Diversity and Distributions (2003) 9, 385 –398


   B I O D I VE R S I T Y R E S E A RC H




Distribution of alien plant species in relation to human
Blackwell Publishing Ltd.




disturbance on the Georgia Sea Islands
J. C. RODGERS III 1 * and K. C. PARKER 2 1Department of Geosciences, Mississippi State
University, Starkville, MS 39762, U.S.A. 2Department of Geography, University of Georgia, Athens, GA
30602, U.S.A.



Abstract. This study investigates the effects of       of accessibility and overall land use. Alien plant
human disturbance and environmental factors          cover was appreciably greater in severely disturbed
on the distribution of alien plant species on the       sites than in less disturbed sites on all islands and
Georgia Sea Islands (GSI), USA. We sampled the        within both habitats. However, the difference
absolute cover of native and alien plant species on      between disturbance categories was much less
two tourist islands (St. Simons Island and Jekyll       pronounced in the primary dunes where human
Island) and on two protected National Wildlife        disturbance agents do not mitigate the harsh
Refuge Islands (Blackbeard Island and Wassaw         environmental conditions of this habitat (salt
Island). On each island, vegetation composition        spray and saline soils). Alien plant abundance on
and environmental variables (soil properties and       the GSI is evidently more dependent upon the
salt spray) were measured in two habitats that        availability of disturbed ground than the degree
differed substantially in their degree of environ-      of accessibility or overall island development. It
mental stress, the more exposed primary dune and       appears that human disturbance increases alien
the more sheltered and inland maritime forest.        cover in general, but in environments where the
Sites were further stratified within each habitat       stress levels are not mitigated, human disturbance
into areas that had different levels of human dis-      does little to foster alien invasions.
turbance. Many alien species were present on all
islands and the absolute cover of alien species was      Key words. Alien plants, biological invasions,
not significantly different among islands even         environmental stress, Georgia Sea Islands, human
though they varied substantially in their degree       disturbance.




                               shown that there may be a positive association
INTRODUCTION
                               between species richness and alien invasibility
The traits that make a community invasible by         (Levine, 2000; Brown & Peet, 2003). Communi-
alien species have been widely discussed in the        ties that have high species richness often have
ecological literature, yet there is still ambiguity      more available resources, and thus would be
regarding why some communities are more inva-         more readily invaded by all species regardless of
sible than others (Crawley, 1987; Lonsdale, 1999;       native or alien status (Huston, 1994; Lonsdale,
Williamson, 1999). Traditionally it was believed       1999; Stohlgren et al., 1999).
that invasions are less successful in species-rich        The association of invasibility and resource
communities because of intense competition with        availability is applicable to habitats that have
the existing native species (Elton 1958; Tilman,       high levels of environmental stress. For the
1997). More recently, however, research has          purposes of this paper, stress is defined as any
                               physical process that inhibits or constrains plant
* Corresponding author. E-mail: rodgers@geosci.msstate.edu  production (Grime, 1979). It has been shown, for

© 2003 Blackwell Publishing Ltd. http://www.blackwellpublishing.com/journals/ddi               385
386  J. C. Rodgers and K. C. Parker

example, that alien species are more abundant       showed that alien species used roadside soils as
in mesic habitats than in habitats where moisture     corridors to invading xeric habitats. Huenneke
availability is more extreme (Rejmanek, 1989).      et al. (1990) showed that the addition of nutrients
Xeric sites are less conducive to alien seed       on serpentine soils of California elevated alien
germination or establishment, and hydric sites      plant abundances. Burke & Grime (1996) dis-
support strong competitors that exclude invasion.     covered that nutrient enrichment increased the
Similarly, Wiser et al. (1998) showed that alien     number of alien plants within limestone grass-
invasions in the Southern Alps of New Zealand       lands of Great Britain. On the Tiwi Islands of
were more frequent in plots that were more fer-      Australia, alien species were most prevalent on
tile, more sheltered, and at lower elevations.      clay-rich soils but were found in drier sandy soils
Additionally, Harrison (1999) reported that alien     in areas around human settlement (Fensham &
species, common throughout most grasslands of       Cowie, 1998). These results suggest that human
California, were significantly less abundant within    disturbance and environmental stress interact
the nutrient-poor serpentine grasslands. Alien      in controlling alien distributions, but more
invasions in Australia have also been shown to      work is needed to determine the nature of these
be associated with more fertile soils (Armor &      relationships.
Piggin, 1977). In these cases, the physically harsh     The purpose of this paper is to investigate the
habitats had limited alien invasions because envi-    interacting effects of human disturbance and
ronmental stress reduces the ability of species to    environmental stress on the distribution of alien
utilize available resources (Davis et al., 2000).     plant species on the Georgia Sea Islands (GSI).
  Human disturbance, on the other hand, often      Specific questions addressed include: (1) Does
increases resource availability and thus makes      the distribution of alien plant species vary with
disturbed habitats more susceptible to invasion      respect to human disturbance, both at a broad
(Davis et al., 2000). In fact, the long-held notion    scale among islands with different visitation
that alien invasions are tied to human disturbance    patterns, and at a more local scale within islands,
(Hobbs & Huenneke, 1992; Mack & D’Antonio,        among habitats receiving different levels of human
1998; Davis et al., 2000) is possibly related to the   use? (2) Does the distribution of alien plant
changes in available resources. Human distur-       species vary with soil cation concentrations, soil
bance facilitates the invasion of alien plants by     pH, and soil organic matter content? (3) Do
opening and creating available grounds for        habitat types characterized by different degrees
invasion, by severely reducing native species that    of environmental stress and human disturbance
previously excluded alien colonization, by selec-     regime interact in their influence on the abun-
tively eliminating long-term relationships between    dance of alien plants?
organisms, by creating vacant niches, and by chang-
ing the natural disturbance regimes (Orians,
                             STUDY AREA
1986; Mack & D’Antonio, 1998).
  Both human disturbance and environmental        The GSI are a group of 15 barrier islands that
stress affect alien invasions by changing resource    are separated from the mainland of Georgia,
availability. It would seem, then, that the interac-   USA, by saltwater estuaries, tidal creeks, and
tion of these two factors would play an impor-      sounds (Fig. 1). Although the smaller islands and
tant role in the distribution of alien species. For    the parts of larger islands that currently face the
example, Forcella & Harvey (1989) reported that      open ocean are mid-Holocene in age (c. 4000 –
alien plants extend their range from lower        5000 BP), the bulk of the larger islands were
montane zones, which have a warmer climate,        actually part of the mainland during the last
into the higher elevations of Montana, which have     glacial maximum (c. 18 000 BP) (Walker &
a much cooler climate, using roadsides as avenues     Coleman, 1987). They became isolated when sea
of migration. Kuhn & Zedler (1997) similarly       level rose following the melting of the ice sheets.
found that aliens have invaded Californian coas-     Because they were once joined to continental
tal marshes along storm drains and sewage         North America, the biota of these barrier islands
spills where human-induced freshwater runoff       is considered to be a subset of the mainland
has lowered salinity levels. Greenberg et al. (1997)   (Ehrenfeld, 1990).

