Forest Ecology and Management 260 (2010) 2079–2087 Contents lists available at ScienceDirect Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco The potential for gene flow from exotic eucalypt plantations into Australia’s rare native eucalypts Robert C. Barbour a,b , Sascha L. Wise a,b , Gay E. McKinnon a , René E. Vaillancourt a,b , Grant J. Williamson a , Brad M. Potts a,b,∗ a b School of Plant Science, University of Tasmania, Private Bag 55, Hobart, Tasmania 7001, Australia Cooperative Research Centre for Forestry, University of Tasmania, Private Bag 55, Hobart, Tasmania 7001, Australia a r t i c l e i n f o Article history: Received 29 July 2010 Received in revised form 26 August 2010 Accepted 30 August 2010 Keywords: Gene flow Biological invasion Genetic pollution Genetic contamination Pollen dispersal Introgression Off-site impacts Forest management Plantation forestry a b s t r a c t Hybridisation through pollen dispersal from exotic plants is increasingly recognised as a threat to the genetic integrity of native plant populations. Its genetic impact can be greater in rare taxa, due to their vulnerability to pollen swamping by more abundant congeners. We assessed the likelihood of pollen dispersal from exotic eucalypt plantations into all of Australia’s rare native eucalypts, and conducted a case study of Eucalyptus perriniana, which is rare in Tasmania. The Australia-wide study involved spatial analyses of the locations for each rare species superimposed on distributions of eucalypt plantations, which were combined with known taxonomically based reproductive barriers. Of the 74 nationally listed rare eucalypt taxa, 22 had locations within 10 km of plantations of the same genus, and eight were within 1 km. These eight proximal taxa are considered priorities for monitoring. In the most extreme case, 30% of point locations originating from herbarium records and field surveys for Eucalyptus conglomerata were within 1 km of exotic plantations. In the case study, E. perriniana revealed considerable reproductive compatibility with adjacent recently established Eucalyptus nitens plantations. However, F1 hybridisation between these species was limited, with 0.2% of the 18,625 seedlings grown from 100 single-tree open-pollinated seedlots being hybrids. For now, the probability of exotic gene flow into E. perriniana appears to be low, however this probability is likely to increase as more E. nitens flowers in the surrounding landscape. These studies suggest that understanding the breeding system and biology of these populations may reveal surprising resistance to such exotic hybridisation as well as identifying high risk situations to focus conservation management. © 2010 Elsevier B.V. All rights reserved. 1. Introduction Hybridisation between species or between naturally allopatric populations is now recognised as being an important contributor to the evolution (Abbott et al., 2003; Ellstrand and Schierenbeck, 2000) and invasion (Ellstrand et al., 2010; Gaskin and Schaal, 2002; Vila et al., 2000) of many exotic plant species. Movement of pollen and its potential genetic and ecological consequences became a public concern following the deployment of genetically modified (GM) crops, as escapes of transgenes into non-transgenic populations were reported (Stewart et al., 2003). These studies highlighted the importance of considering pollen movement and its impacts more generally (Ellstrand, 1992; Ellstrand and Elam, 1993). Gene flow from non-GM crops or exotic species into native populations ∗ Corresponding author at: School of Plant Science, University of Tasmania, Private Bag 55, Hobart, Tasmania 7001, Australia. Tel.: +61 3 6226 2641; fax: +61 3 6226 2698. E-mail address: B.M.Potts@utas.edu.au (B.M. Potts). 0378-1127/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2010.08.049 through hybridisation and introgression has been recognised as posing a risk to the genetic integrity and survival of “pure” species populations (Ellstrand, 1992; Rhymer and Simberloff, 1996; Wolf et al., 2001). These risks have the potential to be higher in the case of rare species and small populations, due to their vulnerability to reproductive swamping (i.e. source/sink effects, Ellstrand and Elam, 1993; Field et al., 2009; Levin et al., 1996; Wolf et al., 2001) and sensitivity to compounding adverse environmental conditions (Rhymer, 2008). Eucalypts represent a central component of the Australian biota, being the dominant tree in the majority of non-arid woodland and forest communities (Williams and Woinarski, 1997). There are nearly 900 eucalypt taxa (CPBR, 2006), which are endemic to Australia and its surrounding islands. Within these, 74 taxa are currently listed as endangered or vulnerable at the national level in Australia and require special attention. A notable characteristic of eucalypts is their propensity for interspecific hybridisation (Potts and Wiltshire, 1997). Under natural circumstances eucalypts are often observed to hybridise (Griffin et al., 1988; Potts and Wiltshire, 1997), however, the degree to which this occurs is limited by pre- 2080 R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 mating barriers such as spatial isolation and flowering asynchrony (Keatley et al., 2004; Potts and Wiltshire, 1997), and post-mating crossing-incompatibilities. The latter prevents successful hybridisation between the three genera of eucalypts (Angophora, Corymbia, and Eucalyptus) and also between the major subgenera within Eucalyptus (Griffin et al., 1988; Potts et al., 2003). At least eight eucalypt species and a few artificial hybrids are being deployed in industrial hardwood plantations in Australia (National Forest Inventory, 2006). This plantation estate has expanded dramatically over the last two decades, and now covers nearly 1 million ha in Australia (950,000 ha, Gavran and Parsons, 2009). Its expansion is likely to continue due to increasing public pressure to reduce native forest harvesting, reduce the Australian trade deficit in forest products and to develop carbon off-set industries (National Forest Inventory, 2006). Due to the often weak barriers to hybridisation within eucalypts, there is concern over the potential for pollen-mediated gene flow from plantations into native eucalypt populations (Kanowski et al., 2005; Potts