Biol Invasions (2010) 12:2363–2372 DOI 10.1007/s10530-009-9649-3

ORIGINAL PAPER

Variation in the phenology and abundance of flowering by native and exotic plants in subalpine meadows Brook J. Wilke • Rebecca E. Irwin

Received: 7 March 2009 / Accepted: 10 November 2009 / Published online: 21 November 2009 Ó Springer Science+Business Media B.V. 2009

Abstract The timing and abundance of flower production is important to the reproductive success of angiosperms as well as pollinators and floral and seed herbivores. Exotic plants often compete with native plants for space and limiting resources, potentially altering community floral dynamics. We used observations and a biomass-removal experiment to explore the effects of an invasive exotic flowering plant, Linaria vulgaris, on community and individual species flowering phenology and abundance in subalpine meadows in Colorado, USA. Invasion by L. vulgaris was associated with a shift in both the timing and abundance of community flowering. Invaded plant communities exhibited depressed flowering by 67% early in the season relative to uninvaded communities, but invaded

Electronic supplementary material The online version of this article (doi:10.1007/s10530-009-9649-3) contains supplementary material, which is available to authorized users. B. J. Wilke (&) W. K. Kellogg Biological Station & Department of Crop and Soil Sciences, Michigan State University, Hickory Corners, MI 49060, USA e-mail: [email protected]

sites produced 7.6 times more flowers than uninvaded sites once L. vulgaris began flowering. This increase in flowers at the end of the season was driven primarily by prolific flowering of L. vulgaris. We also found lower richness and evenness of resident flowering species in invaded plots during the period of L. vulgaris flowering. At the species level, a common native species (Potentilla pulcherrima) produced 71% fewer flowers in invaded relative to uninvaded plots, and the species had reduced duration of flowering in invaded relative to uninvaded sites. This result suggests that L. vulgaris does not simply alter the flowering of subordinate species but also the flowering of an individual common species in the plant community. We then used observational data to explore the relationship between L. vulgaris density and resident floral production but found only partial evidence that higher densities of L. vulgaris were associated with stronger effects on resident floral production. Taken together, results suggest that a dominant invasive plant can affect community and individual-species flowering. Keywords Flowering
Global environmental change
Invasive species
Linaria vulgaris
Phenology

B. J. Wilke
R. E. Irwin Rocky Mountain Biological Laboratory, Crested Butte, CO 81224, USA

Introduction

R. E. Irwin Department of Biology, Dartmouth College, Hanover, NH 03755, USA

Flower production in angiosperms is critical to plant reproduction and fitness as well as pollinator and

123

2364

floral and seed herbivore reproduction and fitness. Variation in the timing, duration, and abundance of flower production can affect the potential number of mates (Rathcke and Lacey 1985), pollination success (Waser 1978; Gross and Werner 1983), seed production (Bishop and Schemske 1998; Totland 1999), and seed dispersal and predation (Primack 1987; Brody 1997). Moreover, fluctuations in flowering phenology and abundance can affect organisms that rely on flowers for food and shelter. For example, pollinators rely on plants for nectar and pollen resources to maintain reproduction, and the behavior and abundance of floral and seed herbivores may be affected by permanent change or temporary fluctuations in flowering phenology (e.g., Augsperger 1980). Although evolutionary forces have certainly shaped patterns of flowering (Rathcke and Lacey 1985; Kochmer and Handel 1986), plants may also express plastic variation in the timing and abundance of flowering due to ecological factors, including abiotic and biotic factors (Inouye et al. 2002; McCall and Irwin 2006; Post et al. 2008). Many of these ecological factors that affect the phenology and abundance of flowers may be influenced by global environmental change, including species invasions (Vitousek 1994). Invasive, exotic plants affect the physical and chemical characteristics of invaded ecosystems. Consequently, invasive plants can affect the timing and rate of native plant growth (D’Antonio et al. 1998; Ridenour and Callaway 2001), but little is known about how invasive plants affect the timing and abundance of flower production for native species and for invaded communities. Invasive plants could affect the phenology and abundance of flowers in their new range in at least three non-mutually exclusive ways. First, invasive plants, by usurping space from natives, may affect community flowering phenology and abundance. A dominant invader that flowers prolifically may add to the overall flower abundance in invaded communities. Many exotic plants are introduced as horticultural species because of their large floral display, so prolific flowering by invasives may be common. Alternatively, for dominant invaders that produce few flowers, the converse pattern may be evident, with invaded communities having lower community flower abundance. Second, a dominant invader may affect community flowering phenology by blooming at different times of the season than displaced

