Oecologia (2009) 161:57–66 DOI 10.1007/s00442-009-1354-5
POPULATION ECOLOGY - ORIGINAL PAPER
Herbivory and population dynamics of invasive and native Lespedeza Michele R. Schutzenhofer Æ Thomas J. Valone Æ Tiffany M. Knight
Received: 27 December 2007 / Accepted: 8 April 2009 / Published online: 15 May 2009 Ó Springer-Verlag 2009
Abstract Some exotic plants are able to invade habitats and attain higher fitness than native species, even when the native species are closely related. One explanation for successful plant invasion is that exotic invasive plant species receive less herbivory or other enemy damage than native species, and this allows them to achieve rapid population growth. Despite many studies comparing herbivory and fitness of native and invasive congeners, none have quantified population growth rates. Here, we examined the contribution of herbivory to the population dynamics of the invasive species, Lespedeza cuneata, and its native congener, L. virginica, using an herbivory reduction experiment. We found that invasive L. cuneata experienced less herbivory than L. virginica. Further, in ambient conditions, the population growth rate of L. cuneata (k = 20.4) was dramatically larger than L. virginica (k = 1.7). Reducing herbivory significantly increased
fitness of only the largest L. virginica plants, and this resulted in a small but significant increase in its population growth rate. Elasticity analysis showed that the growth rate of these species is most sensitive to changes in the seed production of small plants, a vital rate that is relatively unaffected by herbivory. In all, these species show dramatic differences in their population growth rates, and only 2% of that difference can be explained by their differences in herbivory incidence. Our results demonstrate that to understand the importance of consumers in explaining the relative success of invasive and native species, studies must determine how consumer effects on fitness components translate into population-level consequences. Keywords Congeneric comparison Demography Enemy release hypothesis Fitness Matrix model
Introduction Communicated by John Silander.
Electronic supplementary material The online version of this article (doi:10.1007/s00442-009-1354-5) contains supplementary material, which is available to authorized users. M. R. Schutzenhofer T. J. Valone Department of Biology, Saint Louis University, 3507 Laclede Avenue, St. Louis, MO 63103, USA T. M. Knight Department of Biology, Washington University in St. Louis, 1 Brookings Drive, Campus Box 1137, St. Louis, MO 63130, USA Present Address: M. R. Schutzenhofer (&) McKendree University, 701 College Road, Lebanon, IL 62254, USA e-mail:
[email protected]
An important problem in plant ecology is understanding the role of consumers, such as herbivores, in determining patterns of abundance and distribution. A large number of individual-level studies have shown that herbivores can have strong deleterious effects on aspects of plant fitness, such as survival, growth, and reproduction (Hendrix 1988; Crawley 1989; Strauss 1991). However, individual-level effects do not always translate into population-level effects (Leimu and Lehtila 2006). Therefore, it is important to build on the much smaller but growing literature that addresses how individual-level effects of herbivores contribute to population growth rates, using demographic matrix models (Horvitz and Schemske 2002; Leimu and Lehtila 2006; Maron and Crone 2006). Herbivores can have dramatic effects on the population dynamics of rare
123
58
and invasive plant species (Maron and Crone 2006), and thus demographic studies on herbivory are applicable to conservation and management. Consumers, or lack thereof, are thought to play an important role in plant invasions. The enemy release hypothesis, which proposes that introduced species are successful because they left their co-evolved natural enemies behind, is one of the most cited explanations for the success of invasive species (Crawley 1997; Maron and Vila 2001; Keane and Crawley 2002). Despite the assumption that herbivory affects population success, few studies have examined the degree to which lack of herbivory or seed predation contributes to the high population growth rates of invasive plant species (Shea and Kelly 1998; Shea et al. 2005; but see DeWalt 2006; Jongejans et al. 2006). Invasive species are often observed to receive less herbivory than native plant species in the invasive range (reviewed by Liu and Stiling 2006). Studies that compare native and invasive plant species typically measure individual-levels effects, such as the incidence of damage and/ or plant fitness (Agrawal and Kotanen 2003; Colautti et al. 2004; Agrawal et al. 2005; Vila et al. 2005). These comparative studies often use related plant species that occur in the same habitat, which minimizes differences due to evolutionary history and environmental variation, so that differences in herbivory incidence or plant fitness can be attributed to their native or invasive status (Agrawal and Kotanen 2003; Carpenter and Cappuccino 2005). To date, it is not known whether the observed differences in herbivory between native and invasive plant species cause differences in their population growth rates. In our study, we experimentally reduced herbivory and compared the demographic vital rates (e.g., stagespecific rates of survival, growth, and fecundity) as well as the resulting population growth rates of two cooccurring species of Lespedeza: L. cuneata, an invasive exotic perennial that can dominate old-fields and prairies (Price and Weltzin 2003), and L. virginica, a native congener with similar life-history and habitat requirements (Clewell 1966; M.R.S., personal observation). To fully evaluate the effect of herbivory on the population growth rates of these species, we: (1) measured differences in their incidence of herbivory, (2) experimentally reduced herbivory, (3) quantified demographic vital rates, and (4) projected their population growth rates using matrix models to determine the contribution of herbivory to differences in the population growth rates of these species. We expected ambient herbivory to be lower on the invasive relative to its native congener and for herbivory to negatively affect fitness. In ambient conditions, we also expected the invasive species to have a higher population growth rate than the native. If differences in herbivory incidence completely explained the difference
123
Oecologia (2009) 161:57–66
in the population growth rates of these species, then we expected these species would have similar population growth rates when herbivory was experimentally reduced. Alternatively, if factors other than herbivory caused differences in population growth between native and invasive species, then the differences in population growth would remain similar when herbivory was reduced.
