Kew Bulletin

, Volume 65, Issue 4, pp 583–594 | Cite as

Aliens or natives: who are the ‘thugs’ in British woods?

  • R. H. Marrs
  • M. G. Le Duc
  • S. M. Smart
  • K. J. Kirby
  • R. G. H. Bunce
  • P. M. Corney


The invasion of native habitats by exotic, or alien, plant species has received considerable attention recently from policy, research, and practical conservation management perspectives. However, a new hypothesis for species dynamics in Britain suggests that a small number of aggressive native plant species (termed ‘thugs’) may have an equal, or greater, impact on native species and habitats than exotic species. Here, we examine this hypothesis using multivariate techniques with field-layer cover data collected during a country-wide survey of British woodlands. Multivariate analysis of these data identified a north-south gradient on the first axis, and that 20 of the 25 National Vegetation Classification woodland types were sampled within the study. The most abundant field-layer species included three of the proposed native ‘thugs’, i.e. Rubus fruticosus, Pteridium aquilinum and Hedera helix in addition to the native woodland indicator species Mercurialis perennis. Variation partitioning was used to compare the relative importance of native field-layer ‘thug’ species with invading alien shrub and tree species relative to other environmental drivers. The variation in the field-layer data-set explained by the three native ‘thug’ species was significant, but they explained a relatively small proportion of the variation relative to other environmental variables (climate, soil, management factors etc.). They did, however, explain almost four times as much variation as the three alien species that were significantly correlated with field-layer species composition (Acer pseudoplatanus, Impatiens glandulifera, Rhododendron ponticum). The results of this analysis suggest that the field-layer of British woodlands is impacted as much by native ‘thug’ species, as it is from ‘aliens’. Concern about the impact of these native ‘thug’ species has been reported previously, but their impact has not previously been compared to the impact of invading aliens. It is hoped that this analysis will do two things, first to act as a sound baseline for assessing any changing balance that should occur in the future, and second, to prompt both ecologists and conservationists to develop woodland management policies based on sound science.

Key Words

Detrended correspondence analysis National Woodland Survey native species resource assessment variation partitioning woodland field-layer species woodland herb 


Invasive species have been recognised as posing a major threat to the conservation of biodiversity worldwide (Millennium Ecosystem Assessment 2005). Most concerns are related to the impact of neophytes, i.e. those species that are not native to the area and have invaded in the recent past (CBD 2002; IUCN 2000; Liu et al.2005; Peterson et al.2003; Pysek et al.2003; Richardson et al.2005, inter alia). However, the conservation of ecosystems may also be significantly affected by native species that develop high biomass and necromass and reduce the abundance of other species, in some places becoming almost mono-cultures (Bobbink & Willems 1987; Pakeman & Marrs 1992). Indeed, Pearman (2004a) has argued for a shift in the way in which we view invading plant species, suggesting that aggressive native plant species may have an equal, or even greater, impact on other native plants and ecosystems than neophytes (Pearman & Lockton 2004). Pearman (2004b) identified four native species which he termed ‘thugs’: Hedera helix L. spp. (here taken to include H. helix subsp. helix and H. helix subsp. hibernica (G. Kirchn.) D. C. McClint. (McAllister & Rutherford 1990), Pteridium aquilinum (L.) Kuhn, Rubus fruticosus L. agg. and Urtica dioica L. All of these species have been identified as affecting the richness or composition of valued communities (Kirby & Woodell 1998; Marrs & Watt 2006; McAllister & Rutherford 1990). We have retained the ‘thugs’ descriptor for this group of plants for this paper; they have been described either as stress-tolerator/competitors (H. helix, R. fruticosus) or competitors (P. aquilinum, U. dioica) by Grime et al. (1988).