                      © 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
                                    Alien plants on Georgia Sea Islands   387

                                 Island and St. Simons Island are tourist islands
                                 that are accessible by vehicle via a causeway and
                                 have many hotels, golf courses, and residential
                                 areas. In 1997, several million people visited
                                 these islands (Jekyll Island Convention and
                                 Visitors Bureau, personal communication, April,
                                 1999 and Brunswick and the Golden Isles of
                                 Georgia Visitor Centre, personal communication,
                                 April 1999). Both Jekyll and St. Simons Island
                                 are heavily developed, but they each contain
                                 pockets of maritime forest and dunes that have
                                 little human use. For example the Calm Creek
                                 Park on Jekyll Island and the Fort Frederica
                                 National Monument site on St. Simons Island
                                 are areas where few people venture off-trail into
                                 the surrounding forest. Similar pockets of dunes
                                 with little human use exist in areas far from
                                 accesses to public beaches. Blackbeard Island and
                                 Wassaw Island, on the other hand, are National
                                 Wildlife Refuges that have very little human
                                 development and are managed through the US
                                 Fish and Wildlife Service (USFWS). Blackbeard
                                 Island was acquired by the US Navy in the 1800s
                                 as a source of live oak timber. The island was
Fig. 1 The Georgia Sea Islands used in this study.        protected and uninhabited throughout the 19th
                                 and 20th centuries. In 1940 the island was donated
                                 to the USFWS to be managed as a National Wild-
  Primary foredune habitats on the GSI are rel-         life Refuge (Savannah Coastal Refuge, 2003a).
atively stressful for plants. Stresses characteristic      Wassaw Island was owned by one family for over
of these sites include poor soils, due to lower         100 years, during which time it remained mostly
organic matter content and higher salinity; salt         undeveloped. In 1969, the majority of the island
spray from ocean winds; local-scale sand move-          was donated to the USFWS to be managed as a
ments that cause burial and sand blasting of veg-        National Wildlife Refuge, excluding a small par-
etation; high winds; broad-scale changes in island        cel of land in the centre of the island that is still
geomorphology; and susceptibility to washover          privately owned (Savannah Coastal Refuge,
events (Ehrenfeld, 1990; Stallins, 2001). The GSI        2003b). Today both Blackbeard Island and Was-
also contain maritime forest habitats that are inland      saw Island are only accessible by boat and special
and less exposed to the harsh coastal conditions.        permission from the USFWS is needed to visit
These areas are more protected from salt spray          these islands. Approximately 10 000 people visit
and have richer, moister soils. However, available        them annually (Savannah Coastal Refuge, 2003a).
light and access to open ground is more limiting         Both islands are mostly undeveloped; however,
in the forest than in the primary foredunes.           numerous vehicle trails exist in both the forests
  Most of the GSI have been inhabited at some          and dunes that are used by USFWS personnel,
time in the past and they have a long history of         hunters, hikers, and sea-turtle volunteers.
human use by Native Americans, and more
recently, by people of European and African
descent (Vanstory, 1956). Today, the GSI have          METHODS
varying degrees of accessibility and human use.
                                 Field Sampling
Four islands were selected for study to represent
both highly visited islands as well as those that        We sampled the absolute cover of both native
have been granted some type of protection. Jekyll        and alien herbaceous and woody plant taxa

© 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
388  J. C. Rodgers and K. C. Parker

(< 1.5 m in height) on each of the four study       around the random point in each of the four
islands. On each island, we established 20        quadrats. From these data, we calculated the
sampling sites. Ten of the sites were located in     absolute cover of each plant taxon by dividing
primary foredune habitats. The primary dunes       the total number of contacts per species by the
presumably have a high degree of environmental      total possible number of contacts (160 total
stress created from greater exposure to salt       contacts per site: 10 points along the sampling
spray and less fertile soils (Ehrenfeld, 1990). The    frames repeated in 4 cardinal directions, repeated
remaining 10 sites were located in the inland       in 4 quadrats). Plant identification and nomen-
maritime forest habitats where vegetation is less     clature follow Hitchcock (1950), Radford et al.
exposed to the harsh maritime conditions. The       (1968), USDA (1971), and Duncan & Duncan
forest habitats in this study were approximately     (1987). Soil samples (c. 100 g) were also collected
100 m−500 m inland from the shoreline and         from the surface (c. 5 cm depth) and returned to
were separated from the primary dune habitats       the University of Georgia for chemical analysis.
by a secondary dune habitat (an older genetic       Field sampling occurred during October and
sequence of dunes that lie landward of the pri-      November of 1997, a time when many of the
mary dune and are characterized by having more      coastal plant species are in flower (Wilbur Duncan,
woody vegetation).                    University of Georgia, Department of Botany,
  The sites on each island were further stratified    personal communication September, 1996).
according to land use characteristics. The land       We were interested in investigating the degree
use categories were divided into habitats that      to which salt spray varied across the islands.
have a high degree of human disturbance (severely     However, a detailed salt spray research project
disturbed sites) and habitats where human distur-     was beyond the scope of this study because it
bance is minimal (less disturbed sites). The crite-    would have involved analysis of many meteoro-
rion for designating a site as severely disturbed     logical and oceanographic variables. Instead, we
was that the habitat had a high degree of human      initiated a small-scale pilot study during one day
use and visitation. In this study, roadsides, picnic   (and thus during one meteorological setting) on
areas, hiking and vehicle trails, and residential     Wassaw Island in October of 1998. For this pilot
and tourist areas were selected as severely dis-     study, we developed our own methodology. We
turbed sites, with sampling occurring within 2      set out traps consisting of 3 layers of cheesecloth
metres of these types of areas. Only sites that      encased in 17.6-cm diameter, plastic hoops and
were frequently and recently disturbed were        attached to 1.5-m stakes. Five traps were placed
included in the sample so that variation among      at each of three locations along a 150-m linear
sites in the time since last disturbance was       transect (0 m, 50 m, and 150 m). This transect
minimal. Less disturbed sites, on the other hand,     spanned from a primary dune to the interior for-
were at least 30 metres away from disturbed areas     est. Traps were orientated towards the oncoming
and showed no visible signs of severe distur-       sea breezes and they were exposed to winds for
bance. Despite differences in their overall visit-    exactly six hours. Cheesecloth from the exposed
ation and land use, the four islands all support     traps was then collected and returned to the Uni-
sites with these levels of human modification       versity of Georgia Geomorphology Laboratory
within each of the habitat types.             to be analysed for sodium concentration.
  At each site, we sampled vegetation in a 10-m
X 10-m plot with the point intercept method
                             Laboratory Methods
(Goodall, 1957). Each plot was subdivided into
four equal quadrats and within each quadrat, we      Soil samples were analysed for cation concentra-
randomly located a sampling point and measured      tion, pH, and organic matter content. Soil cation
the absolute cover with a 1-m long, 10-pin sam-      concentrations (p.p.m. of dry weight as plant
pling frame placed 1 m away from the sample        available nutrients) of sodium and the plant
point. Taxa that were intercepted by the sam-       nutrients boron, calcium, iron, potassium,
pling pins were recorded at 10 cm intervals along     magnesium, manganese, molybdenum, phosphorus
the sampling frames. Sampling of vegetation was      and zinc were measured with an inductively cou-
repeated in each of the four cardinal directions     pled argon plasma emission spectrometer (ICP)