Materials and methods Study system Lespedeza cuneata (Fabaceae) is an Asian species that was introduced to the United States in 1896, and has since spread across the entire eastern portion of the United States through deliberate plantings for erosion control and forage (Donnelly 1955). It threatens grasslands and rangelands (Price and Weltzin 2003) and is listed by several U.S. states (Kansas, Colorado) as a noxious invader (USDA 2008). Lespedeza virginica is a widespread perennial legume native to the eastern United States, northern Mexico, and Eastern Canada. It can be found in a variety of habitats, ranging from open forests, prairies, and glades, to roadsides (Clewell 1966). Its habitat requirements are comparable to those of L. cuneata, and these species co-occur in the same habitats at our study site (M.R.S., personal observation). Both L. cuneata and L. virginica exhibit a heteromorphic flowering system, producing both chasmogamous (CH) and cleistogamous (CL) flowers. Both species flower in July–September and set seed in September–October at our study site. CH flowers are primarily insect pollinated (Ansley 1960), whereas CL flowers do not open and are therefore self-pollinated. Previous research suggests that CH and CL flowers can differ significantly in fitness (Berg and Redbo-Torstensson 1999). Seeds of both L. cuneata and L. virginica are likely to remain viable in the soil and form soil seed banks, but research on their seed banking ability is sparse. For L. cuneata, authors have noted that seed banks occur, but these studies have not provided details on the longevity of these seeds (Wheeler and Hill 1957; Stevens 2002). Experimental design We conducted an herbivory reduction experiment on native and invasive Lespedeza, and measured demographic vital rates. In May 2005, we located and tagged naturally occurring L. cuneata and L. virginica at Washington University’s Tyson Research Center (24 km west of St. Louis, Missouri, USA), which consists of 809 ha of oak-hickory forest and open fields. The chosen plants were randomly
Oecologia (2009) 161:57–66
located within the property in no particularly distinctive populations. For each species, we categorized individuals into four stage classes: small (1 branch), medium (2–5 branches), large (6–10 branches), and extra-large ([10 branches). These stage classes represent relative discrete differences in fertility (approximately double the seed production between each stage class), and for the smaller stages, survivorship. We randomly chose 30 plants in each stage class for inclusion in our study (2 species 9 4 stage classes 9 2 treatments 9 15 plants/stage class/treatment/ species = 240 plants). In 2005, we could not find any L. virginica plants in the extra-large stage class. Plants of each stage class were randomly assigned to one of two treatments: ambient herbivory (control) and herbivory reduction. Herbivory reduction was achieved through use of a chemical insecticidal spray (Ortho IsotoxTM), at a diluted concentration of 38 ml/l, in midMay and in mid-July in an amount sufficient to coat the upper and lower side of all leaves. Ambient herbivory plants were sprayed with water. A minimum of 10 cm separated all study plants to minimize possible spillover effects of treatments and were distributed throughout approximately 162 ha within the forest edge habitats of Tyson. These treatments manipulated aboveground damage due to leaf chewers, miners and piercers. While we observed very little other damage (e.g., pathogens, vertebrate herbivory) on either Lespedeza species, we cannot rule out the presence or effects of cryptic (e.g., belowground) enemies (Reinhart et al. 2003; Callaway et al. 2004). Once seeds were mature, we surveyed tagged individuals for herbivore damage and measured reproduction. Herbivore damage was quantified by visually assessing percent leaf damage (chewing and mining) on a random subset of 60–90 leaves for each plant using a standard scale which consisted of 5% intervals. Although visual estimates of herbivory are less precise than other methods, the same researcher (M.R.S.) estimated all plant damage to eliminate any systematic variation among species and treatments. We counted the number of seeds present on each plant and categorized the seeds as either CL or CH based on morphological differences in the size and shape of fruits (McKee and Hyland 1941). Herbivore damage data were arcsine transformed to achieve normality. Both herbivore damage and CL and CH seed production were compared across species (2 levels) and treatments (2 levels) with twoway ANOVA using StatisticaTM (ver. 5.1) (Statsoft 1999). In May 2006, we returned to all marked individuals to measure survival and subsequent stage class transitions from 2005 to 2006. For each species and stage class, we used Pearson’s chi-square analysis to test whether stage transitions (a discrete response variable) differed across herbivory treatments.
