B. davidii is a multi-stemmed shrub or small-tree that is native to China and has been introduced as an ornamental world-wide, first to Europe (1890s) and then later to t...
B. davidii is a multi-stemmed shrub or small-tree that is native to China and has been introduced as an ornamental world-wide, first to Europe (1890s) and then later to the Americas, Australia, New Zealand, and some parts of Africa. Since that time, B. davidii has naturalized within sub-oceanic climates in the temperate and sub-Mediterranean zones.
The full potential of this species has yet to be realized; however, it is already considered problematic (i.e. out-competing native, agricultural, and forestry species) in northwestern and northeastern USA and Canada, throughout New Zealand, and in central Europe. B. davidii is tolerant of a broad range of environmental conditions, capable of prolific seed production, grows rapidly, and has a short juvenile period. Due to its popularity, nurseries continue to distribute plants capable of setting seed. Garden residents as well as escapees serve as satellites, which then spread the species on to disturbed and wild lands and this is a cause for concern.
Currently, Missouri Botanic Garden (2010) assigns Buddelja to Loganiaceae, while Flora Europaea (2010) places it in Buddlejaceae.
The genus Buddleja was named by Linné (1737) to honor the English amateur botanist Reverend Adam Buddle. The species was named after Father David who collected and returned specimens of the Chinese flora and fauna to Adrien René Franchet at the Paris Musée National d’ Historie Naturelle (Bean, 1970; Tallent-Halsell and Watt, 2009).
Since it was first named, there have been several spellings of the genus Buddleja such as Buddleia,Buddlea, Buddleaea, Budlea, Buddleya and Budleia (Coats and Creech, 1992; Tallent-Halsell and Watt, 2009). Based on International Rules of Nomenclature the spelling adopted by Linné in 1753 and 1754 must be retained (Sprague, 1928) and thus Buddleja is spelt with a ‘j’ (Tallent-Halsell and Watt, 2009).
Seven subspecies of B. davidii and 90 B. davidii cultivars have been described as of 2008 (Stuart, 2006).
B. davidii is a shrub or small multi-stemmed tree that has a great degree of morphological and physiological plasticity (Miller, 1984; Shi et al., 2006). Descriptions of B. davidii may vary slightly depending on the environment (Tallent-Halsell and Watt, 2009). B. davidii is semi-deciduous: leaves are shed in the autumn and immediately replaced with a set of new, smaller leaves that persist until the following spring (Miller, 1984; Tallent-Halsell, 2008). In general, stems are four-angled. Sub-orbicular to ovate stipules are present and range in length from 1-6 mm (Leeuwenberg, 1979; Webb et al., 1988; Wu and Raven, 1996). Leaves are opposite, usually ovate and shortly petiolate. The upper surface of leaves is dark-green and glabrous or free of hairs, whereas the lower surface is white-tomentose (Binggeli, 1998) with stellate and glanduliferous hairs (Leeuwenberg, 1979; Webb et al., 1988; Wu and Raven, 1996). Leaves range from 5-20 cm long and 1-7 cm wide (Wu and Raven, 1996). Leaf margins are finely toothed (Leeuwenberg, 1979).
B. davidii inflorescences appear at the terminal end of branches arranged in indeterminate corymbose-panicles that can extend up to 30 cm long (Findley et al., 1997). The hermaphroditic flowers in the wild are commonly lilac and purple whereas flowers of cultivars range from white to yellow and red (Stuart, 2006). The calyx is narrowly campanulate, villous, usually four-lobed and 3 mm long (Iwatsuki et al., 1993). The corolla is made up of four petals that are fused for three-quarters of their length into a corolla-tube (Tallent-Halsell and Watt, 2009), which is approximately 5-8 mm long. The flowers are zygomorphic, possessing four stamens adnate slightly above the middle of the corolla-tube (Iwatsuki et al., 1993). The interior of the flower is orange with a series of yellow nectar guides leading to the interior of the tube (Tallent-Halsell and Watt, 2009). The superior ovary is bilocular (Leeuwenberg, 1979).
