Oecologia (2010) 164:399–409
DOI 10.1007/s00442-010-1667-4
POPULATION ECOLOGY - ORIGINAL PAPER
Naturalization of plant populations: the role of cultivation
and population size and density
Mark S. Minton • Richard N. Mack
Received: 10 August 2009 / Accepted: 17 May 2010 / Published online: 9 June 2010
Ó Springer-Verlag 2010
Abstract Field experimentation is required to assess the
effects of environmental stochasticity on small immigrant
plant populations—a widely understood but largely unexplored aspect of predicting any species’ likelihood of naturalization and potential invasion. Cultivation can mitigate
this stochasticity, although the outcome for a population
under cultivation nevertheless varies enormously from
extinction to persistence. Using factorial experiments, we
investigated the effects of population size, density, and
cultivation (irrigation) on the fate of founder populations
for four alien species with different life history characteristics (Echinochloa frumentacea, Fagopyrum esculentum,
Helianthus annuus, and Trifolium incarnatum) in eastern
Washington, USA. The fate of founder populations was
highly variable within and among the 3 years of experimentation and illustrates the often precarious environment
encountered by plant immigrants. Larger founder populations produced more seeds (P \ 0.001); the role of founder
population size, however, differed among years. Irrigation
resulted in higher percent survival (P \ 0.001) and correspondingly larger net reproductive rate (R0; P \ 0.001).
Communicated by John Silander.
Electronic supplementary material The online version of this
article (doi:10.1007/s00442-010-1667-4) contains supplementary
material, which is available to authorized users.
M. S. Minton R. N. Mack
School of Biological Sciences, Washington State University,
Pullman, WA 99164, USA
Present Address:
M. S. Minton (&)
Smithsonian Environmental Research Center,
Edgewater, MD 21307, USA
e-mail: [email protected]
But the minimum level of irrigation for establishment,
R0 [ 1, differed among years and species. Sowing density
did not affect the likelihood of establishment for any species. Our results underscore the importance of environmental stochasticity in determining the fate of founder
populations and the potential of cultivation and large
population size in countering the long odds against naturalization. Any implementation of often proposed postimmigration field trials to assess the risk of an alien species
becoming naturalized, a requisite step toward invasion, will
need to assess different sizes of founder populations and
the extent and character of cultivation (intentional or
unintentional) that the immigrants might receive.
Keywords Environmental stochasticity Field trials Founder population Invasion Propagule pressure
Introduction
The vast majority of introductions of non-indigenous plant
species fail to produce naturalized populations and even
fewer of these populations give rise to plant invasions
(Rejmanek et al. 2005). Predicting which species will
become naturalized, much less invasive, from amongst a
huge pool of immigrant species has long proved daunting
(Smith et al. 1999; National Research Council 2002); this
dilemma is the product of a litany of unknowns, including
immigrant species’ specific donor range(s), genetic composition and ecology, propagule pressure (i.e., size of
founder populations and number of immigrations, sensu
Lockwood et al. 2005), and the environment they will
encounter in a new range (National Research Council 2002
and references therein). Nevertheless, the incentive to
gauge accurately the range of circumstances under which a
123
400
non-indigenous species could become persistent is driven
by enormous environmental, economic and even human
health consequences (Pimentel et al. 2005; Page et al.
2006).
Such prediction is made especially difficult because
stochastic processes and local circumstances in the new
range can greatly affect any population’s performance
(Crawley 1989). Environmental stochasticity (sensu Lande
1988; Simberloff 1988) is omnipresent and has proven
particularly difficult to characterize fully because it
encompasses a wide range of asynchronously operating
factors of varying influence (Menges 1991; Lande 1993;
Higgins et al. 2000). High environmental stochasticity
potentially increases the probability of decline and possible
extinction, regardless of population size (Lande and Orzack
1988; Menges 1992; Kolar and Lodge 2001).
If the odds against naturalization—a requisite step to
invasion (Kolar and Lodge 2001)—are so long, why do
some immigrant populations persist? Through the use of
theoretical population models, Menges (1992) and Grevstad
(1999b) concluded that the probability of extinction
declines as the size of the founder population and the
number of introduction attempts increase. These theoretical
results have been borne out in practice, whether for biological control (Grevstad 1999a), forestry (Perry 1998),
avian introductions (Veltman et al. 1996), or conservation
(Guerrant 1996).
Cultivation, defined as, ‘‘The bestowing of labour and
care upon a plant, so as to develop and improve its qualities’’ (Oxford English Dictionary, 2nd edn, 1989) (e.g.,
irrigation), appears essential for the persistence of many
immigrant plant populations, as it can buffer the population
from vagaries of the environment until the population
reaches a size beyond which negative stochastic events
may not be as devastating (Menges 1991; Mack 1995,
2000). The high percentage of species that become naturalized coincident with some environmental modification,
including cultivation, provides a strong circumstantial link
between cultivation and naturalization (Lonsdale 1994;
Mack and Erneberg 2002). Yet any direct quantitative
assessment of this link, e.g., by following a founder population and its descendants in the new range, is largely
unexplored (Bannister 1965; Rouget and Richardson
2003).
Post-immigration field cultivation trials could provide
an independent line of evidence about the factors that
enhance or thwart naturalization. If populations of an
introduced species experience repeated local extinction
even under cultivation, does this outcome reliably indicate
an exceedingly low likelihood of the species’ eventual
persistence in the new range, e.g., the fate of frost-sensitive
species at high latitudes (Woodward 1990)? Alternatively,
do replicate populations with a net reproductive rate
123
Oecologia (2010) 164:399–409
(R0) [ 1 in consecutive years of cultivation suggest the
odds for persistence are considerably higher, i.e., by consistently producing propagules that could disperse into the
new range? Such field trials could strengthen our understanding of the role of cultivation in naturalization (Mack
2000, 2005) as well as augment our general understanding
of environmental stochasticity for plants (Simberloff
1988). Furthermore, developing a field screening protocol
that could reliably detect the range of these population
outcomes could fulfill an often-proposed component of
alien plant species assessments: experimental post-entry
evaluation (FAO 1996; Walton 2001; Jefferson et al. 2004).
