25
Ecological Monographs, 69(1), 1999, pp. 25­46
1999 by the Ecological Society of America
EXOTIC PLANT SPECIES INVADE HOT SPOTS OF
NATIVE PLANT DIVERSITY
THOMAS J. STOHLGREN,1,5
DAN BINKLEY,2,3
GENEVA W. CHONG,1
MOHAMMED A. KALKHAN,2
LISA D. SCHELL,2
KELLY A. BULL,2
YUKA OTSUKI,2
GREGORY NEWMAN,2
MICHAEL BASHKIN,3
AND YOWHAN SON4
1U.S. Geological Survey, Midcontinent Ecological Science Center, Fort Collins, Colorado 80525-3400 USA
2Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, Colorado 80523-1499 USA
3Department of Forest Sciences, Colorado State University, Fort Collins, Colorado 80523-1499 USA
4Department of Forest Resources, Korea University, Seoul 136-701, Korea
Abstract. Some theories and experimental studies suggest that areas of low plant species
richness may be invaded more easily than areas of high plant species richness. We
gathered nested-scale vegetation data on plant species richness, foliar cover, and frequency
from 200 1-m2
subplots (20 1000-m2
modified-Whittaker plots) in the Colorado Rockies
(USA), and 160 1-m2
subplots (16 1000-m2
plots) in the Central Grasslands in Colorado,
Wyoming, South Dakota, and Minnesota (USA) to test the generality of this paradigm.
At the 1-m2 scale, the paradigm was supported in four prairie types in the Central
Grasslands, where exotic species richness declined with increasing plant species richness
and cover. At the 1-m2 scale, five forest and meadow vegetation types in the Colorado
Rockies contradicted the paradigm; exotic species richness increased with native-plant
species richness and foliar cover. At the 1000-m2 plot scale (among vegetation types), 83%
of the variance in exotic species richness in the Central Grasslands was explained by the
total percentage of nitrogen in the soil and the cover of native plant species. In the Colorado
Rockies, 69% of the variance in exotic species richness in 1000-m2 plots was explained
by the number of native plant species and the total percentage of soil carbon.
At landscape and biome scales, exotic species primarily invaded areas of high species
richness in the four Central Grasslands sites and in the five Colorado Rockies vegetation
types. For the nine vegetation types in both biomes, exotic species cover was positively
correlated with mean foliar cover, mean soil percentage N, and the total number of exotic
species. These patterns of invasibility depend on spatial scale, biome and vegetation type,
spatial autocorrelation effects, availability of resources, and species-specific responses to
grazing and other disturbances. We conclude that: (1) sites high in herbaceous foliar cover
and soil fertility, and hot spots of plant diversity (and biodiversity), are invasible in many
landscapes; and (2) this pattern may be more closely related to the degree resources are
available in native plant communities, independent of species richness. Exotic plant invasions
in rare habitats and distinctive plant communities pose a significant challenge to
land managers and conservation biologists.
Key words: biodiversity; Central Grasslands (USA) plant diversity patterns; exotic plant invasions;
exotic species richness; native plant diversity; Rocky Mountains (USA) plant diversity patterns; spatial
autocorrelation; species composition overlap; species-specific responses; vegetation sampling, multi-
scale.
INTRODUCTION
It has been hypothesized that exotic species might
more easily invade areas of low species diversity than
areas of high species diversity. This pattern was first
observed by Darwin (1859) and phrased as an hypothesis
by Elton (1958). Areas of high species diversity
Manuscript received 1 April 1997; revised 25 November
1997; accepted 22 December 1997; final version received 23
February 1998.
5 Address correspondence to Thomas J. Stohlgren, U.S.
Geological Survey, Natural Resource Ecology Laboratory,
Colorado State University, Fort Collins, Colorado 80523-
1499, USA. E-mail: toms@nrel.colostate.edu
should use limiting resources more completely, preventing
invasion by a potentially competing species
(MacArthur and Wilson 1967, Tilman 1982, 1997, McNaughton
1983, Pimm 1991). Areas of low diversity
are thought to be highly invasible because, according
to food-web theory (Post and Pimm 1983, Drake 1990)
and coevolution theory (Pimm 1984), they have simple
patterns of inter-specific interactions and use resources
less completely.
Some field observations support the notion that species-rich
areas are less prone to invasion. Fox and Fox
(1986) report an inverse relationship between the percentage
of exotic species in a flora and the number of
native species in western Australia. A global review
26 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
by Rejma´nek (1996a) shows that at continental scales
Elton's hypothesis is generally supported. Species-rich
areas in the tropics contain fewer exotic species than
do relatively species-poor areas in temperate latitudes
in North and South America, and in Africa and Europe.
However, Rejma´nek also showed that the data do not
strongly support the mechanism of the species-richness
effect: there was no significant relationship of latitude
to percentage exotic species between 50 N and 50 S
latitude. Little evidence supports the idea that native
species richness was directly responsible for greater
resistance to invasion (Rejma´nek 1996a). Mack
(1996a) postulated that biotic barriers (e.g., native parasites,
predators, grazers) may be partly responsible for
the unsuccessful naturalization of exotic plants in the
tropics and elsewhere.
Some theorists have readily accepted Elton's hypothesis.
MacArthur and Wilson (1967) and Pimm
(1991) concluded that fewer habitats are likely to be
unexploited where more species are present. Several
mathematical models suggest that areas of high species
diversity should be resistant to invasion (Turelli 1981,
Post and Pimm 1983, Rummel and Roughgarden 1983,
Case 1990). Case (1990:9610) asked ``Why do speciesrich
communities repel invasion?'' His theoretical
models were based on randomly constructed, stable,
Lotka-Voltera competitive communities, and the models
showed that the probability of invader success decreased
with the size and structure of the community.
The key aspect of the mathematical models was that
colonization declined in the face of ``many, strongly
interacting species.'' Law and Morton (1996) also presented
a model supporting Elton's (1958) hypothesis.
Other theorists have argued that highly diverse communities
are intrinsically unstable, with some species
dropping in and out routinely (May 1973). In such a
system, we might envision some native species dropping
out and exotic species replacing them. Alternatively,
Huston and DeAngelis (1994) demonstrate theoretically
that a large number of species can coexist as
a result of biogenic small-scale heterogeneity and interactions
among organisms for spatially and temporally
variable resources. Under this scenario, exotic
plant species might invade and coexist with high numbers
of native plant species as long as light, water, and
nutrients were not limiting.
Some recent experimental evidence strongly supports
Elton's hypothesis and the proposed effect of species
richness. In a seed-addition experiment in mature
oak savanna in Minnesota (USA), Tilman (1997) found
that invasibility correlated negatively with plant species
richness (n 60 1-m2
plots). Another experiment
in recently planted 3 3 m plots in the same vegetation
type had similar results (J. Knops, personal communication).
These plot-level results also are consistent
with several studies that showed species-poor islands
are particularly prone to invasions (Elton 1958, Vitousek
et al. 1996). In light of this convincing evidence
across multiple spatial scales, it is tempting to accept
the general paradigm that exotic species tend to invade
and eventually dominate areas of low diversity.
Broad acceptance of this paradigm is premature and
dangerous. First, field observations and experiments
have produced inconsistent results. Some field experiments
in California (USA) showed that species-rich
areas of annual grasslands were more easily invaded
than species-poor areas (Robinson and Quinn 1988,
Robinson et al. 1995). Timmins and Williams (1991)
found that the number of weeds in New Zealand's forest
and scrub reserves did not correlate with the number
of native species. Rejma´nek et al. (1996a) found no
relationship between species richness and invasibility
in a forest reserve in Uganda. Further, several fairly
species-rich areas in the world (e.g., those with Mediterranean
climates) have been readily invaded; perhaps
invasions are facilitated by enormous and continuous
seed sources via human transportation and land-use
practices (Vitousek et al. 1996, Rejma´nek 1996a). Recent
studies also show that species-rich riparian zones
in some areas in the northwestern United States are
more prone to invasion than upland forests that are
species poor (DeFerrari and Naiman 1994). These observations
raise the possibility that species-rich communities
that receive some level of disturbance (e.g.,
fire, herbivory) may have more resources available, at
least temporarily, for invading species (Robinson et al.
1995).
It is humbling to note that community ecologists do
not yet understand causes and patterns of native species
richness. The comprehensive review by Palmer (1994;
150 citations) counted 120 plausible hypotheses attempting
to explain variation in species-richness patterns.
The inclusion of exotic species may further complicate
this situation. Proof of Elton's paradigm that
exotic species invade low-diversity areas would require
consistent observations of native species richness and
invasibility patterns that cross several spatial scales and
a broad range of vegetation types.
A comprehensive evaluation of the relationship of
native and exotic species patterns has been hampered
by poor vegetation-sampling methods and a lack of
standardized approaches over large areas. A recent test
of commonly used rangeland sampling techniques
(e.g., point-transects, Parker 1951; quadrat methods,
Daubenmire 1959), revealed that small quadrats and
linear techniques failed to capture about half the native
and exotic plant species in tallgrass prairie, shortgrass
steppe, northern mixed prairie, and mixed grasslands
(Stohlgren et al. 1998).
We analyzed field data from two biomes in the central/western
United States, the Rocky Mountains and
the Central Grasslands, to (1) assess patterns of exotic
plant invasions in nine vegetation types (two biomes)
using a standardized, multi-scale vegetation sampling
technique; (2) evaluate the relationship of native species
richness and foliar cover to the invasion of exotic
February 1999 27EXOTIC PLANT INVASIONS
FIG. 1. Map of study sites.
plant species at multiple spatial scales (e.g., subplot,
plot, vegetation-type, and biome scales); and (3) determine
whether invasion patterns are a function of
scale, community­environment relationships (e.g., soil
texture, nitrogen, and carbon), spatial autocorrelation
effects, and/or species-specific responses to grazing
and other disturbances.
STUDY AREAS AND METHODS
Study areas
In the Central Grasslands (USA), we had four study
locations (Fig. 1, Appendix): (1) shortgrass steppe at
the Central Plains Experimental Range (Nunn, Colorado:
annual mean precipitation 354 mm, annual
mean maximum temperature 17.39 C, annual mean
minimum temperature 0.06 C); (2) mixed-grass prairie
at the High Plains Experiment Station (Cheyenne,
Wyoming: annual mean precipitation 389 mm, annual
mean maximum temperature 14.39 C, annual
mean minimum temperature 0.61 C); (3) northern
mixed prairie in Wind Cave National Park (Hot
Springs, South Dakota: annual mean precipitation
608 mm, annual mean maximum temperature 15.67 C,
annual mean minimum temperature 0.78 C); and (4)
tallgrass prairie in Pipestone National Monument
(Pipestone, Minnesota: annual mean precipitation
656 mm, annual mean maximum temperature 12.78 C,
annual mean minimum temperature 0.39 C). Except
for the tallgrass prairie location, which had no grazing,
we randomly established four 20 50 m sample plots
at each location that encompassed various grazing intensities
including lightly grazed, heavily grazed, and
protected habitats (Appendix). At each sampling plot
we established a modified Whittaker multi-scale vegetation
plot as described in Methods, below.