Flowering is asynchronous. Each panicle consists of individual flowers that mature acropetally from the base to the top of the inflorescence (Findley et al., 1997). Individual flowers last for 1-3 days, whereas a panicle may persist for >2 weeks (Findley et al., 1997).
The seeds are contained in cylindrical two-valved capsules that are brown, narrowly ellipsoid to narrowly ovoid 5-9 x 1.5-2 mm, acute at the apex, narrowed towards the base, smooth or with stellate hairs (Wu and Raven, 1996; Wilson et al., 2004b). Seeds are brown, thread-like, and long-winged at both ends. They range in size from 3-4 x 0.5 mm with the centre slightly thickened (Norman, 2000). Approximately 500-100 seeds are arranged tightly packed with their long sides aligned with the axis of the capsule (Miller, 1984).
B. davidii is fast-growing and has been reported to increase 0.5-2 m in height annually (Owen and Whiteway, 1980; Watt et al., 2007). Seedling stem diameter can increase annually by as much as 5.6 cm-1 (Watt et al., 2007).
The life span of B. davidii is variable although in general individuals do not live for more than 20 years, often dying from stem rot (Smale, 1990; Binggeli et al., 1998). However, plants >30 years have been recorded (i.e. based on tree rings and historical aerial photographs) in New Zealand (Bellingham et al., 2005).
Plant Type
Top of page
Broadleaved Perennial Seed propagated Shrub Tree Vegetatively propagated Woody
B. davidii is native to central and southwestern China at elevations up to 3500 m, occurring naturally in the following Provinces: Gansu, Guangdong, Guangxi, Guizhou, Hubei, Hunan, Jiangsu, Jiangxi, Shaanxi, Sichuan, Xizang, Yunnan and Zhejiang (Wu and Raven, 1996). The species can be found both on mountainous slopes and in lowlands.
In Europe, B. davidii has naturalized in most Central European countries. Its occurrence has been noted as extending from the Mediterranean in the south to Bergen, Norway, and from Spain in the west to Bulgaria in the east (Sheppard et al., 2006; Ebeling et al., 2008b; GBIF, 2009). The species has been found naturalized in South Africa, Zambia, Zimbabwe (Aluka African Plants, 2009), in Cameroon (GBIF, 2009) and in several Asian countries, such as South Korea (Stuart, 2006).
In North America, B. davidii occurs along the eastern coastline, and as far inland as Tennessee. On the west coast, the species occurs from California to British Columbia, Canada (Reichard, 1996; NatureServe, 2009; Tallent-Halsell and Watt, 2009). In South and Central America the species have been recorded for Peru, Ecuador, Bolivia, Columbia, Panama, and Mexico (GBIF, 2009; DJ Kriticos, CSIRO Entomology, Australia, personal communication, 2009).
Within Australasia, the species has been recorded as occurring in all Australian states apart from Western Australia and Northern Territory (DJ Kriticos, CSIRO Entomology, Australia, personal communication, 2009). In New Zealand, B. davidii has spread throughout both islands with the most invasive populations occurring in the North Island (Esler, 1988; Webb et al., 1988; Gibb, 1994; MS Watt, Scion, Christchurch New Zealand, personal communication, 2009).
The distribution in this summary table is based on all the information available. When several references are cited, they may give conflicting information on the status. Further details may be available for individual references in the Distribution Table Details section which can be selected by going to Generate Report.
In 1869, Father David sent specimens of B. davidii to Franchet (Franchet, 1884, 1888). Specimens of the same species from I’ch’ang Province, China were collected by Henry and named by William Botting Hemsley in 1887 (Anon., 1925). Unaware of Franchet’s description, Hemsley called the plant Buddleja variabilis Hemsley (Hemsley, 1889). The name was eventually reversed 25 years later, due to the discovery of Franchet’s original description. However, B. variabilis is still listed as a synonym of B. davidii.