123

B. J. Wilke, R. E. Irwin

residents (Godoy et al. 2009). Third, competition between invasive and native plants for resources could alter the timing and abundance of flowers of native species and communities. It is likely that not all native flowering plants will respond in the same direction or with equal magnitude to the effects of an invasive plant. Species-specific differences may be strongly dependent on strategies of plant growth and life history as well as levels of resource limitation. The goal of this study was to investigate the effects of an invasive, exotic flowering plant (Linaria vulgaris) on community and individual species flowering phenology and abundance. We focused on a flowering invader in a high-elevation temperate zone in southwestern Colorado, USA where flowering occurs in a discrete four-month season. First, we used an observational study and a biomass removal experiment to examine how a plant invasion, and annual removal of the invasive species, affected flowering phenology and abundance of the entire flowering community (resident plants and the invader) and the resident flowering community only. In addition, we focused on a single common native species to assess whether potential differences in community flowering were driven by the abundance of resident flowering plants or their per-plant flower production. We predicted that L. vulgaris would increase community flower abundance, but that prolific flowering by the invader would come at a cost of reduced flowering by resident species on both a per-site and per-plant basis. Second, we predicted that annual removal of L. vulgaris would lessen its impact on community floral characteristics. Finally, we used observational data to explore the relationship between L. vulgaris density and resident floral production, reasoning that higher L. vulgaris density would be associated with stronger effects on the resident flowering community.

Materials and methods Study system Linaria vulgaris (Scrophulariaceae) is an herbaceous weed native to southeastern Europe and southwestern Asia and that was introduced to North America in the mid 19th century (Kadrmas and Johnson 2002). It is found throughout western North America, particularly

Variation in the phenology and abundance of flowering

in the Rocky Mountains, and is now considered a noxious pest in natural areas and rangelands, especially in disturbed open habitats (Lajeunesse 1999; Pauchard et al. 2003). Linaria vulgaris is also considered a noxious weed in other areas worldwide (Saner et al. 1995; Carpenter and Murray 1998). Linaria vulgaris is a rhizomatous perennial, reproducing sexually as an obligate outcrosser (Arnold 1982; Docherty 1982) and vegetatively by producing adventitious shoots from both the main and lateral roots. Inflorescences produce (mean ± 1 SE) 26.8 ± 2.6 yellow, zygomorphic, typically bee-pollinated flowers for 3–6 weeks in mid-to-late summer (early July to early September at our study site). Ramets produce a wide range of seed crops, ranging from 0 to 6,000 seeds per ramet per year, and genets can produce up to 30,000 seeds annually (Saner et al. 1995). In addition, plants in their second year of growth can produce as many as 250 shoots through vegetative reproduction (Saner et al. 1995). We studied populations of L. vulgaris and resident plants in subalpine meadows around the Rocky Mountain Biological Laboratory (RMBL) in Gothic, Gunnison County, Colorado, USA, elevation 2900 m. The meadows were partially invaded by L. vulgaris, resulting in a mosaic of patches characterized by the presence or absence of L. vulgaris. Abundant native angiosperms in the meadows exhibited a range of life histories, reproductive timing, and growth forms (both shallow- and tap-rooted species) and included Potentilla pulcherrima (Rosaceae), Heliomeris multiflora (Asteraceae), Linum lewisii (Linaceae), Collomia linearis (Polemoniaceae), Lathyrus leucanthus (Fabaceae), Erigeron speciosus (Asteraceae), and Galium boreale (Rubiaceae). Three naturalized exotic plant species were present at the study site: Cirsium arvense (Asteraceae), Taraxacum officinale (Asteraceae), and Tragopogon dubias (Asteraceae). However, flowers from these plants only made up 0.2% of total flower abundance and removing them completely from the analyses (described below) did not affect the results. We also examined flowering phenology and abundance of a single native species, P. pulcherrima, which is an herbaceous perennial that commonly grows with L. vulgaris. Potentilla pulcherrima is a common native flowering plant in this system, making up one-third of floral abundance in meadows without L. vulgaris (Burkle 2008). In addition, the flowering of P. pulcherrima is resource