                      © 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
                                    Alien plants on Georgia Sea Islands    389

at the University of Georgia Chemical Analysis          transformations (Zar, 1996) of absolute alien
Laboratory. Before analysis, soil samples were          cover so that data more closely conformed to a
prepared with the Mehlich-1, or double acid           normal distribution.
extraction procedure (Risser & Baker, 1990). Soil           was also used to test for significant dif-
pH was measured with a hand-held pH meter            ferences in salt spray among the three locations
using a 1:1 soil to distilled water mixture (USDA,        along the transect. Additionally,  was used
1992). Soil organic matter content was measured         to test for significant differences in soil sodium
using loss on ignition (Dean, 1974).               concentrations between dune and forest habitats.
  For the salt spray analysis, we soaked the          Spearman correlation analysis was also used to
exposed cheesecloth in 50 mL of distilled water         identify soil variables that were significantly related
for 24 h and then had the resulting solution ana-        to the cover of both alien species and the more
lysed at the Chemical Analysis Laboratory with          common native species. Common native species
ICP analysis. Five blank samples (unexposed           were defined as those that occurred in at least
cheesecloth that was soaked in distilled water for        half of the plots for each habitat. Analyses were
24 h) were also analysed with the salt spray sam-        performed separately for dune and forest habitats.
ples to determine initial cation concentrations.
The soil cation data and the salt spray data were
both calibrated before analysis by subtracting the        RESULTS
mean concentration of the corresponding cations
                                 Alien Species Encountered on the GSI
from blank reagents.
                                 The alien species that were found in the GSI
                                 study plots are reported by habitat and by island
Data Analysis
                                 in Table 1. This list is not a comprehensive alien
From the cover of individual plant species, we          flora for the entire GSI, but rather it is a list of
calculated the total absolute cover of both native        those species encountered within the sampling
and alien species within each site. Alien/native         frames.
status was determined from published floras,             The most common alien species and the ones
such as Small (1903), Hitchcock (1950), USDA           found in both habitats and across most islands
(1971), Godfrey & Wooten (1979, 1981), and            were grasses (Table 1). Most alien grass species
Sanders (1987).                         were found within forest and dune environments,
  We calculated species richness (no. species per        but Echinochloa crusgalli, Eremochloa ophiuroides,
100 m2 plot) for both native and alien species for        Secale cereale, and Setaria viridis were only
each plot. We then averaged the individual plot         found in forest sites. No alien grasses were
richness values for each island, for both habitats        found only in the dunes when examined across
within each island, and for each disturbance type        all islands. Jekyll Island had the highest number
within both habitats. In addition, we calculated         of alien grasses (9 species), followed by Wassaw
the percentage of the flora comprised by alien          Island (7 species), St. Simons Island (6 species),
species for each plot. Means of the percentage          and Blackbeard Island (4 species).
alien flora were determined for each island, for           There were substantially fewer alien forbs
each habitat, and for each disturbance type           (Table 1), and the distribution of these forbs was
within habitats.                         more island- and habitat-specific. St. Simons
  We used a three-factor nested analysis of           Island had the highest number of alien forbs
variance (; Sokal & Rohlf, 1981) to test the        (3 species), followed by Jekyll Island (2 species)
research hypothesis that human disturbance,           and Wassaw Island (1 species). No alien forbs were
habitat type, and island type have an effect on         encountered in Blackbeard Island study plots.
the distribution of alien plant taxa on the Georgia       Lonicera japonica and Lantana camara were the
Sea Islands. This analysis tests for significant         only alien forb species found in the dune habitat,
differences in the cover of alien plants among          and these taxa were only present on St. Simons
islands, among habitats within islands, and           Island. Sapium sebiferum and Lantana camara
among disturbance categories within habitats           were the only nongrass alien plant species found
within islands. For all tests, we used arcsine          in both less disturbed and severely disturbed sites.

© 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
390  J. C. Rodgers and K. C. Parker

Table 1 Presence of alien taxa encountered in study plots by habitat (primary dune and forest) and by island
(BB = Blackbeard Island, WS = Wassaw Island, LK = Jekyll Island, ST = Simons Island)

                                    Primary Dune       Forest

                        Origin         BB  WS   JK   ST  BB  WS   JK  ST

Grasses
Cynodon dactylon (L.) Pers.           Africa         *   *   *   *      *   *   *
Digitaria sanguinalis (L.) Scop.        Eurasia /              *   *      *   *
                        Mediterranean
Echinochloa crusgalli (L.) Beauv.        Eurasia /                       *
                        Mediterranean
Eleusine indica (L.) Gaertn.          Africa         *      *   *      *   *   *
Eremochloa ophiuroides (Munro) Hack.      South-east Asia                    *
Paspalum notatum Flugge.            Tropical America       *   *   *   *   *   *   *
Poa annua L.                  Eurasia /                    *   *
                        Mediterranean
Secale cereale L.                Eurasia /                           *
                        Mediterranean
Setaria viridis (L.) Beauv.           Eurasia /                           *
                        Mediterranean
Sorghum halepense (L.) Pers.          Eurasia /              *   *         *
                        Mediterranean
Sporobolus poiretii (Roem. & Shult.) Hitch.   Tropical America                       *   *
Forbs
Commelina communis L.              Asia                                *
Lantana camara L.                Tropical America              *
Lonicera japonica Thunb.            Asia                    *
Medicago lupulina L.              Europe /                           *
                        Mediterranean
Richardia brasiliensis Gomez          Tropical America                       *
Sapium sebiferum (L.) Roxb.           Asia                          *