A national conservation policy for rare and threatened ecosystems must, therefore, reflect the potential threats from both types of invasive species (neophyte and native) across the country. In order to develop such a policy it is essential to have good information on the status of the ecosystem under threat with respect to all types of invasive species, and this must be derived using rigorous sampling methods. Here, we provide a baseline assessment of the relative potential effects of aggressive species, separated into ‘thugs’ and ‘aliens’, on the field-layer plant species of native woodlands at the Great Britain scale, to provide a national assessment of risk. This was done by analysing data from the National Woodland Survey (NWS, Bunce & Shaw 1973) in 1971, which used statistically rigorous methods to assess the woodland resource. These methods were derived to provide an unbiased sampling of the entire broad-leaved woodland resource of Great Britain in 1971; evidence demonstrating that this survey is a representative sample has already been reported by Kirby et al. (2005). First, we measured the relative abundance of the most abundant field-layer species; this assessed the overall rank-abundance of native and alien ‘thugs’ relative to other woodland species. Second, we used multivariate analysis to identify the major gradients of variation in floristic composition of the field layer within Great Britain and we related this to the National Vegetation Classification (NVC, Rodwell 1991). This was done to provide an overall description of the woodland resource and to ensure it provided a reasonably comprehensive coverage. Lastly, we used Variation Partitioning (VP, Borcard et al.1992) to identify the relative importance of ‘aliens’ and ‘thugs’ compared to other potential environmental drivers (climate, soils, deer grazing and management) that might influence floristic diversity. Lastly, it was hoped that this work would provide a baseline to assess future change. It must, however, be emphasised that this study is a comparative one, assessing potential threats to the conservation of the woodland ground flora at the national scale. We accept that at a site scale, there may be substantive conservation problems associated with both ‘thugs’ and ‘aliens’.


Survey methods

103 sites were selected objectively from an approximate 10% sample of the woodland resource in Great Britain (n = 2,463), which was surveyed for the Nature Conservation Review (Ratcliffe 1977). Within each site 16 plots, each 200 m2, were placed randomly and vegetation assessed between June and October 1971 using a strict protocol (Bunce & Shaw 1973). The cover of field-layer species (used as response data here) was recorded in each plot, along with a number of environmental variables (see Table 1), which were supplemented by many more derived variables (detailed in Corney et al.2004, 2006). Diameter at breast height (dbh) and the number of stems of all woody species (trees, saplings, shrubs) were also measured within each plot. All taxa were identified to species, and where there was some dispute over nomenclature, taxa were amalgamated (e.g., Viola reichenbachiana Jord. ex Boreau, V. riviniana Rchb.) (Corney et al.2004, 2006). Similarly, woody species were allocated to tree and shrub categories based on a pre-determined list. Tree and saplings were differentiated on a dbh basis. The exact protocol is voluminous and is available in Kirby et al. (2005).
Table 1.

Description of environmental variables used in the multivariate analysis, along with data transformations and software packages used. These variables were either collected during NWS 1971 or generated for the present analysis. S = data source.

Variable group


Number of variables



Site boundary type

Site level presence or absence of different boundary types



Site form, shape

Site area (ha) *; Perimeter (m) *; Two shape indices: area index (Aw/Pw); perimeter index (Pw/Pc)




National grid Easting and Northing (km) of plot position *; Average site altitude (m) b, †; Distance (km) from plot to nearest coast


1, 2

Woodland climate

Forest wind climate; Annual accumulated temperature; Moisture deficit



Deer census data

Deer species census data; presence of 6 species in 10 km grid squares



Climatic seasonal 1961 – 1990 LTA dataa

Mean daily max temp (°C); mean daily min temp (°C); days of ground frost; mean cloud cover (Oktas converted to%); total bright sunshine (h); number of days per month having a rainfall ≥ 1 mm (rain days); number of days per month having a rainfall ≥ 10 mm (wet days); total precipitation (mm)



Climatic yearly 1961 – 2000 LTA data

Annual extreme temp range (°C); number of growing degree days; growing season length; maximum number of consecutive dry days in a year; greatest five day precipitation total in a year (mm); mean rainfall amount (mm) on rain days





Soil pH; LOI (%) c; Cumulative depth (cm) to bottom of 5 soil horizons (A00; A0; A1; A2; B)




Plot slope (deg); Southerly aspect d; Westerly aspect e; Plot altitude (m) ; Distance (m) from plot to nearest site boundary *


1, 2

Ground cover

Cover (%) c of 6 ground cover types (including bryophyte cover, litter, dead wood and bare ground)



Plot level descriptors

Including signs of management (7); habitats (46); signs of animal activity (13)



Soil classification

Proportion of one of 26 soil subgroups present in each plot c, f



Woody spp.

Species basal area

Live and dead (l/d) basal area (cm2) of: tree (≥ 5 cm diameter) (75/47), sapling (< 5 cm diameter) (64/42) and shrub (51/29) species present within each plot g




Total plot statistics

Total tree, sapling & shrub basal area (cm2 plot−1) g, live & dead; Density of tree, sapling & shrub stems (plot−1), live & dead





Month during which the site was surveyed. Used as dummy variables.