2365

limited around the RMBL (Burkle 2008). Because P. pulcherrima can tolerate long, dry periods and short, wet periods by exhibiting rapid reproductive cycles (Stinson 2004), it may use similar strategies to tolerate potential competition for resources with an invasive species. Field plots We established 24 plots, each 4-m2. The plots were divided among treatments in a two-by-two factorial design (six plots per treatment): invasion status (L. vulgaris present or absent) and plant biomass removal (pull or control). For invasion status, 12 of the plots were selected in meadows where L. vulgaris was present at a range of densities (109–429 ramets m-2). The other 12 plots were chosen in areas uninvaded by L. vulgaris (hereafter referred to as uninvaded plots) that had similar slope and aspect as invaded plots. Plots invaded and uninvaded by L. vulgaris were assigned to pull or control treatments at random. For the biomass removal treatment, in half of the invaded plots, L. vulgaris was eliminated annually by pulling out all of the stalks and as much of the root system as possible. Proportional amounts of resident biomass were removed by pulling in half of the uninvaded plots to control for soil disturbance, changes in light availability, and biomass removal. Removal in the uninvaded plots was conducted by pulling primarily non-reproductive biomass. The proportion of dry biomass removed was not significantly different between uninvaded and invaded sites (F1,10 = 0.02, P = 0.88). To avoid affecting long term population dynamics, we pulled biomass from the plots in late-August of each year (from 2000 to 2004) following peak biomass and seed production of most residents, but before L. vulgaris went to seed. All plots were established between 2000 and 2001 and were in exposed, un-shaded meadows within an 180,000 m2 area. Individual plots were greater than 4 m apart. Plots did not differ significantly in aspect, soil pH, or soil percent organic matter at the start of this study (unpublished data). Does Linaria vulgaris affect flowering phenology and abundance? We measured flowering phenology and abundance in the summer of 2004. We counted the number of

123

2366

B. J. Wilke, R. E. Irwin

flowering plants in the plots three times per week, from 10 June to 8 September. During each census, we also recorded the number of flowers per plant on up to ten randomly selected plants per species, excluding grasses. Because some of the species reproduced through rhizomes, it was difficult to identify individual plants without disrupting roots. Therefore, determination of a single plant was done based on above-ground morphological characteristics and then kept consistent throughout the study. We used the total flowering plants and mean flowers per plant to calculate the total number of flowers per plot. After plant biomass removal (described above) on 17 August 2004, we discontinued plant and flower counts in the removal plots. Therefore, comparisons between uninvaded and invaded plots after 17 August were only made for plots that did not have biomass removed. Data analyses We divided the effects of L. vulgaris on flowering characteristics into three categories. All analyses were conducted using PROC MIXED in SAS (SAS Institute 2004). (1)

(2)

We used the maximum number of flowers (all species summed) per plot and the date of maximum flowers to provide information about the maximum floral resources in a community. We used factorial analyses with invasion status (L. vulgaris present or absent), plant removal (pull or control), and their interaction as factors and maximum number of flowers and the date of maximum flowering as response variables; one set of analyses included L. vulgaris and one excluded L. vulgaris to characterize changes in resident species only. We tested how the abundance, richness, and evenness of flowering species varied over the season to assess the effects of L. vulgaris on community flowering phenology. For each plot, we calculated flower abundance as the sum of all flowers at each census period (i.e., the total number of flowers per plot per census), and species richness as the total number of species in bloom per census. We calculated flowering species evenness (E) for each plot per census using the formula: E = H/ln(R), where H is the

123

(3)

Shannon-Wiener diversity index (Pielou 1966) and R is flowering species richness. We analyzed how community flower abundance (logtransformed), richness, and evenness varied over the flowering season and among treatments using factorial repeated-measures analyses. The first order autoregressive model was used as the variance–covariance structure for flowering abundance and richness. Data for evenness were analyzed using a compound symmetry model as the variance–covariance structure. These models were chosen because they had the lowest AIC values. We compared treatment means for each census using t-tests when main effects or interactions were significant. We focused on the common native species P. pulcherimma, which occurred in 20 of the 24 plots (five plots per treatment), to test how L. vulgaris affected individual-species patterns of flowering phenology and abundance that may have been masked at the flowering community level, especially if declines in flower production by some residents were compensated for by increases in flower production by others. The effects of invasion status and plant removal on P. pulcherrima flowering (log-transformed) were analyzed using a factorial repeated-measures design. The first order autoregressive model with heterogeneous variances by day was used as the variance–covariance structure. We used twoway ANOVAs to compare flowers per shoot, duration of flowering (total number of days P. pulcherrima was in bloom per plot), and number of flowering shoots among treatments.

Is Linaria vulgaris cover related to resident flower abundance? If L. vulgaris affects resident flowering, we reasoned that an increase in L. vulgaris density (or percent cover) would result in increased effects on resident flowering; thus, we tested whether there was a relationship between L. vulgaris percent cover and resident flower abundance. During each flower census, we visually estimated the above-ground percent cover of L. vulgaris four times per invaded site using a 625-cm2 quadrat. We calculated mean percent cover of L. vulgaris from the three census periods that

Variation in the phenology and abundance of flowering

coincided with highest community flower production (18 June, 15 July, and 10 August). We tested the prediction that L. vulgaris percent cover would negatively affect resident floral abundance by regressing percent cover of L. vulgaris on resident flower abundance using PROC REG (SAS Institute 2004); we used separate regressions for each date.