                              within plots, with a significant difference occur-
Alien and Native Species Richness
                              ring between disturbance types within habitats
Alien species richness varied little among the       within islands (Table 3). Across all islands and
islands, with mean values ranging from 0.4 on        habitats, severely disturbed sites had an average
Jekyll Island to 0.7 on Wassaw Island (Table 2).      alien cover of 17.9%, whereas less disturbed sites
Native species richness was also similar among       had an average alien cover of only < 1%. Anthro-
the four islands, ranging from 7.4 on Blackbeard      pogenic disturbance significantly increased the
Island to 8.6 on St. Simon’s Island. The percent-      cover of alien species overall, but the magnitude
age of flora composed of alien species, however,       of the increase depended upon the habitat type.
showed a different pattern. The two protected        Forest sites showed a greater difference in cover
islands had a greater percentage of alien species      between severely disturbed and less disturbed
[Blackbeard Island (7.6%) and Wassaw Island         sites; indeed, the highest mean cover of alien
(9.2%)] than Jekyll Island (6.0%) and St. Simons      plants was in severely disturbed forests (Fig. 2a).
Island (5.6%).                       The difference in alien cover between disturbance
                              categories was less pronounced for primary dune
                              habitat.
Spatial Variation in Alien Plant Cover
                                The cover of alien plants was not significantly
Spatial variation across the study area was evi-      different among islands, even though Wassaw
dent in the absolute cover of alien plant species      and Blackbeard Island have much lower human

                       © 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
                                    Alien plants on Georgia Sea Islands    391

Table 2 Summary of native and alien plant species occurrence on the GSI by island, habitat, and disturbance type

                               Mean Species Richness (± SD)
                      Disturbance                      Mean Percent
Island         Habitat      category      Native        Alien     Flora Alien (± SD)

Blackbeard Total                        7.4 ± 2.6      0.5 ± 0.7    7.6 ± 2.6
Blackbeard       Dune Total               7.9 ± 2.7      0.2 ± 0.4    1.7 ± 3.7
Blackbeard       Dune        Severely      9.2 ± 2.9      0.2 ± 0.4    1.4 ± 3.2
Blackbeard       Dune        Less        7.0 ± 1.8      0.2 ± 0.4    2.0 ± 2.1
Blackbeard       Forest Total              6.9 ± 2.9      0.8 ± 0.7   13.5 ± 14.5
Blackbeard       Forest       Severely      5.6 ± 2.4      1.2 ± 0.8    6.8 ± 2.8
Blackbeard       Forest       Less        8.2 ± 3.0      0.4 ± 0.5    1.7 ± 3.7
Wassaw Total                          8.4 ± 2.3      0.7 ± 1.2    9.2 ± 16.4
Wassaw         Dune Total               7.2 ± 2.3      0.1 ± 0.3    0.9 ± 2.9
Wassaw         Dune        Severely      8.4 ± 2.3      0.2 ± 0.4    1.8 ± 8.9
Wassaw         Dune        Less        6.0 ± 1.6      0.0 ± 0.0    0.0 ± 0.0
Wassaw         Forest Total              7.8 ± 3.4      1.3 ± 1.4   16.5 ± 20.6
Wassaw         Forest       Severely      6.6 ± 3.9      0.2 ± 0.4   30.2 ± 21.2
Wassaw         Forest       Less        9.0 ± 2.4      0.1 ± 0.3    2.9 ± 6.4
Jekyll Total                          8.2 ± 2.9      0.7 ± 1.1    6.0 ± 8.9
Jekyll         Dune Total               7.1 ± 2.9      0.4 ± 0.7    4.4 ± 8.4
Jekyll         Dune        Severely      8.2 ± 3.7      0.8 ± 0.8    8.8 ± 3.9
Jekyll         Dune        Less        6.0 ± 2.4      0.0 ± 0.0    0.0 ± 0.0
Jekyll         Forest Total              9.3 ± 2.6      1.0 ± 1.4    7.6 ± 9.6
Jekyll         Forest       Severely       11 ± 2.5      2.0 ± 1.2   15.3 ± 7.9
Jekyll         Forest       Less        7.6 ± 1.1      0.0 ± 0.0    0.0 ± 0.0
St. Simons Total                       8.56 ± 2.4      0.6 ± 0.9    5.6 ± 9.3
St. Simons       Dune Total               9.6 ± 2.4      0.7 ± 2.3    8.9 ± 6.8
St. Simons       Dune        Severely      9.8 ± 2.6      0.8 ± 0.8    6.9 ± 6.7
St. Simons       Dune        Less        9.4 ± 2.4      0.6 ± 0.9    4.9 ± 7.4
St. Simons       Forest Total              7.5 ± 1.9      0.5 ± 1.1    5.3 ± 11.7
St. Simons       Forest       Severely      8.0 ± 2.34     1.0 ± 1.4   10.7 ± 15.3
St. Simons       Forest       Less        7.0 ± 1.4      0.0 ± 0.0    0.0 ± 0.0



Table 3 Nested- table of effects of islands, habitats within islands, and disturbance type within habitats
within islands

Source of Variation        DF      Sum of Squares        Mean Square    F Value     P-value

Island               3      1082.02            360.67      0.281      0.838
Habitat (within Island)       4      5138.32           1284.58      0.565      0.695
Disturbance (within         8     18187.05           2273.38      8.108      0.001
habitat (within Island))
Error               64      17944.39            280.38

Note: boldface effects were significant at P < 0.05.


visitation and experience a lower intensity of          from values for the other islands. St. Simons Island
overall human use (Table 3, Fig. 2b). Jekyll           had the lowest total cover of alien species
Island had the highest mean alien cover overall         (2%), which was lower than both protected
(9.5% across all habitats and disturbance cate-         islands (Blackbeard Island = 5.8% and Wassaw
gories), but this was not significantly different         Island = 4.8%).

© 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
392  J. C. Rodgers and K. C. Parker




                              Fig. 3 Mean (±SE) sodium concentration (ppm) of
                              Georgia Sea Island soils by habitat and by disturbance
                              category. Shaded bars on the right represent severely
                              disturbed sites and open bars on the left represent less
                              disturbed sites.