1 (5)


Data sources


National Woodland Survey 1971 data



Land-Form PANORAMA DTM contour maps; OS DIGIMAP service (EDINA 2002)



ESC decision support system (Ray 2001); 1961 – 1990 LTA data



Biological Records Centre, Centre for Ecology and Hydrology, Monks Wood; 1911 – 1980 data



UK Meteorological Office 5 km × 5 km grid baseline datasets


a Weighted arithmetic average of Meteorological Office monthly data to generate 32 seasonal variables

b Arithmetic average of plot data by site

c Arcsin transformation (Sokal & Rohlf 1995)

d Transformation of aspect (a, degrees), s = sin(a x π/180)/2, giving a southerly aspect

e Transformation of aspect (a, degrees), w = |sin((a x π/180)-π/2)/2)|, giving a westerly aspect

f Where more than one subgroup occurred in each plot, the complex was assumed to comprise equal amounts of each constituent subgroup

g DBH (cm) data converted to basal area (π r2). Square root transformation of basal area data; 0.5 added to all variates (Sokal & Rohlf 1995)

* ERDAS IMAGINE version 8.5 (ERDAS 2001)

MAP MANAGER version 6.2 (ESRI 2001), ARCVIEW GIS version 3.2a (ESRI 2000)

MINDIST2 (Le Duc et al.2000)

Statistical analysis

All multivariate analyses were performed in Vegan (Oksanen 2003) in the R statistical environment (R Development Core Team 2004). Initial analysis of these data using ordination showed extensive distortion in ordinations caused by the inclusion of plots with a non-woodland flora. These outliers were identified using an iterative approach, using a combination of ordination, regression and classification techniques (Corney et al.2006), and once removed the dataset was reduced from 1648 to 1438 plots (87% of those surveyed). The geographical distribution of these plots remained extensive (Fig. 1). The field-layer cover data (ln(y + 1) transformed) were initially analysed using detrended correspondence analysis (DCA) in DECORANA (ter Braak & Šmilauer 2002); species occurring only once in the dataset were removed. The significant environmental variables were correlated with the DCA analysis using function ‘envfit’ in Vegan with 10,000 permutations. For simplicity only the six most significant variables are presented here.
Fig. 1

Distribution of sites surveyed in the 1971 National Woodland Survey of Great Britain. Circle size represents the number of plots from each site included in the present analysis after removing outliers (Corney et al.2006).

The National Vegetation Classification class (NVC, Rodwell 1991) for each plot was determined using TABLEFIT (Hill 1996) and the distribution of each class plotted on the DCA plot as a bivariate standard deviational ellipse (Vegan function ‘ordiellipse’).

Assessment of rank-abundance

In order to assess the relative cover of ‘aliens’ and native ‘thug’ species the overall mean value of each species present in more than 10 quadrats was computed along with the standard error computed using the ‘boot’ function in R with 10,000 permutations (Crawley 2007).

Variation partitioning (VP)

The previously published VP data-structure for these data (Corney et al.2006), was re-analysed. First, the large number of environmental variables were allocated into three primary sets {Site-level variables}, {Plot-level variables} and {Woody species variables}, and within each set the forward selection procedure in Vegan (‘cca’ function and the AIC statistic) was used to select only those variables that were significantly correlated with the species dataset. This procedure starts by selecting the significant variable that reduces the AIC statistic the most, it then adds the next most significant one until the AIC-statistic achieves its lowest value; essentially it selects only those variables within any given set that are significant in explaining the variation in the species dataset. Second, following the recommendations of Legendre & Gallagher (2001) and Peres-Neto et al. (2006) the Vegan function ‘varpart’ was used for the VP with transformed species data (ln x + 1); this procedure quantifies the variation of the species dataset explained by each set and that shared with all the other sets. Thereafter, and where appropriate, the variance was partitioned between the significance sets and tested using Redundancy Analysis (RDA) and a permutation test with 499 permutations. All testable fractions were significant (p < 0.001).

The three primary variable sets were then further subdivided as follows: {Site-level variables} into {Climate}, {Spatial} and {Boundary/Deer} variables, {Plot-level variables} into {Geo-spatial}, {Management} and {Biotic} variables, and {Woody species variables} into {Tree}, {Sapling} and {Shrub} variables (Fig. 2). The VP procedure was then re-run in different combinations to decompose the explained variation between the different sub-sets (Fig. 2). A new ‘thugs’ dataset, based on three native ‘thug’ species (Hedera helix, Pteridium aquilinum, Rubus fruticosus) was removed from the species dataset and added to the {Plot} set of predictor variables. Two further analyses were done using these modified datasets, the first with the ‘thug’ variables removed from the plot-set, and the second with the ‘thugs’ included. Thereafter, to assess the relative importance of ‘alien’ species, a dataset was created including data for all alien species from all strata of the survey (field-layer, shrubs, saplings, trees). The species detected in the survey were: Acer pseudoplatanus L., Impatiens glandulifera Royle, Prunus laurocerasus L., Rhododendron ponticum L., Rhododendron L. spp. and Robinia pseudoacacia L. Forward selection as described above was used to generate a revised group containing three ‘alien’ species that were significantly correlated with the field-layer species composition, these were A. pseudoplatanus, I. glandulifera, and R. ponticum. VP was then used to compare the variation in field-layer species accounted for by the ‘thugs’ and this ‘aliens’ group.
Fig. 2