59
Measuring germination rates To project population growth rate, we needed data on germination rates for each species and seed type (CH, CL). In November 2005, we collected CL and CH seeds from naturally occurring L. cuneata and L. virginica plants at our study site. We introduced 25 seeds of each type (CH, CL) for each species into separate 5-cm-diameter plastic pots, which contained local soil sieved to remove any other Lespedeza seeds; these pots contained holes in the bottom for drainage and were buried into the soil in an old field location at our study site to mimic natural conditions. We replicated each seed type by species combination 25 times, for a total of 100 pots and 2,500 seeds. The number of germinates were recorded the following spring (May 2006). Remaining seeds were destructively removed from all remaining pots (some pots/seeds were lost due to rodent activity), pooled together, and tested for viability with Tetrazolium. Tetrazolium turns red with dehydrogenase or other enzyme activity that generates redox equivalents, indicating seed viability (Moore 1973). We used one-way ANOVA to compare germination rates of CH and CL seeds for each species, and Pearson’s chi-square analysis to compare seed viability between seed types for each species. Matrix model We created a population projection matrix for each species and herbivory treatment using demographic transitions collected from 2005 to 2006. Based on the survivorship and seed set of first year plants, our model places seeds that germinate in the first year (seedlings) into the small stage class. Our model distinguishes between CH and CL seed types, as these seed types differ in their germination and viability (see ‘‘Results’’). Our model makes several assumptions that could affect our population projections. First, we assume that seeds only persist in the seed bank for a single year, and that 1-yearold seeds have the same viability and germination rates as new seeds. This provides a conservative estimate of population growth rate, since seed banks for both species may be more extensive. However, our results should not qualitatively differ from those assuming a longer-lived seed bank, since we found population growth rate to be highly insensitive to changes in the vital rates of seeds in the seed bank (see ‘‘Results’’). Second, we had limited power to detect rare transitions in the matrix, such as the skipping of size classes due to rapid growth (e.g., plants moving from the small to the large size class without entering the medium size class) or dramatic decreases in size. Third, our germination results were obtained without the presence of competition, and thus may overestimate seedling establishment rates. Finally, our demographic model assumes
123
60
Oecologia (2009) 161:57–66
density-independent vital rates, and does not explicitly consider density-dependent effects of herbivores. However, since we did not find strong effects of herbivory independent of density (see ‘‘Results’’), we anticipate that the role of density dependence would not qualitatively alter our findings. When we initiated the experiment in 2005, there were no L. virginica plants in the extra-large stage class. However, plants in the reduced herbivory treatment transitioned into the extra-large class in 2006 (see ‘‘Results’’). We used empirical values from L. cuneata to parameterize the stage transition probabilities of L. virginica plants in the extralarge stage class. We used a simple linear regression to determine the relationship between number of branches and CL and CH seed production for L. virginica and then calculated the expected number of CL and CH seeds that would be produced by plants in the extra-large stage class. Demographic matrices were calculated separately for each species and for each herbivory treatment. The dominant eigenvalue of each matrix is the asymptotic population growth rate, k. When k is \1, the population declines, and when k is[1, the population grows exponentially. We used bootstrap resampling (a type of Monte Carlo) to calculate 95% confidence intervals around k for each species and herbivory treatment. For each species and treatment, our dataset includes the stage of each individual in 2005 and in 2006 and the CL and CH seed production in 2005. We created 1,000 bootstrap datasets and projected k for each bootstrap dataset (Caswell 2001); each dataset was created by sampling individuals with replacement from the original dataset until the bootstrap dataset was equal in size to the original. Randomization tests (n = 1,000 permutations for each pairwise combination) were used to test whether k was significantly different between herbivory treatments for each species (Caswell 2001). We quantified an elasticity matrix for ambient herbivory plants of both L. cuneata and L. virginica. Elasticities quantify the proportional change in population growth rate that would result from a small proportional change in a matrix element (de Kroon et al. 1986; Caswell 2001). We conducted retrospective analyses (life table response experiment, LTRE) to determine the contribution of each matrix element to the difference in k between L. cuneata (C) and L. virginica (V) in the ambient herbivory treatment (Eq. 1). The contribution of a matrix element, aij, depends on both the magnitude of differences in that element between the species,and the sensitivity of k to changes in that matrix element sij ¼ dk=daij : X kc kv ffi acij avij sij ð1Þ ij
An average matrix, which had matrix elements that were midway between the values for each species was created.
123
The sensitivities used in the LTRE were from this average matrix (Caswell 2001). Since we find that only one stage class within L. virginica was significantly altered by our herbivory treatments (see ‘‘Results’’), we did not conduct retrospective analyses among herbivory treatments for either species. Contribution of herbivory We assessed the degree to which herbivory explained the difference in k between L. cuneata and L. virginica. We calculated the difference in k between species in the ambient treatment and then in the herbivory reduction treatment, and then calculated the percent change in these differences. If k differs between these species in the ambient but not reduced herbivory treatment, then herbivory explains 100% of the difference in k between L. cuneata and L. virginica.
Results Herbivory treatments The mean incidence of herbivory was significantly different between the species (F1,183 = 25.3, P \ 0.001) and herbivory treatments (F1,183 = 17.4, P \ 0.001), and there was a significant interaction (F1,183 = 11.9, P = 0.001), as the herbivory reduction treatment more strongly affected L. virginica than L. cuneata. We found that the invasive L. cuneata experienced an average of \1.0% ambient herbivory, which was significantly lower than the 10.0% average level of herbivory experienced by the native L. virginica (Fig. 1a). The herbivory reduction treatment successfully reduced herbivory on both L. cuneata (79.1% reduction from ambient) and L. virginica (81.3% reduction from ambient) (Fig. 1a). We observed that the reduction treatment did not lead to similar levels of herbivory for the two species; L. virginica received more damage than L. cuneata in the reduced herbivory treatment. However, the herbivory reduction was able to successfully reduce the level of herbivory in L. virginica to levels comparable to that of L. cuneata in ambient conditions. Germination rate and seed viability For L. cuneata, CH seeds had a significantly higher germination rate (26%) than CL seeds (17%) (F1,46 = 6.67, P = 0.01). Tetrazolium tests revealed that out of recoverable seeds, 100% of the CH and 94% of the CL seeds were viable; viability was not significantly different between the seed types (v2 = 1.3, P = 0.246). For L. virginica,
Oecologia (2009) 161:57–66
61
(a)
plants in any of the other stage classes, or the number of CH or CL seeds produced by plants in any stage class (Table 1).