The Colorado Rockies site was a 754-ha portion of
the Beaver Meadows area in Rocky Mountain National
Park, Estes Park, Colorado, USA (annual mean precipitation
361 mm, annual mean maximum temperature
13.78 C, annual mean minimum temperature 1.39 C,
2500­3000 m elevation; Fig. 1). The five vegetation
types included lodgepole pine (Pinus contorta), aspen
(Populus tremuloides), ponderosa pine (Pinus ponderosa),
wet meadow, and dry meadow. Most of the area
is moderately grazed by elk and deer. About 9 ha of
the ponderosa pine community was prescribed-burned
(September 1994) and was excluded in this analysis.
We overlaid a 48 32 unit grid (north to south) over
the aerial photograph and randomly selected (computer
random number generator) four sample points in each
of the five plant communities (Appendix; see Stohlgren
et al. [1997a] for details).
Methods
The modified-Whittaker plot measured 20 50 m
and was placed with the long axis parallel to the environmental
gradient (Stohlgren et al. 1995; Fig. 2).
Nested in each plot were 10 0.5 2 m (1-m2
) subplots
systematically spaced along the inside border, two 2
5 m (10-m2
) subplots in alternate corners, and a 5
20 m (100-m2
) subplot in the plot center. Foliar cover
for each species in the understory and percentage bare
ground were estimated to the nearest percent in the 10
1-m2
subplots. We were able to estimate cover to the
nearest percent by carefully training to recognize a 10
10 cm area (i.e., 1% of the 1-m2
subplot). As each
28 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
FIG. 2. Modified-Whittaker sampling design.
species was recorded, the observer mentally counted
how many 10 10 cm areas were covered by that
species. Species that occupied 1% in a subplot were
recorded as 0.5% cover. Cumulative plant species were
noted in the 10-m2
and 100-m2
subplots, and the 1000m2
plots. Ancillary data recorded for each plot included:
Universal Transverse Mercator location and elevation
from a global-positioning system, slope, and as-
pect.
Each site was sampled as close to the phenological
maximum (peak biomass) as possible. Sites in the Colorado
Rockies (Beaver Meadows, Rocky Mountain National
Park) were sampled between 1 June and 15 July
1995. We sampled the shortgrass steppe from 1 to 3
July 1996; the mixed-grass prairie from 25 to 27 June
1996; the northern mixed prairie from 10 to 13 July
1996; and the tallgrass prairie from 24 to 27 July 1996.
Plant species that could not be identified in the field
were collected and identified at the Colorado State University
herbarium; 156 of the 1704 plant specimens
encountered (9.1% of the total specimens) could not
be identified to species due to inappropriate phenological
stage or missing flower parts.
In the Colorado Rockies, five soil samples were taken
in each modified-Whittaker plot (in the center of subplots
1, 4, 7, and 10, and in the plot center) and pooled
into one plastic bag. For the Central Grassland, five
soil samples were taken at each of the four plot corners
and in the center of the plot, and pooled into one plastic
bag. For each sample, surface litter, if present, was
removed, and the top 15 cm of soil was sampled. Samples
were air-dried for 48 h, sieved with a standard
number 10 (2-mm pore size) sieve, ground in a standard
three-ball grinder, and then oven-dried at 55 C for 24
h. Samples were analyzed for percentage total carbon
and nitrogen using a LECO-1000 CHN analyzer (LECO
Corporation, Saint Joseph, Missouri, USA) (following
the methods of Carter 1993), and for particle size based
on the standard hydrometer method (Gee and Bauder
1986).
Soils in the Colorado Rockies also were analyzed
for available nitrogen using two methods. Ion-exchange
resin bags (Binkley and Hart 1989) were constructed
from nylon stocking material and 14 mL of
cation resins (Sybron IONAC C-251, H form [Sybron
Chemicals, Birmingham, New Jersey, USA]) and 14
mL of anion resin (IONAC ASB-1POH, OH form).
Ten resin bags were placed 20 mm below the mineral
soil surface at 5-m intervals along a single diagonal
transect in each 20 50 m plot. Resin bags were installed
in early July, retrieved in early October of 1996,
and then extracted with 50 mL of 2 mol KCl/L solution.
Concentrations of ammonium-N and nitrate-N were determined
by standard flow-injection colorimetry. Damage
by animals in some plots reduced the number of
resin bags retrieved.
The second index of N availability in the Colorado
Rockies used on-site incubations of intact soil cores
(200-mm-diameter core of 0­200 mm depth mineral
soil) in plastic bags. Ten cores were collected at the
same points along the transects, and then mixed, and
subsampled to provide a single sample for determining
initially extractable ammonium and nitrate (10 g of soil
extracted with 100 mL of 2 mol KCl/L). Another 10
cores were sampled, replaced intact in plastic bags and
allowed to incubate from mid-July to mid-August (30
d) in 1996. After incubation the 10 samples/plot were
mixed and subsampled, and then extracted for determination
of ammonium and nitrate (by flow-injection
colorimetry).
Statistical analysis
All statistical analyses were conducted with SYSTAT
(Wilkinson 1996), and P 0.05 was used to determine
significance in all tests. Analysis of variance
(ANOVA) was used to compare the number and foliar
cover of native and exotic species in 1-m2 subplots for
the four vegetation types in the Central Grasslands and
the five vegetation types in the Colorado Rockies.
Analysis of variance is robust to violations of the assumptions
of homogeneous variance and normality as
long as sample sizes are nearly equal (Zar 1974). Where
a significant effect of vegetation type was detected,
Tukey's test was used to contrast mean values for significant
differences.
Multiple regression was used to determine the relationship
of the number and cover of native plant species
in 1-m2
subplots to the number and cover of exotic
plant species, with separate analyses for the vegetation
types in the two biomes. We tested the significance of
each predictor with t values (i.e., against the null hypothesis
that the slope 0). We repeated the analysis
for the two vegetation types with the greatest species
richness and cover of exotic species in each biome.
We used stepwise forward multiple regressions to
assess the ability of native species richness and cover,
and soil characteristics to predict exotic species richness
in the two biomes. Soil characteristics included
total N, total C, and percentages of sand, silt, and clay
(percentage sand was not added into the model to reduce
multi-collinearity). Available N replaced total N
in the model for the Colorado Rockies to provide better
information on resource availability. The forward lin-
February 1999 29EXOTIC PLANT INVASIONS
TABLE 1. Results of ANOVA of number and foliar cover of native and exotic species (means, with 1 SE in parentheses) in
four vegetation types in the Central Grasslands (USA). Means with different lowercase superscript letters within columns
are significantly different at P 0.05, Tukey's test.
1-m2 subplots
Natives
No. Cover
Exotics
No. Cover
1000-m2 plots
Number
Native Exotic
Total
%N %C
Vegetation type
No. of
natives
No. of
exotics
Vegetation type
Mixed-grass
prairie 10.6a 45.9a 0.08a 0.04a 35.2 3.0a 0.17a 1.68a 56 8
(0.5) (3.3) (0.06) (0.03) (2.6) (1.2) (0.01) (0.12)
Shortgrass
steppe 8.5b 57.6a 0.08a 0.04a 25.8 1.0a 0.04a 0.94a 36 3
(0.4) (2.6) (0.04) (0.02) (1.7) (0.4) (0.01) (0.08)
Northern
mixed prairie
6.5c 31.3b 3.05b 28.6b 32.0 8.0b 0.32b 3.54b 59 14
(0.5) (3.1) (0.14) (3.1) (5.0) (0.0) (0.04) (0.55)
Tallgrass prairie
10.0ab 57.6a 2.2c 20.4b 37.8 8.2b 0.39b 4.24b 62 15
(0.7) (4.6) (0.16) (3.3) (8.4) (2.4) (0.02) (0.29)
Statistics
F 13 13 179 40 2 17 27 23
P 0.001 0.001 0.001 0.001 0.157 0.001 0.001 0.001
R2 0.20 0.20 0.78 0.44 0.34 0.81 0.87 0.85
 n 40 subplots per site.
 n 4 plots per site.
ear regression models included only variables meeting
the P 0.15 criterion. Data distributions that were
strongly skewed were transformed prior to analysis.
Log10 transformations were used on available N and N
mineralization potential.
We assessed the relationship between total foliar
cover (i.e., in the herbaceous layer) and native and
exotic species richness with linear and second-order
polynomial regression models. Strong nonlinear relationships
depicting the ``hump-backed'' or unimodal
curve of plant diversity relative to foliar cover (e.g.,
Grime 1973a, Bond 1983, Huston 1994, Huston and
DeAngelis 1994) would make the interpretation of linear
models problematic.
Two sample t tests were used to compare the foliar
cover and frequency (1-m2
subplots) of two exotic
grasses in (1) grazed (by bison, elk, and deer) and
ungrazed (fenced since 1960) areas in the northern
mixed prairie and (2) two plots last burned in 1985 and
1990, respectively, in the tallgrass prairie. Linear regression
was used to correlate disturbance class to the
number and cover of exotic species at the plot level
(1000 m2
) and vegetation-type level (combined data
from four 1000-m2
plots).
Kriging was used to assess possible spatial-autocorrelation
similarities in the patterns of native and exotic
species richness and foliar cover relative to soil characteristics
for the 1000-m2
plots in the Colorado Rockies.
Kriging uses a local estimator in the vicinity of the
point and the autocorrelation structure of the data to
produce a contour map (Legendre and Fortin 1989). In
this study we used kriging (SYSTAT version 6.0) to
smooth contour maps of native and exotic plant diversity
and cover and soil fertility.
Finally, we used Jaccard's coefficient (J) to measure
the overlap between two complete species lists (Krebs
1989) as follows:
J A/(A B C)
where A the number of species found in both paired
sites, B species in site 1 but not in site 2, and C
species in site 2 but not in site 1. A comparison of
species lists for two sites resulting in a similarity coefficient
of 1.0 would indicate complete overlap (i.e.,
identical species lists), while a value of 0.0 would indicate
no overlap. The mean Jaccard's coefficient for
each vegetation type was calculated from all possible
pairwise comparisons between plots.
RESULTS
Plant species richness and cover at the Central
Grasslands and Colorado Rockies sites
The species richness and foliar cover of native and
exotic plant species varied greatly within and between
the two biomes. In the Central Grasslands, exotic species
richness in 1-m2
subplots was significantly greater
in the northern mixed and tallgrass prairie sites than
in the other prairie types (Table 1). The northern mixed
prairie site also had the lowest native species richness
and cover. The tallgrass prairie site had the highest
native species richness and cover (in 1-m2
subplots)
and relatively high exotic species richness and cover.
Despite long-term cattle grazing at the shortgrass
30 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
TABLE 2. Results of ANOVA of number and foliar cover of native and exotic species in (means, with 1 SE in parentheses)
five vegetation types in Rocky Mountain National Park, Colorado (USA). Means with different lower-case superscript
letters within columns are significantly different at P 0.05, Tukey's test.
