Oecologia

, Volume 179, Issue 2, pp 495–507 | Cite as

Effects of non-native Melilotus albus on pollination and reproduction in two boreal shrubs

  • Katie V. Spellman
  • Laura C. Schneller
  • Christa P. H. Mulder
  • Matthew L. Carlson
Plant-microbe-animal interactions – original research

Abstract

The establishment of abundantly flowered, highly rewarding non-native plant species is expected to have strong consequences for native plants through altered pollination services, particularly in boreal forest where the flowering season is short and the pollinator pool is small. In 18 boreal forest sites, we added flowering Melilotus albus to some sites and left some sites as controls in 2 different years to test if the invasive plant influences the pollination and reproductive success of two co-flowering ericaceous species: Vaccinium vitis-idaea and Rhododendron groenlandicum. We found that M. albus increased the pollinator diversity and tended to increase visitation rates to the focal native plant species compared to control sites. Melilotus albus facilitated greater seed production per berry in V. vitis-idaea when we added 120 plants compared to when we added 40 plants or in control sites. In R. groenlandicum, increasing numbers of M. albus inflorescences lowered conspecific pollen loads and percentage of flowers pollinated; however, no differences in fruit set were detected. The number of M. albus inflorescences had greater importance in explaining R. groenlandicum pollination compared to other environmental variables such as weather and number of native flowers, and had greater importance in lower quality black spruce sites than in mixed deciduous and white spruce sites for explaining the percentage of V. vitis-idaea flowers pollinated. Our data suggest that the identity of new pollinators attracted to the invaded sites, degree of shared pollinators between invasive and native species, and variation in resource limitation among sites are likely determining factors in the reproductive responses of boreal native plants in the presence of an invasive.

Keywords

Fruit set Invasive species Ledum palustre ssp. groenlandicum Seed set Rhododendron groenlandicum Vaccinium vitis-idaea 

Introduction

Non-native plants that invade flowering plant communities can have diverse effects on the reproductive success of native plants. Competitive effects of non-native plant introductions on pollination of native species occur through two often co-occurring mechanisms: (a) decreases in pollen quantity, and (b) decreases in pollen quality (Waser 1978). Non-native plant invasions can reduce pollen quantity delivered to native plants by decreasing visitation rates to the native species as a result of pollinator preference for the invasive flowers (Waser 1983; Campbell 1985; Brown et al. 2002; Kandori et al. 2009). Non-native species can decrease pollen quality if they increase the amount of self- or heterospecific pollen being delivered (Morales and Traveset 2008). Conversely, some studies have demonstrated that the presence of invasive plants increased overall pollinator abundances or visitation rates to the entire community, which increased pollen quantity, fruit yield and seed production (Moragues and Traveset 2005; Tepedino et al. 2008). Other studies have found no measurable impacts of invasive plants on the pollination and reproduction of native plants (Bartomeus et al. 2008a). The range of effects invasive plants can have on the pollination of native plants suggests that unique attributes within the plant and pollinator communities can have a large influence on the impact of the invasive plants.

The abundance of invasive plant flowers within the flowering plant community at a site changes pollen flow to the native plants and subsequent reproductive success (Muñoz and Cavieres 2008; Molina-Montenegro et al. 2008; Flanagan et al. 2010). The relative influence an invasive plant has on the reproduction of native plants may also vary with other factors that would be expected to affect plant reproduction, such as habitat type, weather, and inter- or intra-specific competition with other native flowers. Pollinator visitation rates can differ among different habitat types due to differences in conspecific and heterospecific flower abundances, amount of shade, or availability of pollinator nesting sites (Westrich 1996; Gathmann and Tscharntke 2002; Westphal et al. 2003; Bartomeus et al. 2010). Differences in weather between years or between sites can change pollination services by influencing the time available for pollinator flight activity (Kuchko 1988; Tuell and Isaacs 2010), by affecting which types of pollinators are active (Corbett et al. 1993), and by directly affecting resources for flower and fruit growth and maintenance (Jacquemart 1997; Krebs et al. 2009).