B.davidii seeds were first introduced to Europe from Russia by traders (Bean, 1970); however, these seeds produced an inferior form (Bean, 1970; Coats and Creech, 1992). Hemsley (1889) reported that seeds from Pa-tung (Hubei Province, China) were sent to England ca. 1889, but these did not produce flowering plants. Seeds from Tatsienlu, China that were introduced to Louis DeVilmorin of France in 1893 by Jean André Soulié (Herberman, 1919) produced superior plants that were considered suitable for horticulture (Cox, 1986). In 1896, seeds from these Tatsienlu specimens were sent to Kew Gardens, UK (Coats and Creech, 1992).
Further collections of B. davidii seeds were sent from Mt. O’mei Shan, China in 1896 by Father Paul Guillaume Farges (PlantExplorers, 2009) and in the following year by Henry from I-ch’ang, China. Between 1907 and 1910, Wilson collected seeds in the Hubei and Sichuan Provinces, China from which the common garden-variety B. davidii descended (Rehder, 1927; Bean, 1970).
B. davidii naturalized on a significant scale in the 1930s in parts of Europe, after the destruction of cities during World War II. Bombed sites and building rubble were suitable colonization habitats and therefore dense B. davidii thickets established on these sites (Kreh, 1952; Kunick, 1970; Owen and Whiteway, 1980; Miller, 1984; Coats and Creech, 1992; Tallent-Halsell and Watt, 2009). In the 1950s and 1960s, B. davidii became a popular garden shrub, which further contributed to its spread when it escaped from cultivation and naturalized in the wild (Owen and Whiteway, 1980; Miller, 1984; Tallent-Halsell, 2008).
B. davidii has the potential for further expansion (Kriticos et al., 2007; Ebeling et al., 2008b; DJ Kriticos, CSIRO Entomology, Australia, personal communication, 2009). Areas most at risk include Eastern Europe, South Africa, Western Australia and South America. It is most likely that further distribution will be attributed to the horticultural industry. Several of the known cultivars of B. davidii show invasive potential (Anisko and Im, 2001; Moller, 2003; Ream, 2006). Rapid maturation, millions of wind-dispersed seeds and a high rate of germination will positively contribute to range expansion.
B. davidii can spread along rail tracks where seeds are either carried on the locomotives or blown and drawn along in the slipstream of trains (Miller, 1984; Tallent-Halsell and Watt, 2009). Abandoned railway lines, where weeds are not controlled, expedite the spread of B. davidii when they grow into productive thickets in the railway corridors. Automobiles have been found to physically disperse B. davidii seeds (von der Lippe and Kowarik, 2007). Germinants have been observed in the mud stuck to machinery, especially that of gravel mines in floodplains (N Tallent-Halsell, Southwest Ecosystem Services, Las Vegas, USA, personal communication, 2009). Furthermore, B. davidii can spread along sea coasts, floodplains and riparian corridors, limestone quarries, and road and forest edges (Tallent-Halsell and Watt, 2009).
B. davidii thrives on a wide range of soil types. The species is able to establish on calcium based building debris and masonry walls (Owen and Whiteway, 1980), on soils that are high in sand, nutrient poor and in high calcareous substrates (Miller, 1984; Godefroid et al., 2007). Humphries and Guarino (1987) reported B. davidii to be able to flourish in calcium-deficient soils. Moreover, B. davidii is capable of colonizing areas with a pH of 6.0 to 8.91 (Miller, 1984; Godefroid et al., 2007).
Naturalized B. davidii is considered to out-compete native, agricultural, and forestry taxa. It competes with the plantation species Pinus radiata (Richardson et al., 1996) in New Zealand for light (i.e. the thick stands of B. davidii impede germination and growth of seedling and saplings). Ream (2006) and Leach (2007) reported the replacement of riparian native Salix ssp. and Populus spp. by B. davidii in Oregon and Washington, USA. Although B. davidii colonizes disturbed sites, whether it alters successional trajectories over the long term is yet undetermined (Tallent-Halsell, 2008).