c
Average % herbivory
Ambient Reduced
Effects of herbivory reduction treatment on population dynamics
d 2
a b
20
Population growth rate (λ)
(b)
35
a
a
30 25 20 15 10 5 0
b L. cuneata
c
L. virginica
Species Fig. 1 a Incidence of herbivory on Lespedeza cuneata (invasive) and L. virginica (native) in two herbivory treatments: ambient (control) and reduced (sprayed with insecticide). Values are means ± 1 SE. b Population growth rate (k) for Lespedeza cuneata (invasive) and L. virginica (native) in ambient and reduced herbivory conditions. Values are means with 95% bootstrap confidence intervals. Bars with different letters are significantly different (P \ 0.05)
germination rate of CH and CL seeds were not significantly different (F1,39 = 0.847, P = 0.363). Both CH seeds and CL seeds had relatively low germination rates (2% and 3%, respectively). CH seeds had greater viability (97%) than CL seeds (88%) (v2 = 23.9, P \ 0.001).
The projected population growth rate of L. cuneata was very high (k = 20.4) (Fig. 1b), indicating that this population can increase in abundance more than 20-fold in a single year. The population growth rate of L. cuneata did not differ between the herbivory treatments (P = 0.643). Our elasticity analysis revealed the population growth rate of L. cuneata is most sensitive to changes in the fecundity of plants in the smallest stage class, and relatively insensitive to all other matrix elements (Table 2). Lespedeza virginica had a population growth rate of 1.7 in the ambient herbivory treatment. Reducing herbivory significantly increased the population growth rate to 2.0 (P = 0.043) (Fig. 1b). The population growth rate of L. virginica was also most sensitive to changes in the fecundity of plants in the smallest stage class (Table 2). However, in contrast to L. cuneata, the elasticities of L. virginica were more evenly distributed among several matrix elements. Life table response experiment Lespedeza cuneata had a much higher population growth rate compared to its native congener, L. virginica. LTRE reveals that this is due primarily to the higher fecundity of plants in the smallest stage class of L. cuneata compared to L. virginica (Table 2). The transition from small plant to small plant (i.e., small plants producing seeds which germinate and become small plants in the next year) was 20.8 for L. cuneata and only 1.0 for L. virginica (Fig. 2). Further, k is highly sensitive to changes in the small plant to small plant transition (Table 2).
Effects of herbivory reduction treatment on vital rates
Contribution of herbivory to differences between species
Herbivory reduction did not significantly affect any of the vital rates (stage transitions, stage-specific fertility) of L. cuneata (Table 1; Fig. 2a). A reduction in herbivory significantly affected the stage transitions of L. virginica large plants (Table 1; Fig. 2b); in the reduced herbivory treatment, large plants were more likely to advance to a previously non-existent extra-large stage class than in the ambient treatment. However, the herbivory reduction treatment did not affect the stage transitions of L. virginica
We found that under ambient conditions, L. cuneata had a population growth rate that was larger than that of L. virginica (difference = 18.7). Additionally, under herbivory reduction, L. cuneata still maintained a growth rate that was substantially larger than L. virginica (difference = 18.3). Overall, we estimate that herbivory (enemy release) accounts for only 2% of the difference in the population growth rates of these species {[(18.7 – 18.3)/ 18.7] 9 100}.