To address experimentally these general topics we asked
these questions with four non-native species during 3
consecutive years.
1. What is the minimum population size, level of cultivation, or sowing density for establishment, where
establishment is gauged as the R0 C 1?
2. To what degree do species and treatments show
differences in the probability of germination, survival
and fecundity?
3. To what extent do patterns of survival and fecundity
vary from year to year?
4. And in a general sense, to what degree can answers to
these questions predict whether an immigrant population would withstand the environmental stochasticity
in a potential new range?
By sharpening our understanding of the link between
cultivation and plant naturalization, these experiments
could augment the current regulatory approaches (FAO
1996, 2001) governing release of species into new ranges.
Materials and methods
Study species
Four non-native annual species, Echinochloa frumentacea
Link [Echinochloa crus-galli (L.) Beauv. var. frumentacea
(Link) W. Wight], Fagopyrum esculentum Moench,
Helianthus annuus L., and Trifolium incarnatum L., which
differ markedly in habit and ecological amplitude, were
employed in the study. Our research questions needed to be
addressed without introducing new taxa in the region
through our experimentation. Consequently, we chose species that had already been introduced deliberately in eastern
Washington but differ widely in the extent to which they
occur outside cultivation (Hitchcock and Cronquist 1996).
Echinochloa frumentacea, barnyard millet (Cascade Seed
Wholesale, Spokane, Washington), is a close relative of the
invasive plant Echinochloa crus-galli (Holm et al. 1977).
Plants are typically self-fertilized, although wind pollination
Oecologia (2010) 164:399–409
does occur (Maun and Barrett 1986). E. crus-galli and its
close relatives occur widely throughout eastern Washington
(Hitchcock and Cronquist 1996). F. esculentum, buckwheat
(Cascade Seed Wholesale), is insect pollinated and flowers
until killed by frost or desiccation (Wheeler 1950; Wilson
and Myers 1954). Buckwheat has been collected outside
cultivation in eastern Washington [Kahlotus, Washington,
Davis s.n. 18 September 1981], but it is not reported in the
regional flora (Hitchcock and Cronquist 1996). H. annuus,
annual sunflower, is insect pollinated and grows vigorously
in sand to clay soils (Robinson 1978). The domesticated
annual sunflower we sowed (McKay Seed, Moses Lake,
Washington) is persistent in fields, partially because its fruits
often dehisce during harvest; it is now commonly collected
throughout eastern Washington (Hitchcock and Cronquist
1996). The winter annual T. incarnatum, crimson clover
(Brocke and Sons, Kendrick, Idaho), is often employed for
roadside stabilization and forage (Knight 1985). In addition
to being self-compatible, it is insect pollinated and outcrosses frequently. T. incarnatum occasionally escapes in
disturbed sites in the region (Hitchcock and Cronquist
1996).
Experimental setting and design
All experiments were conducted in a repeatedly plowed
field at the USDA Western Regional Plant Introduction
Station Central Ferry Farm, 24.5 km northwest of Pomeroy, Washington, USA. This facility borders the Snake
River in Garfield County (46.67°N, 117.75°W, elevation
198 m) and occurs within the Agropyron spicatum–Poa
secunda habitat type (Daubenmire 1970). The highest air
temperatures usually occur in July or August, which are
typically the driest months (National Oceanic and Atmospheric Administration 2003; Online Resource S1). The
area receives average annual precipitation of 402 mm; the
majority falls from November to March.
Each experiment consisted of one replicate founder
population of each species and treatment (see below for
descriptions of the treatments) assigned randomly in six
blocks (n = 6) introduced on 10–11 June 1998, 15–16
May 1999, and 21–22 May 2000. Seeds sown each year
in the experiments were randomly drawn from the same
commercial seed stocks obtained in 1998; seeds were
stored under low humidity at 20°C until sown. (Percent
seed germination was measured repeatedly through the
course of the experiments; negligible year to year variation occurred.) Seeds of each species were re-mixed
repeatedly to ensure that each drawn sample was
representative.
The above-ground biomass of all populations was harvested yearly, placed in paper bags, and oven dried (60°C;
3 days) to a constant weight. Seeds were separated from
401
extraneous material by mechanical threshing, sieving and
winnowing. If a population produced B200 seeds, seeds
were hand counted; for populations that produced [200
seeds, seed production was estimated, based upon weight
where the width of the yearly 95% confidence intervals for
each species was \5% of the mean.
Founder population size
To assess the effect of population size on the establishment
of a founder population, 18 populations of each species
were introduced within six blocks. Each block consisted of
one replicate population of 50, 250 or 1,000 seeds of each
species sown at a constant density (1,000 seeds m-2)
within the center 0.05 m2 (22.5 9 22.5 cm), 0.25 m2
(50 9 50 cm), or 1.00 m2, respectively, of each 1.90-m2
plot (1.38 9 1.38 m) (4 species 9 3 treatments 9 1 replicate per block). Routine mechanical tilling maintained a
1.40-m plant-free inter-plot border; hand weeding removed
without disrupting the soil surface the few extraneous
seedlings that emerged within the plots. The entire area
was irrigated overhead with impact sprinklers (approximately 2.0 l h-1 m-2; 2–3 times week-1 for 3–4 h) to
maintain the soil at field capacity.