et al., 2003; Salt et al., 2005; Strauss, 2001; Wardell-Johnson et al., 1997). This concern is based on the fact that genetic material used for stocking plantations is typically established well out of its native range, as locally exotic species, hybrids, provenances or genotypes (Barbour et al., 2008a; Potts et al., 2003). Considerable work has been conducted to assess the potential for exotic gene flow from the major temperate plantation eucalypts in Australia, Eucalyptus globulus and Eucalyptus nitens. This work has demonstrated that F1 hybridisation involving plantations and their adjacent native species can occur (Barbour et al., 2008b, 2003, 2005a,b, 2006b, 2002) and that F1 hybrid seedlings are establishing in the wild at some locations (Barbour et al., 2008b, 2003). In addition, work in the subtropical and tropical regions of Australia has identified a potential for such gene flow (Barbour et al., 2008a; Kanowski et al., 2005). An important aspect of assessing the off-site risks from planting Eucalyptus that has not been addressed is the likelihood of exotic gene flow to the rare eucalypt species of Australia. The aim of this study was to assess the likelihood of pollenmediated gene flow from eucalypt plantations into all nationally listed endangered or vulnerable (herein referred to as rare) eucalypt species of Australia to better focus resources to the protection of species most at risk. This assessment was conducted through the integration of spatial analyses of the proximity of plantations to each native species, and the knowledge of the reproductive barriers that exist amongst them. Following this, a detailed case study was conducted into the likelihood of exotic gene flow into a rare Tasmanian species, Eucalyptus perriniana, from surrounding E. nitens plantations, by assessing levels of F1 hybridisation between the two species and their reproductive biology. 2. Materials and methods 2.1. Spatial assessment All 74 eucalypt species or subspecies classified as endangered or vulnerable in 2009 by the Australian Government (Department of the Environment, Water, Heritage and the Arts, www.deh.gov.au/cgi-bin/sprat/public/sprat.pl) were assessed for their spatial distribution relative to eucalypt plantations. Three species on the list, Eucalyptus aquatica, Eucalyptus olivacea and “Eucalyptus sp. Howes Swamp Creek” were not included in the analyses as they were no longer regarded as legitimate taxa and therefore could not be cross-referenced with other resources (e.g. CPBR, 2006). Point location data for each species were provided by the relevant state agencies (see “Acknowledgements”). These data point locations originated from herbarium records and field surveys, and their number per species ranged from 1 to 357. Some populations were represented by more than one point location, but on average we assume that the number of point locations is consistent with the size of a population and the number of populations for a species. Outliers that fell well outside the distributional range of each species (as described in CPBR, 2006) were removed from the analyses. The original data was recorded in a number of different projections and datums, and was projected to GDA94. The precision of the original collection location also varied, but the majority of records had a locational accuracy of better than 100 m. The distribution of eucalypt plantations throughout Australia current to 2008, was provided as GIS shape files by the National Forest Inventory. These files delineated the taxa used in plantations at the genus level only. According to the National Forest Inventory (2006), eucalypt plantations in Australia are typically established using E. globulus (subgenus Symphyomyrtus, 64% of the total area), E. nitens (Symphyomyrtus, 20%), Eucalyptus pilularis (subgenus Eucalyptus) and Eucalyptus grandis (Symphyomyrtus) (combined at 4%), Eucalyptus dunnii (Symphyomyrtus, 3%), Corymbia (species and hybrids from section Politaria, 2%) and other minor taxa such as Eucalyptus cladocalyx (Symphyomyrtus), Eucalyptus occidentalis (Symphyomyrtus) and Eucalyptus cloëziana (Idiogenes) (6%; National Forest Inventory, 2006; see Fig. 1). The subgeneric classification within genus Eucalyptus follows Brooker (2000). Spatial analysis was conducted using ESRI ArcGIS 9.2 software. Eucalypt plantations were extracted from the National Forest Inventory raster, and converted to polygons for further analysis. The shortest distance of each point location to a plantation of the same genus or subgenus was calculated, as was the area of that plantation. Buffers of 10 km, 1 km and 100 m width were generated around plantation polygons, and the proportion of point locations of each species that fell within the buffers was calculated. 2.2. Field and glasshouse-based work assessing E. perriniana The case study focusing on E. perriniana was conducted using the Strickland population in southern Tasmania (42◦ 21′ S, 146◦ 39′ E, see Rathbone et al., 2007). E. perriniana (subgenus Symphyomyrtus) is not listed as a threatened or vulnerable species at the national level, but is listed as rare in Tasmania under the Threatened Species Protection Act (1995). In Tasmania, this species consists of three small populations with an estimated total number of individuals of approximately 1000, the largest of these populations being at Strickland (Rathbone et al., 2007; Wiltshire and Reid, 1987). The conservation value of these populations has increased following recent genetic analyses showing the Tasmanian populations of E. perriniana to be distinct from the mainland populations (Rathbone et al., 2007). While populations of some rare eucalypt species are genetically depauperate, this does not appear to be the case for E. perriniana as nuclear microsatellite diversity within its populations is comparable to that observed in many widespread eucalypt species (Rathbone et al., 2007). Recently, a two-stage planting program using E. nitens has taken place in the landscape surrounding the Strickland E. perriniana population (Fig. 2). Eucalyptus nitens is exotic to the island of Tasmania, as it naturally exists as disjunct populations in Victoria and New South Wales (Pederick, 1979). The first planting of E. nitens at Strickland took place in 1998, with its closest point to the E. perriniana