig. 3 and Fig. 4A). The number of flowers in each inflorescence varies greatly within big-bracted dogwoods, from a few dozen to over one hundred. The hybrids also show intermediacy in the number of flowers: C. × rutgersensis (30–50 flowers/head) from parents C. florida (20–30) and C. kousa (20–50), and C. × elwinortonii (55–80) from parents C. kousa (20–60) and C. nuttallii (70–100).

In C. florida, the individual flowers develop into single-seeded drupes, while in C. kousa the fused flowers develop into single-seeded druplets that are fused into a multiple, berry-like fruit (Fig. 5). The nearly always sterile hybrid C. × rutgersensis may produce parthenocarpic fruit displaying intermediate characteristics (Fig. 5). Fruits containing seeds swell and develop into individual drupes or drupelets. This is the only example we know of where a hybrid has been created between parents with single and multiple fruit types. The hybrid between Jatropha curcas L. and J. integerrima Jacq., formed from the crossing of plants with large drupaceous fruits (J. curcas) and small deeply lobed capsules, displayed an intermediate fruit shape between two different fruit types as well (Rupert et al. 1970; Sujatha and Prabakaran 2002).

Success of hybrids. Ornamental plants play an important role in society, providing aesthetic value, shade, wildlife habitat and food, and soil stabilization. As popular ornamental trees in temperate and sub-tropical regions worldwide, improved cultivars of big-bracted dogwoods are desired. Demand for novel, vigorous, and disease-resistant plant material is high; however, limited genetic variability can exist for some traits. For example, there are only a few cultivars of C. florida that express resistance to powdery mildew (Windham et al. 2003, Windham and Witte 1998) and the floral bracts of C. kousa and C. nuttallii lack the dark red of the most successful C. florida cultivars (Cappiello and Shadow 2005, Dirr 2009). Orton’s use of interspecific hybridization to develop the novel plants described here (Cornus × rutgersensis and Cornus × elwinortonii) resulted in the successful development of cultivars with enhanced aesthetic qualities and improved disease resistance.

Upon its introduction to the US from Asia, dogwood anthracnose devastated natural stands of C. florida, a plant species highly susceptible to this fungal disease. For example, mortality rates as high as 86% occurred in a ten year period in Connecticut (Holzmueller et al. 2006). Cornus nuttallii is also highly susceptible to dogwood anthracnose. The Asian dogwood C. kousa occurs sympatrically with the causal agent of dogwood anthracnose Discula destructiva in Asia, and most cultivars of C. kousa have a high level of tolerance or resistance to this disease (Hibben 1990, Ranney et al. 1995). Because of results from field evaluations and the C. kousa parentage, all of Orton’s hybrids were believed to be highly resistant to this disease at the time of their commercial release. The Stellar® Series and Jersey Star® releases came at a time when disease incidence was high in the United States. However, Ranney et al. (1995) showed that not all of the Rutgers hybrids maintained resistance over the years, although some still displayed tolerance.

Powdery mildew, believed to be introduced from Asia, is less devastating to natural stands of C. florida. Instead, this disease has strongly impacted the nursery industry, raising production costs and reducing aesthetic appeal. Cultivars of C. florida display little resistance to this fungal disease. Of more than 100 available cultivars of C. florida (Santamour and McArdle 1985), only five (‘Jean’s Appalachian Snow’, ‘Karen’s Appalachian Blush’, ‘Kay’s Appalachian Mist’, ‘Appalachian Joy’ and ‘Cherokee Brave’) display high levels of tolerance or resistance to powdery mildew (Li et al. 2009, Ranney et al. 1995). Again, cultivars of C. kousa generally show high levels of tolerance (Li et al. 2009, Ranney et al. 1995). Due to Orton’s selection of parents, several cultivars of C. × rutgersensis (e.g., Stellar Pink®, Aurora®, Stardust®, Celestial®, and Constellation®) are resistant to powdery mildew (Li et al. 2009).

Scientific naming of horticultural plants. Crucial to communication in all parts of our lives is the naming of objects and phenomena. We need words to tell other people what we are talking about, and the words need to have uniform and clear meanings. For botany, our scientific names form such a uniform language that is universal and used in fields including biodiversity inventories, phytochemistry, horticulture, crop plants, and other scientific and/or economic endeavors. Many scientific plant names are listed in the International Plant Names Index (http://www.ipni.org) and in other resources such as floras, dictionaries, The Plant List (http://www.theplantlist.org/), RHS Plant Finder (http://www.rhs.org.uk/plants/), Encyclopedia for Life (http://eol.org), Wikipedia (http://wikipedia.com). Unfortunately, many misspelled, outdated, unpublished, illegitimate, and invalid names are still in use worldwide in popular literature, websites, and non-taxonomic publications, especially for commonly cultivated and medicinal plants (Bennett and Balick 2014 for examples, see Struwe 2014).

It can be argued that we do not need formal scientific names for all artificially created hybrid plants, since cultivar and trademark names exist and names of cultivated plants follow The International Code of Nomenclature for Cultivated Plants (ICNCP, Brickell et al. 2009). However, names of hybrids following the International Code for algae, fungi, and plants may be useful when cataloging species diversity, natural or human-made, and linking hybrids with their parental species. Cornus × rutgersensis is a name already in use on a global scale, but was never proposed formally according to the rules of the ICN. Validating this name is the simplest way to provide an acceptable and useful name to the horticultural community. Since the second hybrid, C. × elwinortonii, is also a commonly grown and well-known hybrid in gardens, to propose it formally is also useful. Even if self-propagating seedlings from these hybrids are not known, we do know that viable seeds are sometimes produced; making it is possible that spontaneous progeny will arise in the future

The hybrids Cornus × rutgersensis (C. florida × C. kousa) and Cornus × elwinortonii (C. kousa × C. nuttallii) were developed at Rutgers University by Dr. Elwin Orton, and are good examples of controlled hybrid crosses showcasing intermediate morphological and phenological characteristics for leaf size, inflorescence bud structure, flowering time, and fruit structure. The horticultural success of big-bracted dogwood hybrids in the nursery and landscape industry can largely be attributed to their inherent disease resistance and enhanced aesthetic qualities that represent novel intermediate phenotypes between their parent species.

The authors gratefully thank Elwin Orton for providing breeding and morphological information regarding these hybrids, Bobbi Angell for providing the line drawings, Kanchi Gandhi at Harvard University Herbaria (and IPNI.org) for nomenclatural expertise and advice, and John Capik for editing. This project was partially funded by the New Jersey Agricultural Experiment Station at Rutgers University. LS was funded by US Department of Agriculture award USDA/NJAES-NJ17112.