Relative to other forest types, boreal forest ecosystems tend to have fewer flowering species, smaller pollinator pools, and shorter flowering periods (Kevan et al. 1993). These factors could intensify the potential negative or positive impacts of non-native plant invasions on the reproductive success of neighboring native plant species (cf. Carlson et al. 2008). Despite the fact that the boreal forest is one of the largest terrestrial biomes on Earth [a third of the world’s forested land (Shugart et al. 1992)], to date we could find only a single study on the impact of invasive plants on the pollination and reproductive success of native plants in the boreal forest that has been published in English in the peer-reviewed literature (see Totland et al. 2006). This single study found that the experimental outplanting of non-native Phacelia tanacetifolia strongly decreased pollinator visitation to a native boreal plant species (Melampyrum pratense) but did not change fruit set or seed production (Totland et al. 2006). High resource limitations on plant reproduction in boreal forest habitats may explain why the substantial change in pollinator visitation to the native plant did not lead to a change in reproductive success. Without stigmatic pollen load data, however, the Totland et al. (2006) study could not disentangle the relative influence of pollen limitation and resource limitation on plant reproduction in their field sites.

Compared to other places, low levels of anthropogenic disturbance and cold climate have limited the introduction and survival rates of non-native plants in the boreal forest (Sanderson et al. 2012). However, the number of non-native species occurring within Alaska increased by 46 % between 1941 and 2006 (Carlson and Shephard 2007). Increases in the number and extent of non-native species in Alaska may be attributed in large part to increases in human population and associated disturbances [e.g., more roads, resource extraction (Walker and Walker 1991; US Census Bureau 2010; Carlson and Shephard 2007)] and increased influx of propagules via imported agricultural and horticultural commodities (Conn et al. 2008a). Climatic shifts in Alaska such as warmer winters (Serreze et al. 2000) and longer growing seasons (Myneni et al. 1997) have also increased the likelihood of invasive plant success.

A few species, such as Melilotus albus Medik., have spread rapidly throughout the state, primarily along road corridors (AKEPIC 2014). M. albus is one of the few non-native species in Alaska that has also spread widely into naturally disturbed areas such as river floodplains (Conn et al. 2008b; Spellman and Wurtz 2011) and wildfire scars (Spellman et al. 2014). M. albus is native to Eurasia and was introduced to Alaska in 1913 as a potential cold-hardy forage and nitrogen-fixing crop (Irwin 1945), and now occurs throughout Alaska from as far south as Metlakatla (55.122ºN, −131.561ºW) to north of Coldfoot (67.286ºN, −150.171ºW) at the base of the Brooks Mountain Range (AKEPIC 2014). M. albus can reduce native seedling recruitment along glacial river floodplains by directly competing with native plants for light (Spellman and Wurtz 2011). Additionally, this species offers considerable nectar and pollen resources to floral visitors (Peterson 1989; Malacalza et al. 2005; Tepedino et al. 2008) with an extremely high number of flowers per plant [up to 350,000 flowers per plant (Royer and Dickinson 1999)], particularly in comparison to native boreal insect-pollinated plants that offer less pollen and nectar rewards. As a result, M. albus invasions could also alter plant communities by changing the pollination and reproductive success of native boreal plants.