All plots of the same species within a block were harvested on the same day or consecutive days. In all 3 years,
plots of F. esculentum, H. annuus, and T. incarnatum were
harvested in either September or October, depending on the
phenology of seed maturation. Mature inflorescences of
E. frumentacea were harvested in early September 1998 to
prevent their removal by red-winged blackbirds (Agelaius
phoeniceus); however, subsequent flowering and fruit
maturation appeared to be stimulated by the initial harvest,
and a final harvest was conducted from 20 October to 1
November. Maun and Barrett (1986) report that clipping
shoot tips of E. crus-galli during flowering and seed maturation can stimulate lateral buds and additional flowering.
In 1999 and 2000, plots of barnyard millet were harvested
only in October. Care was taken to ensure that the site of
any block was employed only once during the 3 years of
experimentation.
Founder population density
Effect of density of the founder population on establishment was evaluated by sowing six founder populations
(250 seeds each) of each species distributed within six
blocks. Within each block seeds were sown into the center
of 0.05, 0.25, and 1.00 m2 of each plot, resulting in sowing
densities of 5,000, 1,000, and 250 seeds m-2, respectively
(4 species 9 3 treatments 9 1 replicate per block). All
populations were sown, maintained, and harvested by the
same protocol as described above.
123
402
Irrigation
Four treatments were imposed on populations of the four
species to assess the effect of irrigation, an important
component of cultivation in an arid climate. A split-plot
design was employed in each of six blocks. Each block
(approximately 10.6 m 9 10.6 m) contained four plots;
one irrigation treatment [ambient precipitation only (control), 15, 30 and 60 min of irrigation] was randomly
assigned to each plot. Each species was randomly assigned
to one of four 0.25-m2 (50 9 50 cm) sub-plots within each
plot. Plots were 3 m apart, and subplots were 1.3 m apart.
A replicate population of 250 seeds was sown into each
subplot.
In 1998, the irrigation was applied independently overhead 2–3 times week-1 to each irrigation plot, which
contained the four species’ sub-plots, using two opposing
Rain Bird A-17-APC1 shrub spray heads on 0.5-m-tall,
1.25-cm PVC risers that were 3 m apart, delivering
approximately 3.0 l h-1 m-2. In 1999 and 2000, the plots
were irrigated every third day until harvest with drip-irrigation to minimize wind drift. Water delivery was regulated by automatic valves, each connected to an
independent channel in a series of irrigation clocks. Driplines consisted of four 1.10-m-long sections of Aquapore
1.59-cm soaker hose (approximately 1.5 l h-1 m-2); each
hose encircled a species’ subplot. Extraneous seedlings
were removed routinely by hand. Harvests were conducted
on 27 September 1998, 23 September 1999, and 5 October
2000; seeds were dried and cleaned, as described above.
Sowing protocol each year was the same as described
above.
Performance measures and establishment criteria
R0 (the average number of seeds produced per seed sown)
gauged the fate of each immigrant seed population. The
inflection point between increasing and decreasing populations is R0 = 1, where R0 [ 1 signals an increasing
population (Caswell 1989). With different sizes of founder
populations, the resulting different denominators confound
the interpretation of R0; consequently, we employed the
total number of seeds produced per founder population in
assessing the role of the population size on establishment.
Determining a fixed quantitative criterion for establishment
is difficult for highly stochastic processes (Mollison 1986).
Two measures, the maximum R0 C 1 and the more conservative mean R0 C 1 of each treatment combination,
were employed to gauge the potential for establishment.
Differences in R0 among treatments could result from
either differences in germination and survival or differences in reproductive output. Survival was measured as the
proportion of seeds that produced adult plants. Fecundity is
123
Oecologia (2010) 164:399–409
considered here as the mean number of seeds produced per
individual.
Statistical analyses
Survival, fecundity and R0 were analyzed independently
using ANOVA for each experiment. In each experiment
[25% of all founder populations died without any fecundity. As a result, the statistical assumptions of normality
and homogeneity of variance could not be satisfied. Consequently, the data were rank transformed, and the
ANOVA was conducted using the ranks (Zar 1996).
Results from each year were analyzed independently when
an interaction between year and any other variable was
significant. Multiple comparisons of the treatments were
conducted with Fisher’s protected least significant difference, and differences were considered significant at B0.05
(Zar 1996). Proc GLM in SAS, version 9.1 (SAS Institute)
was used for all statistical analyses.
Results
Founder population size
The total number of seeds produced per founder population
was greater for correspondingly larger founder populations,
although significant differences in species’ seed production
occurred among years (Fig. 1; Online Resource S2). Founder populations with 1,000 seeds always produced more
seeds (P \ 0.05) than did the smallest founder populations
among all four species. Differences in seed production
among species are skewed by the greater seed production of
E. frumentacea (P \ 0.05; Fig. 1). Nevertheless, much
variability occurred in seed production for all species, among
years and replicates. All populations of T. incarnatum were
destroyed completely in 1998; in other years many crimson
clover populations increased, and one produced nearly
10,000 seeds. Even within years, some replicate populations
became extinct, while other replicate populations produced
R0 [[ 1 (e.g., in 1998, R0 among replicates E. frumentacea
populations ranged from 0 to[120,000 seeds; Fig. 1a).