1-m2 subplots 1000-m2 plots
Vegetation
type
Natives
No. Cover
Exotics
No. cover
Number
Native Exotic
Total
%N %C
Resinbag
N
Soilcore
N
No. of
natives
No. of
exotics
Vegetation type
Aspen 9.7a 43.0a 1.13a 6.0a 58.2 5.5a 0.10 3.45 0.79 14.95 125 11
(0.9) (5.7) (0.3) (1.3) (4.8) (1.8) (0.01) (0.80) (0.35) (9.45)
Wet
meadow 8.2a 83.4b 1.0a 6.7a 40.2 4.2 0.37 5.61 0.54 5.76 79 8
(0.4) (4.3) (0.2) (1.9) (8.9) (1.0) (0.15) (1.86) (0.29) (2.59)
Dry
meadow 12.1b 44.5a 0.6b 0.6b 45.2 2.0 0.10 1.50 0.85 4.17 72 5
(0.6) (1.8) (0.1) (0.1) (3.2) (0.7) (0.02) (0.52) (0.20) (1.59)
Ponderosa
pine 4.6c 14.7c 0.3b 0.2b 37.0 2.0 0.10 2.95 0.39 1.46 69 5
(0.5) (2.3) (0.1) (0.1) (3.2) (0.7) (0.04) (0.80) (0.10) (0.62)
Lodgepole
pine 2.0c 7.6c 0.2b 0.3b 35.4b 1.4b 0.17 3.83 0.72 0.89 69 3
(0.3) (1.9) (0.1) (0.2) (3.6) (0.2) (0.04) (0.59) (0.33) (0.29)
Statistics
F 48.5 71.9 9.2 9.7 3.2 3.1 2.2 2.3 0.4 1.8
P 0.001 0.001 0.001 0.001 0.040 0.045 0.12 0.11 0.78 0.19
R2 0.50 0.60 0.16 0.17 0.45 0.44 0.37 0.38 0.15 0.34
 n 40 subplots per site.
 n 4 plots per site.
steppe and mixed prairie sites, exotic species richness
and cover were significantly lower than at the other
two sites. For the 1-m2 subplots, the site effect (i.e.,
the percentage variation explained by the ANOVA
model) was highly significant for all variables tested,
and it was greater for exotic species richness and cover
than for native species richness and cover (Table 1).
At the 1000-m2 plot scale, no significant differences
were found in the number of native species at the four
sites, but the northern mixed prairie and the tallgrass
prairie had greater exotic species richness than mixed
grass prairie and shortgrass steppe (Table 1).
In the Colorado Rockies, the dry-meadow type had
the greatest native species richness at the 1-m2
scale
(Table 2). Exotic species richness and cover at the 1-
m2
scale was significantly greater in the wetter vegetation
types (e.g., the aspen and wet-meadow types)
compared to the drier, coniferous-forest types (Table
2). Again, the site effect was highly significant for all
variables tested at the 1-m2
scale, but it was greater for
native species richness and cover than for exotic species
richness and cover. The vegetation types with the
highest and lowest native species richness (the drymeadow
and lodgepole pine types) did not differ significantly
in the species richness and cover of exotic
species at the 1-m2
scale. Total herbaceous foliar cover
was lowest under the dense canopy of lodgepole pine
stands and highest in the open wet meadows (Table 2).
At the 1000-m2
scale, the aspen type surpassed the
dry-meadow type in native species richness and widened
the margin in exotic species richness. The lodgepole
pine type gained native plant species more rapidly
than exotic plant species with increasing spatial scale
compared to the other four vegetation types.
Effects of scale on evaluating exotic
plant species invasions
The nested plot design (Fig. 2) allowed us to evaluate
the effects of plot size on the number (percentage) of
exotic species in an area. In a homogeneous environment,
the percentage of exotic species might be fairly
constant. We found that vegetation types in both biomes
had large variability in the percentage of exotic species
at different spatial scales. The mixed-grass prairie type
in the Central Grassland showed only 1.7% exotic species
at the 1-m2
scale and 12.5% exotic species (8 of
64 plant species) identified in the four 1000-m2
plots
(Table 1). The rate that exotic species were encountered
exceeded the rate native species were added to the species
pool as additional plots were established (i.e., as
scale increased) due to the patchy distribution of the
few exotic species. The opposite was true of the northern
mixed-prairie type, which had a greater spatial mixing
of exotic species throughout the study area resulting
in similar percentages of exotic species at multiple spatial
scales. Some exotic species (e.g., Poa pratensis,
Bromus japonicas) had frequencies of 90% in the
forty 1-m2
subplots.
In the Colorado Rockies, the percentage of exotic
species was more consistent across spatial scales (Table
2). About half the exotic species recorded for a vegetation
type could be found in a 1000-m2
plot. In each
February 1999 31EXOTIC PLANT INVASIONS
of the five vegetation types, the percentage of exotic
species was greater at the 1-m2
scale than at the 1000-
m2
scale. In four out of five vegetation types, however,
the percentage of exotic species was less at the 1000-
m2
scale than at the 4000-m2
scale (four plots combined).
This suggests exotic species distributions vary
by vegetation type and may be patchier at some scales
than at others.
Relationships of species richness and
total foliar cover
At the 1-m2
scale for the four vegetation types in the
Central Grasslands, only two significant hump-backed
relationships were found between native species richness
and total foliar cover, and none were found for
exotic species richness (Fig. 3). For native species richness
in the mixed-grass and tallgrass prairie types,
where the relationships were significant, only 18% and
13% of the variance, respectively, could be explained
by the second-order polynomial model. The few data
points with high foliar cover commonly had high leverage
in the nonlinear models. For all sites combined,
positive linear regressions and upsloping nonlinear regressions
best described the highly variable relationship
between native and exotic species richness and
total foliar cover (Fig. 3).
The ``hump-back'' model (Grime 1973a, Bond 1983)
was more evident for native and exotic species richness
and total foliar cover at the 1-m2 scale for vegetation
types in the Colorado Rockies (Fig. 4). However, there
was considerable variability in the scattergrams. The
non-linear relationships were stronger for forest types
than for meadow types, and stronger for native species
than for exotic species. The total variance explained
by non-linear models was 15% except for the aspen
type or when all types were combined.
At the 1000-m2 scale in the Central Grasslands, linear
and nonlinear regressions of native or exotic species
richness to total foliar cover were not significant (n
16 plots, R2
0.09 in all cases; Fig. 5). At the 1000-
m2
scale in the Colorado Rockies, the nonlinear regression
was weaker than at the 1-m2
scale for native
species richness, and linear and nonlinear regressions
of exotic species richness to total foliar cover were
identical and essentially linear (n 20 plots). Exotic
species richness was positively linearly correlated with
total cover at the 1000-m2
scale in the Colorado Rockies
and for both biomes combined, and weakly correlated
(r2
0.38, P 0.004) in the Central Grasslands.
All six graphs were at least weakly linearly correlated
(P 0.2; Fig. 5).
We concluded that the linear multiple regressions
used in the following sections to compare biome-level
responses are probably more representative than nonlinear
models because: (1) very little variance was explained
by nonlinear models, in particular with respect
to exotic species richness (Figs. 3 and 4); (2) nonlinear
models were heavily influenced by the leverage of a
few high-cover data points; and (3) linear regressions
best described responses for native and exotic species
at larger spatial scales in both biomes for typical ranges
in foliar cover (Fig. 5).
Predicting invasibility with data on native species
richness and cover
In both biomes, native species richness and foliar
cover at the 1-m2
scale was significantly correlated with
species richness and foliar cover of exotic species (i.e.,
P 0.001), but only 11­22% of the total variation was
explained by these predictors (Table 3). For the four
vegetation types in the Central Grasslands (combined
data from the 1-m2
subplots), exotic species richness
was strongly, negatively correlated with both the number
and cover of native plant species. The cover of
exotic species in the Central Grassland sites also significantly
declined with increasing numbers of native
plant species. These results strongly support the paradigm
that exotic species can more easily invade areas
of low species diversity.
The opposite relationship was found for the five vegetation
types in the Colorado Rockies: exotic species
richness was strongly, positively correlated with both
the number and cover of native plant species (Table 3).
The cover of exotic species also was positively correlated
with the number and cover of native species,
although the relationship was greater between the cover
of exotic and native species.
We examined the two vegetation types with the most
exotic species at the 1-m2 scale in both biomes to evaluate
the consistency of regression results above. For
the northern mixed-prairie and tallgrass prairie types,
exotic-species richness strongly declined with increasing
numbers of native plant species (Table 4). However,
the cover of native species was a weaker predictor of
exotic species richness in the tallgrass prairie type, and
a nonsignificant predictor in the northern mixed-prairie
type. In the aspen type in the Colorado Rockies, the
number of exotic species was strongly, positively associated
with the number of native plant species, but
the cover of native species had a slightly negative affect
on exotic species richness. The wet-meadow type
showed opposite results: exotic species richness was
strongly positively associated with the cover of native
plant species, while the number of native species was
statistically unrelated to exotic species richness (Table
4).
Predicting invasibility with data on
native vegetation and soils
Data on soil characteristics were pooled at the plot
scale (1000-m2
) to better predict patterns of invasion
at the landscape scale (i.e., among vegetation types and
study sites). Stepwise multiple regressions showed that
the percentage of total N in the soil (``soil total %N'';
a measure of soil fertility; positive relationship) and
the cover of native species (negative relationship) ex-
32 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
FIG. 3. Relationship (scattergram) of native and exotic species richness to total species cover for 1-m2 subplots in four
prairie vegetation types of the Central Grasslands (USA). Coefficients of determination and significance levels of linear and
nonlinear regressions are shown where they are significant. In the top eight panels each black dot represents data from a 1m2
subplot (for a total of 40 1-m2 subplots per panel); the bottom two panels include all 160 subplots from the four regional
grasslands that comprise the Central Grasslands.
February 1999 33EXOTIC PLANT INVASIONS
plained 83% of the variation in exotic-species richness
among the Central Grassland sites (Table 5). For exotic
species cover 73% of the variation was explained by
soil %clay (positive relationship) and native species
cover (negative relationship). Because the plots were
clumped within study sites, and the sites themselves
were widely scattered in the Central Grasslands, the
regression analyses should not be used to infer similar
relationships at the landscape scale.
In contrast, the plot-scale data of the five vegetation
types within a 754-ha area in the Colorado Rockies
was ideal for predicting invasion patterns on the landscape.
Here, 69% of the variance in exotic species richness
(at the 1000-m2
scale) could be explained by the
positive correlations with the number of native species
and total soil carbon (a measure of organic matter;
Table 5). For exotic species cover, 68% of the variation
among vegetation types was explained by soil total %C,
log10(available N), and the cover of native species. Total
soil %C was strongly correlated with %N (n 35,
r 0.88, F 118.0, P 0.001), and 47.2% of the
variance in log10(available N) could be explained by
log10(%clay) (n 20, t 3.66, P 0.002) and
log10(nitrogen mineralization potential) (n 20, t
2.195, P 0.042). The positive interrelationships between
exotic species richness/cover and soil resources
is clear.