In many of the instances of M. albus invasion documented in or adjacent to burned boreal forest (Villano and Mulder 2008), the understory is dominated by Vaccinium vitis-idaea L. (lingonberry or lowbush cranberry) and Rhododendron groenlandicum (Oeder) Kron and Judd (formerly Ledum palustre ssp. groelandicum; Labrador tea), two abundant insect-pollinated Ericaceous plant species that have broad circumboreal distributions (Hultén 1968). These species are of cultural, subsistence, and economic importance (Garibaldi 1999; Quiner 2005; Holloway 2006; Nelson et al. 2008). Because both species overlap with M. albus in habitat (Villano and Mulder 2008), flowering times (personal observation), and pollinator communities (Turkington et al. 1978; Eckardt 1987; Davis et al. 2003; Dlusski et al. 2005; Tepedino et al. 2008), we chose to focus on V. vitis-idaea and R. groenlandicum in this study. Within interior Alaska, bumblebees (Bombus spp.), syrphid flies (Syrphidae), and solitary bees (Adrena spp.) are the pollinators that carry the greatest amount of V. vitis-idaea pollen (Davis et al. 2003), but other pollinator guilds carry its pollen as well [e.g., Lepidopterans, other flies, beetles (Davis 2002)]. R. groenlandicum is visited by pollinators in all the aforementioned guilds (personal observation). M. albus has generalist flowers visited by a wide range of species, including solitary bees, bumblebees, wasps, flies, butterflies, and moths (Coe and Martin 1920; Turkington et al. 1978; Tepedino et al. 2008). V. vitis-idaea and R. groenlandicum are self-compatible (Jacquemart and Thompson 1996; Jacquemart 1997; Wheelwright et al. 2006). Both species, however, have decreased fruit and seed set when insect pollinators are excluded and increase fruit set and seed production when they are supplemented with outcross pollen (Hall and Beil 1970; Fröborg 1996; Jacquemart and Thompson 1996; Davis 2002; Wheelwright et al. 2006). This was confirmed in interior Alaska, where flowers from which pollinators were excluded showed a 79 and 18 % reduction in fruit set for V. vitis-idaea and R. groenlandicum, respectively, compared to flowers open to insect pollination (C. P. H. Mulder and K. V. Spellman, unpublished data).

We conducted a preliminary observational study during the summer of 2010 to compare insect pollinator visitation to native plants and V. vitis-idaea fruit set at sites with and without M. albus along the roadside. These sites were located throughout interior Alaska, along the Steese, Elliot, and Dalton highways. The abundance of insect pollinators observed at sites with flowering M. albus was approximately two times higher than at sites without M. albus (L. C. Schneller, unpublished data). The sites with M. albus also had a greater proportion of V. vitis-idaea flowers setting fruit compared to sites without M. albus present (49 ± 20 % in sites with M. albus, 16 ± 7 % in sites without M. albus; C. P. H. Mulder, unpublished data). However, we could not attribute these changes in pollinator activity and V. vitis-idaea fruit set to the presence of M. albus, as site conditions that favor M. albus establishment may also favor higher pollinator activity, and promote greater abundance of native flowers and greater fruit set. To disentangle potential confounding environmental effects, we conducted a controlled M. albus addition experiment, which we report here. Specifically, we ask these questions:

  1. 1.

    Does M. albus addition alter V. vitis-idaea and R. groenlandicum pollination and reproduction?

     
  2. 2.

    Does the abundance of M. albus vary the effect it has on pollination and reproduction of these native plants?

     
  3. 3.

    How important is the influence of M. albus relative to other factors expected to influence native plant reproduction such as weather and number of native flowers?

     

Materials and methods

Study area

During the growing seasons in 2011 and 2012, we located boreal forest sites within the Bonanza Creek Boreal Long-Term Ecological Research Program research areas near Fairbanks, Alaska (Bonanza Creek Experimental Forest, 64.709ºN, −148.326ºW, and Caribou and Poker Creeks Research Watershed, 65.141ºN, −147.457ºW). Sites were selected to contain flowering V. vitis-idaea and R. groenlandicum, and primarily occurred in two habitat types: (a) “mixed” sites that contain deciduous tree species (Betula neoalaskana Sarg. and/or Populus tremuloides Michx.) and white spruce (Picea glauca (Moench) Voss); and (b) black spruce (Picea mariana Mill.) sites (Appendix 1). The mixed sites tended to occur on gentle hill slopes (3–10 % grade) or at the tops of hills with understory vegetation composed primarily of the two focal species, Vaccinium uliginosum, Rosa acicularis, Viburnum edule, Salix spp., Alnus viridis, Geocaulon lividum, and Cornus canadensis. The black spruce sites occurred in low-lying areas with minimal slope (0–3 % grade) and understory vegetation composed primarily of the two focal species, Vaccinium uliginosum, Rubus chamaemorus, Salix spp., and moss species. The mixed deciduous and white spruce stands had greater average canopy cover than the black spruce sites (56 % in mixed deciduous-spruce sites and 17 % in black spruce sites) and greater abundances of native flowers (~1.8 times the number of flowers per square meter).