Chromosomal analyses indicate that the basic chromosome number of B. davidii is 2n=76 (4x) (Moore, 1960; Iwatsuki et al., 1993; Chen et al., 2007). Breeding programmes began as early as 1920 when W. van de Weyer developed interspecific hybrids (Buddleja globasa x Buddleja magnifica) (Moore, 1960; Wilson et al., 2004a). However, only cultivars that have been bred in Europe and the USA since the 1930s have economic value to the horticultural industry (Albrecht, 2004). In addition to the more than 90 B. davidii cultivars (Stuart, 2006), there are two hybrids: (1) Buddleja davidii x globosa that is classified as B. x weyeriana and characterized by yellow to orange flowers;and (2) Buddleja fallowiana x davidii named Buddleja davidii ‘Lochinch’ (Wigtownshire, Scotland). B. davidii ‘Lochinch’ was thought to be a sterile alternate to B. davidii; however, field observations revealed that it produces abundantly by seeds (EPPO, 2005).
Since the late 1990s several hybridization programmes have been initiated (Wilson et al., 2004a). They have focused on altering flower and leaf colour, inflorescence morphology, growth habit, and developing sterile plants and new hybrids (Lindstorm et al., 2002; Lindstorm and Burkett, 2004; Wilson et al., 2004a; Podaras, 2005).
Reproductive Biology
B. davidii flowering and fruiting normally occur when the plant reaches 2 years old (Miller, 1984; Watt et al., 2007), although anecdotal information indicates that it may occasionally occur in the first year (Kreh, 1952; Owen and Whiteway, 1980; Esler, 1988; S Ebeling, UFZ Helmholtz Centre for Environmental Research, Germany, personal observation, 2009). Flowering is initiated in response to long days of summer (Moore, 1960). The flowering period has been found to extend from late spring to the middle of autumn in the northern hemisphere (Wu and Raven, 1996) and from early summer to late summer, and occasionally as late as mid-autumn, in the southern hemisphere (Webb et al., 1988).
B. davidii does not self-pollinate and therefore depends on insect pollinators (Miller, 1984; Norman, 2000). Due to flower morphology and abundant nectar, butterflies may be sufficient pollinators although bees, hummingbirds and other insects visit the flowers (Miller, 1984; Houghton et al., 2003; Owen and Whiteway, 1980).
A single mature B. davidii individual can produce millions of seeds; however, estimates of the number of seeds produced vary (100,000 to 3,000,000) among cultivars (Miller, 1984; Brown, 1990; Wilson et al., 2004b; Thomas et al., 2008c). Seed formation and ripening typically occur within 3 weeks after flowering, but are retained in the capsules throughout winter (Miller, 1984). During arid periods, the valves of the capsule dry and curl outward (Miller, 1984). This exposes the seeds to the air and enables dispersal by air movement (Miller, 1984; Stuart, 2006). Seed release stops with increasing humidity by closing capsules until conditions dry again. Once released, the majority (95%) of seeds from one individual were dispersed 10 m or further from the parent (Miller, 1984). B. davidii seeds are also reported to be water-dispersed (Miller, 1984; Webb et al., 1988; Brown, 1990).
B. davidii has a short-lived seed bank. In the laboratory, seed viability remained high, up to 2.5 years, but declined rapidly afterwards (Miller, 1984).
Seeds of B. davidii do not demonstrate innate dormancy (Miller, 1984). Light, temperatures above 6°C (optimal 25°C; Jay, 2006), and a base water potential of -1.8 MPa are required for germination (Tallent-Halsell and Watt, 2009).
A comparison of native and invasive populations of B. davidii in Europe showed strong evidence for increased plant vigour in the introduced range: plants in invasive populations were taller, had thicker stems, larger inflorescences and heavier seeds than plants in native populations (Ebeling et al., 2008a). Moreover, herbivory was substantially reduced in invasive populations (Ebeling et al., 2008a).
Common garden experiments in the invasive range in Europe comparing invasive populations did not provide evidence for local adaptation to climatic conditions (S Ebeling, UFZ Helmholtz Centre for Environmental Research, Germany, personal observation, 2009).