123
62
Oecologia (2009) 161:57–66
Table 1 Mean (across both herbivory treatments) vital rates of plants in each stage classes for Lespedeza cuneata and Lespedeza virginica Vital rates
L. cuneata
v2 or F value
df
P
L. virginica
Small seed set
100.8 (CL)
0.97
1
0.34
10.6 (CL)
1.3
1
0.26
14.2 (CH)
1.08
1
0.32
5.8 (CH)
2.5
1
0.12
Small–dead
0.22
1.22
2
0.54
0.07
3.82
2
0.15
small–small
0.17
Small–medium
0.61
Medium seed set
258.9 (CL)
0.48
1
0.50
36.4 (CL)
0.93
1
0.34
33.9 (CH)
0.78
1
0.39
13.1 (CH)
1.2
1
0.28
Medium–dead
0.09
3.33
3
0.34
0
0.19
2
0.91
Medium–small
0.04
0.22
Medium–medium
0.79
0.72
Medium–large Large seed set
0.08 873.9 (CL)
3.11
1
0.09
0.06 87.2 (CL)
0.05
1
0.82
109.9 (CH)
0.001
1
0.97
56.4 (CH)
0.53
1
0.47
Large–dead
0.04
0.91
3
0.82
0
6.57
2
0.04
Large–medium
0.17
0.18
Large–large
0.61
0.65
Large–extra large
0.17
Extra large seed set
1993 (CL)
0.01
1
0.92
n/a
n/a
n/a
n/a
292.6 (CH)
2.32
1
0.15
Extra large–large
0.27
1.03
1
0.31
n/a
n/a
n/a
n/a
Extra large–extra large
0.73
n/a
n/a
n/a
n/a
n/a
n/a
Germination rate
0.471 (CL) 0.357 (CL)
df
P
0.60 0.33
0.18
n/a n/a
n/a
n/a
0.883 (CH) Seed viability
v2 or F value
0.032 (CL) 0.025 (CH)
n/a
n/a
0.297 (CH)
n/a
0.846 (CL) 0.808 (CH)
Chi-square and ANOVA results shown for each stage class to test for differences in stage transitions and CH and CL seed production (respectively) across herbivory treatments CH Chasmogamous, CL cleistogamous
Discussion We found that the incidence of herbivory was greater on the native L. virginica than on its invasive congener L. cuneata, and that herbivory affected plant fitness. Herbivory was found to explain 100% of the difference in one component of fitness (growth of plants in the large stage class); L. virginica plants were able to advance to the extra-large stage class in the reduced herbivory treatment, but this was very rare under ambient herbivory conditions. A few previous studies have shown strong contributions of herbivory to differences in plant fitness in similar herbivory reduction experiments on native and exotic congeners. For instance, herbivory explained 100% of the difference in fitness (above-ground biomass) between native and exotic Lepidium (Agrawal et al. 2005), and 17% of the difference in fitness (seed survivorship) in native and exotic Senecio (Blaney and Kotanen 2001). Our study is the first to examine the population-level effects of herbivory on native and invasive congeners. We
123
find that L. cuneata has a much higher population growth rate than L. virginica in both ambient and reduced herbivory treatments, demonstrating that herbivory does not explain the differences in the relative success of these species. Despite the large effects of herbivory on the stage transitions of large plants, these individual-level effects translated into relatively small population-level effects, as only 2% of the differences in population growth rates between native and invasive Lespedeza can be explained by herbivory. The relatively small population-level effects of herbivory on L. virginica can be explained by our elasticity analysis. The population growth rate of L. virginica is relatively insensitive to changes in the stage transitions of large plants, which were the vital rates influenced by herbivory reduction. Elasticity analysis revealed that L. cuneata was highly sensitive to changes in the fertility of small plants. This is unusual for perennial plant species, which are typically most sensitive to the survival of plants in larger stage classes (Silvertown et al. 1993). However, this sensitivity
Oecologia (2009) 161:57–66
63 Table 2 Elasticity analysis and life table response experiment (LTRE) for L. cuneata and L. virginica plants in ambient herbivory conditions
(a) Lespedeza cuneata 10.2
3.8
CH SB
1.18 0.49 0.26
Small
19.0
0.08
52.4 175.7
Med.
0.17
0.17
Large
0.27
411.7 165.0 376.3
(b) Lespedeza virginica 10.3 4.57
L. virginica
Small
0.946
0.355
15.9
Medium
0.022
0.069
1.9
Large
0.000
0.014
0.039
Extra large
0.000
0.013
0
0.000
0.010
0.046
0.72 0.06
0.33
Small
1.45 3.45
Med.
CH seed bank
n/a
1.0
8.68
CL SB
Contribution
Transitions of: 44.4
0.03
Summed elasticities L. cuneata
XL
48.9
CH SB
Summed vital rates
Fertility of:
0.17
CL SB
0.73
0.61
0.79
20.8 0.61
AH: 0.12 HR: 0.24
AH: 0.82 HR: 0.47 AH: 0.06 HR: 0.29
Large
n/a
n/a
XL
0.02
CL seed bank
0.007
0.095
0.171
Small
0.022
0.134
0.612
Medium
0.001
0.132
0.012
Large
0.000
0.060
0
Extra large
0.000
0.014
0
n/a
29.8 71.4
n/a
Fig. 2 Annual life cycle graphs for a Lespedeza cuneata and b Lespedeza virginica, explicitly incorporating mating system: cleistogamous (CL) and chasmogamous (CH) seed types. Circles represent stage classes and arrows represent transitions (fecundity, growth) from one year to the next. Values are mean transitions; averaged across herbivory treatments for vital rates not significantly influenced by the herbivory treatment and shown separately (in bold) for transitions significantly influenced by the herbivory treatment: AH ambient herbivory (control), HR herbivory reduction (for statistical analyses, see Table 1). Extra-large L. virginica plants were not available for demographic study
result may be explained by L. cuneata’s high population growth rate. In such rapidly growing populations, reproduction early in life allows this species to maintain a rapid generation time (see also Parker 2000). Other factors besides herbivory must explain the difference in population growth rate between L. cuneata and L. virginica. Our LTRE analysis revealed that higher fertility of small plants explains the much higher growth rate of L. cuneata compared to L. virginica. Lespedeza cuneata produces many more seeds and its seeds have higher germination compared to L. virginica. The environmental factors or species traits that contribute to these demographic differences are the key to understanding the relative success of this species. Lespedeza cuneata seeds germinate immediately, whereas a large proportion of L. virginica’s seeds remain in the seed bank. When environmental conditions are favorable, as they were during our study, this should allow L. cuneata to have a more rapid generation time and thus faster population growth rate. However, L. virginica may be better able to recover from a catastrophic event through germination of seeds out of the