Results Does L. vulgaris affect flowering phenology and abundance? Maximum flowers per plot When all species (both residents and L. vulgaris) were included in the analyses, peak community flowering occurred 23 days later in invaded (12 August) compared to uninvaded (21 July) plots (F1,20 = 36.86, P \ 0.0001). Moreover, maximum flowers per plot was significantly higher in invaded compared to uninvaded plots (F1,20 = 4.84, P = 0.04), with 97% of the total flowers from L. vulgaris in invaded plots. When L. vulgaris was excluded from the analysis, peak community flowering of residents occurred 11 days earlier in invaded compared to uninvaded plots (F1,20 = 4.81, P = 0.04), and invaded plots contained 184% fewer flowers at peak flowering than uninvaded plots (F1,10 = 9.88, P = 0.01). In all analyses (here and below), we found no effect of biomass removal and no interaction between L. vulgaris invasion and biomass removal on flowering phenology, abundance, or richness (Supplementary Electronic Appendix 1). The lack of effect of L. vulgaris removal on patterns of flowering in invaded plots may be because up to three years of L. vulgaris pulling had no effect on the number of L. vulgaris stems in 2004 (F1,10 = 1.61, P = 0.23). Flower abundance, richness, and evenness throughout the season Considering only resident flowers, uninvaded plots contained 198% more flowers than invaded plots across the season (F1,20 = 5.03, P = 0.008). We also found a significant interaction between invasion status and day of the year when L. vulgaris was included in the analysis (F38,648 = 2.60, P \ 0.0001), suggesting

2367

that the effect of invasion status on flower production was not consistent across the season. When we compared flower production at each census day, we found that uninvaded plots produced three times more flowers than invaded plots in mid-July (15–24 July), but invaded plots produced 7.6 times more flowers than uninvaded plots from mid-August to early September (10 August to 7 Sept.; Fig. 1). After 31 July, over 95% of the flowers in the invaded plots were from L. vulgaris (Fig. 1). We found no difference in the total richness (L. vulgaris included) of flowering species between uninvaded and invaded plots across the season (F1,20 = 1.87, P = 0.19; Fig. 2a). But, when we excluded L. vulgaris from the analysis, we found more resident flowering species across the season in uninvaded compared to invaded plots (F1,20 = 5.03, P = 0.04), indicating that L. vulgaris displaced at least one resident species (Fig. 2a). However, the same resident species was not displaced across all plots. For flowering species evenness, we found a significant interaction between invasion status and day when L. vulgaris was included (F38,660 = 3.36, P \ 0.0001). From mid June to late July (10 June–29 July), flowering species evenness was 22% higher in invaded plots compared to uninvaded plots, but from late July to early September (31 July–9 Sept), flowering species evenness was 86% higher in uninvaded compared to invaded plots (Fig. 2b). When excluding L. vulgaris from the analysis, we also found a significant interaction between invasion status and day (F38,660 = 1.79, P = 0.003). From 31 July to 9 Sept, flowering species evenness was 39% higher in uninvaded plots relative to invaded plots when L. vulgaris was excluded from the analysis (Fig. 2b). Individual species level Uninvaded plots contained an average of 2.3 times more P. pulcherrima flowers than invaded plots (F1,16 = 3.87, P = 0.067; Fig. 3), with no main effect of plant removal (Supplementary Electronic Appendix 1). We found significant effects of day (F28,341 = 18.89, P \ 0.0001) and an interaction between day by invasion status (F26,341 = 1.65, P = 0.03) for P. pulcherrima flowering. We also found a significant interaction between day and plant

123

2368

B. J. Wilke, R. E. Irwin

1400

*

Invaded included Invaded(L. (L.vulgaris) vulgaris included)

Total flowers plot -1

1200 1000

Invaded vulgaris)excluded) excluded Invaded(L. (L vulgaris Uninvaded Uninvaded

** *

*

* * **

800 600

** * * *

400

** * *

200 0 6/10 6/15 6/20 6/25 6/30 7/5 7/10 7/15 7/20 7/25 7/30 8/4 8/9 8/14 8/19 8/24 8/29 9/3 9/8

Date Fig. 1 Mean flower abundance per plot is plotted throughout the flowering season for invaded and uninvaded plots. Invaded plots including and excluding L. vulgaris flowers are plotted separately to depict the proportion of flowers produced by L. vulgaris relative to the whole community. Triangles represent invaded plots (including L. vulgaris), squares represent uninvaded plots, and open circles represent invaded plots with

(a)