The total number of seeds produced is a function of survival and fecundity, consequently these parameters were
analyzed independently (Online Resource S2). The general
pattern of total seed production is largely driven by year to
year variation in fecundity. For example, E. frumentacea
populations, which produced seeds in all years, still displayed large variability in fecundity (i.e., 1998, range 2,310–
20,217 seeds plant-1; 1999, range 405–6,965 seeds plant-1;
2000, range 687–42,991 seeds plant-1). Populations of
F. esculentum, H. annuus and T. incarnatum were totally
destroyed through a combination of low survival and failure
Oecologia (2010) 164:399–409
403
Fig. 2a, b Survival (%) and fecundity (seeds plant-1) for founder
populations of 50 (open bars), 250 (hatched bars) and 1,000 seeds
(filled bars) in 1998. Histograms depict raw means (n = 6), bars
represent the range (minimum, maximum), and letters indicate
significant differences [Fisher’s LSD (P \ 0.05)] in ranked means
for the species 9 founder size population interaction in a survival and
species differences in b fecundity. See Fig. 1 for species names; see
Online Resource S2 for statistical details
Fig. 1 Total number of seeds produced by founder populations of 50
(open bars), 250 (hatched bars) and 1,000 seeds (filled bars) in
a 1998, b 1999 and c 2000. Histograms depict raw means (n = 6),
bars represent the range (minimum, maximum), and letters indicate
significant differences [Fisher’s least significant difference (LSD)
(P \ 0.05)] in ranked means within years for species. See Online
Resource S2 for statistical details. EF Echinochloa frumentacea, FE
Fagopyrum esculentum, HA Helianthus annuus, TI Trifolium
incarnatum
of even surviving plants to reproduce. For example, in 1998
founder populations of E. frumentacea died through low
survival of the founders (Fig. 2a), while populations of
F. esculentum, H. annuus and T. incarnatum went extinct
through a lack of reproduction (Fig. 2b). In contrast, replicate populations of E. frumentacea always produced seeds,
provided at least one parent survived (results not shown).
Founder population density
Effect of sowing density on R0 varied among years, species
and sowing densities for all species (Fig. 3; Online
Resource S3). Populations of E. frumentacea produced a
larger R0 than did populations of other species in all years,
and in 1998 all T. incarnatum populations and most
H. annuus populations went extinct. One replicate founder
population of E. frumentacea did not survive in 1998 and
1999, although the majority of E. frumentacea introductions produced R0 [[ 1 (Fig. 3). In 1998, populations at
the lowest sowing density, 250 seeds m-2, produced a
smaller R0 compared to populations at the other densities
(P \ 0.05; Fig. 3a). During 1999 and 2000 the sowing
density did not affect R0 for any founder population
(P [ 0.05), although differences among species were
highly significant in all 3 years (Fig. 3). Seed production
among populations of T. incarnatum with the same sowing
density were highly variable in 1999 and 2000 (Fig. 3b, c).
The fate of H. annuus was highly dependent upon the year.
All populations produced R0 [ 1 in 2000; in 1998 and
1999, however, all populations produced a R0 \ 1 (Fig. 3),
and many went extinct. Unlike the other species, no population of F. esculentum in the density trials produced a
R0 [ 1 in any year.
123
404
Fig. 3 Net reproductive rate (R0) in a 1998, b 1999, and c 2000 of
founder populations sown at low density 250 seeds m-2 (open bars),
medium density 1,000 seeds m-2 (hatched bars), and high density
5,000 seeds m-2 (filled bars). Histograms depict raw means (n = 6),
bars represent the range (minimum, maximum), and letters refer to
significant differences [Fisher’s LSD (P \ 0.05)] in ranked means
within each year for species. See Fig. 1 for species names; see Online
Resource S3 for statistical details
Irrigation
Two-way interactions were significant (P \ 0.05) between
species, irrigation treatments and year of introduction
(Fig. 4; Online Resource S4). In 1998 and 1999, the
interaction between species 9 irrigation was highly significant (P \ 0.01; Fig. 4a, b). In 1998, irrigation resulted
in a R0 greater than the value in the control treatment for
E. frumentacea and F. esculentum (P \ 0.05); nevertheless, only E. frumentacea populations produced a R0 [ 1
(Fig. 4a). In 1999, the mean R0 was [1 in at least one
treatment level for each species under irrigation, although
the minimum amount of irrigation to produce R0 [ 1
123
Oecologia (2010) 164:399–409
Fig. 4a–c R0 in populations irrigated for 0 min (open bars), 15 min
(hatched bars), 30 min (filled bars), or 60 min (cross-hatched bars)
per irrigation event (see text for details). Histograms depict the raw
means (n = 6), bars represent the range (minimum, maximum), and
letters refer to significant differences between ranked means by
Fisher’s LSD (P \ 0.05) for the species 9 irrigation interaction in
a 1998 and b 1999 and species differences c 2000. See Fig. 1 for
species names; see Online Resource S4 for statistical details
differed among species (Fig. 4b). Larger values for R0
resulted among all species in 2000 under irrigation compared with R0 among the controls. A significant (P \ 0.05)
species effect resulted from the larger R0 produced by
populations of E. frumentacea compared with the R0 for
other species. R0 varied substantially among and within
species and years (Fig. 4). In each year at least one replicate population of each species was totally destroyed, but a
minimum of one population of each species produced
R0 [ 1 at least once in the 3 years. For example, the R0 in
1998 for E. frumentacea populations with 30 min irrigation
ranged from 0 to 118 (Fig. 4).
Survival and fecundity varied (P \ 0.01) due to
interactions between species, irrigation, and year (Online
Oecologia (2010) 164:399–409
405
Fig. 5 Survival (%) in a 1998
and b 1999 and fecundity (seeds
plant-1) in c 1998 and d 1999 of
founder populations irrigated
for 0 min (open bars), 15 min
(hatched bars), 30 min (filled
bars), or 60 min (cross-hatched
bars) per irrigation event (see
text for details). Histograms
depict the raw means (n = 6);
bars represent the range
(minimum, maximum). Letters
indicate significant [Fisher’s
protected LSD (P \ 0.05)]
differences in ranked means.