Experimental design

In 2011, we selected 17 sites placed greater than 300 m apart to minimize pollinator movement between sites. Nine sites were in mixed deciduous and spruce sites and eight were in black spruce forest sites. Sites were circular and extended 40 m in all directions from the site center. M. albus did not occur at any of these sites. Eleven sites were randomly assigned to have 40 greenhouse-grown flowering second-year M. albus plants added to the site center (Mel 40), and six were control sites (no M. albus added). The sites contained one or both of the focal native species, with 16 sites containing V. vitis-idaea and 15 sites containing R. groenlandicum (Table 1).
Table 1

Number of sites for each species and in different habitat types in 2011 and 2012 by treatment

Species

Habitat type

Year and treatment

2011

2012

Control

Mel 40

Control

Mel 40

Mel 120

Vaccinium vitis-idaea

Mixed

4

5

4

2

5

Black spruce

2

5

2

4

1

Total

6

10

6

6

6

Rhododendron groenlandicum

Mixed

2

5

2

2

3

Black spruce

2

6

2

4

1

Total

4

11

4

6

4

MEL 40 forty M. albus plants added, MEL 120 one hundred and twenty M. albus plants added

In 2012, we discontinued the site that did not have V. vitis-idaea and added two new sites to bring the total number of sites to 18. To address the influence of invasive plant patch size on the reproductive success of our focal species, we added a higher M. albus addition level to our design in 2012 (120 plants added; Mel 120). The 18 sites were allocated to each of the three treatment levels: control, Mel 40, and Mel 120 (six sites each). To compare years directly, we retained the same treatments in six of the sites (three control and three Mel 40 sites). The remaining three control sites, three Mel 40 sites and six Mel 120 sites were randomly assigned. We assigned sites without respect to habitat type, but had multiple sites of each treatment in each with the exception of a single black spruce site that received the Mel 120 treatment (Table 1).

M. albus was added to the Mel 40 or Mel 120 sites at the time that V. vitis-idaea and R. groenlandicum flower buds emerged, but had not yet opened. M. albus were grown in the greenhouse in conetainer pots (7 cm diameter at the top, 22 cm in length); each pot contained one individual with five to 181 inflorescences [mean (±SE) of 49 ± 18 flowers per inflorescence]. Either 40 or 120 pots were placed in the center of the site in holes of similar dimensions so that the top of each pot was flush with the ground surface. M. albus density was 15 plants m−2, resulting in circular patches approximately 2.6 and 8 m2 in size. The range in number of inflorescences added to each site was 334–942 (16,366–46,158 total flowers) for Mel 40 sites and 1068–1608 (52,332–78,792 total flowers) for Mel 120 sites. The addition levels we used (Mel 40 and Mel 120) were comparable to the patch sizes and stem densities found within burned boreal forest in interior Alaska, which are typically in the earliest of invasion stages where they occur (Villano and Mulder 2008). Once flowers of focal native species had dropped their petals (18–28 days after M. albus addition), M. albus was removed from the sites. To prevent accidental introductions, we removed any immature seeds that appeared on the M. albus plants throughout the duration of the experiment. Sites were also visited a year following the experiments to confirm that no M. albus plants were present.