7

Flowering species richness

Fig. 2 The richness (a) and evenness (b) of flowering species are plotted throughout the flowering season for invaded (L. vulgaris included or excluded) and uninvaded plots. Invaded plots including and excluding L. vulgaris flowers are plotted separately. Symbols and asterisks as in Fig. 1

exclusion of L. vulgaris from the analysis (here and Fig. 2). Days in which the invaded (including L. vulgaris) and uninvaded plots differed significantly in flower abundance are marked with an asterisk. The arrows represent peak flowering dates for invaded plots (including and excluding L. vulgaris) and uninvaded, plots

6 5 4 3 2

Invaded(L.(L. vulgaris) included Invaded vulgaris included) Invaded(L.(L.vulgaris ) excluded Invaded vulgaris excluded)

1

Uninvaded Uninvaded

0

Flowering species evenness

(b) 0.8 0.7

**

0.6 0.5 0.4

**

*** *** * *

0.3 0.2 0.1 0 6/10 6/15 6/20 6/25 6/30 7/5 7/10 7/15 7/20 7/25 7/30 8/4

Date

123

8/9 8/14 8/19 8/24 8/29 9/3

9/8

Variation in the phenology and abundance of flowering

2369

removal (F27,341 = 2.47, P = 0.0001). Comparisons between uninvaded and invaded plots by day showed significantly more P. pulcherrima flowers in uninvaded compared to invaded plots from late June to mid-July (26 June–20 July; Fig. 3). The timing of peak P. pulcherrima flowering did not differ between invaded and uninvaded plots, but uninvaded plots contained an average of 133% more P. pulcherrima flowers at peak flowering than invaded plots (F1,16 = 4.13, P = 0.059). On a per stem basis, neither the presence of L. vulgaris nor the plant removal treatment affected P. pulcherrima flowers per stem (Supplementary Electronic Appendix 1; Table 1). However, there were two times more P. pulcherrima stems in uninvaded than invaded sites (F1,16 = 3.60, P = 0.076), and duration of P. pulcherrima flowering in each plot was an average of 7.4 days longer in uninvaded compared to invaded plots (F1,16 = 4.94, P = 0.04; Table 1).

Flowers per stem

Flowering duration (days)

Invaded

2.01 (0.15)

50.5 (2.07)

Uninvaded

2.20 (0.11)

57.9 (2.55)

Invaded removal

2.26 (0.14)

56.2 (2.31)

Uninvaded removal

1.96 (0.11)

52.2 (2.75)

Data shown are means with standard errors in parentheses

L. vulgaris percent cover (b ± SE = -4.09 ± 1.40; F1,10 = 8.57, P = 0.02), as we predicted. Finally, later in the season when L. vulgaris was flowering (10 August), we found a significant positive relationship between L. vulgaris percent cover and resident flower abundance (b ± SE = 3.06 ± 0.17; F1,10 = 271.17, P \ 0.0001).

Discussion

Is Linaria vulgaris cover related to resident flower abundance? The relationship between L. vulgaris percent cover and resident floral abundance in the invaded plots was not consistent across the flowering season. Early in the season (18 June), we found no significant relationship between L. vulgaris percent cover and resident flower abundance (F1,10 = 1.85, P = 0.20). However, in the middle of the flowering season (15 July), resident flower abundance was negatively related to

This study extends our understanding of the array of effects that invasive plants can have in their new range, in this case via changes in the phenology and abundance of community- and species-level flowering. The presence of the invasive plant Linaria vulgaris was associated with impacts on flowering patterns by the entire flowering community, partially supporting our initial predictions. Invaded plots exhibited a temporal delay in peak flowering, with more total

300

*

Invaded 250

P. pulcherrima flowers plot-1

Fig. 3 Potentilla pulcherrima flower abundance is plotted across the flowering season. Triangles represent invaded plots while squares represent uninvaded plots. Asterisks as in Fig. 1