population being 600 m. Following this, further plantings took place in the surrounding landscape (Fig. 2). At the time of the study, the first planting was eight and half years old and had flower buds present and seed capsules from flowering in previous years, while the second stage of plantings was still reproductively immature. As of early 2008, approximately 562 ha of plantations existed within 10 km of the E. perriniana Strickland population but no plantation was within 500 m. R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 2081 Fig. 1. Spatial distribution of Australia’s rare eucalypt (Corymbia and Eucalyptus) species relative to industrial eucalypt plantations. Species that have point locations that fall within 10 km (see Table 1) are shown in bold. Areas under plantations (dark shading) have been enhanced for the purposes of the figure. Superscripts indicate the conservation status of each species (E , endangered; V , vulnerable). In February 2007, open-pollinated seed capsules were collected from 100 E. perriniana trees situated across the entire Strickland population to quantify the levels of F1 hybridisation with the E. nitens plantation. In addition, five trees of each other Symphyomyrtus species at the site (Eucalyptus dalrympleana, Eucalyptus rodwayi and Eucalyptus rubida), and ten trees of E. nitens, were sampled to assist in the verification of hybrid progenies identified within the E. perriniana seedlots. Capsules were separated into age-groups corresponding to the previous two flowering seasons (i.e. two years) of seed production. This was possible because eucalypts often retain capsules from many flowering seasons, and these can be differentiated by their position on a branch (Barbour et al., 2002). All trees sampled had one-year-old capsules and 48% of trees also had two-year-old capsules. The open-pollinated seed of each tree and capsule age (seedlot) was spread evenly over the surface of potting mix within 17 cm × 49 cm × 34 cm boxes, and germinated and grown under glasshouse conditions. A replicate of each seedlot was sown in two glasshouses, with the position of seedlot within replicate being random. Once seedlings reached ten nodes of development (following Barbour et al., 2003; Potts and Reid, 1988) they were morphologically screened to identify putative hybrids. The morphological characteristics of the E. perriniana × nitens F1 hybrids had been previously characterised using artificial hybrids (Barbour et al., 2005a), allowing for their rapid and confident visual identification. Nevertheless, verification of the identified E. nitens hybrids was conducted (see below). Any putative hybrids between native species at the site were also identified by their intermediate mor- 2082 R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 recorded was from the same cohort of seed (eucalypt seed displays limited post-dispersal dormancy/survival). Since, after disturbance, eucalypts can regenerate from seed or vegetatively from below ground lignotubers, the frequency of each process was assessed. The existence of seedlings in the unburnt northern area of the population was also assessed through field surveys at this time. A portion of the northern section of the population was subsequently burnt in 2008 and surveyed in 2010 for the presence of E. perriniana and hybrid seedlings. 2.3. Verification of hybrids Fig. 2. Aerial photograph of the landscape surrounding the population of Eucalyptus perriniana at Strickland, Tasmania. The area established with E. nitens plantations has been indicated, as well as the reproductive status of these plantations at the time of the study. The native eucalypt species observed to hybridise with E. perriniana are components of the eucalypt forest surrounding the E. perriniana population and co-occur with E. perriniana at the margins of the population. phology between parental species. The screening process involved the recording and removal from the boxes of most pure species seedlings to prevent overcrowding. All putative hybrids and diverse samples of seedlings of pure E. perriniana (n = 48), E. nitens (29) and putative parents of the native hybrids (12 E. rubida, 18 E. dalrympleana, 20 E. rodwayi) were retained and labelled for later verification. The frequency of each hybrid type was calculated for each species, tree and capsule age. Flowering time assessments of E. perriniana and E. nitens were conducted over the 2007–2008 (7 assessments), 2008–2009 (8) and 2009–2010 (5) summer periods. Twenty trees of both species were assessed. The E. nitens trees were selected widely through the mature plantation, based on having greater than 100 flower umbels, while the E. perriniana were selected from the 100 trees sampled for their seed, based on having the highest number of flower buds. The percentage of the total flower bud crop that was in flower was documented every 10–15 days over the flowering period (Barbour et al., 2006b). Field assessments of the regeneration strategy of E. perriniana were conducted two years after a fire that took place in the southern part of the population. This small fire occurred six months after the 2007 seed collection, ensuring that any seedling recruitment Hybrid verification involved two approaches: seedling morphometric and molecular genetic analyses. The morphometric analysis involved the assessment of each putative natural or exotic hybrid seedling, two pure species seedlings from each E. perriniana tree found to produce E. nitens hybrids, four pure species seedlings from each tree of the other native species, and three pure species seedlings from the E. nitens trees sampled. Twenty-seven leaf, stem, branching and apical bud characteristics were assessed (see Barbour et al., 2008b for a description of each trait). These characteristics were analysed using a canonical discriminant analysis (PROC DISCRIM of SAS, version 9.1, Cary, NC, USA), using the maternal species as the grouping factor. Molecular genetic verification involved the use of five microsatellites used by Rathbone et al. (2007) in a study of E. perriniana. DNA was extracted following Rathbone et al. (2007). To serve as a reference, 127 E. nitens trees (one tree per family, from four different provenances (Pederick, 1979): Macalister = 22 families; Rubicon = 20; Toorongo = 81, Thompson Valley = 4), representative of the gene pool used in Tasmanian plantations, were fingerprinted. DNA was also extracted from 25 putative E. perriniana × nitens F1 hybrids that arose from thirteen E. perriniana trees. The E. perriniana reference DNA samples were those from Rathbone et al. (2007) collected from all three Tasmanian populations (Espies Craig = 10 trees, Strickland = 11, Hungry Flats = 36). PCR of four loci (EMCRC2, EMCRC7, EMCRC11 and EMBRA10) was conducted simultaneously using a Qiagen Multiplex PCR kit (Qiagen Pty Ltd.). A fifth locus, EMCRC8, was amplified separately following Rathbone et al. (2007). The amplified PCR products were sized on a Beckman Coulter, CEQTM 8000 Genetic Analysis System (Fullerton, California, USA). Binning of alleles was done according to Rathbone et al. (2007). Data analysis was undertaken using the Bayesian approach implemented in STRUCTURE 2.2 (Pritchard et al., 2000) to assign the probabilities (membership fraction) of the hybrids belonging to the a priori defined pure parental populations. 3. Results 3.1. Spatial assessment of plantations across Australia Of the 74 nationally listed rare eucalypt taxa of Australia, 22 were found to have point locations (representing an individual or population) within 10 km of a eucalypt plantation of the same genus (Table 1). Looking specifically at each genus, no point locations for the four rare Corymbia species were found within this distance of Corymbia plantations. No known Angophora species are established as plantations, so neither of the two rare Angophora were found to be at risk. All 22 rare eucalypt taxa found within 10 km of the same-genus eucalypt plantations, therefore, belonged to genus Eucalyptus; four from subgenus Eucalyptus, one from subgenus Minutifructus, and 17 from subgenus Symphyomyrtus. Four taxa (Eucalyptus merrickiae, Eucalyptus nicholii, Eucalyptus raveritiana and E. rubida subsp. barbigerorum) were expected to have a low level of exotic gene flow. These were taxa with less Table 1 Spatial assessment of rare native Eucalyptus species relative to eucalypt plantations across Australia. In total 74 taxa were assessed, and those presented were found to have point locations within 10 km of plantations of the same genus. Conservation status n point locations Within 10 km Within 1 km Within 100 m E. alligatrix ssp. limaensisS,37 E. argutifoliaS,62 E. cadensS,38 E. conglomerataE,10 E. crenulataS,36 E. crispataS,67 E. glaucinaS,21 Vulnerable Vulnerable Vulnerable Endangered Endangered Vulnerable Vulnerable 11 20 32 34 30 8 301 4,026 10,338 9,284 370,425 76,542 26,350 11,125 360 3,825 4,459 727 5,820 6,927 96 100.0 55.0 46.9 76.5 73.3 12.5 55.8 9.1 0.0 0.0 29.4 0.0 0.0 5.3 E. gunnii ssp. divaricataS,33 E. johnsonianaE,70 E. kabianaS,11 E. lateriticaE,71 E. leprophloiaS,68 E. merrickiaeS,45 E. morrisbyiS,34 E. nicholiiS,20 Endangered Vulnerable Vulnerable Vulnerable Endangered Vulnerable Endangered Vulnerable 71 45 7 14 7 34 30 153 18,630 10,121 1,672 13,101 24,419 32,689 7,469 68,878 5,335 922 727 2,478 2,433 2,877 909 7,992 12.7 35.6 100.0 28.6 14.3 2.9 73.3 2.0 E. paludicolaS,35 E. parvulaS,30 E. raveretianaM,5 E. rubida ssp. barbigerorumS,17 E. strzeleckiiS,32 E. subereaE,72 E. tetrapleuraS,18 Endangered Vulnerable Vulnerable Vulnerable Vulnerable Vulnerable Vulnerable 28 92 57 15 45 20 308 5,907 31,828 93,422 99,811 12,294 13,645 8,790 0 4,265 9,972 9,242 120 2,478 1,584 78.6 16.3 1.8 6.7 46.7 25.0 80.2 Distance to plantations (m) Mean % of point locations in proximity to plantations Min. Mean area (ha) of closest plantations Major plantation species planted in regiona 0.0 0.0 0.0 0.0 0.0 0.0 0.3 10 21 11 8 17 321 248 0.0 2.2 14.3 0.0 0.0 0.0 6.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 17 91 4 29 12 4 8 16 3.6 0.0 0.0 0.0 11.1 0.0 0.0 3.6 0.0 0.0 0.0 0.0 0.0 0.0 158 15 51 39 309 30 13 glo, nit cla glo, nit dun, gra, cor, pil glo, nit cla clo, cor, dun, gra, pil nit cla dun, gra, cor, pil cla cla cla, glo, sal glo, nit clo, cor, dun, gra, pil glo nit cor, camXgra nit glo, nit Cla clo, cor, dun, gra, pil R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 SpeciesSubgenus,Fig.1 map ref. Eucalypt name codes: cam, Eucalyptus camaldulensis; cla, E. cladocalyx; clo, Eucalyptus cloeziana; cor, taxa from section Politaria of genus Corymbia; dun, E. dunnii; glo, E. globulus; gra, E. grandis; nit, E. nitens; pit, E. pilularis; sil, E. saligna. Subgenenera are abbreviated as superscript S, Symphyomyrtus, E, Eucalyptus and M, Minutifructus. a National Forest Inventory (2006) and plantation industry representatives pers. com. 2083 2084 R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 than 10% of their point locations within 10 km and no point location within 1 km of plantations. Eight taxa were identified as the most likely to be subject to exotic gene flow, as they had point locations that were within 1 km of plantations, i.e. Eucalyptus alligatrix subsp. limaensis (360 m), Eucalyptus conglomerata (727 m), Eucalyptus glaucina (96 m), Eucalyptus johnsoniana (922 m), Eucalyptus kabiana (727 m), Eucalyptus morrisbyi (909 m), Eucalyptus paludicola (0 m) and Eucalyptus strzeleckii (120 m). Assuming that the number of point locations was correlated with the number of populations, the taxa most at risk would be those with the larger fraction of their point locations within proximity of plantations. Two taxa had 100% of their distribution within 10 km of plantations, E. alligatrix subsp. limaensis and E. kabiana, and another two taxa had more than 10% of their distribution within 1 km of plantations, E. conglomerata and E. strzeleckii. Two regions of Australia displayed the highest numbers of rare eucalypt taxa within 10 km of plantations. These were the planting zone north of Perth in Western Australia (six taxa), and the northern New South Wales/Southern Queensland region (four taxa). The rare E. alligatrix ssp. limaensis, E. cadens, E. crenulata, E. gunnii ssp. divaricata, E. merrickiae, E. morrisbyi, E. paludicola, E. parvula, E. rubida ssp. barbigerorum and E. strzeleckii were found within 10 km of the major temperate Australian plantation species (E. globulus and E. nitens). DNA analysis of the morphologically identified E. perriniana × nitens F1 hybrids provided further evidence of their hybridity. All hybrid seedlings displayed