The current distribution of B. davidii indicates the core distribution to be in warmer humid regions including temperate, Mediterranean and subtropical climates (Tallent-Halsell and Watt, 2009). Therefore, it is assumed that the distribution into cooler continental climates is limited (Krivanek and Pysek, 2006; DJ Kriticos, CSIRO Entomology, Australia, personal communication, 2009). An eco-physiological study of the frost hardiness in B. davidii revealed no local adaptation to minimum temperatures (Ebeling et al., 2008b). This is in line with results of the combined niche model that did not detect a niche shift between the species’ native range in China, and its invasive range in Europe and North America. Furthermore, the niche model showed that the potential invasive range of B. davidii is still not completely occupied suggesting that climatic conditions are currently not limiting the further spread of this species (Ebeling et al., 2008b).
Ebeling et al. (2008b) are consistent with the results of Kriticos et al. (2007; DJ Kriticos, CSIRO Entomology, Australia, personal communication, 2009) based upon the process-based distribution model that predicted that stress from cold combined with insufficient thermal accumulation, excluded B. davidii from areas in China as well as prohibited the spread of the species into most of Canada, the Russian Federation, Scandinavia and northern inland USA. These results may vary depending on the cold hardiness of the genotypes of B. davidii (plant breeding in B. davidii has focused on enhancing cold hardiness; Albrecht, 2004; Podaras, 2005).
In contrast, the distribution model by Kriticos et al. (2005, 2006, 2007; DJ Kriticos, CSIRO Entomology, Australia, personal communication, 2009) found that heat stress excluded the distribution of the species from most low lying areas within the tropics, southern and central USA, low lying northern regions in South America, most African regions north of the equator (with the exception of Ethiopia and the coastal fringe of the Mediterranean), and interior and coastal regions in Australia north of Brisbane.
Dry stress appears to exclude B. davidii from the dry steppe and desert climatic regions throughout the world (Kriticos et al., 2007).
The process-based distribution model developed by Kriticos et al. (2007; DJ Kriticos, CSIRO Entomology, Australia, personal communication, 2009) projected potential for further expansion. Areas most at risk included Eastern Europe, South Africa, Western Australia, and a broad coastal strip of land within South America from Rio de Janeiro in the north to central Argentina in the south.
Nitrogen (N) and water are important resources limiting plant growth. Invasiveness in plants can increase by increasing N- and water-utilization efficiencies (Feng et al., 2007). The invasion of B. davidii might be facilitated by having high concentrations of N and phosphorus relative to other woody species (Cornelissen et al., 1996; Bellingham et al., 2005; Feng et al., 2007; Thomas, 2007). In addition, B. davidii are able to assimilate nitrate in their leaves rather than in the roots and stems (Al Gharbi and Hipkin, 1984). B.davidii has been found to allocate more leaf N to photosynthesis and consequently has a higher photosynthetic capacity than several other woody species (Feng et al., 2007).
In a comparative study of 80 woody species of the UK and north Spain, Cornelissen et al. (1996) found that seedlings of B. davidii have the second highest specific leaf area (SLA), which increased more rapidly than most other species as the plants matured (Feng et al., 2007; Thomas, 2007). The rapid growth of B. davidii might be explained by maintaining leaf area irrespective of form at low N (Humphries and Guarino, 1987). The survival of seedlings at the cotyledon to first leaf stage is strongly affected by soil matric potential. Humphries et al. (1982) found that B. davidii seedlingsseem to be more tolerant of a reduction in water supply than the native, Betula pendula.
Harley and Harley (1987) reported B. davidii as non-mycorrhizal. However, Dickie et al. (2007) found the presence of arbuscular mycorrhizae (AM) in B. davidii in New Zealand, the UK and North America.
Several types of phytochemical compounds had been isolated from B. davidii. Yu (1933) isolated a flavonol glycoside from leaves that he called buddleoflavonoloside. This is identical to linarin (Baker et al., 1951), which was isolated from the flowers as well as free aglycone acacetin. Fan et al. (2008) found linarin in B. davidii leaves, which has acetylcholinesterase-inhibitory properties.