Species-specific responses to grazing and community
responses to disturbance
The 16 1000-m2 plots in the Central Grassland sites
varied in disturbance class (Appendix), but at the 1000m2
scale there were no significant relationships between
disturbance and the number of exotic species (r
0.065, F 0.060, P 0.810) or foliar cover of exotic
species (r 0.123, F 0.213, P 0.651). However,
we documented many site-specific and species-specific
responses to different fire and grazing regimes in the
northern mixed prairie (Table 6).
Smooth brome (Bromus inermis), for example, averaged
9.3% cover in the tallgrass prairie area that was
burned in the spring of 1986, 1987, 1988, and 1990,
and it was found in 10% of the subplots (i.e., one subplot).
In the area that was burned in spring of 1986,
1987, 1988, and 1989 (but not in 1990), the cover of
smooth brome dropped to 3.8%, but it was found in
50% of the subplots. Quackgrass (Agropyron repens)
was far more consistent in the two areas. Cover and
frequency both doubled in the more recently burned
site.
Two sample t tests revealed significant species-specific
responses to grazing at the northern mixed-prairie
site in Wind Cave National Park, South Dakota (Table
6). The foliar cover of B. japonicus was 3 times greater
in areas grazed by bison, elk, and deer compared to
ungrazed sites. Likewise, the frequency (1-m2
subplots)
of B. japonicus (annual, seed-dispersed) was nearly
twice as great in grazed vs. ungrazed sites. However,
long-term cessation of grazing appeared to have little
effect on the widespread distribution of a common, sodforming,
perennial exotic species, Kentucky bluegrass
(Poa pratensis) at the same sites (Table 6).
Landscape-scale patterns of plant species
richness and soils
The sampling design at the Colorado Rockies site
allowed for an analysis of spatial patterns of vegetation
and soils in the 754-ha area (Fig. 6). Spatial interpolation
(kriging maps) identified a high degree of spatial
autocorrelation between the hot spots of native plant
species richness and cover. Likewise, there was high
spatial autocorrelation between native and exotic species
richness, and between the hot spots of exotic species
richness and the areas with the greatest exotic species
cover. Only one sampled hot spot of native species
richness (Fig. 6: x 28 and y 20) had not been
significantly invaded (in terms of exotic species cover).
Patterns of native and exotic species richness and
cover were also spatially autocorrelated with two important
soil characteristics: %clay (a measure of waterholding
capacity) and total %N (a measure of soil fertility).
However, at least one hot spot of native and
exotic species richness and cover (x 5, y 20­27)
matched up poorly with %clay and even more poorly
with %N. This was an area with clearings and stands
of aspen in ponderosa pine forest. The clearings and
aspen allowed light penetration to the forest floor, and,
hence, higher understory species richness than the
closed-canopy sites nearby.
Soil characteristics varied considerably among sites
and vegetation types (Tables 1 and 2). In the Central
Grasslands, mean values for total nitrogen in soils followed
the following progression: tallgrass prairie
northern mixed prairie k mixed-grass prairie shortgrass
steppe (Table 1). Mean values for %clay in soils
followed the same progression: tallgrass prairie (35.3
2.5%) northern mixed prairie (32.3 2.9%) k
mixed-grass prairie (20.9 2.0%) shortgrass steppe
(14.7 1.5%).
For the vegetation types in the Colorado Rockies,
mean values for available nitrogen in soils followed
the following progression: wet meadow k aspen k
lodgepole pine dry meadow ponderosa pine, and
nitrogen mineralization potential followed the progression:
aspen wet meadow lodgepole pine
ponderosa pine k dry meadow (Table 2).
Biome-scale patterns of native and exotic species
richness and cover
Despite differences in subplot-scale to plot-scale patterns
of invasion in the Central Grasslands and Colorado
Rockies, a consistent and alarming picture of exotic
plant invasion emerged at the biome scale (Fig.
7). Areas of high native species richness were invaded
more heavily than areas of low species richness. The
tallgrass and northern mixed-prairie types in the Cen-
34 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
February 1999 35EXOTIC PLANT INVASIONS
FIG. 5. Relationship of native and exotic species richness to total herbaceous cover for 1000-m2 plots in the Central
Grasslands, Colorado Rockies, and combined sites. Coefficients of determination and significance levels of linear and nonlinear
regressions are shown where they are significant. Lowercase letters were used for the Central Grassland types (t tallgrass
prairie, n northern mixed prairie, m mixed-grass prairie, and s shortgrass steppe); capital letters were used for the
Colorado Rockies vegetation types (L lodgepole pine, A aspen, P ponderosa pine, W wet meadow, and D dry
meadow).

FIG. 4. Relationship (scattergram) of native and exotic species richness to total species cover for 1-m2 subplots in five
vegetation types in the Colorado Rockies (USA). Coefficients of determination and significance levels of linear and nonlinear
regressions are shown where they are significant.
tral Grasslands also had the highest soil fertility and
annual mean precipitation. The aspen and wet-meadow
types, the two most invaded vegetation types in the
Colorado Rockies landscape, had the greatest amounts
of available nitrogen in the soils. There were strong
linear relationships between the number of native species
and the foliar cover of native species in the two
biomes (Central Grasslands: r 0.47, F 45.6, P
0.001; Colorado Rockies: r 0.50, F 66.1, P
0.001), and even stronger relationships between the
number of exotic species and the foliar cover of exotic
species in the two biomes (Central Grasslands: r
0.64, F 109.3, P 0.001; Colorado Rockies: r
0.75, F 248.8, P 0.001). Nonlinear regressions
described these relationships less strongly. At landscape-
and biome-scales, areas of high native species
richness in our study area may be heavily invaded by
exotic species (Fig. 7).
To further assess patterns of invasibility within and
across the two biomes, we combined the data for additional
regression analysis. We found that, for the nine
vegetation types in both biomes, mean foliar cover (i.e.,
36 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
TABLE 3. Multiple regression of the number and foliar cover of native plant species on the number or cover of exotic species
in 1-m2 subplots in the Central Grasslands and Colorado Rockies study areas.
Study areas Independent variable
Dependent variable
No. of exotic species
Coefficient t P
Dependent variable
Cover exotic species
Coefficient t P
Central Grasslands Constant 3.040 9.927 0.001 3.589 15.413 0.000
No. native spp. 0.126 3.651 0.001 0.094 2.755 0.007
Cover native spp. 0.012 2.364 0.019 0.005 1.094 0.277
Colorado Rockies Constant 0.015 0.128 0.898 0.589 0.673 0.502
No. native spp. 0.069 4.721 0.001 0.185 1.683 0.094
Cover native spp. 0.005 2.351 0.020 0.052 3.202 0.002
Note: For example, if y Central Grasslands number of exotic species, then y 3.04 0.126 (no. native species)
0.012 (cover native species).
 For Central Grasslands: F 17.4, df 2, 157, P 0.001, R2 0.182. For Colorado Rockies: F 26.0, df 2, 197,
P 0.001, R2 0.209.
 For Central Grasslands: F 11.0, df 2, 157, P 0.001, R2 0.216. For Colorado Rockies: F 12.3, df 2, 197,
P 0.001, R2 0.111.
TABLE 4. Multiple regression of the number and foliar cover of native plant species on the number of exotic species in 1m2
subplots in the two most invasive sites in each study area (biome).
Study area Vegetation types Independent variable
Dependent variable
No. of exotic species
Coefficient t P
Central Grasslands Northern mixed grass prairie Constant 3.556 9.995 0.001
No. native spp. 0.109 1.956 0.058
Cover native spp. 0.007 0.779 0.441
Tallgrass prairie Constant 3.719 10.751 0.001
No. native spp. 0.091 2.244 0.031
Cover native spp. 0.011 1.802 0.080
Colorado Rockies Aspen§ Constant 0.619 1.699 0.098
No. native spp. 0.234 6.341 0.001
Cover native spp. 0.008 1.428 0.162
Wet meadow Constant 0.544 0.819 0.418
No. native spp. 0.004 0.068 0.946
Cover native spp. 0.018 2.929 0.006
 Wind Cave site: df 2, 37, P 0.159, R2 0.10.
 Pipestone site: df 2, 37, P 0.001, R2 0.38.
§ Rocky Mountain National Park: df 2, 37, P 0.001, R2 0.54.
Rocky Mountain National Park: df 2, 37, P 0.019, R2 0.19.
of the herbaceous layer, excluding tree cover) was
strongly positively correlated with the total number of
exotic species, mean cover of exotic species, and mean
soil total %N (Fig. 8). For the nine vegetation types,
mean %N in the soil was also strongly positively correlated
with the total number of exotic species and the
mean cover of exotic species. At this spatial scale, there
was also a strong positive correlation between the total
number and mean cover of exotic species (Fig. 8).
DISCUSSION
First, we discuss our results in terms of comparisons
to other studies (field observations, mathematical models,
and controlled experiments). Then we discuss the
application of these results to address ecological theory.
Finally, we discuss the implications of this research
to land managers and conservation biologists.
Comparisons to other studies
At the landscape and biome scales, species-rich plant
communities in our study area were particularly vulnerable
to invasion by exotic species (Fig. 7), contrary
to the classic paradigm. Detecting weed invasions in
species-rich communities is not unprecedented.
Bridgewater and Backshall (1981) reported a positive
correlation between diversity and exotic plant invasions
in natural communities in the South West Australian
coast. Planty-Tabacchi et al. (1996) reported
higher invasibility in species-rich riparian zones in
southwest France and in the northwestern United States
compared to upland sites: species-rich sites had greater
percentages of exotic species in the flora. The authors
attributed this to intermediate-disturbance regimes and
habitat variability in riparian zones leading to high native
and exotic species richness, while succession and
canopy closure in forests reduced understory diversity
over time. Exotic species invasions of species-rich riparian
zones is also well documented (Fox and Fox
1986, Macdonald et al. 1986, Crawley 1987, Malanson
1993, DeFerrari and Naiman 1994). In a study of herbarium
specimens for a coastal foothills landscape in
California, Knops et al. (1995) found that the number
February 1999 37EXOTIC PLANT INVASIONS
TABLE 5. Stepwise multiple regression of native species richness and cover, and soil characteristics as predictors of exotic
species richness and cover in 1000-m2 plots in the Central Grasslands and Colorado Rockies.
Study area
Dependent
variable
Independent
variable Coefficient F P
Central Grasslands No. exotic spp. Soil total %N 24.10 57.38 0.001
Cover native spp. 0.06 5.74 0.032
Cover exotic spp. Soil % clay 1.22 29.17 0.001
Cover native spp. 0.23 3.41 0.088
Colorado Rockies No. exotic spp.§ No. native spp. 0.147 31.88 0.001
Soil total %C 0.502 5.66 0.030
Cover exotic spp. Soil total %C 1.70 15.89 0.001
Log10 (available N) 3.28 5.73 0.030
Cover native spp. 0.20 26.05 0.001
Note: % clay is a measure of water-holding capacity; %N is a measure of soil fertility.
 n 16 plots, R2 0.83.
 n 16 plots, R2 0.73.
§ n 17 plots, R2 0.69.
n 17 plots, R2 0.68.