a close to 50% membership fraction to both E. perriniana and E. nitens, again verifying the paternal contribution of E. nitens. Over the three seasons assessed, the E. perriniana population was observed to flower from the beginning of January to mid February, peaking in mid-January to early February depending upon the year. The flowering time of E. perriniana virtually completely overlapped that of the E. nitens plantation (overlap in 2007–2008, 5 out of 5 times E. perriniana was observed to flower; 2008–2009, 6 out of 7; 2010, 4 out 4). An assessment undertaken in the second year of flowering observations also found that 20 out of 200 surveyed E. nitens trees were reproductive, and of these 20, the estimated number of flower umbels per tree was 115 ± 15. Assessments of the patterns of regeneration in the E. perriniana population following the fire found 310 mature trees that had been burnt, and all of these were resprouting from lignotubers two years post-fire. No regeneration through seedling establishment was recorded in the first burnt area, and only one seedling (identified by its small stature and apparent lack of a developed underground lignotuber) was recorded in the unburnt area despite extensive searches. While seedling establishment was rare, 26 seedlings were observed which had established following the second fire, and all were pure E. perriniana (Matthew Larcombe, pers. com.). 3.2. E. perriniana case study In total, 18,625 E. perriniana seedlings were assessed in order to estimate the level of hybridisation with the adjacent E. nitens plantation. The average level of hybridisation with E. nitens in openpollinated progenies of the 100 E. perriniana trees was 0.28%. All hybrids with E. nitens were produced by 20 of these trees. Summed across both years of capsule collection, nine trees produced levels higher than 1%, and one reached 4.3%. While the tree with the highest levels of hybridisation was located in the middle of the northern section of the population, no spatial structure was evident in the distribution of the levels of hybridisation across the population. Assessment of the variation in hybridisation between the capsule ages found only one out of the 48 trees that had twoyear-old capsules to be hybridising with E. nitens (0.02% of the 4315 seedlings assessed), while 19 of the 100 trees bearing one-year-old capsules (0.36% of the 14,310 seedlings assessed) produced hybrids. The two-year-old capsules were pollinated when the adjacent plantation was seven years old, while the one-year-old capsules were pollinated one year later. Levels of hybridisation involving E. nitens were found to be lower than those involving the native species growing immediately adjacent to the E. perriniana. This natural hybridisation averaged 1.69% (two-year-old capsules, 2.32%; oneyear-old capsules 1.50%; 71 E. perriniana trees hybridising) and reached 14.6% for one tree summed across both years. All three sympatric native eucalypts at the site (E. dalrympleana, E. rodwayi and E. rubida) were involved in producing natural hybrids with E. perriniana. Morphometric assessment of the E. perriniana × nitens F1 hybrid seedlings relative to their pure parental species seedlings demonstrated strong discrimination between all three cross-types (Fig. 3a). This discrimination was also seen between the three other native species at the site, E. dalrympleana, E. rodwayi and E. rubida, although E. dalrympleana and E. rubida abutted in multivariate space. The exotic hybrids were found to deviate strongly away from the native species plane of variation in the three dimensional plot, and lie intermediate between the E. perriniana and E. nitens seedlings. In comparison, the natural hybrids that were identified did not deviate in their morphology towards E. nitens, and were intermediate in character with E. dalrympleana, E. rodwayi or E. rubida, indicating the paternal contribution of all three species. The 4. Discussion Despite the broad overlap of the geographic distribution of both Australia’s rare eucalypt taxa, and the current eucalypt plantation estate (Fig. 1), only 22 of the 74 endangered or vulnerable taxa were found within 10 km of a plantation. The four species predicted to have highest likelihood of exotic gene flow are E. alligatrix subsp. limaensis, E. conglomerata, E. kabiana, and E. strzeleckii; either because they had more than 10% of their distribution within 1 km of plantations, or due to having 100% of their distribution within 10 km of plantations. The extent to which pollen from plantations will reach populations of these species and allow reproductive contact therefore needs to be addressed. These taxa should be the initial focus of more detailed risk assessment. Patterns of pollen dispersal within eucalypt genera are poorly understood (Barbour et al., 2008a; Potts et al., 2003), however, the majority of deposition is expected to occur within a few hundred metres of plantation boundaries, although the dispersal kernel is likely fat-tailed (Barbour et al., 2005b; Linacre and Ades, 2004; Sampson and Byrne, 2008; Mimura et al., 2009). Pollen movement from E. nitens plantations has been assessed through the collection of open-pollinated seed from native Eucalyptus ovata trees (Barbour et al., 2005b). In that study, 13% of seedlings raised from seed taken from E. ovata trees on the plantation boundary were hybrids with E. nitens (reaching as high as 56% for one tree), but by 200 m away from the plantations hybridisation level had dropped to 1% and continued at this level to the extent of the studied areas (1600 m). Patterns of pollen movement will, however, be greatly affected by the particular animal vectors involved in dispersal (Barbour et al., 2008a; House, 1997), landscape factors such as the degree of forest fragmentation (Byrne et al., 2008; Mimura et al., 2009; Sampson and Byrne, 2008; Sork and Smouse, 2006), and the mosaic of species that exist within the native forest (Griffin, 1989). Source/sink ratios and the proportion of the rare species distribution within proximity of plantations will also affect the probability and consequences of exotic hybridisation (Potts et al., 2003). Nevertheless, isolation distances of multiple kilometres are expected to be effective at significantly decreasing the probability of hybridisation