Venn diagrams illustrating the percentage of total variation explained (both unique and shared components), in the woodland field-layer composition attributable to nested sets of environmental variables through variation partitioning. All testable fractions were significant (p < 0.001).


Assessing the woodland community composition

The DCA analysis produced eigenvalues of 0.528 and 0.325, and gradient lengths of 5.8 and 5.0 for the first two axes respectively. The species plot (Fig. 3A) showed a transition along axis one from those with low scores that tend to be shade-tolerant woodland species, associated with fertile, base-rich soils (e.g., Glechoma hederacea L., Circaea lutetiana L. and woodland specialist species Mercurialis perennis L. and Galium odoratum Scop.); through a region with species typical of moderately open woodland (Veronica chamaedrys L., Viola reichenbachiana, Deschampsia cespitosa (L.) P. Beauv.), suggesting increasing illumination, intermediate fertility and neutral to slightly acid soils, through to species with high scores which tend to be associated with acidic, infertile soils and well-lit habitats such as heathland communities and moorlands (Pteridium aquilinum, Vaccinium myrtillus L.). The gradient on axis two appears to be related to moisture, with species often found on moderately damp soils (e.g., Hedera helix, and the woodland specialists Hyacinthoides non-scripta (L.) Chouard, Lamiastrum galeobdolon (L.) Ehrend. & Polatschek) with low scores and species typical of water-saturated flushes, wet woodlands and poorly-aerated soils (Chrysosplenium oppositifolium L., Filipendula ulmaria (L.) Maxim.) with high scores. The relationship with the six most significant environmental variables shows a clear north-south gradient from right to left along axis one, and the more base-rich soils being found in the upper right hand quadrant (Fig. 3B).
Fig. 3

DCA plot produced from analysis of 1438 plots in the National Woodland Survey of Great Britain in 1971. Plots of A the 62 most abundant species and B the six most significant environmental variables are shown. Species abbreviations follows Hill (1996) and are coded: Woodland specialists = Open image in new window; Geographically-restricted woodland specialists = Open image in new window;Other woodland and non-woodland species = X; a full species list is available Corney et al.2006).

The allocation of the woodland plots to NVC classes showed that the National Woodland Survey (NWS) sampled 20 of the 25 woodland NVC types, although three were detected in only one plot. Of the missing classes, four were Salix-dominated shrub communities (W1, W2, W3, W20), and the fifth was the W18 Pinus sylvestris-Hylocomium splendens community, typical of Scottish native pine woods (Rodwell 1991). The distributions of the 20 classes detected are overlain on the DCA ordination (Fig. 4). W8, the diverse Fraxinus-Mercurialis woodlands and W21, Crataegus-Hedera scrub, both associated with base-rich, fertile soil conditions and low field-layer light, are found at the low end of axis one, whereas W17 (open, Atlantic, bryophyte-rich oak woods) and W16 (oak-birch-heaths of more acid, less fertile sites) appear at the positive end. W14 woods, relatively xeric Fagus L. woodlands, are found at the low end of axis two, whereas W7 and W19, wetter lowland Alnus Mill. and upland Juniperus L. woodlands, are at the positive end. The largest ellipse (W25, Pteridium-Rubus under-scrub) is located in the centre of the biplot, and almost entirely overlaps with W10 (Quercus-Pteridium-Rubus woodland), the lowland analogue of upland W11, which tends to be wetter and frequently has a more open canopy than its lowland counterpart.
Fig. 4

The National Vegetation Classes (NVC) detected in the 1438 plots National Woodland Survey of Great Britain in 1971 and plotted in the same DCA ordination space shown in Fig. 3. The NVC classes (described in Rodwell 1991) are plotted as bivariate standard-deviational ellipses. NVC classes not detected are detailed in the top panel.