TABLE 6. Foliar cover and frequency (mean 1 SE) of two
exotic plant species in 1-m2 subplots in the Central Grasslands
in (A) two burned areas in tallgrass prairie, and in
(B) areas grazed by native ungulates (bison, elk, deer) and
in ungrazed/exclosed areas (fenced since 1960) in northern
mixed-grass prairie.
A) Tallgrass prairie site
Parameter and species
Burned last
in 1990
Burned last
in 1989
Foliar cover (%)
Bromus inermis 9.3 2.4 3.8 1.8
Agropyron repens 0.1 0.1 0.05 0.01
Frequency (%)
Bromus inermis 10.0a 7 0.05b 17
Poa pratensis 20.0 13 10.0 10
B) Northern mixed prairie
Parameter and species Grazed Ungrazed
Foliar cover (%)
Bromus japonicus 8.9a 2.3 2.6b 1.1
Poa pratensis 15.7 2.0 18.6 3.3
Frequency (%)
Bromus japonicus 97.5 2 50.0 12
Poa pratensis 92.5 4 90.0 7
Note: Means with different lowercase superscript letters
within rows are significantly different, two-sample t test.
of exotic species generally increased with the number
of native species in a community type, except in a
recently disturbed site. Here, however, the riparian type
had fewer exotic species than nearby woodland and
grassland types.
Only at the subplot and plot scales, and only in the
Central Grasslands, did our results agree with the classic
paradigm and with the experimental results of Tilman
et al. (1996): species-rich areas appeared more
resistant to invasion than species-poor areas. However,
we found the opposite relationship (at the same 1-m2
spatial scale) in the Colorado Rockies (Tables 2­4),
and both study areas at the biome scale showed increased
invasions with increased species richness (Fig.
7). The inconsistency of the relationship between high
species richness and low invasibility at multiple spatial
scales undermines the general application of Elton's
hypothesis to other sites and spatial scales.
Our results in the Central Grasslands showed highly
varying effects of disturbance on diversity. Long-term
grazing produced variable effects on native and exotic
species richness and cover. The shortgrass steppe and
mixed-grass prairie sites had been moderately to heavily
grazed by cattle for 100 yr (and by bison before
then) and were comparatively low in diversity, yet they
were not invaded (Table 1, Fig. 7). These vegetation
types were also comparatively low in soil fertility (%N
and %C; Table 1, Fig. 8). The northern mixed-prairie
sites that received moderate grazing by native ungulates
were heavily invaded. Ungrazed, but recently
burned sites in the tallgrass prairie were heavily invaded.
These results suggest that factors other than
grazing, such as soil fertility (Table 5, Fig. 8) and fire
(Table 6), may facilitate invasion of exotic species in
some systems. Other studies clearly demonstrate the
role of disturbance in enhancing and maintaining species
diversity in grasslands. McNaughton's (1983)
comprehensive analysis of the Serengeti ecosystem
found that grazing enhances diversity at local and landscape
scales (also see McNaughton 1979). However,
preliminary data from meadows in Rocky Mountain
National Park showed no significant difference in the
plant species richness of grazed and long-ungrazed areas
(Stohlgren et al. 1997b). Whisenant and Uresk
(1990) showed that burning in mixed-grass prairie
communities could reduce Bromus japonicus, but that
fire also negatively affected some native plant species,
while positively influencing other native plant species.
At plot and landscape scales, areas of high soil fertility
(total %N and %C, and available N) and waterholding
capacity (%clay) were particularly invasible
(Tables 1, 2, and 5). For the Colorado Rockies site, the
major hot spot in the center of the kriging map for
native species cover (Fig. 6) mapped neatly over the
maps of total %N and %clay in this landscape. We
38 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
FIG. 6. Kriging maps of the number and foliar cover of native and exotic plant species, percentage clay in the top 15
cm of soil, and total soil N in the 754-ha Colorado Rockies study area.
found strong positive linear relationships between species
richness and foliar cover for native and exotic plant
species in both biomes. The negative relationship found
in the multiple regression between available nitrogen
and the cover of exotic species (Table 5) contrasts with
the other soils analyses that show positive relationships
between exotic species richness (and cover) and soil
fertility (total %N), organic matter (total %C), and
%clay. Because simple regressions were weaker between
available nitrogen and the cover of exotic species
at plot scales, one possible explanation is that areas
high in exotic species cover may use more nitrogen
than areas with similar cover of native species (i.e., the
statistical analysis describes the post-invasion environment
rather than the pre-invasion environment). Although
we did not measure productivity directly, our
results are supported by experiment results by Tilman
et al. (1996) who found that more diverse communities
are more productive, and contradicted by Tilman
(1997) who found higher plant species richness in plots
with low available nitrogen. Fox and Fox (1986) reported
that in New South Wales, Australia, the inci-
February 1999 39EXOTIC PLANT INVASIONS
FIG. 7. Relationship of native species richness to number
of exotic species at vegetation-type and biome scales (based
on combined species lists from 1000-m2 plots).
dence of invasion was higher in sites with higher rainfall.
Our field data (Table 5, Figs. 6, 7, and 8), would
lead us to conclude that fertile soils probably enhance
native and exotic species richness and cover, given typical
disturbance regimes (Hobbs and Huenneke 1992).
Species-specific responses to disturbance cannot be
ignored, as they thwart sweeping generalizations about
the effects of disturbance or grazing related to plant
species richness and diversity. The examples of Bromus
japonicus and Poa pratensis (Table 6) showed that exotic
species cannot be viewed as a group of species
with similar responses to the same level of disturbance.
Moore (1959) and Pierson and Mack (1990) also found
that exotic species invaded grazed and ungrazed sites
in unpredictable ways. As long as species-specific responses
are unpredictable (Hobbs and Huenneke 1992),
generalizations about exotic species richness may be
as illusive and elusive as theories about the causes of
native species richness patterns (Palmer 1994).
Issues of scale also prevent many sweeping generalizations
about patterns of invasibility. Small-scale experiments
and field observations may produce results
that are scale dependent (Tables 1 and 2). The experimental
plots used by J. Knops (personal communication;
see also Tilman et al. [1996]) identified a strong
relationship between species richness and cover (productivity)
and a negative relationship between species
richness and the number of exotic species in small plots
( 10 m2
) for one study area. Our data show that the
negative relationship between species richness and the
number of exotic species can be exactly the opposite
at the same 1-m2
scale in another biome (Tables 1­5),
and even opposite in the same biome at a larger spatial
scale (Fig. 7). Care must be taken not to extrapolate
the results of small-scale experiments and observations
to the landscape, region, and world.
Spatial autocorrelation may be an important and often
overlooked component of exotic plant invasions.
Our field observations and kriging maps in the Colorado
Rockies suggested that sites that were close to hot
spots of exotic plant species richness and cover (Fig.
6) appeared to have more seedlings of exotic species
than similar sites further away from the hot spots. Many
studies have shown exotic plants spreading from foci
(e.g., Mack 1981, Kareiva 1990, Bergelson et al. 1993).
Rejma´nek (1996a) also showed that islands closer to
continents had more exotic species than islands further
away. In fact, he suggested that the distance factor was
likely more important than the native species richness
factor in the invasion of islands.
Application of results to ecological theory
We now can reexamine the general hypothesis of
Elton (1958) that species-rich communities are less easily
invaded than species-poor communities. At subplot
(1-m2) and plot (1000-m2) scales, our results in the
Central Grasslands strongly support Elton's hypothesis.
However, our results from the Colorado Rockies at subplot,
plot, and vegetation-type scales strongly refute
the hypothesis. Our biome-scale results (Fig. 7) combined
with results from riparian zones (Fox and Fox
1986, Macdonald et al. 1986, Crawley 1987, Malanson
1993, DeFerrari and Naiman 1994, Planty-Tabacchi et
al. 1996) suggest that many species-rich areas may be
particularly invasible.
Our results more closely follow the theory of May
(1973) that diverse systems are more likely to vary in
time and space, which may be termed ``unstable,'' and
the models of Huston and DeAngelis (1994), which
showed that many species can coexist in spatially heterogeneous
areas as long as nutrients and light are not
limiting. May's theory suggested that high rates of species
turnover may provide a mechanism for low ``stability''
(defined here as persistence of native species
assemblages and resistance to invasion). Locally rare
species (species with 1% cover) are a major component
of species richness in grassland systems (Stohlgren
et al. 1998), and they are highly spatially variable
(also see McNaughton 1983, Coffin and Lauenroth
1989). Throughout the Central Grasslands and Colorado
Rockies, roughly half the plant species found in
each 1000-m2
plot had 1% foliar cover. Furthermore,
we found a higher percentage of exotic species in vegetation
types with lower overlap in species composition
(Fig. 9). We hypothesize that low species overlap
among plots would make it difficult for species-rich
communities to develop stable community­environ-
40 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
FIG. 8. Relationships of mean foliar cover (herbaceous layer only) and mean soil %N to exotic species richness and
cover by vegetation type (data combined from four 1000-m2 plots per vegetation type). Also shown is the relationship between
exotic species richness and cover. Lowercase letters were used for the Central Grassland types (t tallgrass prairie, n
northern mixed prairie, m mixed-grass prairie, and s shortgrass steppe); capital letters were used for the Colorado
Rockies vegetation types (L lodgepole pine, A aspen, P ponderosa pine, W wet meadow, and D dry meadow).
ment relationships. Even areas with similar species
richness are likely to have different mixes of species
and probably different levels of unused (available) resources
so that ``resource shifts'' (Fox and Fox 1986)
may be commonplace.
The models of Huston and DeAngelis (1994) provided
an additional framework for evaluating invasions
based on resource transport and supply rates, and on
the distinction between local and regional effects. In
contrast to transport-limited competition models that
February 1999 41EXOTIC PLANT INVASIONS
FIG. 9. Relationship of species overlap to percentage of
exotic species at vegetation-type and biome scales (based on
the mean similarity, J, of pairwise comparisons of species
lists from 1000-m2 plots).
FIG. 10. Relationship of number of transient species to
percentage of exotic species at vegetation-type and biome
scales. Transient species are plant species that do not occur
in all sample plots within a vegetation type.
assumed a complete mixing of resources and organisms
(e.g., Tilman 1982), Huston and DeAngelis showed that
if these assumptions were relaxed, the local coexistence
of many potential competitors (i.e., native and exotic
species) was likely. The low species-composition overlap
within vegetation types (Fig. 9) and the variation
in species richness and cover at the 1-m2
scale (Figs.
1 and 2) and at multiple spatial scales (Tables 1 and
2) may create and maintain non-equilibrium conditions
presumed by Huston and DeAngelis (1994: 972).
The key advancement of Huston and DeAngelis
(1994) was the distinction between local and regional
effects. We found that relationships at one scale, such
as the relationship between native and exotic species
richness, can completely reverse at a larger scale (Tables
1­3). Thus, the processes controlling local native
and exotic species richness, frequency, and foliar cover
(Table 5) may be different from the processes controlling
regional plant diversity (Figs. 7 and 8). If diversity
increases with spatial scale, such as covering
more heterogeneous landscapes and environmental gradients
(Fig. 6), then the growing regional pool of potential
competitors (i.e., native and exotic plant species
with various life-history traits) could help maintain local
and regional heterogeneity (Huston 1979, Huston
and Smith 1987).