                              of sodium in the soil and the higher amount of
                              salt spray suggest that the primary dune habitat
                              was more stressful in these environmental prop-
                              erties than the forest habitat.
Fig. 2 Mean (±SE) absolute cover of alien plants by
habitat (a) and by island (b). The shaded bars on the
right indicate the mean cover of disturbed sites and    Relationships of alien cover to
the open bars on the left indicate mean cover of the    environmental concentrations
less disturbed sites.
                              The  results indicated a significant differ-
                              ence (P < 0.05) between the dune and forest soils
                              for all variables except molybdenum and zinc
Spatial Variation in Environmental Variables
                              (Table 4). Forest soils had significantly higher
Salt spray and soil salinity were significantly       organic matter and manganese contents, whereas
higher in the primary dune habitats than in the      dune soils had significantly greater pH values and
forest habitat (P < 0.001) when analysed with       concentrations of boron, calcium, iron, potassium,
. During the salt spray investigation, winds     magnesium, sodium and phosphorus.
were shifting from westerly to easterly flow          The correlations between soil properties and
and were blowing between 5 and 10 mph. Under        alien cover showed responses that differed
these reasonably calm weather conditions, the       between dunes and forests (Table 5). Alien cover
mean accumulation of airborne sodium from         in dunes was significantly negatively correlated
the five traps in the primary dune habitat, at       with soil pH and positively correlated with organic
0 m along the transect, was 190 mg Na/cm−2/h        matter content. Alien cover in the forests sites,
[standard deviation (SD) = 45 mg Na/cm−2/h].        however, showed the opposite trends. Additionally,
Sodium concentrations dropped off substantially      six of the nine plant nutrients tested were signif-
at 50 m along the transect to a mean of 40 mg       icantly positively related to the alien cover in for-
Na/cm−2/h (SD = 96 mg Na/cm−2/h). In the forest      est sites, whereas none of these nutrients was
habitat, at 150 m along the transect, sodium con-     linked to alien cover in the dunes. Lastly, alien
centrations were approximately one-eighth of        cover was negatively correlated with soil sodium
that measured in the primary dune (mean = 25 mg      concentrations in the dune sites but was not
Na/cm−2/h and SD = 60 mg Na/cm−2/h).            significantly related to sodium in the forest sites.
  The primary dune also had significantly           Native species cover was less responsive to the
greater concentrations of sodium in the soil        soil variables (Table 5). Forbs and grasses in the
(Fig. 3). The average concentration in the dune      dunes were significantly correlated with only
habitat (145.4 p.p.m.) was three times as high as     molybdenum. In the forest plots, the native her-
in the forest (33.8 p.p.m.). The greater abundance     baceous species cover was significantly negatively

                       © 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
                                     Alien plants on Georgia Sea Islands   393

Table 4 Means (± SE) for the soil properties examined by habitat type

Soil Property                      Dune                     Forest

pH*                            7.1 (± 0.1)                 4.6 (± 0.1)
Organic matter (g)*                    0.5 (± 0.8)                 7.1 (± 0.8)
B (p.p.m.)*                        13.1 (± 4.9)                 1.55 (± 4.9)
Ca (p.p.m.)*                      3563.2 (± 182.0)               466.4 (± 182.0)
Fe (p.p.m.)*                       148.0 (± 6.4)                 76.0 (± 6.4)
K (p.p.m.)*                        45.6 (± 3.9)                 27.7 (± 3.9)
Mg (p.p.m.)*                       112.3 (± 8.4)                 83.8 (± 8.4)
Mn (p.p.m.)*                        9.0 (± 2.9)                 16.0 (± 2.9)
Mo (p.p.m.)                        0.03 (± 0.2)                 0.09 (± 0.2)
Na (p.p.m.)*                       145.4 (± 6.2)                 33.8 (± 6.2)
P (p.p.m.)*                       1904.7 (± 74.9)                163.04 (± 74.9)
Zn (p.p.m.)                        2.8 (± 0.4)                 3.7 (± 0.4)

* indicates that means are significantly different between habitats (P < 0.05).


Table 5 Spearman rank order correlation coefficients (r s) between soil properties examined and cover of both
alien and native plant species by habitat type

                    Alien Plant Cover                Native Plant Cover

Soil Property             Dune           Forest         Dune          Forest

pH                   − 0.61†           0.40†         0.19          0.08
Organic matter (g)            0.32*          − 0.30†        − 0.12         − 0.33*
B                   − 0.18           0.40*        − 0.26         − 0.18
Ca                   − 0.23           0.40*        − 0.27         − 0.01
Fe                    0.17           0.31*        − 0.01          0.01
K                    0.21          − 0.07         − 0.10         − 0.06
Mg                   − 0.21          − 0.06         − 0.10         − 0.18
Mn                   − 0.004           0.43†        − 0.19          0.04
Mo                    0.07          − 0.28         − 0.36*          0.17
Na                   − 0.34*           0.16         − 0.20         − 0.03
P                   − 0.15           0.67†        − 0.24          0.05
Zn                    0.27           0.38*        − 0.16         − 0.14

* Values that are significantly different between habitat groups at P < 0.05 are indicated with *; at P < 0.001
with†.

correlated with organic matter, the same response        both dune and forest habitats, which suggests
as the alien species. But unlike the alien species,       that the establishment of these species is not
native forest forbs and grasses showed no signif-        limited by a lack of dispersal capabilities. Many
icant correlations with any of the soil nutrients.        of the alien grass species that originated from
                                 the Old World, such as Digitaria sanguinalis and
                                 Sorghum halepense, have coexisted with people
DISCUSSION                            for thousands of years and are well adapted to
                                 human disturbance (Baker, 1986). These species
Individual Alien Species
                                 are suited to invade and occupy disturbed areas on
Grasses were the most commonly occurring             the GSI. The alien grasses of tropical origin, such
alien plants found on the GSI. They occurred on         as Cynodon dactylon and Eremochloa ophiuroides,
both protected and tourist islands and within          have been introduced in the south-eastern United

© 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
394  J. C. Rodgers and K. C. Parker