Within each site, 25 circular plots were established for each of the occurring focal species (1-m2 plot for V. vitis-idaea, and 1.77-m2 plot for R. groenlandicum) ranging from 1 to 40 m from the site center. Five plots were placed within five distance ranges from the site center: 1–2, 3–5, 8–10, 15–20 and 25–40 m. Within these plots, five V. vitis-idaea or five R. groenlandicum ramets were marked for tracking fruit set and seed production. In the 1- to 2-m distance category, focal plants were always selected outside of the M. albus patch to avoid plants where the root systems may have been damaged during the M. albus transplanting. This study focuses on whole-site impacts of M. albus on focal native species reproduction, and spatial variation of the effects within sites will be discussed in a forthcoming paper.

Pollinator activity and community

In 2011, we observed insect pollinator activity in 15 of the 17 sites (four control sites and 11 Mel 40 sites). We did not observe pollinator activity in 2012. Pollinator observations occurred between 8 a.m. and 6 p.m. during calm, rain-free periods from 6 to 18 June 2011. One focal plot per distance category was randomly chosen for a 2-m × 2-m pollinator observation of 15 min, for a total of five observations per site. For each observation, we counted the total number of open flowers and then recorded pollinator landings on open flowers of focal species within the plot. Observed pollinators were grouped into categories (butterflies, wasps, bumblebees, solitary bees, syrphid flies, and non-syrphid flies) for field identification. We calculated visitation rates using insect landings per number of flowers per hour of observation in each plot within each of the 15 sites used for observations (12 sites with V. vitis-idaea present and 14 sites with R. groenlandicum present). We used four pairs of sites (the four control sites each paired with a Mel 40 site that was observed on the same day) to assess differences in the pollinator community composition between the treatments. Due to the overall low number of pollinator sightings in these sites, we pooled observations in the four control sites and the four Mel 40 sites to calculate Simpson’s diversity index and proportional similarity (Brower and Zar 1984).

Pollination

To measure pollen deposition, we collected V. vitis-idaea and R. groenlandicum stigmas from randomly selected open flowers near each of our marked focal plants. We did not take stigmas from marked plants to avoid interfering with fruit set. Three (in 2011) or five (in 2012) V. vitis-idaea and five R. groenlandicum stigmas were collected from each of the 25 plots in each site ~14 days after the M. albus was added. The stigmas were mounted on microscope slides and stained with a basic fuchsin gel (Kearns and Inouye 1993). Each pollen tetrad (in the case of R. groenlandicum and V. vitis-idaea pollen) or pollen grain (in the case of other species) on the stigma was identified to genus (using anther vouchers we collected from all flowering species at the sites as a reference) and counted under a compound light microscope. The proportion of heterospecific pollen grains on the stigmas was low (3.2 ± 0.3 % for V. vitis-idaea and 0.8 ± 0.2 % for R. groenlandicum), so only conspecific pollen loads on the stigmas were used in the final analysis.

Flowers were considered to be “well pollinated” when they had ten or more pollen tetrads on the stigma. We selected this threshold because and fruit production was greatest when the pollen loads exceeded this pollen level for both focal species in control sites (Appendix 2). Further, the maximum seed number per fruit in V. vitis-idaea control plots across the 2 years of study was 47, indicating that greater than ten pollen tetrads was necessary for maximum seed production. Few stigmas had zero pollen grains on them, making the presence or absence of pollen inadequate for the detection of variation in the proportion of flowers that were pollinated.

Fruit set and seed production

We calculated percent fruit set as the percentage of flower buds on marked plants at the beginning of our experiment that produced fruit by the end of the growing season. To determine seed production per fruit in V. vitis-idaea we dissected up to five berries per marked plant and counted the number of seeds produced under a dissecting microscope. For R. groenlandicum, which has minute seeds that are released as the fruit ages and dries, we dried inflorescences at 65 °C until the fruits opened and released the seeds. The weight of the seeds was divided by the number of fruits on the inflorescences to derive seed mass per fruit.