For each stage class, the elasticity of fertility considers the summed elasticities for CH and CL seed types and the elasticity of stage transitions considers the summed elasticities of all possible stage transitions. The contribution of each summed vital rate indicates the LTRE contribution of that vital rate to the observed difference in k between L. cuneata and L. virginica CH Chasmogamous, CL cleistogamous
seed bank. A notable example of the important role of seed banks during catastrophes was documented by Kalisz and McPeek (1992) in the native winter annual Collinsia verna. The presence of the seed bank contributed little to k during favorable years, but was critical during devastating flood years. To test whether the differences in seed longevity give L. virginica an advantage over L. cuneata in stochastic environments, future research would need to collect longterm demographic data on these species to document their performance during and after catastrophic years, the frequency of catastrophic years, and the longevity of their seed banks. This information could be used to parameterize a stochastic model with more stage classes (for seed transitions). The values of k documented in this study for the two species and herbivory treatments are useful as a comparative tool, but the values of k are less useful as a future projection of the population sizes of these Lespedeza species for several reasons. First, there are reasons to suspect that our field germination experiment led to inflated values of germination rates, and therefore k. We used outdoor seed baskets to measure natural germination rate, and our goal was to keep the abiotic (hydrology, weather conditions) and biotic (soil microbes) conditions as natural as possible. However, the seed baskets prevented seedlings from competing as strongly with grass competitors, since grass was not present in the pots, and therefore we may
123
64
have overestimated early seedling survivorship. Through simulations, we explored how reducing the early seedling survivorship would affect our projections. When early seedling survivorship is reduced 90%, L. cuneata has k = 4.0 (in both control and reduced herbivory treatments), and L. virginica has k = 1.09 in the control treatment and k = 1.24 in the reduced herbivory treatment (S1 in Electronic Supplementary Material). Thus, qualitatively, the main conclusions of the study (i.e., there are significant differences in the population growth of these species, and herbivory explains a small fraction of that difference) appear to be robust, even if early seedling survivorship is overestimated by a large degree. Second, even if the k values reported here are accurate, these high rates of population growth, in particular for the invasive L. cuneata, would not be sustained for very long because the population would rapidly reach carrying capacity or more complicated population dynamics (e.g., Pardini et al. 2009). Third, our k value summarizes one environment and does not examine the role of temporal stochasticity. Nevertheless, our results suggest that in ideal environmental conditions, high levels of population growth can occur in L. cuneata, which may, in part, explain its invasiveness. The short-term nature of our demographic study raises two concerns that could affect the accuracy of our comparisons between these species and herbivory treatments. First, herbivory might have carry-over effects such that fitness consequences of the herbivore reduction experiment become magnified with repeated years of reduction. If this is true, then the effect sizes of the herbivory removal experiment might be underestimated for L. virginica. Ehrle´n (2003) experimentally removed herbivores for 3 years and measured the effects on the demography and population growth rate of a native perennial plant (Lathyrus vernus). He found that the short-term results did not differ from the long-term results, suggesting that the demographic effects of herbivory reduction manifest quickly for this species which is in the same plant family (Fabaceae) as our study species. However, without longer term data, we cannot be sure that carry-over effects are not important in these Lespedeza species. Second, our model currently assumes that the seed bank only lasts a year. While data are lacking, Lespedeza seeds are assumed to have seed banks that can persist for several years. This assumption is most important for the native L. virginica, since few of its seeds were observed to germinate in their first year. We expect that little germination from the seed bank for several years into the future would have a minor effect on the population growth rate, since seed bank transitions have low elasticity. However, if L. virginica seeds have a pulse of high germination in their second or third year in the seed bank, this would likely cause a significant increase in its population growth rate, and as a
123
Oecologia (2009) 161:57–66