States as lawn grasses because they are heat-       the alien species encountered in our study plots,
tolerant and can quickly spread (Hitchcock,        especially the grasses, were present on most of
1950). These growth characteristics would explain     the islands and within both dune and forest hab-
why they are also abundant in the severely dis-      itats. Furthermore, the absolute cover of these
turbed sites across all islands. Like the Eurasian     species showed no significant difference among
grasses, the tropical grass species seem to be well    islands even though the islands differed substan-
suited to exploiting the disturbed ground on the GSI.   tially in their level of development. The cover of
  There were substantially fewer alien forbs in      alien plants on the tourist islands was not signif-
the study plots. Those that were present are nox-     icantly different from that of the more remote
ious weeds that occur in many different habitats      and protected National Wildlife Refuge islands
besides barrier islands. Commelina communis        when examined across both habitat types and
and Medicago lupulina, for example, are weedy       disturbance types. These results differ somewhat
species that are found throughout the US          from those of Lonsdale (1999), who reported that
Coastal Plain from Texas to Maine (Duncan &        natural refuges usually have approximately half
Duncan, 1987). Each alien forb was restricted to      the number of non-native species than the sur-
a particular island and to a particular habitat.      rounding unprotected areas support. It is unclear
This may indicate more limited dispersal capabil-     when alien species arrived to each of the islands,
ities than the alien grasses. Lonicera japonica is     but many alien species may have become estab-
an alien species that is found in many areas of      lished on the islands prior to their respective
North America. It is a problematic weed in the       development or protection. If propagules were
eastern US (Robertson et al., 1994) and it has       already present within areas subsequently desig-
been classified as a ‘severe threat’ by the Florida     nated as nature reserves, disturbances common to
Exotic Pest Council (Floridata, 2002). It was       both developed and protected islands, such as con-
only found in a few dune sites on St. Simons        struction and use of roadways, may have facilitated
Island, but its seeds are dispersed by birds.       the dispersal of aliens within natural reserves.
Because this species is such a threat elsewhere        The abundance of alien species, however, was
and because it has the capability of quickly        strongly related to the degree of disturbance.
spreading, L. japonica may warrant additional       Sites subjected to human disturbance included
research to monitor changes in distribution.        areas that experienced the removal of existing
  Perhaps the most threatening alien species       vegetation and, through changes in available
found on the GSI was Sapium sebiferum (tallow       light levels and disruption of soil, alteration of
tree). This species can cause widespread changes      the physical environment. Alien species cover was
in forest structure by forming monospecific         substantially greater in severely disturbed sites in
stands at the exclusion of native trees species. It    both the primary dune and forest habitats. Addi-
can tolerate shade and will grow through existing     tionally, the  results showed that alien
canopies (Jones & McLeod, 1989). S. sebiferum       cover among islands was not significantly differ-
also spreads very aggressively by vegetative means     ent, which indicates that this contrast between
(producing suckers and resprouting from the        disturbed and less disturbed sites was similarly
stump) and by producing hundreds of seeds         evident on all islands. Therefore, the establish-
per year. In this study, S. sebiferum was only       ment of alien species on the GSI seems to be
found on Wassaw Island forest sites, but it was      more related to the existence of human-modified
growing in both severely disturbed and less        areas rather than overall island land use.
disturbed sites. It is likely that the seeds will be     The results from this study examined the cur-
carried to the other islands, and thus its dis-      rent cover of alien species in different settings,
tribution should also be carefully studied.        but it did not determine what factors actually led
                              to these differences. It is unclear whether the
                              physical disturbance (e.g. construction and use
Alien species on islands with different levels
                              of roadways) led to the higher invasibility of
of human use
                              disturbed sites, or whether the roadways allowed
The abundance of alien species on the GSI was       for the influx of more alien plant propagules.
not related to overall island land use. Many of      Propagule pressure (Lonsdale, 1999) is a major

                       © 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
                                    Alien plants on Georgia Sea Islands   395

influence on habitat invasibility and future           more important control of alien cover in the
research may need to determine how this factor          primary dunes than the soil nutrients examined
interacts with both human disturbance and the          in this study. The maritime forest understorey
physical constraints of the environment. Alien          had much lower values of salt spray and soil
species have been reported to be more extensive         sodium, and cover of alien species was signifi-
in disturbed areas on other US barrier islands          cantly positively correlated with soil concentra-
(Miller & Jones, 1967; Anderson & Alexander,           tions of six of the nine essential plant nutrients.
1985; Klotz, 1986; Looney et al., 1993). The           Native forb and grass cover showed no relation-
invasion of alien taxa on more remote oceanic          ship with these factors. Perhaps in the absence
islands has also been facilitated by the availability      of saline stress, the cover of alien plants in the
of human disturbed sites (Cronk, 1980; Lorence          maritime forest responded more closely to varia-
& Sussman, 1986; Kloot, 1987; & Dean et al.,           tion in soil fertility, with greater cover in the
1994). Merlin & Juvik (1992) found that on            more fertile sites. It appears, then, that the
islands with similar area and proximity, alien          presence of saline conditions limits alien plant
cover was highest on islands that had been            growth in the primary dune habitats to a much
grazed. These studies suggest that although dis-         greater degree than in the maritime forest.
persal of alien taxa to invaded islands is impor-          The maritime forest is not a completely stress-
tant, the availability of suitable sites created by       free habitat because plants on the forest floor
human modification is just as crucial in deter-          compete for space and available light. On the
mining the abundance of alien species (D’Antonio         GSI, human disturbance dramatically increased
& Dudley, 1995). The results from this study           alien plant cover perhaps by opening available
further substantiate this claim.                 ground. In contrast to the dune habitat where
                                 disturbance does not ameliorate the stressful
                                 saline conditions, human disturbance in the forest
Alien plant species by habitat
                                 habitat may decrease stress levels, thus making them
The cover of alien plants on the GSI did not           more conducive to invasion by non-native taxa.
differ between habitats within islands. However,          These results are somewhat similar to previous
it was evident that alien cover showed different         studies that have reported enhanced alien inva-
responses within the different habitats when           sions in stressful habitats that have been dis-
examined across all islands. The suites of varia-        turbed (Kuhn & Zedler, 1997; Fensham & Cowie,
bles correlated with alien cover in each habitat         1998; Harrison, 1999). In most of those studies,
suggest environmental controls for the patterns         human disturbance served to mitigate the stresses
observed. Primary dunes had a higher abundance          of the environment (e.g. diluting saline water,
of salt spray and a greater concentration of           changing soil textures, or adding calcium to
sodium in the soil than maritime forests. Soil          serpentine soils) that had evidently limited alien
salinity is considered stressful for most plant         growth previously. In contrast, stress mitigation
species because it reverses plant root osmotic          ssociated with human disturbance is not evident
potentials, antagonizes the uptake of essential         on the primary dunes of the GSI. Trampling and
plant nutrients, and decreases ability of seed to        other mechanical alterations of the dune environ-
germinate (Ehrenfeld, 1990). Alien cover was           ment do not diminish the stresses associated with
negatively correlated with sodium in the primary         salt spray, poor soils, or sand burial. In the forest
dunes, indicating that the dune sites with the          habitats of the GSI, however, human disturbance
highest alien cover occurred where soil sodium          often opens sites, making them more susceptible
concentrations were lowest. Native dune cover,          to alien invasion by increasing light levels. There-
however, was not significantly correlated with soil        fore, human activity enhances alien occurrence
sodium. Moreover, alien cover in the primary           primarily in less saline habitats of the GSI (e.g.
dunes was not correlated with concentrations of         forest habitats), where the disturbance mitigates
any of the nine essential plant nutrients, even         previously limiting factors. Where human activity
though these nutrients varied spatially within this       instead does not ameliorate environmental stresses
habitat. Therefore, the stresses associated with         (e.g. primary dunes), disturbance does little to
exposure to higher salinity may have served as a         facilitate the establishment of the alien taxa.