Environmental covariates

We measured weather and vegetation variables that we expected to influence our response variables. Temperature and relative humidity (RH) during the experimental period (time of M. albus addition to time of M. albus removal in the site or nearest M. albus addition site for control sites) were obtained every 30 min using a HOBO-Pro data logger (Onset Computer, Cape Cod, MA) fixed 0.5 m above the ground surface at the center of each site. Number of hours of rain was estimated as number of hours with RH ≥ 100 %. Tree canopy cover was estimated for each of the 25 plots using a convex spherical crown densiometer (model A; Forest Densiometers, Bartlesville, OK) on the north and south edges of the plot. We visually estimated the percent shrub cover present above the V. vitis-idaea (up to 1 m in height) in each plot. R. groenlandicum was the tallest understory plant in the plots where it occurred, so shrub cover over this species was not a factor. To provide an estimate of flower abundance and richness, we counted the number of open flowers and flower buds present for each insect-pollinated species within the 25 plots at the time of M. albus addition.

Analysis

To test for differences in pollinator visitation rates between control and Mel 40 treatments, we used a non-parametric Wilcoxon rank sum test on site-level averages across five plots per site. All our other response variables were calculated as averages of 25 focal plant plots per site and met the assumptions of normality and constant variance. We performed the statistical analyses on the plant and environmental data using SAS version 9.1 (SAS Institute, Cary, NC) and the pollinator data using R version 2.14.2 (R Development Core Team 2012).

To determine the influence of the M. albus addition treatment and year on V. vitis-idaea and R. groenlandicum pollination, fruit set and seed production, we conducted multivariate ANOVA (MANOVA) using site-level means of response variables for V. vitis-idaea and R. groenlandicum. The response variables in the multivariate models included number of conspecific pollen grains delivered to stigmas (hereafter “conspecific pollen”), percent flowers receiving ten or more pollen grains or tetrads (hereafter “percent pollinated flowers”), percent flowers setting fruit (hereafter “percent fruit set”), and number of seeds per fruit for V. vitis-idaea or seed mass per fruit for R. groenlandicum (hereafter “seeds or seed mass per fruit”). We conducted several tests to disentangle treatment, year and site effects. We first tested for a treatment effect in each year individually. We then tested for year, treatment, and year by treatment interactions using both years but excluding the sites that received the Mel 120 treatment in 2012 (since this treatment was not imposed in 2011). Finally, we evaluated site and year effects using only sites where the treatment remained the same across both years of the experiment. Separate ANOVAs were then run to assess which individual response variable responded most strongly to treatment and year effects, and to allow us to disentangle the relative roles of pollen limitation and resource limitation on our focal species’ reproductive responses. This resulted in 32 ANOVA tests (4 response variables × 4 ANOVA tests × 2 species). We found six significant tests (Table 2) compared to the 1.6 expected by chance under α = 0.05, and believe we can interpret our results with only a small risk of committing a type I error. We also treated the abundance of M. albus inflorescences added to each site as a continuous variable and used linear regression to determine if the number of M. albus inflorescences present influenced our four pollination and reproduction response variables.
Table 2

ANOVA results for models testing for Melilotus albus addition treatment (Trt) effects for four individual response variables (total pollen on stigmas, percent flowers pollinated, percent flowers setting fruit, and number of seeds per fruit for V. vitis-idaea or seed mass per fruit for R. groenlandicum)

Focal species

Data set

Source of variation

df

Error df

Response variables

Conspecific pollen F

% Flowers pollinated F

% Fruit set F

Seeds or seed mass per fruit F

V. vitis-idaea

2011

Trt

1

14

0.01

1.04

0.03

0.13

2012

Trt

2

15

0.27

0.02

1.09

6.30**

2011 and 2012 excluding Mel 120

Year

1

24

0.01

1.20

12.06**

1.77

Trt

1

 

0.19

0.45

1.59

0.04

Year × Trt

1

 