result would minimize the difference in k between L. virginica and L. cuneata. Invasive species can have enormous environmental and economic impacts (Pimentel et al. 2000; Colautti et al. 2006). Therefore, it is critical that we understand the mechanisms that allow invasive species to achieve rapid population growth rates (e.g., Parker 2000) if we are to develop appropriate management regimes to control the spread of these species. Matrix models have been created for several invasive plant species; however, few studies have explicitly considered factors that might regulate population growth (Parker 2000; but see Jacquemyn et al. 2005). A few studies have examined the effects of consumers on the population dynamics of invasive plant species, and these have important implications for management of these species through biological control. Jongejans et al. (2006) documented that enemies regulate the population growth rate of Carduus nutans in its native range (Eurasia); thus, biological control might be successful in its invasive range, which spans several continents. The invasive biennial, Senecio jacobaea, is capable of high rates of population growth in its invasive range in North America, but has been projected to decline towards extinction with the use of two herbivorous biocontrol agents (McEvoy and Coombs 1999). If the enemy release hypothesis is found to explain the success of an invasive species, then one application of this result might be a biological control program. We suggest that population modeling would aid in both identifying enemy release as an important mechanism of invasiveness and in determining the likely success of biological control. To date, most biological control research has focused on identifying appropriate specialist herbivores to introduce (Buckingham 1994; Julien and Griffiths 1998). Much less attention has been given to investigating whether enemy damage will reduce the population growth rate of an invasive species to a desirable level. However, recent studies have used simulated herbivory and matrix modeling to determine how effective a leaf-chewing insect would have to be to cause the desired decrease in the population growth of invasive plant species (Raghu and Dhileepan 2005; Schutzenhofer and Knight 2007). To date, studies examining the population dynamics of invasive plant species (reviewed by Ramula et al. 2008) and those using ecological experiments to test the mechanisms responsible for the success of invasive species have proceeded separately. Since our research focuses on understanding the factors responsible for the growth and spread of an invasive species, an estimate of population growth rate is an appropriate response variable to examine mechanisms surrounding population success. Specifically, our study compares native and invasive congeners to investigate the relative role of herbivory in explaining their
Oecologia (2009) 161:57–66
different levels of population success. Our results indicate that herbivory is relatively unimportant in explaining the difference in population growth rate between these species. Future comparisons between congeneric native and invasive species might find a greater role of herbivory, particularly in cases where the native species receives higher ambient levels of herbivory than that observed here for L. virginica. In addition, future experimental tests of the enemy release hypothesis should also consider the population growth rate of species in their native and invasive ranges. Species that are strongly regulated by specialist enemies in their native range may be more likely to show high population growth in their invasive range in the absence of those enemies (Maron and Vila 2001; Wolfe 2002; Joshi and Vrieling 2005). We show that differences in herbivory explain a trivial amount of the difference in the population growth rates of the invasive plant, L. cuneata, and its co-occurring native congener, L. virginica. Our study highlights the importance of measuring fitness and using models to project population growth in herbivory reduction experiments. In our case, the native and invasive species differed in their incidence of herbivory, and herbivory was found to influence one component of plant fitness. However, herbivory did not explain the substantially higher population growth rate of the invasive species relative to its native congener in the year examined. The population dynamics of these native and invasive Lespedeza species were most sensitive to changes in fertility early in life. Thus, factors that explain the higher fertility of L. cuneata compared to L. virginica should explain its relative success. Acknowledgments We thank J.M. Chase and E.A. Pardini for discussion and comments on the manuscript, C.D. Melm, J. Mueller, and K. Smyth for field assistance, the Tyson Research Center for allowing us to conduct this project and for logistical support. This research was funded by St. Louis University, the Tyson Research Center’s Crescent Hills Research Fund, and Washington University, St. Louis, MO. The authors declare that this work was conducted in compliance with the laws of the United States.
References Agrawal AA, Kotanen PM (2003) Herbivores and the success of exotic plants: a phylogenetically controlled experiment. Ecol Lett 6:712–715 Agrawal AA, Kotanen PM, Mitchell CE, Power AG, Godsoe W, Klironomos J (2005) Enemy release? An experiment with congeneric plant pairs and diverse above- and belowground enemies. Ecology 86:2979–2989 Ansley W (1960) The influence of time and frequency of cutting on persistence, forage yield, seed yield, and seed type of sericea lespedeza. MS thesis, Auburn University, Auburn Berg H, Redbo-Torstensson P (1999) Offspring performance in three cleistogamous Viola species. Plant Ecol 145:49–58