Three iridoid glycosides: aucubin; catapol; and methylcatapol, have been isolated from leaves (Paris and Chaslot, 1955; Duff et al., 1965). Yoshida et al. (1978a, b) obtained five sesquiterpenes, named buddledins A, B, C, D, and E from the roots of B. davidii. Buddledins A, B and C were found to have a caryophyllene skeleton and to be piscicidal (Yoshida et al., 1976). Houghton (1984) reported the presence of coniferaldehyde and related lignan-type compounds in the stem of the species. B. davidii extracts demonstrated antifungal activity against soil fungi (Houghton et al., 2003).
Associations
In Germany, B. davidii is associated within the Urtico-Sambucetalia nigrae ordo nov., an order that contains nitrophile shrubs on waste ground, disposal sites, railways and eutrophicated agricultural land (Schubert et al., 2001). B. davidii is also associated with Ailanthus altissima (tree of heaven) and Rubus armeniacus (Himalayan blackberry) (Schubert et al., 2001) and assorted Salix species (Kunick, 1970).
The impact that B. davidii has on plant communities over the long-term is yet undetermined (Bellingham et al., 2005; Tallent-Halsell, 2008). Miller (1984) did not reveal intraspecific competition nor predictable development sequences of vegetation associated with B. davidii. Conversely, studies by Williams (1979) and Smale (1990) in Urewera National Park, North Island, New Zealand revealed that B. davidii quickly displaced primary native colonizers on floodplains. However, this accelerated the reforestation process back to native forest in floodplains when undisturbed (Smale, 1990).
B. davidii leaves are palatable to cattle and goats, but apparently not to deer (Gillman, 1998). Additionally, leaves appear to be palatable to polyphagous insects such as slugs, snails and aphids, but also to the glasshouse whitefly [Trialeurodes vaporariorum] and red spider mite [Tetranychus urticae] (Miller, 1984; Gillman, 1998). Specialized insects feeding on B. davidii have been identified: the weevils, Gymnaetron tetrum, Cleopus japonicus, and Mecysolobus erro; a dipteran leaf miner (Amauromyza verbasci) and a leaf bug (Campylomma verbasci) (Tallent-Halsell and Watt, 2009).
The stem borer, M. erro is distributed in several provinces of China (M Kay, Environmental Risk Management Authority New Zealand (ERMA), New Zealand, unpublished data).
The defoliator, C. japonicus has been identified in Southeast Asia, but appears to be restricted to the same distributional limits as B. davidii. Interestingly, other members of the genus (e.g. Cleopus pulchellus, Cleopus solani) occur throughout Europe (M Kay, Environmental Risk Management Authority New Zealand (ERMA), New Zealand, unpublished data).
B. davidii has evolved strategies to survive defoliation. In comparison to foliated plants, Watt et al. (2007) found high defoliation induced increased light-use efficiency, biomass allocation to leaves, specific leaf area, and reduced rates of leaf loss. Partially defoliated B. davidii plants have also been found to have greater leaf size and retain leaves for longer periods, than foliated plants (Thomas et al., 2008b). However, defoliation does appear to reduce seed number and mass per plant (Thomas et al., 2008c). Despite the relatively strong compensatory response to defoliation that B. davidii exhibits, repeated herbivory over several growing seasons is likely to negatively impact growth. B.davidii has been found to remobilize nitrogen (N) for new spring growth from older leaves with little contribution of N from woody tissue, even when they are substantially defoliated (Thomas et al., 2008a).
The release of B. davidii seeds takes place in early spring (northern hemisphere) or late autumn (southern hemisphere), during dry periods when capsules open (Miller, 1984). Miller (1984) reported that 95% of the seeds fall outside of a 10 metre radius, whereas Ream (2006) determined dispersal distances of up to 14 metres depending on structure of the habitat. The maximum dispersal distance by wind has not been determined.
B. davidii seeds are also reported to be water-dispersed, especially along sea coasts, flood plains and riparian corridors (Miller, 1984; Webb et al., 1988; Brown, 1990), where they can be washed downstream and establish new populations (ISSG, 2009).