0.28

0.71

1.39

0.19

Sites where Trt remains same in 2011 and 2012

Site

5

5

2.09

0.82

1.39

0.66

Year

1

 

0.36

0.06

7.67*

1.09

R. groenlandicum

2011

Trt

1

13

1.15

0.53

0.10

2.02

2012

Trt

2

11

1.30

2.00

0.41

1.53

2011 and 2012 excluding Mel 120

Year

1

21

1.17

1.31

0.12

24.58***

Trt

1

 

0.35

0.25

0.04

0.13

Year × Trt

1

 

0.54

0.26

0.04

4.57*

Sites where Trt remains same in 2011 and 2012

Site

4

4

2.50

1.74

7.01

6.88

Year

1

 

0.00

1.13

0.07

20.55*

* p < 0.05, ** p < 0.01, *** p < 0.001

To identify the relative importance of M. albus and the other environmental variables in explaining differences in pollination, fruit set, and seed production between sites, we calculated Akaike’s information criterion (AIC) variable importance values for seven or eight site-level abiotic and biotic covariates using multiple linear regression on the four response variables for each of the focal species. The covariates included the number of M. albus inflorescences added to site, percent tree canopy cover per plot, percent shrub cover per plot (for V. vitis-idaea only), number of conspecific flowers per plot, number of all flowers per plot, flower richness per plot, and mean temperature and number of hours of rain during the addition experiment at the site. We included all possible models and ranked them using AIC adjusted for small sample size (AICc) (Burnham and Anderson 2002), then calculated cumulative AICc weights (\(0 \le \mathop \sum \nolimits \omega_{i} \le 1\)), or importance values, for each biotic or abiotic variable (Burnham and Anderson 2002; Arnold 2010). We considered importance values >0.55 as indicative of well-supported variables. Average parameter estimates for each well-supported variable were calculated using the set of best-supported models (those within two AICc units of the model with the lowest AICc score) to assess the direction of the response of each focal species to the parameter. Since canopy cover was lower in black spruce sites than in mixed deciduous-spruce sites, and it had an important positive relationship with pollination and seed production variables in V. vitis-idaea (Table 4), we re-ran the multiple linear regression analysis described above separately for the two habitat types to determine the relative importance of M. albus in the two habitat types.

Results

Effects of M. albus on pollinator activity and community

V. vitis-idaea

Three out of four control sites without M. albus had no pollinator visitation to V. vitis-idaea (mean = 0.0001 visits/flower per hour), while Mel 40 treatment sites had a range of visitation rates between 0 visits/flower per hour and 0.46 visits/flower per hour (mean = 0.113 visits/flower per hour) (Fig. 1a). This effect was not statistically significant (W = 12.5, p = 0.544).
Fig. 1

Pollinator visitation rates to Vaccinium vitis-idaea (a) and Rhododendron groenlandicum (b) flowers in sites without Melilotus albus added (Control) and in sites with 40 M. albus plants added (Mel 40). Each box plot shows the 1st quartile, median (dark line), and 3rd quartile, and whiskers show the minimum and maximum value range for pollinator visitation rates. Minimum, 1st quartile, and median values were equal to 0 in all cases

R. groenlandicum

Three out of four control sites without M. albus had no pollinator visitation to R. groenlandicum (mean = 0.0006 visits/flower per hour), while Mel 40 treatment sites had a range of visitation rates between 0 visits/flower per hour and 0.034 visits/flower per hour (mean = 0.0076 visits/flower per hour) (Fig. 1b). This effect was not statistically significant (W = 15, p = 0.458).

Pollinator community

The pollinator guilds had 40 % proportional similarity between the pooled control and Mel 40 site pairs. More pollinator guilds visited the focal species in the Mel 40 sites than control sites. Butterflies, syrphid flies, other types of flies, and wasps were only present in the Mel 40 treatment sites, while bumblebees and solitary bees were present in control and Mel 40 sites. Only bees visited V. vitis-idaea, while a variety of pollinator guilds visited R. groenlandicum. The pollinator guild-level Simpson’s D was 0.49 in the control sites and 0.77 in the Mel 40 sites.