See Fig. 1 for species names;
see Online Resource S4 for
statistical details
Resource S4). In 1998, irrigation increased survival
among populations of E. frumentacea, F. esculentum and
T. incarnatum (Fig. 5a); the low R0 values (Fig. 4a) were
due to the lack of reproduction among all four species in
at least the control treatment. Fecundity of E. frumentacea and F. esculentum founder populations was greater
with irrigation (P \ 0.05; Fig. 5c). Irrigation affected
survival in 1999 (Fig. 5b), but species differences
accounted for much of the variability in fecundity
(Fig. 5d). Fecundity differences in 2000 were also due to
species differences, and irrigation only affected survival
(data not shown).
Repeated introductions
In all three experiments and across all treatments, E.
frumentacea produced many more seeds than were sown.
In contrast, total seed production across all founder populations of the other three species often did not exceed the
sum total of all seeds sown (Table 1). E. frumentacea
consistently displayed the highest percentage of introductions that resulted in at least one adult plant. In all
three experiments, R0 was \1 for F. esculentum, despite
63–80% of its founder populations producing at least one
individual.
Discussion
Results of these experiments form four central findings.
First, the formidable barriers to a population’s persistence,
as gauged by R0, can be mitigated either by large size of the
founder population or enhanced survival through cultivation. Second, the fates of a species’ small founder populations differ substantially not only due to year to year
environmental variation but also among replicate populations in the same year. Additionally, the relative effect of
population size and cultivation on R0 is dependent upon
environmental stochasticity. Finally, species differ radically in their potential for persistence under the treatments
imposed here (e.g., E. frumentacea is distinctive for its
high R0 among most replicates). This final result is particularly relevant to the development of field trials to detect
alien species with the potential for naturalization (Davis
et al. 2010).
Mitigating barriers to persistence
The low likelihood that an immigrant population will
persist reflects the combined severity of abiotic and biotic
hazards, particularly for small founder populations
(Simberloff 1989; Crawley et al. 1993; Guerrant 1996).
123
406
Table 1 Total seed production
of each species across all
treatments and years for each
experiment
Oecologia (2010) 164:399–409
Species
Total seeds
sown
Total seeds
produced
Replicates producing
C1 plant (%)
Echinochloa frumentacea
23,400
2,598,197
89
Fagopyrum esculentum
23,400
2,097
78
Helianthus annuus
23,400
55,024
61
Trifolium incarnatum
23,400
40,317
57
E. frumentacea
13,500
2,759,978
96
F. esculentum
13,500
535
80
H. annuus
13,500
47,703
56
T. incarnatum
13,500
18,526
67
Founder size experiment (n = 54)
Founder density experiment (n = 54)
Irrigation experiment (n = 54)
E. frumentacea
n Total number of replicate
founder populations per species
18,000
525,476
79
F. esculentum
18,000
3,773
63
H. annuus
18,000
5,290
53
T. incarnatum
18,000
4,257
76
The probability of persistence increases through: (1) an
increase in the number of propagules, their points of
introduction and the number of introduction attempts; (2)
alteration of the environment; or some combination of (1)
and (2) (Guerrant 1996; Mack 2000).
Theoretical models (Lande 1988, 1993) and simulations
(Lande and Orzack 1988; Menges 1991; Grevstad 1999b)
indicate that extinction varies inversely to the size of the
immigrant population. Our multi-year field results with
multiple species strengthen these predictions and are also
consistent with the results of other experimental plant
introductions (Panetta and Randall 1994). The likelihood of
persistence also rises as the number of introduction events
increases (Simberloff 1989; Guerrant 1996; Veltman et al.
1996; Grevstad 1999b). This outcome is clearly illustrated
by both H. annuus and T. incarnatum: both species produced a majority of replicate populations with R0 \ 1;
nevertheless, both produced in toto more seeds than were
sown. Each species’ apparent ability to persist was due to
several replicate populations with high fecundity.
The largest impact of cultivation in these experiments
was to reduce hazards in the most environmentally sensitive stage(s), which is usually germination (Harper 1977;
Menges 1992). R0 was greater for all species under irrigation, primarily the product of greater survival compared
with the control. In 1999, whether a population increased
(R0 [ 1) or decreased (R0 \ 1) turned on the level of
survival it experienced; fecundity was not affected by
irrigation. Nonetheless in 1998 all the species had low seed
production as reflected in values for R0, regardless of the
level of irrigation.
The levels of irrigation provided in the irrigation
experiment were much lower than the amount provided in
123
the founder population size and density experiments. Our
intention was to determine if small additions of water
would facilitate establishment and to evaluate the establishment potential of each species under irrigation. For the
most part, survival and R0 were progressively greater for all
species as water availability increased (Figs. 4, 5; Online
Resource S4).
The unpredictable fate of founders
At least one replicate population of almost every combination of species and treatment (e.g., population size,
sowing density, and irrigation level) went extinct during
the 3 years of experimentation—emphasizing that even a
small distance separating populations of the same species
can translate into major differences in their fates. This local
extinction is likely attributable to the environment’s high
risks for seeds and seedlings (cf. Guerrant 1996). Similar
low probabilities of establishment have been documented
in experimental introductions of Melaleuca quinquenervia
(Myers 1983), Emex australis (Panetta and Randall 1994),
and Brassica napus (Crawley et al. 1993) at sites well
within these species’ ecological amplitude.