Table 1 Flowers per stem and flowering duration for Potentilla pulcherrima

Uninvaded Native

* *

200

*

150

100

* **

** *

*

50

0 6/10 6/15 6/20 6/25 6/30 7/5

7/10 7/15 7/20 7/25 7/30 8/4

8/9

8/14 8/19 8/24

Date

123

2370

flowers produced later in the season. Over 95% of the flowers at peak flowering were those of L. vulgaris. Prior to the onset of L. vulgaris flowering, uninvaded plots maintained higher total floral abundance than invaded plots for nearly a two week period, but this pattern quickly reversed when L. vulgaris came into bloom. Although the effects of invasive plants on patterns of species and community flowering phenology and abundance are not well explored, there is some evidence to suggest that the effects of L. vulgaris exhibited here are not atypical. Brown et al. (2002) highlight an example where an invasive species produced more flowers on a per plant basis than a native congener. Thus, if flower-rich invasive plants grow at similar or higher densities than natives, invaded sites may exhibit higher flower production than native sites. Overall, invaded plots had lower flower production of resident species than uninvaded plots, and peak flowering of resident species occurred earlier in invaded plots. However, it is important to note that the periods of increased resident floral abundance in invaded and uninvaded plots overlapped, even though the specific date of peak flowering differed between the two (Fig. 1). Moreover, it was hard to denote a ‘peak’ in flowering of resident species in invaded plots given that the floral abundance of residents was so low. We are unaware of studies, including this one, that have found that invasive plants definitely alter the phenology of native species in invaded relative to uninvaded sites. However, given well-known effects of resource availability on flowering phenology (Holway and Ward 1965; Jackson 1966; Kochmer and Handel 1986), invaders may commonly be associated with plastic and potentially adaptive responses in resident flowering phenology. The presence of L. vulgaris was associated with a reduction in resident flowering richness (by displacing on average one resident species, although not the same resident species across all plots) and evenness, due to the dominance of L. vulgaris when in flower. From these results, we can strongly conclude that L. vulgaris alters the pattern of community flower abundance and phenology by providing a large spike in invader floral resources at the end of the flowering season, significantly later than the peak in resident flower abundance. Likewise, previous studies have shown that invasive plants can reduce native flowering richness in invaded

123

B. J. Wilke, R. E. Irwin

sites. For example, on islands, plant invasions have been implicated in reduced angiosperm richness (Cox and Elmqvist 2000). In addition, lower evenness of resident flowering species in invaded plots during the latter part of the season suggests that L. vulgaris may change the structure of the flowering community above and beyond its effects of being present and replacing resident species. At the individual-species level, uninvaded plots contained two times more P. pulcherrima flowering stems and extended flowering duration than invaded plots. This result suggests that L. vulgaris does not simply displace some of the subordinate species in a flowering community, but is also associated with effects on the flowering duration and abundance of a common native species. From these results, we cannot differentiate whether the change in flowering duration was due to asynchronous flowering among stems or greater duration of individual plant flowering, perhaps because the plants were more robust. Additional observational work would tease apart the mechanisms involved. Variation in individual-species and community flower phenology and abundance associated with plant invasions is likely to have both direct and indirect effects on native plants and animals (White et al. 2006). We highlight three potential effects here. First, reductions in native flower production at the plot level could lead to reduced native seed set (given that flower and seed number are often positively correlated, e.g., Brody and Mitchell 1997) and seedling recruitment (assuming a positive correlation between seed number and seedling recruitment, Clark et al. 2007), which could affect native plant demography. Second, variation in the timing and abundance of flowers may affect interactions among plant species via shared mutualists and antagonists (reviewed in Strauss and Irwin 2004; Bjerknes et al. 2007). The outcome of these multi-species interactions will depend on the degree to which flowering species are relatively rare versus common and the degree to which they compete for vs. facilitate mutualisms and antagonisms (e.g., Feinsinger 1987; Ehrle´n et al. 2002; Tora¨ng et al. 2006). Third, lower evenness of flowering species may affect native insects that rely on flowers for food. Floral richness and evenness may be necessary for the creation and maintenance of a greater range of niches for multiple pollinator functional groups (Potts et al. 2003;

Variation in the phenology and abundance of flowering

Fenster et al. 2004), although we cannot rule out the possibility that increased flowers (and rewards) in invaded vs. uninvaded sites may benefit pollinators (Tepedino et al. 2008), just as highly rewarding flowering crops benefit some pollinators (Westphal et al. 2003). How changes in community-level flowering relate to community-level nectar and pollen reward structure and affects on pollinator performance in invaded vs. uninvaded sites is unknown and warrants further research. Late in the season when L. vulgaris was in full bloom, we observed a positive relationship between L. vulgaris percent cover and resident floral abundance. Three relevant hypotheses to explain this unexpected result are that (1) because we didn’t manipulate L. vulgaris, its presence in the invaded sites may be correlated with some unmeasured factor that positively affected resident species richness or abundance later in the season, (2) L. vulgaris preferentially invades into sites that harbor more lateblooming species, or (3) the presence of L. vulgaris modifies the environment and facilitates some lateblooming species. Rigorously investigating each of these hypotheses requires experimental tests. One caveat should be considered when interpreting the results of this study. We did not experimentally add L. vulgaris into sites. Instead, we relied on a biomass removal treatment to remove L. vulgaris and then controlled for biomass removal by removing resident biomass from uninvaded sites. Biomass removal associated with the eradication of an exotic species can sometimes cause dramatic changes in ecosystem function (Zavaleta et al. 2001), but removal of biomass in this study did not influence flowering phenology or abundance even though biomass removal likely involved increased light opening and disturbance. Biomass removal of L. vulgaris or residents in the late summer had little impact on the following year’s flowering patterns, thus limiting the effects that this treatment had on flowering for most variables. The extensive root system and clonal abilities of L. vulgaris likely prevented our plot scale removal efforts from impacting the vigor of L. vulgaris genets in the study areas. We chose not to remove biomass earlier in the season because of potential significant consequences on long-term resident plant population dynamics. However, because our invaded and uninvaded plots spanned a number of microsites and there was no