B. davidii was introduced to Europe and North America as an ornamental and since that time, various breeding programmes have continued to develop B. davidii hybrids and cultivars. Gardens are a leading source of spreading, naturalized B. davidii populations (Ream, 2006). Currently, B. davidii is widely cultivated and an extremely popular garden plant of economic value to the horticultural industry (Turnbull, 2004; Wilson et al., 2004a). Due to its popularity, the horticultural trade has been recognized as one of the main dispersal pathways for this and other plant invasions (Dehnen-Schmutz et al., 2007). Nevertheless, production nurseries continue to contribute to the spread of B. davidii by serving as satellites that subsequently compound the number of seeds that are dispersed beyond gardens (Ream, 2006).
The negative impact of naturalized B. davidii is competition with plantation pine species. The species has had a substantial and detrimental impact on the growth of plantation species by restricting light availability in a number of countries, including New Zealand (Richardson et al., 1996). In Europe, transportation routes have been negatively effected (Reinhardt et al., 2003), but there is no analysis of the costs caused by the negative economic impact of B. davidii.
In New Zealand, B. davidii is estimated to cost the forestry industry between NZD $0.5 and 2.9 million annually in control costs and loss of production (Kriticos, 2007).
In Urewera National Park, North Island, New Zealand, B. davidii quickly displaces primary native colonizers on floodplains. This accelerates the reforestation process back to native forest in streambeds (Williams, 1979; Smale, 1990).
Impact on Biodiversity
Giuliano et al. (2004) showed that Lepidoptera in urban parks in New York City visited B. davidii more than other plants in the same vicinity, which might affect pollination success of native plant species. In contrast, Pfitzner (1983) revealed that butterflies preferred stands of Lythrum salicaria (purple loosestrife) rather than B. davidii growing in gardens next to the observed habitat.
The horticultural industry benefits greatly from the sale of B. davidii (Turnbull, 2004; Wilson et al., 2004a). Certain cultivars were worth over US$ 200,000/year to Georgia/USA plant growers (Dirr, 1997). To growers outside of Georgia, plants were worth over US$ 1,000,000 annually (CANR, 1996). The value of B. davidii for the horticultural industry has not been assessed in other continents.
Social Benefit
B. davidii is an extremely popular garden plant due to its low maintenance, long flowering season, colourful and fragrant flowers, and its attractiveness to butterflies. B. davidii readily colonizes disturbed and/or abandoned lands rendering them more desirable to the public (Tallent-Halsell and Watt, 2009). As an example, B. davidii colonized an abandoned quarry that had been previously used as an asbestos waste dumping area (Doughty, 2007; Theivam and Allen, 2007; KCGG, 2009). Over time, the resulting B. davidii thickets were so valued for their beauty and wildlife that they were registered as a “Village Green” through the effort of the “Keep Croxley Green” group (KCGG) in the UK and thus guaranteed to be protected from destruction for perpetuity (KCGG, 2009).
Environmental Services
The flowering B. davidii has been closely linked with butterflies, moths and hummingbirds. Several butterfly species have been found on B. davidii (Owen and Whiteway, 1980; Giuliano et al., 2004; Stuart, 2006). Also other insects have been observed as visitors, such as wasps, hornets, lacewings and beetles (Pickens, 1931; Stuart, 2006).
In Eugene/Oregon, USA, B. davidii is prohibited (USDA-NRCS, 2009).
Efforts to curtail the spread of B. davidii in Oregon have proved to be ineffective because only B. davidii and not the cultivated varieties was elevated to the noxious weed quarantine list in 2004 (Ream, 2006).
Public awareness
The Oregon State University extension Service Master Gardener Programme does not recommend B. davidii for butterfly gardens because of its invasiveness (Savonen, 2009).
In New Zealand, B. davidii is listed as noxious by the New Zealand Ministry of Agriculture and Forestry and cannot be propagated, released, displayed or sold under the Biosecurity Act Sections 52 and 53 (Rahman and Popay, 2009).
In the USA, B. davidii is currently listed as a ‘B’ designated noxious weed by the Oregon Department of Agriculture (Ream, 2006).