with plantations (Barbour et al., 2005b; Linacre and Ades, 2004). R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 2085 Fig. 3. Verification of F1 hybrid seedlings found amongst open-pollinated seedlots of the rare Eucalyptus perriniana as a result of pollen movement from Eucalyptus nitens plantations. Verification was conducted through the use of seedling morphometric analyses (E. perriniana n = 46; E. nitens n = 29 exotic hybrids n = 37; native hybrids n = 93) (a) and microsatellite markers (b). The STRUCTURE analysis of the molecular data showed that all putative E. perriniana × nitens hybrid seedlings (n = 25) displayed a close to 50% membership fraction to either of the parents (E. perriniana n = 57; E. nitens n = 127), again verifying the paternal contribution of E. nitens. The reproductive incompatibilities between genera and the major subgenera of Eucalyptus will also play a major role in governing the degree of hybridisation. In the cases of E. conglomerata, E. johnsoniana, E. lateritica and E. suberea, which belong to subgenus Eucalyptus, the risks associated with being in close proximity to plantations can only be realised if the planted species belongs to the same subgenus (Griffin et al., 1988; Potts et al., 2003). Their proximity to E. pilularis will, therefore, be the most notable concern, as this is the major species from subgenus Eucalyptus currently used for plantation establishment (National Forest Inventory, 2006). At present, only one of these rare species (i.e. E. conglomerata) occurs in the same geographic region as E. pilularis plantations. The barriers to inter-subgeneric crossing, however, may be incomplete in the case of the minor subgenera (Stokoe et al., 2001). For example, it may be possible for Eucalyptus raveretiana (subgenus Minutifructus) to hybridise with Symphyomyrtus species (Griffin et al., 1988), due to their close phylogenetic affinities (Whittock et al., 2003) and hybrids between subgenus Eucalyptus and the monotypic subgenus Idiogenes (E. cloëziana—a minor plantation species) have been recorded (Stokoe et al., 2001). Nevertheless, three of these four rare species from the subgenus Eucalyptus would appear to be at no risk as they are likely to be well isolated from any plantations of species from subgenera Eucalyptus or Idiogenes. The vast majority of species used in plantations belong to subgenus Symphyomyrtus, and the majority of rare eucalypts also belong to this subgenus. Despite the interspecific hybridisation recorded within this subgenus (Griffin et al., 1988), barriers often exist that will act to prevent hybridisation between plantation and native species. These barriers include limitations in the amount of pollen released from plantations (Barbour et al., 2008a,b, 2006b; Williams et al., 2006), interspecific flowering asynchrony (Barbour et al., 2006b; CPBR, 2006; Vanden Broeck et al., 2003), crossingincompatibilities (Barbour et al., 2005a; Potts et al., 2003), and limitations in the degree to which exotic hybrids can establish in the maternal habitat (Barbour et al., 2008b, 2003, 2006a). The spatial analysis conducted in the current work, therefore, provides a framework for identifying rare species likely to be most at risk, which should be prioritised for closer on-site assessment of these additional barriers. We have conducted one such assessment for E. perriniana in Tasmania, following initial inspection that identified this species was at potential risk due to the close proximity of exotic plantations. If on-site assessments confirm a species to be at risk, a number of options are available for management (Barbour et al., 2008b). These include increasing spatial isolation following the harvesting of a plantation, coupled with surveys and weeding of any established hybrids. 2086 R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 In the case of the rare E. perriniana, there appear to be few reproductive barriers to hybridisation with the adjacent E. nitens plantations at Strickland. Flowering time assessments found almost complete synchrony between E. perriniana and E. nitens, and previous artificial pollination studies have demonstrated that they are readily cross-compatible (Barbour et al., 2005a, 2006b and unpublished data). At this stage the levels of hybridisation appear to be limited by the small proportion of plantations that have reached reproductive maturity and their spatial isolation from the E. perriniana population. The increase in hybridisation with E. nitens observed over two years is likely to have been caused by the increase in the flowering of the E. nitens plantation over this period. If the flowering in the E. nitens plantation continues to increase as the trees mature this may result in an increase in the level of hybridisation until plantations are harvested. The levels of exotic hybridisation in E. perriniana were notably lower than those involving the native species on the population boundary. Nevertheless, the buffer distance of 500–600 m that is currently maintained is expected to limit the level of hybridisation given current information on pollen dispersal in E. nitens (Barbour et al., 2005b) and other eucalypts (Mimura et al., 2009). The vigorous vegetative regeneration of the mature cohort of E. perriniana following wildfire and low seedling recruitment means that generation replacement is likely to take 100’s of years or more (Tyson et al., 1998), allowing ample opportunity for studying the long-term population dynamics and the adaptive management of the exotic gene flow into this long-lived eucalypt species. While further study is required, the extreme habitat in which E. perriniana grows in Tasmania is also expected to strengthen the current barriers to exotic gene flow. This habitat is subjected to poor drainage through winter and desiccation through summer (Wiltshire and Reid, 1987). This extreme habitat is likely to select against maladapted hybrid genotypes. Indeed, E. perriniana appears to have maintained its genetic integrity through evolutionary time despite ongoing natural interspecific hybridisation and even sporadic establishment of these hybrids at the population margins (Rathbone et al., 2007; Wiltshire and Reid, 1987). Breeding system and genetic considerations in the management of rare species will increasingly need to