Year to year environmental variability had a greater
effect on survival in our experiments than did the treatments—an outcome that is not restricted to our study
species (Crawley et al. 1993; Grevstad 1999a; Tielborger
and Kamon 2000). Consequently, the minimum level of
cultivation required to produce a mean R0 C 1 differed
year to year. For example, in 1998, all F. esculentum, H.
annuus and T. incarnatum introductions went extinct, while
in 1999 and 2000, R0 [ 1 for at least one population under
irrigation for each species. This pattern among years
Oecologia (2010) 164:399–409
indicates that irrigation alone did not compensate for other
facets of annual environmental stochasticity. For example,
most populations of H. annuus and T. incarnatum were
destroyed in 1998 within all three experiments, although in
subsequent years both species produced populations with
R0 [ 1. Knight and Eberhardt (1985) and Menges (1992)
contend that environmental stochasticity should primarily
affect survival and not reproductive output. Our results
indicate that although yearly differences in R0 were greatly
affected by the survival of potential parents, changes in
fecundity (e.g., no reproduction of T. incarnatum and
H. annuus in 1998) were also important.
In some years, the environmental stochasticity for some
factors was so extreme that it greatly affected population
dynamics, despite cultivation. For example, the July–
September monthly mean maximum and minimum temperatures were highest in 1998; these temperatures were
reached in the period with the largest proportion of failed
introductions (Figs. 1, 3, 4; Online Resource S1). An
unusually wet May in 2000 (275 mm) and lower maximum
temperatures (19.3°C) compared to June 1998 (58.7 mm
and 24.0°C) and May 1999 (67.1 mm and 18.5°C), the
month the seeds were sown, likely caused an increase in
germination and survival in 1998 compared to results in
other years (Online Resource S1).
The influence of precipitation is illustrated by the high
survival among founder populations in 2000 compared
with high mortality in June 1998 and May 1999, periods of
unseasonably low precipitation. Nonetheless, aerial application of water in 1998 compared with drip irrigation in
1999 and 2000 may hamper the comparison among these
years’ results. A comparison of populations in 1999 and
2000 indicates similarly high percent survival to that seen
in the founder population size and density experiments.
Yearly differences in survival and reproduction that are
consistent among all populations for 1998 further illustrate
the effects of environmental stochasticity. Percentages of
germination and survival in the irrigation experiment were
similar in 1998 and 1999, even though irrigation systems
differed between these years. Unlike results in other years,
all populations of H. annuus in 1998 failed to reproduce
and subsequently went extinct. Such episodic recruitment
and local extinction are common among many taxa in a
variety of habitats (Hobbs and Mooney 1991; Lichter 2000;
Tielborger and Kamon 2000).
Some replicate populations failed to establish each year,
year to year environmental variation was not, however, the
sole cause. In our experiments, even replicates under high
irrigation or with a large founder population differ in R0 by
2–3 orders of magnitude (e.g., E. frumentacea). Despite the
relationship between population size and R0 [ 1, introductions with all three founder population sizes occasionally produced R0 \ 1 (e.g., H. annuus in 1998). Similar
407
results have been reported for M. quinquenervia in southern Florida, where there was great disparity in the fate of
immigrant seed populations, even under conditions conducive to persistence (Myers 1983). These results collectively underscore the highly varied response to the same
environmental factors that can occur among seemingly
identical populations in close proximity. Clearly, the fate of
founder populations is highly uncertain (Crawley 1989).
Species differences
Experimental outcomes among the four non-native species
varied widely during the 3 years of the study. E. frumentacea displayed the highest potential for persistence, a
likely function of its high seed production and high rate of
populations with at least one reproductive adult, even
without irrigation. Both H. annuus and T. incarnatum
demonstrate the potential to persist under some circumstances, while F. esculentum is unlikely to persist within
this new range in eastern Washington, USA from one
generation to the next, even with irrigation. This gradient
of results agrees with each species’ general frequency of
naturalization (Holm et al. 1977; Knight 1985; Knapp
et al. 2002) and bears close similarity to the species’
frequency of occurrence outside cultivation in the study
region (Hitchcock and Cronquist 1996). Field trials that
incorporate even a few (e.g., three) parameters can
accurately rank differences among species’ fates in a new
range.
Implications for post-entry screening
Related to our pursuit of answers to the four questions
stated above, we have in effect initiated here development
of an experimental protocol by which introduced species
can be screened post-immigration for their likelihood of
naturalization. Consequently, these experiments were
deliberately conducted in a field trial setting, rather than
within natural communities (cf. Burke and Grime 1996;
Von Holle and Simberloff 2005). We knew from the outset
that no three experiments would unequivocally achieve any
of our goals. For instance, much more needs to be explored
about the role that cultivation plays in protecting founder
populations in a new range (Mack 1995, 2000). Nevertheless, the importance, if not primacy of population size,
seems clear (cf. Lockwood et al. 2005; Von Holle and
Simberloff 2005; Maron 2006). The important and often
essential role of initial cultivation (Mack 2000) for naturalization is also supported by these results. Developing a
post-entry experimental screening protocol is an openended goal, although the need for this line of investigation
to complement, rather than replace, other predictive tools,
seems apparent, especially for the ever increasing pool of
123
408
immigrant species for which ecological information is
limited (Weber 2003; Davis et al. 2010).