B. davidii appears on the “most invasive” species list of the Pacific Northwest Pest Plant Council and the native Plant Societies of Oregon and Washington (Savonen, 2009). The Washington State Noxious Weed Control Board listed B. davidii as a Class B noxious weed (WSNWCB, 2009). The California Invasive Pest Plant Council (Cal-IPC) has evaluated B. davidii, but it has yet to be listed (Calflora, 2009). B. davidii is a category 3 watch species in the New York metropolitan region (Brooklyn Botanic Garden, 2009).
In Europe, B. davidii can also be found on several plant watching lists (Tutin et al., 1972; SKEW, 2009).
Control
Cultural control and sanitary measures
Ream (2006) documented the management of B. davidii in different production and retail nurseries in Oregon and discovered that retail nurseries are not the source of B. davidii escapes. Plants are either discarded or severely pruned and stored in enclosed houses for winter protection. Most production nurseries prune plants before seed mature, eliminating the seed source. Where this was not the case, seedlings were found around the nursery (Ream, 2006). Moreover, some of the nurseries prevent the spread of B. davidii and other plants by regular herbicide applications (Ream, 2006).
Physical/mechanical control
Physical removal on a small spatial scale may help in the early stages of invasion. Young shrubs can be dug up, but this method is not recommended for mature plants (Tallent-Halsell and Watt, 2009). Remaining stumps should be treated with glyphosate herbicide (Kaufman and Kaufman, 2007).
In New Zealand, B. davidii is controlled by aerial sowing of cover grasses such as Holcus lanatus (Yorkshire fog)in the autumn, prior to planting, which has been found to effectively suppress the growth of young B. davidii seedlings (Tallent-Halsell and Watt, 2009).
Movement control
Dead-heading is the recommended method to reduce the spread of B. davidii by seeds (Turnbull, 2004; Ream, 2006; Savonen, 2009), but this practice has been linked to reducing the quality of the shrub in subsequent years and increasing the plant’s susceptibility to disease (Warr et al., 2002; Tallent-Halsell and Watt, 2009).
Biological control
In 2006, Cleopus japonicus was introduced and released as a potential biocontrol agent for B. davidii in New Zealand (Zhang et al., 1993; Kritcos, 2006; Watson, 2007). Further releases were made in 2007 and 2008 following careful monitoring of weevil behaviour and establishment (Watson 2008). As of 2009, it was still considered too early to judge the field effectiveness of C. japonicus. A second species under consideration for biological control of B. davidii in New Zealand is the stem weevil, Mecysolobuserro. The adults feed on the tender terminal shoots causing tips to wither and die. Host-range testing of this species is still underway (Kay, 2002).
Chemical control
Glyphosate herbicide without surfactans has been reported to be effective against small shrubs (Ream, 2006), whereas large shrubs with heavy pubescent leaves were less vulnerable to foliar application. Direct and precise application, such as painting cut stumps is more effective than spraying (Ream, 2006; Zazirska and Altland, 2006). Treatment with triclopyr or imazapyr has not been effective (Ream, 2006). In New Zealand, B. davidii is typically controlled in recent clearcut stands using herbicides that are usually aerially applied immediately before (i.e. glyphosate and metsulfuron) and then again after (i.e. terbuthylazine and hexazinone) planting of plantation conifers (Tallent-Halsell and Watt, 2009).
Because stem and root fragments readily regenerate, debris piles should be burned, composted or otherwise treated in such a way to kill all seeds, stems and root fragments (Tallent-Halsell and Watt, 2009).
Although B. davidii colonizes disturbed sites, the impact of long-term establishment is yet undetermined. In New Zealand, it has been observed that in the absence of disturbance, native and non-native trees can overtop B. davidii stands (Bellingham et al., 2005; Tallent-Halsell, 2008). Further research is needed to determine the long-term effect that B. davidii might have on the successional trajectories of native taxa.
Further research is needed to determine how B. davidii might spread under changes in temperature and precipitation (i.e. climate change).
Grierson AJC; Long DG, 2001. Flora of Bhutan including a record of plants from Sikkim and Darjeeling. Volume 2 Part 3. Edinburgh, UK: Royal Botanic Garden, Edinburgh and Royal Government of Bhutan.