Species rank-abundance

The four most abundant species included three of the proposed native ‘thugs’: Rubus fruticosus, Pteridium aqulinum, and Hedera helix in addition to Mercurialis perennis (Fig. 5). Urtica dioica had, at least in 1971, a much lower mean cover and hence is discounted from further discussion as a ‘thug’ here. The two most abundant alien species (Acer pseudoplatanus, Rhododendron ponticum) had a relatively low cover and were ranked 38 and 70 in terms of abundance. The correlations between the abundance of the ‘thug’ species and the species richness in each sample was negative (r = -0.33, df = 1436, P < 0.0001).
Fig. 5

Rank-abundance plot of the most abundant 76 species found in the 1438 plots of the National Woodland Survey of Great Britain in 1971. Standard errors are bootstrap estimates based on 10,000 resamplings.

Variation partitioning

The reworked variance partitioning of Corney et al. (2006) indicated that 41% of the variation in composition of field-layer species could be attributed to the measured sets of variables (e.g., {Site,} {Plot}, {Woody species}, Fig. 2). The subsequent decomposition of this variation (Fig. 2) shows the relative importance of {Climate}, {Spatial}, {Boundary}, {Deer}, {Geo-spatial}, {Biotic} and {Management} sets of variables, as well as the influence of {Woody species} expressed as {Shrubs}, {Saplings} and {Canopy Trees}. Importantly, the amount of variance shared between these sets of variables is also shown at each stage of the decomposition. Thus, whilst {Site}, {Plot} and {Woody species} explain 41% of the variation, individually they accounted for 24.6%, 29.2% and 9.3% respectively. However, when shared explained variation was removed, however, only 9.6%, 12.3% and 1.8% was explained by the three sets respectively. In contrast, the variation accounted for by the native ‘thugs’ and ‘aliens’, albeit significant, was low, and indeed lower than any of the other sub-sets tested. The native ‘thugs’ accounted for 2.1% of the total variation and the ‘aliens’ accounted for 0.7% (Fig. 2); interestingly the effects were almost additive as only 0.1% of the variation was shared between them.


In order to make sound judgements on any conservation policy it is essential to have high-quality information on the communities being investigated. In order to develop national policies it is therefore essential to have an assessment of the state of the “nation’s resource”. Realistically, this can only be achieved by surveying the resource completely (only possible for a limited number of sites) or using a stratified random survey methodology. Here, we used such an approach to survey the semi-natural, native woodland of Great Britain, and were gratified that the survey produced (a) a reasonable geographic coverage of the island, and (b) a good cover of the range of the biotic variation found in British woodlands, i.e. 20 of the 25 woodland types identified in the National Vegetation Classification (Rodwell 1991). Moreover, of the five communities that were not sampled in the survey, four of these were scrub woodland communities, and the fifth (Scottish native pine woods), was excluded because only broadleaved woodlands were sampled. The DCA ordination produced a bio-geographical gradient on the first axis, with the cooler and wetter northern sites at one end and the southern, warmer and drier sites at the other. The second axis appeared to reflect a soils gradient. Thus, the survey did indeed cover a large range of the geographical and biotic variation found in British woods. Most of the woodlands (60%) were classed as ancient woodland, and many are under some form of conservation management agreement.

The most abundant four species found in the field-layer included three of Pearmans’s native ‘thugs’, Rubus fruticosus, Pteridium aquilinum, and Hedera helix plus the native woodland indicator species — Mercurialis perennis. The other proposed ‘thug’ species Urtica dioica had a much lower mean rank abundance (n = 14), and whilst accepting that this species might cause local problems, it was discounted as a major problem at national level in 1971. The inclusion of M. perennis within this group may at first seem surprising. It is often used as an indicator of ancient woodland status in parts of the country, but has also been identified as a species that suppresses other woodland ground flora species (Pigott 1977; Rackham 2006). This species has also been classified as a stress-tolerator/competitor (Grime et al.1988). However unlike the other 'thugs’ it has a very poor mobility (Peterken & Game 1981), which limits its ability to become dominant in recently-created woodland and it often declines when the canopy is opened up, for example under traditional coppice management (Rodwell 1991). For these reasons, we have excluded it from our native ‘thugs’ group here. Reductions in species richness and other changes in community composition associated with mono-specific stands of the three native ‘thug’ species have been reported previously (Kirby & Woodell 1998; Marrs & Watt 2006; McAllister & Rutherford 1990). Corney (2006) has also shown that these four native species have clearly identified realised niches that cover the entire niche range of all other species in British woods.