© 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
396  J. C. Rodgers and K. C. Parker

                              Armor, R.L. & Piggin, C.M. (1977) Factors influ-
CONCLUSION                          encing the establishment and success of exotic
This study indicates that human disturbance         plants in Australia. Proceedings of the Ecological
interacts with physical environmental conditions       Society of Australia 10, 15 – 26.
                              Baker, H.G. (1986) Patterns of plant invasions in
in determining the abundance of alien species.
                               North America. Ecology of Biological Invasions of
On the Georgia Sea Islands, human disturbance        North America and Hawaii (eds H.A. Mooney
enhances the extent of alien cover in general, but      and J.A. Drake), pp. 44 – 57. Springer-Verlag, New
alien cover increases more substantially in the       York, NY.
maritime forest than in the primary dunes, which      Brown, R.L. & Peet, R.K. (2003) Diversity and
are exposed continually to salt spray and have        invasibility of southern Appalachian plant com-
                               munities. Ecology 84, 32 – 39.
soils with higher salinity values. In such habitats
                              Burke, M.J.W. & Grime, J.P. (1996) An experimental
where the limiting environmental factors are not       study of plant community invasibility. Ecology 77,
mitigated, human disturbance does little to facil-      776 –790.
itate alien species invasion.               Crawley, M.J. (1987) What makes a community
  On a more applied level, the spatial pattern of      invasible? Colonization, Succession, and Stability
alien occurrence on the Georgia Sea Islands         (eds A.J. Gray, M.J. Crawley and P.J. Edwards),
underscores the role played by localized anthro-       pp. 429 – 453. Blackwell Scientific Publications,
                               Oxford, England.
pogenic disturbance in the invasion of native       Cronk, Q.C.B. (1980) Extinction and survival in the
plant communities by alien species. Many alien        endemic vascular flora of Ascension Island. Bio-
species were found on all islands, regardless of       logical Conservation 17, 207– 219.
their levels of human use and accessibility, which     D’Antonio, C.M. & Dudley, T.L. (1995) Biological
indicates that the lower visitation of the pro-       invasions as agents of change on islands versus
tected islands does not prevent their invasion by      mainlands. Islands: Biology Diversity and Ecosys-
                               tem Function, Ecological Studies 115 (eds P.M.
alien species. Therefore, the spatial arrangement
                               Vitousek, L.L. Loope & H. Andsersen), pp. 103 –
of disturbed ground appears to have an impor-        121. Springer, New York, NY.
tant influence on the rate and extent of alien       Davis, M.A., Grime, J.P. & Thompson, K. (2000)
invasion across the more protected islands.         Fluctuating resources in plant communities: a
                               general theory of invasibility. Journal of Ecology
                               88, 528 – 534.
ACKNOWLEDGMENTS                      Dean, W.E. Jr (1974) Determination of carbonate
This research was supported by a NSF Geogra-         and organic matter in calcareous sediment and
                               sedimentary rocks by loss on ignition: Compari-
phy and Regional Science Doctoral Disserta-
                               son with other methods. Journal of Sedimentary
tion Improvement Grant (SBR-9701830). Jenny         Petrology 44, 242 – 248.
Cruse-Sanders, Sam Rodgers, Tony Stallins, and       Dean, W.R.J., Milton, S.J., Ryan, P.G. & Moloney, C.L.
John Schaefer provided extensive help in the         (1994) The role of disturbance in the establish-
field. Wilbur Duncan and Tony Stallins provided        ment of indigenous and alien plants at Inaccessi-
help with plant identification. We also wish to        ble and Nightingale Islands in the South Atlantic
thank John Schaefer, John Crawford, and Carolyn       Ocean. Vegetatio 113, 13 – 23.
                              Duncan, W. & Duncan, M. (1987) The Smithsonian
Hodges for providing transportation to Blackbeard      Guide to Seaside Plants of the Gulf and Atlantic
and Wassaw Islands. We are indebted to Albert        Coast, from Louisiana to Massachusetts, Exclu-
Parker, David Leigh, Chris Peterson, and Vernon       sive of Lower Peninsular Florida, P. 409. Smithso-
Meentemeyer for their comments on an earlier         nian Institution Press, Washington, D.C.
draft of the manuscript. Finally, we wish to        Ehrenfeld, J. (1990) Dynamics and processes of barrier
thank the USFWS, the Jekyll Island Authority,        island vegetation. Aquatic Sciences 2, 437– 480.
                              Elton, C.S. (1958) The Ecology of Invasions by Ani-
and Fort Frederica National Historic Park for
                               mals and Plants, p. 182. Methuen, London, UK.
giving us permission to sample on their properties.    Fensham, R.J. & Cowie, I.D. (1998) Alien plant
                               invasions on the Tiwi Islands. Extent, implications
                               and priorities for control. Biological Conservation
REFERENCES
                               83, 55 – 68.
Anderson, L.C. and Alexander, L.L. (1985) The       Floridata (2002). Website, www.floridata.com/ref/L/
vegetation of Dog Island, Florida. Florida Scientist    loni_jap.cfm.
48, 232–251.                       Forcella, F. & Harvey, S.J. (1989) Eurasian weed

                       © 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
                                    Alien plants on Georgia Sea Islands    397