be addressed given the growing number of anthropogenic incursions they face, e.g. climate change, exotic species, habitat modification (Field et al., 2009; Kramer and Havens, 2009). Many rare plant species have been artificially “hemmed in” by vegetation clearing for rural or residential development. These landscape changes reduce any opportunities they may have had for migration with shifting climatic envelopes, resulting in their evolutionary and tolerance responses being their only remaining strategies for survival (Aitken et al., 2008; Carroll and Fox, 2008; Kramer and Havens, 2009). Pollen-mediated gene flow from exotic species may reduce the reproductive capacity of rare species through the production of inviable zygotes or offspring (Lopez et al., 2000; Potts et al., 2003), exacerbating the deleterious effects of small population size such as inbreeding and poor seed quality and further reducing their fitness and evolutionary potential (Wolf et al., 2001). On the other hand, hybridisation may provide genetic diversity to assist in the evolutionary responses of such populations (Ellstrand and Elam, 1993; Levin et al., 1996; Whelan et al., 2006). In either case, greater understanding of the biology of rare species will be important to assess the potential for exotic gene flow and the risks this poses in order to develop strategies to ensure their survival under changing landscape and climatic conditions. However, with long-lived organisms such as forest trees, a full appreciation of the risks from exotic hybridisation will require long-term monitoring which is needed to inform adaptive management strategies that integrate both biological and societal considerations (Kramer and Havens, 2009; Ellstrand et al., 2010). Acknowledgements The authors wish to thank the CRC for Forestry for financial support, Gunns Ltd. and the Department of Primary Industries and Water (Threatened Species Unit) for permission to conduct seed collections and field observations, and Scott Nichols, Anthony Mann, Jnthony Bloomfield and Matthew Lacombe for their assistance. We also thank the National Plantation Inventory for providing the GIS shape files for the distribution of eucalypt plantations, and the following agencies for the point location data for the rare species; Department of Environment and Climate Change (New South Wales), National Herbarium of Victoria, Brisbane Botanic Gardens, State Herbarium of South Australia, Department of Environment and Conservation (Western Australia), and the Department of Primary Industries, Parks, Water and Environment (Tasmania). We thank Dr. Michael Powell for helpful comments on the manuscript. References Abbott, R.J., James, J.K., Milne, R.I., Gillies, A.C.M., 2003. Plant introductions, hybridization and gene flow. Philos. Trans. R. Soc. Lond., Ser. B: Biol. Sci. 358, 1123–1132. Aitken, S.N., Yeaman, S., Holliday, J.A., Wang, T., Curtis-McLane, S., 2008. Adaptation, migration or extirpation: climate change outcomes for tree populations. Evol. Appl. 1, 95–111. Barbour, R.C., Crawford, A.C., Henson, M., Lee, D.J., Shepherd, M., 2008a. The risk of pollen-mediated gene flow from exotic Corymbia plantations into native Corymbia populations in Australia. Forest Ecol. Manage. 256, 1–19. Barbour, R.C., Otahal, Y., Vaillancourt, R.E., Potts, B.M., 2008b. Assessing the risk of pollen-mediated gene flow from exotic Eucalyptus globulus plantations into native eucalypt populations of Australia. Biol. Conserv. 141, 896–907. Barbour, R.C., Potts, B.M., Vaillancourt, R.E., 2003. Gene flow between introduced and native Eucalyptus species: exotic hybrids are establishing in the wild. Aust. J. Bot. 51, 429–439. Barbour, R.C., Potts, B.M., Vaillancourt, R.E., 2005a. Gene flow between introduced and native Eucalyptus species: crossability of native Tasmanian species with exotic E. nitens. Aust. J. Bot. 53, 465–477. Barbour, R.C., Potts, B.M., Vaillancourt, R.E., 2005b. Pollen dispersal from exotic eucalypt plantations. Conserv. Genet. 6, 253–257. Barbour, R.C., Potts, B.M., Vaillancourt, R.E., 2006a. Gene flow between introduced and native Eucalyptus species: early-age selection limits invasive capacity of exotic E. ovata × nitens F1 hybrids. Forest Ecol. Manage. 228, 206–214. Barbour, R.C., Potts, B.M., Vaillancourt, R.E., Tibbits, W.N., 2006b. Gene flow between introduced and native Eucalyptus species: flowering asynchrony as a barrier to F1 hybridisation between exotic E. nitens and native Tasmanian Symphyomyrtus species. Forest Ecol. Manage. 226, 9–21. Barbour, R.C., Potts, B.M., Vaillancourt, R.E., Tibbits, W.N., Wiltshire, R.J.E., 2002. Gene flow between introduced and native Eucalyptus species. New Forest 23, 177–191. Brooker, M.I.H., 2000. A new classification of the genus Eucalyptus L’Her. (Myrtaceae). Aust. Syst. Bot. 13, 79–148. Byrne, M., Elliott, C.P., Yates, C.J., Coates, D.J., 2008. Maintenance of high pollen dispersal in Eucalyptus wandoo, a dominant tree of the fragmented agricultural region in Western Australia. Conserv. Genet. 9, 97–105. Carroll, S.P., Fox, C.W., 2008. Conservation Biology: Evolution in Action. Oxford University Press, New York. CPBR, 2006. EUCLID: Eucalypts of Australia. CSIRO Publishing, Collingwood, Victoria. Ellstrand, N.C., 1992. Gene flow by pollen: implications for plant conservation genetics. Oikos 63, 77–86. Ellstrand, N.C., Biggs, D., Kaus, A., Lubinsky, P., McDade, L.A., Preston, K., Prince, L.M., Regan, H.M., Rorive, V., Ryder, O.A., Schierenbeck, K.A., 2010. Got hybridization? A multidisciplinary approach for informing science policy. BioScience 60, 384–388. Ellstrand, N.C., Elam, D.R., 1993. Population genetic consequences of small population size: implications for plant conservation. Annu. Rev. Ecol. Evol. Syst. 24, 217–242. Ellstrand, N.C., Schierenbeck, K.A., 2000. Hybridization as a stimulus for the evolution of invasiveness in plants? Proc. Natl. Acad. Sci. U.S.A. 97, 7043–7050. Field, D.L., Ayre, D.J., Whelan, R.J., Young, A.G., 2009. Molecular and morphological evidence of natural interspecific hybridization between the uncommon Eucalyptus aggregata and the widespread E. rubida and E. viminalis. Conserv. Genet. 