2371

relationship between invasion status and site aspect, soil pH, or soil organic matter (R. E. Irwin, unpublished data), it seems unlikely that some unknown factor was spuriously correlated with L. vulgaris invasion across all of the invaded plots. Nonetheless, to test the patterns we observed more fully would require working across a larger geographic area or in a common garden or greenhouse to simulate invasion. It is well documented that exotic species impact the structure and function of ecosystems worldwide (Vitousek 1992). This study adds to the collection of literature supporting effects of invaders on native ecosystems, but documents a different mechanism by which invasive species may alter ecosystem structure and function, namely via changes in individualspecies or community flowering phenology and abundance. Given the widespread abundance of invasive plants in some ecosystems, this may be a common way in which invaders affect native systems. The next step is to understand the population, community, and ecosystem consequences of changes in individual species or community flowering phenology and abundance following plant invasions and to experimentally assess the mechanisms involved. Acknowledgments We thank those who aided in data collection and species identification, including E. Bruneau, L. Burkle, B. Frase, S. Elliot, M. Erhart, K. Ritter, E. Schuett, B. Krueger, and M. London. We also thank L. Burkle, K. Button, R. Calsbeek, R. Cox, S. Elliott, G. Mittelbach, J. Mellard, T. Robinson, S. Syswerda, and L. Walters for providing comments on earlier drafts of the manuscript and S. Kravchenko for statistical assistance. The RMBL provided access to facilities and field sites. Funding was provided by the National Science Foundation (DEB-0089643 and DEB-0455348). This is the W. K. Kellogg Biological Station contribution #1544. We thank two anonymous reviewers and the editor of Biological Invasions for substantial guidance during the review process.

References Arnold RM (1982) Pollination, predation and seed set in Linaria vulgaris (Scrophulariaceae). Am Midl Nat 107:360–369 Augsperger CK (1980) Mass flowering in a tropical shrub (Hybanthus prunifolius): influences on pollinator attraction and movement. Evolution 34:475–488 Bishop JG, Schemske DW (1998) Variation in flowering phenology and its consequences for lupines colonizing Mount St. Helens. Ecology 79:534–546 Bjerknes A-L, Totland O, Hegland SJ, Nielsen A (2007) Do alien plant invasions really affect pollination success in native plant species? Biol Conserv 138:1–12

123

2372 Brody AK (1997) Effects of pollinators, herbivores, and seed predators on flowering phenology. Ecology 78:1624–1631 Brody AK, Mitchell RJ (1997) Effects of experimental manipulation of inflorescence size on pollination and predispersal seed predation in the hummingbird pollinated plant Ipomopsis aggregata. Oecologia 110:86–93 Brown BJ, Mitchell RJ, Graham SA (2002) Competition for pollination between an invasive species (Purple loosestrife) and a native congener. Ecology 83:2328–2336 Burkle LA (2008) Bottom-up effects of nutrient enrichment on plants, pollinators, and their interactions. Dissertation, Dartmouth College, Hanover, NH, USA Carpenter A, Murray T (1998) Element stewardship abstract for Linaria genistifolia and Linaria vulgaris. In: vol 2003. The Nature Conservancy Clark CJ, Poulsen JR, Levey DJ, Osenberg CW (2007) Are plant populations seed limited? A critique and meta-analysis of seed addition experiments. Am Nat 170:128–142 Cox PA, Elmqvist T (2000) Pollinator extinction in the Pacific Islands. Conserv Biol 14:1237–1239 D’Antonio CM, Hughes RF, Mack M, Hitchcock D, Vitousek PM (1998) The response of native species to removal of invasive exotic grasses in a seasonally dry Hawaiian woodland. J Veg Sci 9:699–712 Docherty Z (1982) Self-incompatibility in Linaria. Heredity 49:349–352 ˚ gren J (2002) Pollen limitation, seed preEhrle´n J, Kack S, A dation and scape length in Primula farinosa. Oikos 97: 45–51 Feinsinger P (1987) Effects of plant species on each other’s pollination: is community structure influenced? Trends Ecol Evol 2:123–126 Fenster CB, Armbruster WS, Wilson P, Dudash MR, Thomson JD (2004) Pollination syndromes and floral specialization. Annu Rev Ecol Evol Syst 35:375–403 Godoy O, Richardson DM, Valladares F, Castro-Dı´ez P (2009) Flowering phenology of invasive alien plant species compared with native species in three Mediterranean-type ecosystems. Ann Bot 103:485–494 Gross RS, Werner PA (1983) Relationships among flowering phenology, insect visitors, and seed-set of individuals: experimental studies on four co-occurring species of goldenrod (Solidago: Compositae). Ecol Monogr 53: 95–117 Holway JG, Ward RT (1965) Phenology of alpine plants in Northern Colorado. Ecology 46:73–83 Inouye DW, Dodge GJ, Morales MA (2002) Variation in timing and abundance of flowering by Delphinium barbeyi huth (Ranunculaceae): the roles of snowpack, frost, and La Nina, in the context of climate change. Oecologia 130:543–550 Jackson MT (1966) Effects of microclimate on spring flowering phenology. Ecology 47:407–415 Kadrmas T, Johnson WS (2002) Managing yellow and dalmation toadflax. In: University of Nevada Cooperative Extension Kochmer JP, Handel SN (1986) Constraints and competition in the evolution of flowering phenology. Ecol Monogr 56:303–325 Lajeunesse S (1999) Dalmation and yellow toadflax. In: Petroff JK (ed) Biology and management of noxious rangeland