 infestation in western Montana in relation to veg-       Lorence, D.H. & Sussman, R.M. (1986) Exotic spe-
 etation and disturbance. Madroño 30, 102 –109.          cies invasion into Mauritius wet forest remnants.
Godfrey, R.K. & Wooten, J.W. (1979) Aquatic            Journal of Tropical Ecology 2, 147–162.
 Plants of the Southeastern United States: Mono-        Mack, M.C. & D’Antonio, C.M. (1998) Impacts of
 cotyledons. University of Georgia Press, Athens,         biological invasions on disturbance regimes. Trends
 GA.                               in Ecology and Evolution 13, 195 –198.
Godfrey, R.K. & Wooten, J.W. (1981) Aquatic           Merlin, M. & Juvik, J. (1992) Relationships among
 Plants of the Southeastern United States: Dicoty-        native and alien plants on Pacific Islands with and
 ledons. University of Georgia Press, Athens, GA.         without significant human disturbance and feral
Goodall, D.W. (1957) Some considerations in the          ungulates. Alien Plant Invasions in Native Ecosys-
 use of point quadrat methods for the analysis of         tems of Hawaii: Management and Research (eds
 vegetation. Australian Journal of Biology Science        C.P. Stone and L.L. Loope), pp. 597– 624. Univer-
 5, 1– 41.                            sity of Hawaii Press, Honolulu, HI.
Greenberg, C.H., Crownover, S.H. & Gordon, D.R.         Miller, G.J. & Jones, S.B. Jr (1967) The Vascular
 (1997) Roadside soils: a corridor for invasion of        Flora of Ship Island, Mississippi. Castanea 32,
 xeric scrub by nonindigenous plants. Natural           84 – 89.
 Areas Journal 17, 99 –109.                   Orians, S.H. (1986) Site characteristics favoring
Grime, J.P. (1979) Plant Strategies and Vegetation         invasions. Ecology of Biological Invasions in North
 Processes, p. 222. Wiley, Chichester, England.          America and Hawaii (eds H.A. Mooney and J.A.
Harrison, S. (1999) Local and regional diversity in        Drake), pp. 133 –148. Springer-Verlag, New York,
 a patchy landscape: Native, alien, and endemic          NY.
 herbs on serpentine. Ecology 80, 70 – 80.           Radford, A.E., Ahles, H.E. & Bell, C.R. (1968)
Hitchcock, A.S. (1950) Manual of the Grasses of the        Manual of the Vascular Flora of the Carolinas. The
 United States, Revised by Agnes Chase, p. 1051.         University of. North Carolina Press, Chapel Hill,
 Dover Publications, Inc, New York, NY.              NC.
Hobbs, R.J. & Huenneke, L.F. (1992) Disturbance,         Rejmanek, M. (1989) Invasibility of plant commu-
 diversity, and invasion: Implications for conserva-       nities. Biology Invasion: a Global Perspective (eds
 tion. Conservation Biology 6, 324 – 337.             J.A. Drake and H.A. Mooney), pp. 369 – 388.
Huenneke, L.F., Hamburg, S.P., Koide, R., Mooney, H.A.       Wiley and Sons, New York, NY.
 & Vitousek, P.M. (1990) Effects of soil resources       Risser, J.A. & Baker, D.E. (1990) Testing soils for
 on plant invasions and community structure in          toxic metals. Soil Testing and Plant Analysis (ed.
 Californian serpentine grassland. Ecology 71,          R.L. Westerman), pp. 275 – 298. Soil Science Soci-
 478– 491.                            ety of America, Madison, WI.
Huston, M.A. (1994) Biology Diversity: The coexist-       Robertson, D.J., Robertson, M.C. & Teague, T. (1994)
 ence of species on changing landscapes, p. 704.         Colonization dynamics of four exotic plants in a
 Cambridge University Press, Cambridge, UK.            northern Piedmont natural area. Bulletin of the
Jones, R.H. & McLeod, K.W. (1989) Shade tolerance         Torrey Botanical Club 121, 119 –129.
 in seedlings of Chinese tallow tree, American syc-       Sanders, J.D. (1987) Identity of Lantana depressa
 amore, and cherrybark oak. Bulletin of the Torrey        and L. ovatifolia (Vergenaceae) of Florida and the
 Botanical Club 116, 371– 377.                  Bahamas. Systematic Botany 12, 44 – 60.
Kloot, P.M. (1987) The invasion of Kangaroo Island        Savannah Coastal Refuge (2003a). Website: http://
 by alien plants. Australian Journal of Ecology 12,        blackbeardisland.fws.gov/blackbeardnwrx.html.
 263– 266.                           Savannah Coastal Refuge (2003b). Website: http://
Klotz, L.H. (1986) The vascular flora of Wallops          savannah.fws.gov/wassawnwrx.html.
 Island and Wallops Mainland, Virginia. Castanea        Small, J.K. (1903) Flora of the Southeastern United
 51, 306 – 326.                          States. The Author, New York, NY.
Kuhn, N.L. & Zedler, J.B. (1997) Differential effects of     Sokal, J.D. & Rohlf, S. (1981) Biometry, 2nd edn.
 salinity and soil saturation on native and exotic        W.H. Freeman and Co, New York NY.
 plants of a coastal salt marsh. Estuaries 20, 391– 403.    Stallins, J.A. (2001) Soil and vegetation patterns in
Levine, J. (2000) Species diversity and biological         barrier-island dune environments. Physical Geo-
 invasions: relating local process to community          graphy 22, 79 – 98.
 pattern. Science 288, 852 – 854.                Stohlgren, T.J., Binkley, D., Chong, G.W.,
Lonsdale, W.M. (1999) Global patterns of plant           Kalkhan, M.A., Schell, L.D., Bull, K.A., Otsuki, Y.,
 invasions and the concept of invasibility. Ecology        Newman, G., Bashkin, M. & Son, Y. (1999) Exotic
 80, 1522 –1536.                         plant species invade hot spots of native plant
Looney, P.B., Gibson, D.J., Blyth, A. & Cousens, M.        diversity. Ecological Monographs 69, 25–46.
 (1993) Flora of the Gulf Islands National Sea-         Tilman, D. (1997) Community invasibility, recruit-
 shore, Perdido Key, Florida. Bulletin of the Torrey       ment limitation, and grassland biodiversity. Eco-
 Botanical Club 120, 327– 341.                  logy 78, 81– 92.

© 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398
398  J. C. Rodgers and K. C. Parker

USDA (1971) Common Weeds of the United States.         Society of America, Centennial Special Volume 2,
 Dover Publications, Inc. New York, NY.            Boulder, CO.
USDA (1992) Soil Survey Laboratory Method Manual.      Williamson, M. (1999) Biological Invasions. Chapman
 Soil Survey Investigations Report no. 42, Version 2.0.    & Hall, London, UK.
Vanstory, B. (1956) Georgia’s Land of the Golden       Wiser, S.K., Allen, R.B., Clinton, P.W. & Platt, K.H.
 Isles. University of Georgia Press, Athens, GA.       (1998) Community structure and forest invasion by
Walker, H.J. & Coleman, J.M. (1987) Atlantic and        an exotic herb over 23 years. Ecology 79, 2071–2081.
 Gulf Coastal Province. Geomorphic Systems of North     Zar, J. (1996) Biostatistical Analysis, 3rd edn. Prentice
 America (ed. W. Graf), pp. 51–110. Geological        Hall, Upper Saddle River.




                        © 2003 Blackwell Publishing Ltd, Diversity and Distributions, 9, 385 – 398