Comparing thugs and aliens

Alien plant species invasion is perceived to be a major threat and management to control such species can take up a considerable amount of resources (Millennium Ecosystem Assessment 2005; DEFRA 2003). However, in order to assess relative risk, it is essential that the effects of both alien and native species are evaluated together, and indeed compared to the potential impacts of other factors such as climate change. Here we have demonstrated from a baseline survey of British woodlands in 1971 at the national scale, that there is the potential for a small number of native species to have a much larger impact on woodland field-layer species than ‘aliens’, as the native ‘thugs’ accounted for almost four times the variation of the ‘aliens’. As a result, we suggest that the status of all these species be monitored and that their impact considered carefully in conservation management plans. However while both native and alien ‘thugs’ accounted for significant amounts of the variation in field-layer species composition, the amounts were small overall, and much smaller than other environmental variables. With respect to other environmental factors that are likely to influence field-layer species, both climate and management accounted for much more variation than ‘thugs’, native or alien, and it is likely that changes in these will have much greater impacts.

Conclusions and caveats

In the past, a large fraction of conservation management has been based on individual practitioner experience, or bias (Pullin et al.2004; Sutherland et al.2004), and there is a need to develop an evidence-based approach to conservation management practice (Stewart et al.2005). This study attempted to assess the relative potential importance of native and alien ‘thug’ species on woodland ground flora by carrying out a systematic survey at the national scale in Great Britain. Three of the previously-identified native ‘thugs’ explained more of the variation than a group of alien species which included Rhododendron ponticum, the shrub popularly deemed to be the worst alien invader of woodland in Britain. This conclusion reflects the relative risk of these species groups at the national scale; it does of course not imply that alien species do not cause problems at the local scale in individual woods, rather that the impact of these native ‘thug’ species might be an even greater risk threat. We accept that the native ‘thugs’ have co-evolved over millennia with the native flora, but with increasing eutrophication of the British countryside (Smart et al.2005) it is possible that any expansion of these ‘thug’ species will enhance biotic homogenisation processes (Smart et al.2006). Moreover, this work provides a baseline assessment of the woodland resource at the national scale. It can, therefore, be used in the future to assess change in status of both native and alien ‘thugs’ as a result of changes in climate, management, pollution, or spread of alien species.

It would also be worthwhile monitoring the relative abundance of the species identified within the native and alien ‘thugs’ groups in individual woodlands. Where there is an increase in cover of any of these species then there may be declines in other field-layer species, and hence a change in conservation value. Where such effects are damaging the native woodland flora, appropriate experimental research should be undertaken to develop methods for reducing the ‘thug’ species and retaining or restoring rich woodland field-layer flora (Dehnen-Schmutz et al.2004; Edwards et al.2000; Pakeman & Marrs 1992).



PMC received financial support from English Nature and CEH. We thank Dr Hugh McAllister (University of Liverpool) for technical support, Sandra Mather for the illustrations and both David Pearman (BSBI) and Kevin Walker (CEH, Monks Wood), for valuable comments on earlier versions of this manuscript. The National Woodland survey was funded by Countryside Council for Wales, Defra, English Nature, Forestry Commission, JNCC, Scottish Natural Heritage and the Woodland Trust.