123

B. J. Wilke, R. E. Irwin weeds. Oregon State University Press, Corvallis, pp 202– 216 McCall AC, Irwin RE (2006) Florivory: the intersection of pollination and herbivory. Ecol Lett 9:1351–1365 Pauchard A, Alaback PB, Edlund EG (2003) Plant invasions in protected areas at multiple scales: Linaria vulgaris (Scrophulariaceae) in the west Yellowstone area. West N Am Nat 63:416–428 Pielou EC (1966) Measurement of diversity in different types of biological collections. J Theor Biol 13:131–144 Post ES, Pedersen C, Wilmers CC, Forchhammer MC (2008) Phenological sequences reveal aggregate life history response to climate warming. Ecology 89:363–370 Potts SG, Vulliamy B, Dafni A, Ne’eman G, Willmer P (2003) Linking bees and flowers: how do floral communities structure pollinator communities? Ecology 84:2628–2642 Primack RB (1987) Relationships among flowers, fruits and seeds. Annu Rev Ecol Syst 18:409–430 Rathcke B, Lacey EP (1985) Phenological patterns of terrestrial plants. Annu Rev Ecol Syst 16:179–214 Ridenour WM, Callaway RM (2001) The relative importance of allelopathy in interference: the effects of an invasive weed on a native bunchgrass. Oecologia 126:444–450 Saner MA, Clements DR, Hall MR, Doohan DJ, Crompton CW (1995) The biology of Canadian weeds. 105. Linaria vulgaris Mill. Can J Plant Sci 75:525–537 SAS Institute (2004) SAS user’s guide. In: 9.1 edn. SAS Inst., Cary Stinson KA (2004) Natural selection favors rapid reproductive phenology in Potentilla pulcherrima (Rosaceae) at opposite ends of a subalpine snowmelt gradient. Am J Bot 91:531–539 Strauss SY, Irwin RE (2004) Ecological and evolutionary consequences of multispecies plant-animal interactions. Annu Rev Ecol Evol Syst 35:435–466 Tepedino VJ, Bradley BA, Griswold TL (2008) Might flowers of invasive plants increase native bee carrying capacity? Intimations from Capitol Reef National Park, Utah. Nat Areas J 28:44–50 ˚ gren J (2006) Facilitation in an insect Tora¨ng P, Ehrle´n J, A pollinated herb with a floral display dimorphism. Ecology 87:2113–2117 Totland O (1999) Effects of temperature on performance and phenotypic selection on plant traits in alpine Ranunculus acris. Oecologia 120:242–251 Vitousek PM (1992) Global environmental change: an introduction. Annu Rev Ecol Syst 23:1–14 Vitousek PM (1994) Beyond global warming: ecology and global change. Ecology 75:1861–1876 Waser NM (1978) Competition for hummingbird pollination and sequential flowering in two Colorado wildflowers. Ecology 59:934–944 Westphal C, Steffan-Dewenter I, Tscharntke T (2003) Mass flowering crops enhance pollinator densities at a landscape scale. Ecol Lett 6:961–965 White EM, Wilson JC, Clarke AR (2006) Biotic indirect effects: a neglected concept in invasion biology. Divers Distrib 12:443–455 Zavaleta ES, Hobbs RJ, Mooney HA (2001) Viewing invasive species removal in a whole-ecosystem context. Trends Ecol Evol 16:454–459