Intermediate disturbances and grazing, which likely
contribute to the high diversity in many ecosystems
(Caswell 1978, Connell 1978, Petraitis et al. 1989),
may prevent the dominance of one or a few superior
competitors and further contribute to non-equilibrium
conditions, high plant-species turnover (Fig. 9), and
near-continuous resource shifts in time and space.
Thus, it might be difficult to completely monopolize
resources in fertile sites (Case 1990) to maintain stability
(Tilman et al. 1996) and resist invasion (Elton
1958).
Law and Morton (1996) suggested that the key to
Elton's hypothesis is that the richness of the regional
pool of species dictated the complexity that lead to
resistance to invasion. Instead, our field data suggest
that resistance to invasion is heavily compromised by:
(1) the spatial heterogeneity of locally rare species (Fig.
9; see also McNaughton 1993), and (2) the number of
``transient'' plant species in the regional species pool
(Fig. 10). We defined transient species as those plant
species that do not occur in all plots in a vegetation
type. We propose that in non-forested areas the combination
of high species richness, high spatial variability,
high numbers of transient species, and high
turnover rates makes it that much more difficult for a
community to develop the tight interspecies­environment
relationships necessary to monopolize resources.
That is, the number of all possible combinations of
species mixes in any site rises exponentially as new
species are added to the pool (Fig. 10). Despite high
plant species richness and foliar cover, soil resources
may be available to invading plant species (Table 5,
Fig. 8).
How do we reconcile these results with ample evidence
that species-poor islands are unusually susceptible
to invasion (e.g., Elton 1958, Bramwell 1979,
Moulton and Pimm 1986, Loope and Mueller-Dombois
42 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
1989, Vitousek 1990), and planting experiments that
clearly show higher invasibility in species-poor plots
(Tilman et al. 1996, Tilman 1997, J. Knops, personal
communication)? We suggest that Elton's hypothesis is
only appropriate where the species assemblage poorly
monopolizes available resources and that invasibility
may be independent of species richness in many systems.
Invasibility may be far more dependent on resource
availability. The tallgrass prairie, northern
mixed prairie, and wet-meadow sites all had high levels
of light, water, and nutrients. This would explain why
species-poor islands and post-disturbance sites are susceptible
to invasion, and why our forested sites (lodgepole
pine, low light) and low soil-fertility sites (e.g.,
shortgrass steppe and dry meadow, low soil N) are not.
This also would help explain the results of planting
experiments where species richness is manipulated
(Tilman et al. 1996). In Tilman's experiments, the major
effects of low species richness may have been
caused by including atypically low species-richness
treatments (i.e., 2­10 species per 9 m2) based on random
selections of species from a species pool. Even
the lowest diversity grassland plots we studied rarely
had 15 species per 10 m2 (i.e., shortgrass steppe), and
these areas were the least invasible of all those we
investigated. Tilman's (1996:719) results appeared independent
of species richness in plots where species
richness was 5 species per 9 m2. The primary difference
between experimental plots and natural landscapes
is that many natural landscapes are not simple
random selections from a species pool (as in Case 1990,
Law and Morton 1996, Tilman et al. 1996). Rather,
they are nonrandom species assemblages of dominant
and locally rare species with more complex spatial dynamics
and inter-specific relationships that influence
the availability of resources (Lack 1976) and resource
transport rates (Huston and DeAngelis 1994). We suspect
that many low-diversity natural systems (e.g., the
shortgrass steppe, dry meadow, conifer forests; Tables
1, 2, and 5, Figs. 6 and 8) are resistant to invasion
because: (1) they have low levels of total resources
such as soil nutrients, light, or water available to the
herbaceous layer, and (2) the native plant species in
the system monopolize available resources to resist in-
vasion.
Why does there appear to be a ``hump-back'' in the
relationship between cover and diversity for the forested
sites or at 1-m2
scales, but a linear (or simply
scattered) pattern for the non-forested sites or larger
spatial scales (Figs. 3, 4, and 8)? Huston (1992) and
Huston and DeAngelis (1994) speculate that the humpbacked
pattern for the forests may relate to the very
high cover levels leading to substantial reductions in
light availability for understory plants in the forests,
which may lead to reduced diversity. In grasslands,
however, even the ``high'' cover end of the spectrum
may have been too low for the availability of light to
reduce the level of diversity that could be supported
by soil resources (such as water and nutrients). Our
study was not designed to address the hump-backed
model of diversity and productivity (Huston 1992, Huston
and DeAngelis 1994), as we did not measure productivity
directly. At the 1-m2
scale several vegetation
types had maximum values of species richness in the
moderate ranges of foliar cover (Figs. 3 and 4), which
would appear to support the unimodal model. The nonforested
sites we studied rarely had extremely high cover
values, so data may represent the rising limb of the
hump-backed model. Issues of scale were also important.
The significant unimodal relationships at the 1-
m2
scale disappeared or were weaker at the 1000-m2
scale, or weakly linear (Fig. 5 [P 0.2 for all graphs]
and Figs. 3, 4, and 8). Exotic species richness was
positively linearly correlated with total cover at the
1000-m2
scale in the Colorado Rockies and for both
biomes combined, and weakly correlated (r 0.38, P
0.14) in the Central Grasslands (Fig. 5). At landscape
and biome scales, for exotic species richness and cover
we found strong positive linear relationships with total
understory foliar cover and %N in the soil for the nine
vegetation types studied (Fig. 8). Thus, many of the
relationships we explored were strongly scale dependent,
and fits to a particular model at one spatial scale
my be misleading.
Contrary to the findings of Rapson et al. (1997), we
found that the humped-back relationship can be strongly
influenced by plot size (Figs. 3­5; also see Oksanen
1996). The largest quadrat tested by Rapson et al.
(1997) was 2 2 m. Granted, it is difficult to measure
biomass at larger scales, but we advise a multi-scale
approach to such investigations. The 1-m2 quadrat size
endorsed by Grime (1973a, 1997) may not provide a
realistic picture of plant species richness at larger spatial
scales (Fig. 5). Our results suggest that many ecological
relations are scale dependent and vegetationtype
or biome specific, and that many ecological factors
and interactions in addition to herbaceous foliar cover
are of course important in structuring the species diversity
of an ecosystem.
Generalizations and suggested areas of emphasis
We cautiously offer the following generalizations
and suggested areas for future research.
1) Both native and exotic plant species respond favorably
to fertile areas and available resources. Native
and exotic species richness and cover correlate positively
in most vegetation types in our study areas, suggesting
that both groups respond to increased available
resources with increased diversity. Additional studies
are needed to see if areas of low diversity are resource
poor (e.g., shortgrass steppe, the dry-meadow type),
and if invasibility might be limited in more stressful
environments (Cale and Hobbs 1991, Hester and Hobbs
1992, Hobbs and Huenneke 1992). New studies must
be designed to survey and monitor the effects of invasions
across gradients of resource availability, en-
February 1999 43EXOTIC PLANT INVASIONS
vironmental change, and disturbance (Fox and Fox
1986, Swincer 1986).
2) The attribute of ``high species richness'' does not
ensure complete use of all available resources, stability,
or resistance to invasion. Our data show that ``a
few'' strongly interacting species, especially in natural
communities low in resources (e.g., shortgrass steppe,
mixed-grass prairies) may resist invasion, despite longterm
heavy grazing. In contrast to Case (1990), and in
light of Fig. 7, we challenge theorists and field ecologists
to ask the question ``Why are species-rich communities
so invasible?'' Why are species-rich areas of
Cape South Africa, southern Australia (Rejma´nek
1996a), California, and Florida (Vitousek et al. 1996),
riparian zones (DeFerrari and Naiman 1994), and other
hot spots of plant diversity so heavily invaded?
3) Many other factors are likely far more important
indicators of invasion than the attribute of species richness.
Efforts must be made to make experiments (including
species addition and removal experiments) that
investigate species richness as the primary causal agent
more realistic (i.e., more comparable to natural systems).
Models and experiments using random species
mixes may mask the importance of nonrandom species
assemblages and the inter-specific interactions found
in natural systems (Huston 1997). Experiments in protected
environments may not be representative of natural
landscapes that are frequently grazed, burned, and
disturbed. Integrated observational and experimental
studies are needed to link levels of disturbance (Hobbs
and Huenneke 1992), proximity to and abundance of
seeds or reproductive tissues of exotic species (Fig. 6),
interactions with other species (plants and animals;
Newsome and Noble 1986, Mack 1996b), and resource
availability (Huston and DeAngelis 1994, Wedin and
Tilman 1996, Tilman 1997) at meaningful spatial
scales.
4) Species-specific responses to disturbance and
habitat variation deserve far more attention than the
search for sweeping generalizations. Groves (1986),
Smith (1989), Mack (1996a, b), Rejma´nek (1996b) and
others clearly point to the need for species-specific research,
especially for naturalized exotic species, or perhaps
focused on groups with similar life histories (e.g.,
Smith and Huston 1989). We agree with many scientists
(e.g., Moulton and Pimm 1986, D'Antonio 1993, Lodge
1993) that colonist species and their target communities
cannot be studied independently. Thus, integrated studies
of plant demography and ecology are urgently needed
where exotic species invade and dominate.
5) Quantitative inventory and monitoring efforts
should be initiated immediately to assess the status and
trends of invading plants in natural areas across scales.
We must take a leadership role in developing standardized
approaches to inventory and monitor invasive
species and their effects on natural systems (Everitt et
al. 1995, Stohlgren et al. 1997a, b, 1998). Since the
factors influencing exotic species establishment (species
richness; Tables 3, 4, and 5) may be different than
those influencing exotic species cover (and potential
dominance; Table 5, Fig. 8), we suggest that the processes
of initial invasion and naturalization be studied
in separate but integrated ways.
Implications for land managers and
conservation biologists
Parks and biological reserves are important areas to
assess the invasion of exotic plant species within a
region--pristine areas serve as index sites of largescale
environmental change (Vitousek et al. 1996).
While the causes remain unclear, we are concerned that
areas of high native plant species richness and cover
(Tables 1 and 2, Fig. 7) and areas high in soil fertility
(Table 5, Figs. 6 and 8) in natural landscapes may be
highly invasible. Of equal concern is that these sites
that are high in plant diversity are usually hot spots of
biological diversity--rare habitats with distinctive
plant and animal communities. The aspen and wetmeadow
areas in the Colorado Rockies sites, for example,
had higher diversity of vascular plants, birds
(T. Stohlgren, T. Mabee, and J. Woolf, unpublished
data), and butterflies (Simonson 1998) than the other
vegetation types in the area. Tallgrass prairie systems
have high endemism and biodiversity, and even higher
levels of risk from land-use change, fragmentation, and
fire suppression (Leach and Givnish 1996).
We are uncertain as to why many areas identified as
hot spots of endangered species in the United States
(Flather et al. 1994, Dobson et al. 1997) are also hot
spots for invasive plant species (e.g., Hawaii, California,
Florida, tallgrass prairie remnants). Exotic plant
species may accelerate the rapid loss of native species
in tallgrass prairie ecosystems, which have long been
threatened by fragmentation and interruption of landscape-scale
processes (Smith 1981, Chapman 1984,
Leach and Givnish 1996).