Post-entry screening of introduced species has been
proposed, recommended and implicitly adopted in phytosanitary requirements at the regional (Jefferson et al. 2004),
national (Pheloung 2001; Walton 2001; National Research
Council 2002) and international levels (FAO 1996). But
such screening has been rarely, if ever implemented, unless
as a retrospective evaluation of plant introduction trials
originally conducted for commercial purposes (Lonsdale
1994). In contrast, field trials are mandatory in the evaluation of genetically engineered organisms prior to their
commercial release (Snow et al. 2005). Detrimental features, including enhanced weediness in the transgenic plant
or its relatives through hybridization, are specifically
evaluated (Crawley et al. 1993; Snow et al. 2005). Based
on the results here, we contend that field trials would also
prove highly informative for genetically non-transformed
plants, where the goal is to build a neutral, transparent, and
reproducible assessment of the fate of species in a new
range (Panetta et al. 2001; National Research Council
2002).
Acknowledgments We thank R. A. Black, R. M. Hannan, C. L.
Kinter, E. L. Minton, R. R. Pattison and R. B. Pratt for their support
and comments throughout the project. R. Colautti provided helpful
comments on an earlier draft of the manuscript. K. Tetrich and the
USDA Plant Introduction Facility in Central Ferry, Washington
provided essential help. Finally, we are grateful to numerous volunteers, who assisted with the implementation and harvesting of these
experiments. This research was supported by grants to M. S. M. from
the Betty Higinbotham Trust and a Natural Resources Conservation
Grant at Washington State University. Experiments reported in this
work comply with the current laws of the United States of America.
References
Bannister MH (1965) Variation in the breeding system of Pinus
radiata. In: Baker HG, Stebbins GL (eds) The genetics of
colonizing species. Academic Press, New York, pp 353–372
Burke MJW, Grime JP (1996) An experimental study of plant
community invasibility. Ecology 77:776–790
Caswell H (1989) Matrix population models. Sinauer, Sunderland
Crawley MJ (1989) Chance and timing in biological invasions. In:
Drake JA, Mooney HA, di Castri F, Groves RH, Kruger FJ,
Rejmanek M, Williamson M (eds) Biological invasions. A global
perspective. Wiley, Chichester, pp 407–423
Crawley MJ, Hails RS, Rees M, Kohn D, Buxton J (1993) Ecology of
transgenic oilseed rape in natural habitats. Nature 363:620–623
Daubenmire R (1970) Steppe vegetation of Washington. Wash Agric
Exp Stn Tech Bull 62
Davis AS, Cousens RD, Hill J, Mack RN, Simberloff D, Raghu S
(2010) Screening bioenergy feedstock crops to mitigate invasion
risk. Front Ecol Environ. doi:10.1890/090030
FAO (1996) International standards for phytosanitary measures,
import regulations: guidelines for risk analysis. Secretariat of the
123
Oecologia (2010) 164:399–409
International Plant Convention of the Food and Agriculture
Organization (FAO) of the United Nations, Rome
FAO (2001) International standards for phytosanitary measures:
guidelines for phytosanitary certificates. Secretariat of the
International Plant Convention of the Food and Agriculture
Organization (FAO) of the United Nations, Rome
Grevstad FS (1999a) Experimental invasions using biological control
introductions: the influence of release size on the chance of
population establishment. Biol Invasions 1:313–323
Grevstad FS (1999b) Factors influencing the chance of population
establishment: implications for release strategies in biocontrol.
Ecol Appl 9:1439–1447
Guerrant EO (1996) Designing populations: demographic, genetic,
and horticultural dimensions. In: Falk DA, Millar CI, Olwell M
(eds) Restoring diversity: strategies for reintroduction of endangered plants. Island Press, Washington, DC, pp 171–207
Harper JL (1977) Population biology of plants. Academic Press,
London
Higgins SI, Pickett STA, Bond WJ (2000) Predicting extinction risks
for plants: environmental stochasticity can save declining
populations. Trends Ecol Evol 15:516–520
Hitchcock CL, Cronquist A (1996) Flora of the Pacific Northwest.
University of Washington Press, Seattle
Hobbs RJ, Mooney HA (1991) Effects of rainfall variability and
gopher disturbance on serpentine annual grassland dynamics.
Ecology 72:59–68
Holm LG, Plucknett DL, Pancho JV, Herberger JP (1977) The
world’s worst weeds. University of Hawaii, Honolulu
Jefferson L, Havens K, Ault J (2004) Implementing invasive
screening procedures: the Chicago Botanic Garden model. Weed
Technol 18:1434–1440
Knapp AK, Fay PA, Blair JM et al (2002) Rainfall variability, carbon
cycling, and plant species diversity in a mesic grassland. Science
298:2202–2205
Knight WE (1985) Crimson clover. In: Taylor NL (ed) Clover science
and technology. American Society of Agronomy, Madison, pp
491–502
Knight RR, Eberhardt LL (1985) Population dynamics of Yellowstone grizzly bears. Ecology 66:323–334
Kolar CS, Lodge DM (2001) Progress in invasion biology: predicting
invaders. Trends Ecol Evol 16:199–204
Lande R (1988) Genetics and demography in biological conservation.
Science 241:1455–1460
Lande R (1993) Risks of population extinction from demographic and
environmental stochasticity and random catastrophes. Am Nat
142:911–927
Lande R, Orzack SH (1988) Extinction dynamics of age-structured
populations in a fluctuating environment. Proc Natl Acad Sci
USA 85:7418–7421
Lichter J (2000) Colonization constraints during primary succession
on coastal Lake Michigan sand dunes. J Ecol 88:825–839
Lockwood JL, Cassey P, Blackburn T (2005) The role of propagule
pressure in explaining species invasions. Trends Ecol Evol
20:223–228
Lonsdale WM (1994) Inviting trouble: introduced pasture species in
northern Australia. Aust J Ecol 19:345–354
Mack RN (1995) Understanding the processes of weed invasions:
the influence of environmental stochasticity. In: Stirton CH
(ed) Weeds in a changing world. British Crop Protection
Council, symposium proceedings no. 64. Brighton, UK, pp
65–74
Mack RN (2000) Cultivation fosters plant naturalization by reducing
environmental stochasticity. Biol Invasions 2:111–122
Mack RN (2005) Predicting the identity of plant invaders: future
contributions from horticulture. HortScience 40:1168–1174
Oecologia (2010) 164:399–409
Mack RN, Erneberg M (2002) The United States naturalized flora:
largely the product of deliberate introductions. Ann Mo Bot Gard
89:176–189
Maron JL (2006) The relative importance of latitude matching and
propagule pressure in the colonization success of an invasive
forb. Ecography 29:819–826
Maun MA, Barrett SCH (1986) The biology of Canadian weeds. 77.