  1. Bobbink, R. & Willems, J. H. (1987). Increasing dominance of Brachypodium pinnatum (L.) Beauv. in chalk grasslands: a threat to a species-rich ecosystem. Biol. Conserv. 40: 301 – 314.CrossRefGoogle Scholar
  2. Borcard, D., Legendre, P. & Drapeau, P. (1992). Partialling out the spatial component of ecological variation. Ecology 73: 1045 – 1055.CrossRefGoogle Scholar
  3. Bunce, R. G. H. & Shaw, M. W. (1973). A standardized procedure for ecological survey. J. Environm. Managem. 1: 239 – 258.Google Scholar
  4. CBD (2002). VI/23. Alien species that threaten ecosystems, habitats or species In: Annex I. Decisions adopted by the conference of the parties to the Convention on Biological Diversity at its sixth meeting, pp. 249 – 262. UNEP, The Hague, Netherlands.URL:
  5. Corney, P. M. (2006). Assessment of the factors influencing woodland field-layer vegetation composition across Britain. PhD thesis, University of Liverpool, Liverpool.Google Scholar
  6. ____, Le Duc, M. G., Smart, S. M., Kirby, K. J., Bunce, R. G. H. & Marrs, R. H. (2004). The effect of landscape-scale environmental drivers on the vegetation composition of British woodlands. Biol. Conserv. 124: 491 – 505.CrossRefGoogle Scholar
  7. ____, ____, ____, ____, ____ & ____ (2006). Relationships between the species composition of forest field-layer vegetation and environmental drivers, assessed using a national scale survey. J. Ecol. 94: 383 – 401.CrossRefGoogle Scholar
  8. Crawley, M. J. (2007). The R book. Wiley, Chichester.CrossRefGoogle Scholar
  9. DEFRA (2003). Review of non-native species policy; Report of the working group. DEFRA, London.
  10. Dehnen-Schmutz, K., Perrings, C. & Williamson, M. (2004). Controlling Rhododendron ponticum in the British Isles: an economic analysis. J. Environm. Managem. 70: 323 – 332.Google Scholar
  11. EDINA (2002). Digimap online mapping service ( © Crown Copyright. An EDINA Digimap/JISC supplied service.
  12. Edwards, C., Clay, D. V. & Dixon, F. L. (2000). Stem treatment to control Rhododendron ponticum under woodland canopies. Aspects Appl. Biol. 58: 1 – 8.Google Scholar
  13. ERDAS (2001). ERDAS IMAGINE, version 8.5. ERDAS, inc., Atlanta, Georgia,Google Scholar
  14. ESRI (2000). ArcView GIS, version 3.2a. Environmental Systems Research Institute, Redlands, California.Google Scholar
  15. ____ (2001). Map Manager version 6.2. Environmental Systems Research Institute, Redlands, California.Google Scholar
  16. Grime, J. P., Hodgson, J. D. & Hunt, R. (1988). Comparative plant ecology: a functional approach to common British species. Unwin Hyman, London.Google Scholar
  17. Hill, M. O. (1996). TABLEFIT version 1.0, for identification of vegetation types. Institute of Terrestrial Ecology, Huntingdon.Google Scholar
  18. IUCN (2000). Guidelines for the prevention of biodiversity loss caused by alien invasive species. Prepared by the SSC Invasive Species Specialist Group; Approved by the 51st Meeting of the IUCN Council, Gland, Switzerland.
  19. Kirby, K. J. & Woodell, S. R. J. (1998). The distribution and growth of bramble (Rubus fruticosus L. agg.) in British semi-natural woodland and their implications for nature conservation. J. Practical Ecol. & Conserv. 2: 31 – 41.Google Scholar
  20. ____, Smart, S. M., Black, H. I. J., Bunce, R. G. H., Corney, P. M. & Smithers, R. J. (2005). Long-term ecological changes in British broadleaved woodland 1971 –2001. Research Report. 653. English Nature, Peterborough.
  21. Le Duc, M. G., Pakeman, R. J. & Marrs, R. H. (2000). Vegetation development on upland and marginal land treated with herbicide, for bracken (Pteridium aquilinum) control, in Great Britain. J. Environm. Managem. 58: 147 – 160.Google Scholar
  22. Legendre, P. & Gallagher, E. D. (2001). Ecologically meaningful transformations for ordination of species data. Oecologia 129: 271 – 280.CrossRefGoogle Scholar
  23. Liu, J., Liang, S. C., Liu, F. H., Wang, R. Q. & Dong, M. (2005). Invasive alien plant species in China: regional distribution patterns. Diversity Distrib. 11: 341 – 347.CrossRefGoogle Scholar
  24. Marrs, R. H. & Watt, A. S. (2006). Biological Flora of the British Isles: Pteridium aquilinum (L.) Kuhn. J. Ecol. 94: 1272 – 1321.CrossRefGoogle Scholar
  25. McAllister, H. A. & Rutherford, A. (1990). Hedera helix L. and H. hibernica (Kirchner) Bean (Araliaceae) in the British Isles. Watsonia 18: 7 – 15.Google Scholar
  26. Millennium Ecosystem Assessment (2005).
  27. Oksanen, J. (2003). Gradient Analysis of Vegetation (Gravy) software, version 0.0-21.