Even if the mechanisms are poorly understood, many
scientists agree that exotic plant species will eventually
cause a decline in native plant species and ecosystem
diversity (Bock et al. 1986, Billings 1990, D'Antonio
and Vitousek 1992). Short-term effects of invasive
plants include drastically altering fire cycles
(D'Antonio and Vitousek 1992), nutrient cycling (Vitousek
1990), wildlife forage quality (Medina 1988,
Trammell and Butler 1995), and wildlife grazing patterns
(Trammell and Butler 1995). By reducing palatability
and productivity, infestations of the noxious
weed leafy spurge (Euphorbia esula), for example, can
reduce the carrying capacity of pastures for livestock
by 50­75% in some areas (Lym and Messersmith
1987).
Internal feedbacks and spatial-autocorrelation effects
may further accelerate the spread of exotic plants
(Bergelson et al. 1993). For example, Bromus japonicus,
which we found in the northern mixed prairie site,
increases litter buildup and decreases soil evaporation,
44 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
which favors its germination and establishment (Whisenant
1990). Small populations of exotic species can
spread quickly (Mack 1981) so their early detection is
crucial (Stohlgren et al. 1998).
It may become increasingly important to maintain
natural disturbance regimes. Invasion by exotic species
has either already occurred or is inevitable in many
systems. Maximizing the persistence of native species
may depend on maintaining the natural range of disturbances.
Although information is lacking for most
areas, some studies suggest that reducing or exceeding
the frequency, intensity, spatial patterns, or scale of
disturbances will likely lead to a faster replacement of
native species by exotic species (Abrams et al. 1986,
Fox and Fox 1986, Milchunas et al. 1990, Hobbs and
Huenneke 1992, Planty-Tabacchi et al. 1996).
Other factors could simultaneously influence native
and exotic species richness. Nitrogen deposition from
air pollution may further impoverish native diversity,
facilitating or inhibiting the spread of exotic species
(Hobbs and Huenneke 1992). Even successful control
and eradication efforts do not assure that the effects of
the exotic species are erased. Invasive Acacia species
in Cape ecosystems (South Africa) can have long-term
effects on soil nutrient enhancement and nutrient mineralization
patterns (Stock et al. 1995). Care must be
taken in applying control treatments, especially if the
infested areas are hot spots of native plant diversity.
ACKNOWLEDGMENTS
Cindy Villa, Helen Fields, April Owen, and Elizabeth
Smith assisted with the field work in the Colorado Rockies.
Debbie Casdorph, Randy Griffis, Stephanie Neeley, and Mary
Damm assisted with the field work in the Central Grasslands.
The National Biological Service, now the Biological Resources
Division of the U.S. Geological Survey, provided
funding for the research. We received strong support from
Doyle Fredrick, Michael Ruggiero, Norita Chaney, and Paul
Geissler for the Central Grasslands research. Logistical support
was provided by the staff of Rocky Mountain National
Park, the National Resource Ecology Laboratory at Colorado
State University, and the Midcontinent Ecological Science
Center (Biological Resources Division, U.S. Geological Survey).
Ross Rice, Mark Lindquist, Dick Hart, Todd Suess, and
Dave Kenny also provided logistical support. Michael Huston
and James A. Drake provided helpful suggestions on an earlier
version of the manuscript. To all we are grateful.
LITERATURE CITED
Abrams, M. D., A. K. Knapp, and L. C. Hulbert. 1986. A
ten-year record of aboveground biomass in a Kansas tallgrass
prairie: effects of fire and topographic position.
American Journal of Botany 73:1509­1515.
Bergelson, J., J. A. Newman, and E. M. Floresroux. 1993.
Rates of weed spread in spatially heterogeneous environments.
Ecology 74:999­1011.
Billings, W. D. 1990. Bromus tectorum, a biotic cause of
ecosystem impoverishment in the Great Basin. Pages 301­
322 in G. M. Woodwell, editor. The earth in transition:
patterns and processes of biotic impoverishment. Cambridge
University Press, Cambridge, UK.
Binkley, D., and S. C. Hart. 1989. The components of nitrogen
availability assessments in forest soils. Advances in
Soil Science, Volume 10. Springer-Verlag, New York, New
York, USA.
Bock, C. E., J. H. Bock, K. L. Jepson, and J. C. Ortega. 1986.
Ecological effects of planting African lovegrasses in Arizona.
National Geographic Research 2:456­463.
Bond, W. 1983. On alpha diversity and the richness of the
Cape flora: a study in southern Cape Fynbos. Pages 337­
356 in F. J. Kruger, D. T. Mitchell, and J. U. M. Jarvis,
editors. Mediterranean type ecosystems: the role of nutrients.
Springer-Verlag, New York, New York, USA.
Bramwell, D., editor. 1979. Plants and islands. Academic
Press, London, UK.
Bridgewater, P. B., and D. J. Backshall. 1981. Dynamics of
some Western Australian ligneous formations with special
reference to the invasion of exotic species. Vegetatio 46:
141­148.
Cale, P., and R. J. Hobbs. 1991. Condition of roadside vegetation
in relation to nutrient status. Pages 353­362 in D.
A. Saunders and R. J. Hobbs, editors. Nature conservation
2: the role of corridors. Surrey-Beatty, Chipping Norton,
Australia.
Carter, M. R., editor. 1993. Soil sampling and methods of
analysis. Lewis, Boca Raton, Florida, USA.
Case, T. J. 1990. Invasion resistance arises in strongly interacting
species-rich model competition communities.
Proceedings of the National Acadamy of Science 87:9610­
9614.
Caswell, H. 1978. Predator mediated coexistence: a nonequilibrium
model. American Naturalist 112:127­154.
Chapman, K. A. 1984. An ecological investigation of native
grassland in southern lower Michigan. Thesis. Western
Michigan University, Kalamazoo, Michigan, USA.
Coffin, D. P., and W. K. Lauenroth. 1989. The spatial and
temporal variability in the seed bank of a semiarid grassland.
American Journal of Botany 76:53­58.
Connell, J. H. 1978. Diversity in tropical rain forests and
coral reefs. Science 199:1302­1310.
Crawley, M. J. 1987. What makes a community invasible?
Pages 429­453 in A. J. Gray, M. J. Crawley, and P. J.
Edwards, editors. Colonization, succession, and stability.
Blackwell, Oxford, UK.
D'Antonio, C. M. 1993. Mechanisms controlling invasion
of coastal plant communities by the alien succulent Carpobrotus
edulis. Ecology 74:83­95.
D'Antonio, C. M., and P. M. Vitousek. 1992. Biological
invasions by exotic grasses, the grass/fire cycle, and global
change. Annual Review of Ecology and Systematics 23:
63­87.
Darwin, C. 1859. The origin of species. Reprinted by Penguin
Books, London, UK.
Daubenmire, R. F. 1959. Canopy coverage method of vegetation
analysis. Northwest Science 33:43­64.
DeFerrari, C. M., and R. J. Naiman. 1994. A multi-scale
assessment of the occurrence of exotic plants on the Olympic
Peninsula, Washington. Journal of Vegetation Science
5:247­258.
Dobson, A. P., J. P. Rodriguez, W. M. Roberts, and D. S.
Wilcove. 1997. Geographic distribution of endangered
species in the United States. Science 275:550­553.
Drake, J. A. 1990. The mechanics of community assembly
and succession. Journal of Theoretical Biology 147:213­
233.
Elton, C. S. 1958. The ecology of invasions by animals and
plants. Methuen, London, UK.
Everitt, J. H., G. L. Anderson, D. E. Escobar, M. R. Davis,
N. R. Spencer, and R. J. Andrascik. 1995. Use of remote
sensing for detecting and mapping leafy spurge (Euphorbia
esula). Weed Technology 9:599­609.
Flather, C. H., L. A. Joyce, and C. A. Bloomgarden. 1994.
Species endangerment patterns in the United States. General
Technical Report RM-241. United States Department
of Agriculture Forest Service, Fort Collins, Colorado, USA.
February 1999 45EXOTIC PLANT INVASIONS
Fox, M. D., and B. J. Fox. 1986. The susceptibility of natural
communities to invasion. Pages 57­60 in R. H. Groves and
J. J. Burdon, editors. Ecology of biological invasions. Cambridge
University Press, Cambridge, UK.
Gee, G. W., and J. W. Bauder. 1986. Particle size analysis.
Pages 383­411 in A. Klute, editor. Methods of soil analysis.
Part 1. Physical and mineralogical methods. Second edition.
American Society of Agronomy, Madison, Wisconsin,
USA.
Grime, J. P. 1973a. Competitive exclusion in herbaceous
vegetation. Nature 242:344­347.
. 1973b. Control of species density in herbaceous vegetation.
Journal of Environmental Management 1:151­167.
. 1997. The humped-back model: a response to Oksanen.
Journal of Ecology 85:97­98.
Groves, R. H. 1986. Plant invasions of Australia: an overview.
Pages 137­149 in R. H. Groves and J. J. Burdon,
editors. Ecology of biological invasions. Cambridge University
Press, Cambridge, UK.
Hester, A. J., and R. J. Hobbs. 1992. Influence of fire and
soil nutrients on native and non-native annuals at remnant
vegetation edges in the Western Australian wheatbelt. Journal
of Vegetation Science 3:101­108.
Hobbs, R. J., and L. F. Huenneke. 1992. Disturbance, diversity,
and invasion: implications for conservation. Conservation
Biology 6:324­337.
Huston, M. A. 1979. A general hypothesis of species diversity.
American Naturalist 113:81­101.
. 1992. Individual-based forest succession models and
the theory of plant competition. Pages 408­420 in D. L.
DeAngelis and L. J. Gross, editors. Populations and communities:
an individual-based perspective. Chapman &
Hall, New York, New York, USA.
. 1994. Biological diversity: the coexistence of species
in changing landscapes. Cambridge University Press,
New York, New York, USA.
. 1997. Hidden treatments in ecological experiments:
re-evaluating the ecosystem function of biodiversity. Oecologia
110:449­460.
Huston, M. A., and D. L. DeAngelis. 1994. Competition and
coexistence: the effects of resource transport and supply
rates. American Naturalist 144:954­977.
Huston, M. A., and T. M. Smith. 1987. Plant succession: life
history and competition. American Naturalist 130:168­
198.
Kareiva, P. 1990. Population dynamics in spatially complex
environments: theory and data. Philosophical Transactions
of the Royal Society of London 330:175­190.
Knops, J. M. H., J. R. Griffin, and A. C. Royalty. 1995.
Introduced and native plants of the Hastings Reservation,
central coastal California: a comparison. Biological Conservation
71:115­123.
Krebs, C. J. 1989. Ecological methodology. Harper & Row,
New York, New York, USA.
Lack, D. 1976. Island biology illustrated by the land birds
of Jamaica. Blackwell, Oxford, UK.
Law, R., and R. D. Morton. 1996. Permanence and the assembly
of ecological communities. Ecology 77:762­775.