Echinochloa crus-galli (L.) Beauv. Can J Plant Sci 66:739–759
Menges ES (1991) The application of minimum viable population
theory to plants. In: Falk DA, Holsinger KE (eds) Genetics and
conservation of rare plants. Oxford University Press, New York,
pp 45–61
Menges ES (1992) Stochastic modeling of extinction in plant
populations. In: Fiedler PL, Jain SK (eds) Conservation biology:
the theory and practice of nature conservation preservation and
management. Chapman and Hall, New York, pp 255–275
Mollison D (1986) Modelling biological invasions: chance, explanation, prediction. Philos Trans R Soc Lond B Biol Sci 314:675–
693
Myers RL (1983) Site susceptibility to invasion by the exotic tree
Melaleuca quinquenervia in southern Florida. J Appl Ecol
20:645–658
National Oceanic and Atmospheric Administration (2003) Annual
climatological summary. U.S. Department of Commerce,
National Climatic Data Center, Asheville
National Research Council (2002) Predicting invasions of nonindigenous plants and plant pests. National Academy Press, Washington, DC
Page NA, Wall RE, Darbyshire SJ, Mulligan GA (2006) The biology
of invasive alien plants in Canada. 4. Heracleum mantegazzianum Sommier & Levier. Can J Plant Sci 86:569–589
Panetta FD, Randall RP (1994) An assessment of the colonizing
ability of Emex australis. Aust J Ecol 19:76–82
Panetta FD, Mackey AP, Virtue JG, Groves RH (2001) Weed risk
assessment: core issues and future directions. In: Groves RH,
Panetta FD, Virtue JG (eds) Weed risk assessment. CSIRO,
Collingwood, pp 231–240
Perry DA (1998) The scientific basis of forestry. Annu Rev Ecol Syst
29:435–466
Pheloung PC (2001) Weed risk assessment for plant introductions to
Australia. In: Groves RH, Panetta FD, Virtue JG (eds) Weed risk
assessment. CSIRO, Collingwood, pp 231–240
Pimentel D, Zuniga R, Monison D (2005) Update on the environmental and economic costs associated with alien invasive species
in the United States. Ecol Econ 52(3):273–288
Rejmanek M, Richardson DM, Higgins SI, Pitcairn MJ, Grotkopp E
(2005) Ecology of invasive plants: state of the art. In: Mooney
409
HA, Mack RN, McNeely JA, Neville L, Schei PJ, Waage JK
(eds) Invasive alien species: a new synthesis. Island Press,
Washington, DC, pp 104–161
Robinson RG (1978) Production and culture. Sunflower science and
technology. American Society of Agronomy. Crop Science
Society of America, Soil Science Society of America, Madison
Rouget M, Richardson DM (2003) Inferring process from pattern in
plant invasions: a semimechanistic model incorporating propagule pressure and environmental factors. Am Nat 162:713–724
Simberloff D (1988) The contribution of population and community
biology to conservation science. Annu Rev Ecol Syst 19:473–
511
Simberloff D (1989) Which insect introductions succeed and which
fail? In: Drake JA, Mooney HA, di Castri F, Groves RH, Kruger
FJ, Rejmanek M, Williamson M (eds) Biological invasions: a
global perspective. Wiley, Chichester, pp 61–75
Smith CS, Lonsdale WM, Fortune J (1999) When to ignore advice:
invasion prediction and decision theory. Biol Invasions 1:89–96
Snow AA, Andow DA, Gepts P et al (2005) Genetically engineered
organisms and the environment: current status and recommendations. Ecol Appl 15:377–404
Tielborger K, Kamon R (2000) Temporal environmental variation tips
the balance between facilitation and interference in desert plants.
Ecology 81:1544–1553
Veltman CJ, Nee S, Crawley MJ (1996) Correlates of introduction
success in exotic New Zealand birds. Am Nat 147:542–557
Von Holle B, Simberloff D (2005) Ecological resistance to biological
invasion overwhelmed by propagule pressure. Ecology 86:3212–
3218
Walton CS (2001) Implementation of a permitted list approach to
plant introductions to Australia. In: Groves RH, Panetta FD,
Virtue JG (eds) Weed risk assessment. CSIRO, Collingwood, pp
93–99
Weber E (2003) Invasive plant species of the world. A reference
guide to environmental weeds. CABI, Wallingford
Wheeler WA (1950) Forage and pasture crops. A handbook of
information about the grasses and legumes grown for forage in
the United States. Van Nostrand, New York
Wilson HK, Myers WM (1954) Field crop production: agronomic
principles and practices. Lippincott, Chicago
Woodward FI (1990) The impact of low-temperatures in controlling
the geographical distribution of plants. Philos Trans R Soc Lond
B Biol Sci 326:585–593
Zar JH (1996) Biostatistical analysis, 2nd edn. Prentice Hall, Upper
Saddle River
123