10, 881–896. Gaskin, J.F., Schaal, B.A., 2002. Hybrid Tamarix widespread in US invasion and undetected in native Asian range. Proc. Natl. Acad. Sci. U.S.A. 99, 11256–11259. Gavran, M., Parsons, M., 2009. National Plantation Inventory 2009 Update. National Forest Inventory, Bureau of Rural Sciences, Canberra. Griffin, A.R., 1989. Sexual reproduction and tree improvement strategy—with particular reference to Eucalyptus. In: Gibson, G.I., Griffin, A.R., Matheson, A.C. (Eds.), Proceedings of the IUFRO Conference. Breeding Tropical Trees: Population Structure and Genetic Improvement Strategies in Clonal and Seedling Forestry. Oxford Forestry Institute, Oxford, pp. 52–67. R.C. Barbour et al. / Forest Ecology and Management 260 (2010) 2079–2087 Griffin, A.R., Burgess, I.P., Wolf, L., 1988. Patterns of natural and manipulated hybridisation in the genus Eucalyptus L’Herit—a review. Aust. J. Bot. 36, 41–66. House, S.M., 1997. Reproductive biology of eucalypts. In: Williams, J., Woinarski, J. (Eds.), Eucalypt Ecology: Individuals to Ecosystems. Cambridge University Press, Cambridge, pp. 30–56. Kanowski, J., Catterall, C.P., Wardell-Johnson, G.W., 2005. Consequences of broadscale timber plantations for biodiversity in cleared rainforest landscapes of tropical and subtropical Australia. Forest Ecol. Manage. 208, 359– 372. Keatley, M.R., Hudson, I.L., Fletcher, T.D., 2004. Long-term flowering synchrony of box-ironbark eucalypts. Aust. J. Bot. 52, 47–54. Kramer, A.T., Havens, K., 2009. Plant conservation genetics in a changing world. Trends Plant Sci. 14, 599–607. Levin, D.A., Francisco-Ortega, J., Jansen, R.K., 1996. Hybridization and the extinction of rare plant species. Conserv. Genet. 10, 10–16. Linacre, N.A., Ades, P.K., 2004. Estimating isolation distances for genetically modified trees in plantation forestry. Ecol. Model. 179, 247–257. Lopez, G.A., Potts, B.M., Tilyard, P.A., 2000. F1 hybrid inviability in Eucalyptus: the case of E. ovata × E. globulus. Heredity 85, 242–250. Mimura, M., Barbour, R.C., Potts, B.M., Vaillancourt, R.E., Watanabe, K.N., 2009. Comparison of contemporary mating patterns in continuous and fragmented Eucalyptus globulus native forests. Mol. Ecol. 18, 4180–4192. National Forest Inventory, 2006. Australia’s Plantations. Bureau of Rural Sciences, Canberra. Pederick, L.A., 1979. Natural variation in Shining Gum (Eucalyptus nitens). Aust. Forest Res. 9, 41–63. Potts, B.M., Barbour, R.C., Hingston, A.B., Vaillancourt, R.E., 2003. Turner Review No. 6—genetic pollution of native eucalypt gene pools—identifying the risks. Aust. J. Bot. 51, 1–25. Potts, B.M., Reid, J.B., 1988. Hybridisation as a dispersal mechanism. Evolution 42, 1245–1255. Potts, B.M., Wiltshire, R.J.E., 1997. Eucalypt genetics and genecology. In: Williams, J., Woinarski, J. (Eds.), Eucalypt Ecology: Individuals to Ecosystems. Cambridge University Press, Cambridge, pp. 56–91. Pritchard, J.K., Stephens, M., Rosenberg, N.A., Donnelly, P., 2000. Association mapping in structured populations. Am. J. Hum. Genet. 67, 170–181. Rathbone, D.A., McKinnon, G.E., Potts, B.M., Steane, D.A., Vaillancourt, R.E., 2007. Microsatellite and cpDNA variation in island and mainland populations of a regionally rare eucalypt, Eucalyptus perriniana (Myrtaceae). Aust. J. Bot. 55, 513–520. Rhymer, J.M., 2008. Hybridization, introgression, and the evolutionary management of threatened species. In: Carroll, S.P., Fox, C.W. (Eds.), Conservation Biology: Evolution in Action. Oxford University Press, New York, pp. 130–140. 2087 Rhymer, J.M., Simberloff, D., 1996. Extinction by hybridisation and introgression. Annu. Rev. Ecol. Syst. 27, 83–109. Salt, D., Lindenmayer, D., Hobbs, R., 2005. Trees and Biodiversity: A Guide for Australian Farm Forestry. Rural Industries and Development Corporation/Land & Water Australia/FWPRDC/MDBC Joint Venture Agroforestry Program, Kingston, ACT, Australia. Sampson, J.F., Byrne, M., 2008. Outcrossing between an agroforestry plantation and remnant native populations of Eucalyptus loxophleba. Mol. Ecol. 17, 2769–2781. Sork, V.L., Smouse, P.E., 2006. Genetic analysis of landscape connectivity in tree populations. Landscape Ecol. 21, 821–836. Stewart, C.N., Halfhill, M.D., Warwick, S.I., 2003. Transgene introgression from genetically modified crops to their wild relatives. Nat. Rev. Genet. 4, 806–817. Stokoe, R.L., Shepherd, M., Lee, D.J., Nikles, D.G., Henry, R.J., 2001. Natural inter-subgeneric hybridization between Eucalyptus acmenoides Schauer and Eucalyptus cloeziana F. Muell (Myrtaceae) in southeast Queensland. Ann. Bot. 88, 563–570. Strauss, S.Y., 2001. Benefits and risks of biotic exchange between Eucalyptus plantations and native Australian forests. Aust. Ecol. 26, 447–457. Tyson, M., Vaillancourt, R.E., Reid, J.B., 1998. Determination of clone size and age in a mallee eucalypt using RAPDs. Aust. J. Bot. 46, 161–172. Vanden Broeck, A., Cox, K., Quataert, P., Van Bockstaele, E., Van Slycken, J., 2003. Flowering phenology of Populus nigra L., P. nigra cv. italica and P. × canadensis Moench. and the potential for natural hybridisation in Belgium. Silvae Genet. 52, 280–283. Vila, M., Weber, E., Antonio, C.M.D., 2000. Conservation implications of invasion by plant hybridization. Biol. Invas. 2, 207–217. Wardell-Johnson, G., Williams, J.E., Hill, K., Cumming, R., 1997. Evolutionary biogeography and contemporary distribution of Eucalyptus. In: Williams, J., Woinarski, J. (Eds.), Eucalypt Ecology: Individuals to Ecosystems. Cambridge University Press, Cambridge, pp. 92–128. Whelan, R.J., Roberts, D.G., England, P.R., Ayre, D.J., 2006. The potential for genetic contamination vs. augmentation by native plants in urban gardens. Biol. Conserv. 128, 493–500. Whittock, S., Steane, D.A., Vaillancourt, R.E., Potts, B.M., 2003. Molecular evidence shows that the tropical boxes (Eucalyptus subgenus Minutifructus) are overranked. Trans. R. Soc. S. Aust. 127, 27–32. Williams, D.R., Potts, B.M., Neilsen, W., Joyce, K., 2006. The effect of tree spacing on the production of flowers in Eucalyptus nitens. Aust. Forest 69, 299–304. Williams, J.E., Woinarski, J.C.Z. (Eds.), 1997. Eucalypt Ecology: Individuals to Ecosystems. Cambridge University Press, Cambridge. Wiltshire, R.J.E., Reid, J.B., 1987. Genetic variation in the Spinning Gum, Eucalyptus perriniana F. Muell. ex Rodway. Aust. J. Bot. 35, 33–47. Wolf, D.E., Takebayashi, N., Rieseberg, L.H., 2001. Predicting the risk of extinction through hybridization. Conserv. Biol. 15, 1039–1053.