  28. Pakeman, R. J. & Marrs, R. H. (1992). The conservation of bracken Pteridium aquilinum (L.) Kuhn-dominated communities in the UK, and an assessment of the ecological impact of bracken expansion or its removal. Biol. Conserv. 62: 101 – 114.CrossRefGoogle Scholar
  29. Pearman, D. (2004a). Invading aliens — or invading natives? BSBI News 96: 41 – 42.Google Scholar
  30. ____ (2004b). The native plants are restless. Radio 4; Nature (25th October). URL:
  31. ____ & Lockton, A. (2004). Aliens and introductions: A discussion paper. URL:
  32. Peres-Neto P. R., Legendre, P., Dray, S. & Borcard, D. (2006). Variation partitioning of species data matrices: estimation and comparison of fractions. Ecology 87: 2614 – 2625.PubMedCrossRefGoogle Scholar
  33. Peterken, G. F. & Game, M. (1981). Historical factors affecting the distribution of Mercurialis perennis in central Lincolnshire. J. Ecol. 69: 781 – 796.CrossRefGoogle Scholar
  34. Peterson, A. T., Papes, M. & Kluza, D. A. (2003). Predicting the potential invasive distributions of four alien plant species in North America. Weed Sci. 51: 863 – 868.CrossRefGoogle Scholar
  35. Pigott, C. D. (1977). The scientific basis of practical nature conservation: aims and methods of conservation. Proc. Roy. Soc., London, Ser. B, Biol. Sci. 197: 59 – 68.CrossRefGoogle Scholar
  36. Pullin, A. S., Knight, T. M., Stone, D. A. & Charman, K. (2004). Do conservation managers use scientific evidence to support their decision-making? Biol. Conserv. 119: 245 – 252.CrossRefGoogle Scholar
  37. Pysek, P., Jarosik, V. & Kucera, T. (2003). Inclusion of native and alien species in temperate nature reserves: An historical study from Central Europe. Conserv. Biol. 17: 1414 – 1424.CrossRefGoogle Scholar
  38. R Development Core Team (2004). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna.
  39. Rackham, O. (2006). Woodlands. HarperCollins, London.Google Scholar
  40. Ratcliffe, D. A. (1977). A nature conservation review: the selection of biological sites of national importance to nature conservation in Britain. CUP, Cambridge.Google Scholar
  41. Ray, D. (2001). Ecological Site Classification, version 1.7. A PC-based Decision Support System for British Forests. Forestry Commission, Edinburgh.Google Scholar
  42. Richardson, D. M., Rouget, M., Ralston, S. J., Cowling, R. M., Van Rensburg, B. J. & Thuiller, W. (2005). Species richness of alien plants in South Africa: Environmental correlates and the relationship with indigenous plant species richness. Ecosci. 12: 391 – 402.CrossRefGoogle Scholar
  43. Rodwell, J. (ed.) (1991). Woodlands and scrub. CUP, Cambridge.Google Scholar
  44. Smart, S. M., Bunce, R. G. H., Marrs, R., Le Duc, M., Firbank, L. G., Maskell, L. C., Scott, W. A., Thompson, K. & Walker, K. J. (2005). Large-scale changes in the abundance of common higher plant species across Britain between 1978, 1990 and 1998 as a consequence of human activity: Tests of hypothesised changes in trait presentation. Biol. Conserv. 124: 355 – 371.CrossRefGoogle Scholar
  45. ____, Thompson, K., Marrs, R. H., Le Duc, M. G., Maskell, L. C. & Firbank, L. G. (2006). Biodiversity loss and biotic homogenization across human-modified ecosystems. Proc. Roy. Soc. London, Ser. B, Biol. Sci. 273: 2659 – 2665.CrossRefGoogle Scholar
  46. Sokal, R. R. & Rohlf, F. J. (1995) Biometry. 3rd edition. W. H. Freeman and Co., New York.Google Scholar
  47. Stewart, G. B., Coles, C. F. & Pullin, A. S. (2005). Applying evidence-based practice in conservation management: Lessons from the first systematic review and dissemination projects. Biol. Conserv. 126: 270 – 278.CrossRefGoogle Scholar
  48. Sutherland, W. J., Pullin, A. S., Dolman, P. M. & Knight, T. M. (2004). The need for evidence-based conservation. Trends Ecol. Evol. 19: 305 – 308.PubMedCrossRefGoogle Scholar
  49. ter Braak, C. J. F. & Šmilauer, P. (2002). CANOCO Reference manual and CANODRAW for Windows User's guide: Software for Canonical Community Ordination (version 4.5). Microcomputer Power, Ithaca, NY.Google Scholar

Copyright information

© The Board of Trustees of the Royal Botanic Gardens, Kew 2011

Authors and Affiliations

  • R. H. Marrs
    • 1
  • M. G. Le Duc
    • 1
  • S. M. Smart
    • 2
  • K. J. Kirby
    • 3
  • R. G. H. Bunce
    • 4
  • P. M. Corney
    • 1
    • 5
  1. 1.School of Environmental SciencesUniversity of LiverpoolLiverpoolUK
  2. 2.Centre for Ecology & HydrologyUniversity of LancasterLancasterUK
  3. 3.Natural England, Northminster HousePeterboroughUK
  4. 4.AlterraWageningenThe Netherlands
  5. 5.Hyder Consulting UK Ltd, The MillStroudUK

Personalised recommendations