65 Blaney CS, Kotanen PM (2001) Effects of fungal pathogens on seeds of native and exotic plants: a test using congeneric pairs. J Appl Ecol 38:1104–1113 Buckingham G (1994) Biological control of aquatic weeds. In: Rosen D, Bennett FD, Capinera JL (eds) Pest management in the subtropics: biological control, a Florida perspective. Intercept, Andover, pp 413–480 Callaway RM, Thelen GC, Rodriguez A, Holben WE (2004) Soil biota and exotic plant invasion. Nature 427:731–733 Carpenter D, Cappuccino N (2005) Herbivory, time since introduction and the invasiveness of exotic plants. J Ecol 93:315–321 Caswell H (2001) Matrix population models: construction, analysis, and interpretation, 2nd edn. Sinauer, Sunderland Clewell AF (1966) Native North American species of Lespedeza (Leguminosae). Rhodora 68:359–405 Colautti RI, Ricciardi A, Grigorovich IA, MacIsaac HJ (2004) Is invasion success explained by the enemy release hypothesis? Ecol Lett 7:721–733 Colautti RI, Bailey SA, van Overdijk CDA, Amundsen K, MacIsaac HJ (2006) Characterized and projected costs of nonindigenous species in Canada. Biol Invasions 8:45–59 Crawley MJ (1989) Insect herbivores and plant-population dynamics. Annu Rev Entomol 34:531–564 Crawley MJ (1997) Plant-herbivore dynamics. In: Crawley MJ (ed) Plant ecology. Blackwell Science, Oxford, pp 401–474 de Kroon H, Plaisier A, van Groenendael J, Caswell H (1986) Elasticity: the relative contribution of demographic parameters to population growth rate. Ecology 67:1427–1431 DeWalt SJ (2006) Population dynamics and potential for biological control of an exotic invasive shrub in Hawaiian rainforests. Biol Invasions 8:1145–1158 Donnelly ED (1955) The effects of outcrossing on forage and seed yields in sericea lespedeza, L. cuneata. Agron J 47:466–467 Ehrle´n J (2003) Fitness components versus total demographic effects: evaluating herbivore impacts on a perennial herb. Am Nat 162:796–810 Hendrix SD (1988) Herbivory and its impact on plant reproduction. In: Lovett Doust J, Lovett Doust L (eds) Plant reproductive ecology: patterns and strategies. Oxford University Press, Oxford, pp 246–263 Horvitz CC, Schemske DW (2002) Effects of plant size, leaf herbivory, local competition and fruit production on survival, growth and future reproduction of a neotropical herb. J Ecol 90:279–290 Jacquemyn H, Brys R, Neubert MG (2005) Fire increases invasive spread of Molinia caerulea mainly through changes in demographic parameters. Ecol Appl 15:2097–2108 Jongejans E, Sheppard AW, Shea K (2006) What controls the population dynamics of the invasive thistle Carduus nutans in its native range? J Appl Ecol 43:877–886 Joshi J, Vrieling K (2005) The enemy release and EICA hypothesis revisited: incorporating the fundamental difference between specialist and generalist herbivores. Ecol Lett 8:704–714 Julien MH, Griffiths MW (1998) Biological control of weeds: world catalogue of agents and their target weeds, 4th edn. CAB International, Wallingford Kalisz S, McPeek M (1992) Demography of an age-structured annual: Resampled projection matrices, elasticity analyses, and seed bank effects. Ecology 73:1082–1093 Keane RM, Crawley MJ (2002) Exotic plant invasions and the enemy release hypothesis. Trends Ecol Evol 17:164–170 Leimu R, Lehtila K (2006) Effects of two types of herbivores on the population dynamics of a perennial herb. Basic Appl Ecol 7:224–235 Liu H, Stiling P (2006) Testing the enemy release hypothesis: a review and meta-analysis. Biol Invasions 8:1535–1545
123
66 Maron JL, Crone E (2006) Herbivory: effects on plant abundance, distribution and population growth. Proc R Soc Lond B 273:2575–2584 Maron JL, Vila M (2001) When do herbivores affect plant invasion? Evidence for the natural enemies and biotic resistance hypotheses. Oikos 95:361–373 McEvoy PB, Coombs EM (1999) Biological control of plant invaders: regional patterns, field experiments, and structured population models. Ecol Appl 9:387–401 McKee R, Hyland HL (1941) Apetalous and petaliferous flowers in Lespedeza. Am Soc Agron J 33:811–815 Moore RP (1973) Tetrazolium staining for assessing seed quality. In: Heydecker W (ed) Seed ecology; proceedings. Pennsylvania State University Press, University Park, Pennsylvania, pp 347– 366 Pardini E, Drake J, Chase J, Knight TM (2009) Complex population dynamics and control of the invasive biennial, Alliaria petiolata (garlic mustard). Ecol Appl 19:387–397 Parker IM (2000) Invasion dynamics of Cytisus scoparius: a matrix model approach. Ecol Appl 10:726–743 Pimentel D, Lach L, Zuniga R, Morrison D (2000) Environmental and economic costs of nonindigenous species in the United States. Bioscience 50:53–65 Price CA, Weltzin JF (2003) Managing non-native plant populations through intensive community restoration in Cades Cove, Great Smoky Mountains National Park, USA. Restor Ecol 11:351–358 Raghu S, Dhileepan K (2005) The value of simulating herbivory in selecting effective weed biological control agents. Biol Control 34:265–273 Ramula S, Knight TM, Burns J, Buckley Y (2008) General guidelines for invasive plant management based on comparative demography of invasive and native plant populations. J Appl Ecol 45:1124–1133
123
Oecologia (2009) 161:57–66 Reinhart KO, Packer A, Van der Putten WH, Clay K (2003) Plant-soil biota interactions and spatial distribution of black cherry in its native and invasive ranges. Ecol Lett 6:1046–1050 Schutzenhofer MR, Knight TM (2007) Population effects of augmented herbivory on Lespedeza cuneata: Implications for biological control. Ecol Appl 17:965–971 Shea K, Kelly D (1998) Estimating biocontrol agent impact with matrix models: Carduus nutans in New Zealand. Ecol Appl 8:824–832 Shea K, Kelly D, Sheppard AW, Woodburn TL (2005) Contextdependent biological control of an invasive thistle. Ecology 86:3174–3181 Silvertown J, Franco M, Pisanty I, Mendoza A (1993) Comparative plant demography—relative importance of life-cycle components to the finite rate of increase in woody and Herbaceous Perennials. J Ecol 81:465–476 Statsoft (1999) STATISTICA for Windows, 6.0 edn. Statsoft, Tulsa Stevens S (2002) Element stewardship abstract: Lespedeza cuneata (Dumon-Cours.) G. Don: Sericea Lespedeza, Chinese Bush clover (http://tncweeds.ucdavis.edu/esadocs/documnts/lespcun. html, 30 September 2008). The Nature Conservancy Strauss SY (1991) Direct, indirect, and cumulative effects of 3 native herbivores on a shared host plant. Ecology 72:543–558 USDA N (2008) PLANTS Database (http://plants.usda.gov, 28 April 2008). National Plant Data Center, Baton Rouge, LA Vila M, Maron JL, Marco L (2005) Evidence for the enemy release hypothesis in Hypericum perforatum. Oecologia 142:474–479 Wheeler WA, Hill DD (1957) Grassland seeds. Van Nostrand, Princeton Wolfe LM (2002) Why alien invaders succeed: support for the escape-from-enemy hypothesis. Am Nat 160:705–711