Leach, M. K., and T. J. Givnish. 1996. Ecological determinants
of species loss in remnant prairies. Science 273:
1555­1558.
Legendre, P., and M-J. Fortin. 1989. Spatial pattern and ecological
analysis. Vegetatio 80:107­138.
Lodge, D. M. 1993. Biological invasions: lessons for ecology.
Trends in Ecology and Evolution 8:133­136.
Loope, L. L., and D. Mueller-Dombois. 1989. Characteristics
of invaded islands, with special reference to Hawaii. Pages
257­280 in J. A. Drake, F. DiCasti, R. H. Groves, F. J.
Kruger, H. A. Mooney, M. Rejma´nek, and M. H. Williams,
editors. Biological invasions: a global perspective. John
Wiley & Sons, New York, New York, USA.
Lym, R. G., and C. G. Messersmith. 1987. Leafy spurge
control with resulting forage production from several herbicide
treatments. Research Progress Report. Western Society
Weed Science, Fargo, North Dakota, USA.
MacArthur, R. H., and E. O. Wilson. 1967. The theory of
island biogeography. Princeton University Press, Princeton,
New Jersey, USA.
Macdonald, I. A. W., F. J. Powrie, and W. R. Siegfried. 1986.
The differential invasion of southern Africa's biomes and
ecosystems by alien plants and animals. Pages 205­225 in
I. A. W. Macdonald, F. J. Kruger, and A. A. Ferrar, editors.
The ecology and management of biological invasions in
southern Africa. Oxford University Press, Oxford, UK.
Mack, R. N. 1981. Invasions of Bromus tectorum L. Into
western North America: an ecological chronicle. Agroecosystems
7:145­165.
. 1996a. Biotic barriers to plant naturalization. Pages
39­46 in V. C. Moran and J. H. Hoffman, editors. Proceedings
of the IX International Symposium on Biological
Control of Weeds. University of Cape Town, Stellenbosch,
South Africa.
. 1996b. Understanding the processes of weed invasions:
the influence of environmental stochasticity. Pages
65­74 in C. H. Stirton, editor. Weeds in a changing world.
Symposium Number 64. British Crop Protection Council
Brighton, UK.
Malanson, G. P. 1993. Riparian landscapes. Cambridge University
Press, Cambridge, UK.
May, R. M. 1973. Stability and complexity in model ecosystems.
Princeton University Press, Princeton, New Jersey,
USA.
McNaughton, S. J. 1979. Grassland­herbivore dynamics.
Pages 46­81 in A. R. E. Sinclair and M. Norton-Griffiths,
editors. Serengeti: dynamics of an ecosystem. University
of Chicago Press, Chicago, Illinois, USA.
. 1983. Serengeti grassland ecology: the role of composite
environmental factors and contingency in community
organization. Ecological Monographs 53:291­320.
. 1993. Biodiversity and function of grazing ecosystems.
Pages 361­383 in E. D. Schulze and H. A. Mooney,
editors. Biodiversity and ecosystem function. Springer-Verlag,
Berlin, Germany.
Medina, A. L. 1988. Diets of scaled quail in southern Arizona.
Journal of Wildlife Management 52:753­757.
Milchunas, D. G., W. K. Lauenroth, P. L. Chapman, and M.
K. Kazempour. 1990. Community attributes along a perturbation
gradient in a shortgrass steppe. Journal of Vegetation
Science 1:375­384.
Moore, R. M. 1959. Some ecological aspects of the weeds
problem in Australia. Pages 447­449 in Proceedings of the
IVth International Congress of Crop Protection, Volume I.
International Congress of Crop Protection, Hamburg, Ger-
many.
Moulton, M. P., and S. L. Pimm. 1986. Species introductions
to Hawaii. Pages 231­249 in H. A. Mooney and J. A. Drake,
editors. Ecology of biological invasions of North America
and Hawaii. Springer-Verlag, New York, New York, USA.
Newsome, A. E., and I. R. Noble. 1986. Ecological and
physiological characters of invading species. Pages 1­17
in R. H. Groves and J. J. Burdon, editors. Ecology of biological
invasions. Cambridge University Press, Cambridge,
UK.
Oksanen, J. 1996. Is the humped relationship between species
richness and biomass an artifact due to plot size? Journal
of Ecology 84:293­295.
Palmer, M. W. 1994. Variation in species richness: toward a
unification of hypotheses. Folia Geobotanica et Phytotaxonomica
29:511­530.
46 THOMAS J. STOHLGREN ET AL. Ecological Monographs
Vol. 69, No. 1
Parker, K. W. 1951. A method for measuring trend in range
condition in National Forest Ranges. USDA Forest Service,
Washington, D.C., USA.
Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989.
The maintenance of species diversity by disturbance. Quarterly
Review of Biology 64:393­418.
Pierson, E. A., and R. N. Mack. 1990. The population biology
of Bromus tectorum in forests: effects of disturbance,
grazing, and litter on seedling establishment and reproduction.
Oecologia 84:526­533.
Pimm, S. L. 1984. The complexity and stability of ecosystems.
Nature 307:321­326.
. 1989. Theories of predicting success and impact of
introduced species. Pages 351­367 in J. A. Drake, F.
DiCasti, R. H. Groves, F. J. Kruger, H. A. Mooney, M.
Rejma´nek, and M. H. Williams, editors. Biological invasions:
a global perspective. John Wiley & Sons, New York,
New York, USA.
. 1991. The balance of nature? University of Chicago
Press, Chicago, Illinois, USA.
Planty-Tabacchi, A., E. Tabacchi, R. J. Naiman, C. DeFerrari,
and H. De´camps. 1996. Invasibility of species-rich communities
in riparian zones. Conservation Biology 10:598­
607.
Post, W. M., and S. L. Pimm. 1983. Community assembly
and food web stability. Mathematical Bioscience 64:169­
192.
Rapson, G. L., K. Thompson, and J. G. Hodgson. 1997. The
humped relationship between species richness and biomass­testing
its sensitivity to sample quadrat size. Journal
of Ecology 85:99­100.
Rejma´nek, M. 1984. Perturbation-dependent coexistence and
species diversity in ecosystems. Pages 220­230 in P. Schuster,
editor. Stochastic phenomena and chaotic behaviour in
complex systems. Springer-Verlag, Berlin, Germany.
. 1996a. Species richness and resistance to invasions.
Pages 153­172 in H. Orians, R. Dirzo, and J. H. Cushman,
editors. Biodiversity and ecosystem processes in tropical
forests. Ecological Studies Volume 122. Springer-Verlag,
Berlin, Germany.
. 1996b. A theory of seed plant invasiveness: the first
sketch. Biological Conservation 78:171­181.
Robinson, G. R., and J. F. Quinn. 1988. Extinction, turnover
and species diversity in an experimentally fragmented California
annual grassland. Oecologia 76:71­82.
Robinson, G. R., J. F. Quinn, and M. L. Stanton. 1995. Invasibility
of experimental habitat islands in a California
winter annual grassland. Ecology 76:786­794.
Rummel, D. J., and J. Roughgarden. 1983. Some differences
between invasion-structured and coevolutionary-structured
competitive communities. Oikos 41:477­486.
Simonson, S. 1998. Rapids assessment of butterfly diversity:
a method for landscape evaluation. Thesis. Colorado State
University, Fort Collins, Colorado, USA.
Smith, C. W. 1989. Non-native plants. Pages 60­69 in C. P.
Stone and D. B. Stone, editors. Conservation biology in
Hawaii. University of Hawaii Press, Honolulu, Hawaii,
USA.
Smith, D. D. 1981. Iowa prairie­ an endangered ecosystem.
Proceedings of the Iowa Academy of Science 88:7­10.
Smith, T. M., and M. A. Huston. 1989. A theory of the spatial
and temporal dynamics of plant communities. Vegetatio 83:
49­69.
Stock, W. D., K. T. Wienand, and A. C. Baker. 1995. Impacts
of invading N2-fixing Acacia species on patterns of nutrient
cycling in two Cape ecosystems: evidence from soil incubation
studies and 15N natural abundance values. Oecologia
101:375­382.
Stohlgren, T. J., K. A. Bull, and Y. Otsuki. 1998. Comparison
of rangeland vegetation sampling techniques in the Central
Grasslands. Journal of Range Management 51:164­172.
Stohlgren, T. J., G. W. Chong, M. A. Kalkhan, and L. D.
Schell. 1997a. Rapid assessment of plant diversity patterns:
a methodology for landscapes. Environmental Monitoring
and Assessment 48:25­43.
Stohlgren, T. J., M. B. Coughenour, G. W. Chong, D. Binkley,
M. Kalkhan, L. D. Schell, D. Buckley, and J. Berry. 1997b.
Landscape analysis of plant diversity. Landscape Ecology
12:155­170.
Stohlgren, T. J., M. B. Falkner, and L. D. Schell. 1995. A
modified-Whittaker nested vegetation sampling method.
Vegetatio 117:113­121.
Swincer, D. E. 1986. Physical characteristics of sites in relation
to invasions. Pages 67­76 in R. H. Groves and J. J.
Burdon, editors. Ecology of biological invasions. Cambridge
University Press, Cambridge, UK.
Tilman, D. 1982. Resource competition and community
structure. Monographs in population biology. Princeton
University Press, Princeton, New Jersey, USA.
. 1997. Community invasibility, recruitment limitation,
and grassland biodiversity. Ecology 78:81­92.
Tilman, D., D. Wedin, and J. Knops. 1996. Productivity and
sustainability influenced by biodiversity in grassland ecosystems.
Nature 379:718­720.
Timmins, S. M., and P. A. Williams. 1991. Weed numbers
in New Zealand's forest and scrub reserves. New Zealand
Journal of Ecology 15:153­162.
Trammell, M. A., and J. L. Butler. 1995. Effects of exotic
plants on native ungulate use of habitat. Journal of Wildlife
Management 59:808­816.
Turelli, M. 1981. Niche overlap and invasion of competitors
in random environments. 1. Models without demographic
stochasticity. Theoretical Population Biology 20:1­56.
Vitousek, P. M. 1990. Biological invasions and ecosystem
processes: towards an integration of population biology and
ecosystem studies. Oikos 57:7­13.
Vitousek, P. M., C. M. D'Antonio, L. L. Loope, and R. Westbrooks.
1996. Biological invasions as global environmental
change. American Scientist 84:468­478.
Wedin, D. A., and D. Tilman. 1996. Influence of nitrogen
loading and species composition on the carbon balance of
grasslands. Science 274:1720­1723.
Whisenant, S. 1990. Postfire population dynamics of Bromus
japonicus. American Midland Naturalist 123:301­308.
Whisenant, S. G., and D. W. Uresk. 1990. Spring burning
Japanese brome in a western wheatgrass community. Journal
of Range Management 43:205­208.
Wilkinson, L. 1996. SYSTAT, version 6.0. SPSS, Chicago,
Illinois, USA.
Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall, Englewood
Cliffs, New Jersey, USA.
APPENDIX
A table giving the geographic location, vegetation, soil, and disturbance level for each 1000-m2 plot in the study is available
in ESA's Electronic Data Archive: Ecological Archives M069-002.