Effects of M. albus on native plant pollination and reproduction

V. vitis-idaea

In our MANOVA, we found a marginal treatment effect on V. vitis-idaea pollination and reproduction in 2012, and a marginal year effect (Appendix 3). In 2012, Mel 120 sites produced four more seeds per berry on average compared to the control and Mel 40 sites, a significant increase of approximately 15 % (Table 2; Fig. 2d). Fruit set in 2012 was greater than in 2011 for V. vitis-idaea (Table 2; Fig. 2c) and the difference in fruit set in Mel 40 compared to the control sites mean was greater in 2012 than in 2011 [1.1 % increase in 2011, 13.4 % increase in 2012, F(1,14) = 5.26, p = 0.03; Fig. 2c]. We did not detect any differences in the number of conspecific pollen grains on V. vitis-idaea stigmas or the percent well-pollinated flowers between treatments or between years in our ANOVA tests (Table 2; Fig. 2a, b). The number of conspecific pollen grains on stigmas, percent flowers pollinated, fruit set, and seeds per berry in V. vitis-idaea could not be explained by the number of M. albus inflorescences in the site (Table 3).
Fig. 2

V. vitis-idaea conspecific pollen loads on stigmas (a), percent flowers pollinated (b), percent flowers setting fruit (c), and number of seeds per fruit (d) in sites without M. albus (Control), Mel 40 sites, and sites with 120 M. albus plants added (Mel 120) during the summers of 2011 and 2012. Bars are mean ± SE of site-level averages for each treatment in each year. The Mel 120 treatment was only conducted in 2012. *p < 0.05, +p < 0.1 (for differences between treatment means)

Table 3

Linear regression analysis for V. vitis-idaea and R. groenlandicum pollination and reproduction responses to the number of M. albus inflorescences present at a site

Species

Response variable

Model df

Error df

M. albus inflorescences parameter estimate (SE)

F

p

R2

V. vitis-idaea

Conspecific pollen

1

34

−0.0002 (0.002)

0.01

0.94

0.0002

 

% Flowers pollinated

1

34

−0.00003 (0.00005)

0.41

0.52

0.01

 

% Fruit set

1

34

0.001 (0.006)

0.05

0.82

0.002

 

Seeds per fruit

1

34

0.002 (0.002)

2.07

0.16

0.06

R. groenlandicum

Conspecific pollen

1

27

−0.007 (0.003)

3.82

0.06

0.12

 

% Flowers pollinated

1

27

−0.01 (0.006)

4.05

0.05

0.13

 

% Fruit set

1

27

−0.005 (0.005)

0.99

0.33

0.04

 

Seeds per fruit

1

27

0.0004 (0.002)

0.07

0.79

0.003

The number of M. albus inflorescences present in sites where it was added ranged from 334 to 1608 inflorescences (mean 795 ± 85). Individual M. albus plants had up to 181 inflorescences with approximately 50 flowers per inflorescence

Models with p < 0.1 are indicated in italic

R. groenlandicum

We found a highly significant year effect for our R. groenlandicum MANOVA model testing for year, treatment, and interaction effects (Appendix 3). Seed mass per fruit was the individual variable driving this response (Table 2). There was more than double the mean seed mass per fruit in 2012 compared to 2011 (Fig. 3d). In 2011, Mel 40 sites had 42 % greater seed mass per fruit relative to the control sites, while in 2012 the Mel 40 sites had 24 % less seed mass per fruit relative to the control sites (Fig. 3d). The year by treatment interaction was also significant (Table 2). The number of M. albus inflorescences at a site decreased the number of R. groenlandicum pollen tetrads and percent flowers pollinated (a decrease of one tetrad or 1 % flowers pollinated for every 100 inflorescences added; Table 3). Fruit set could not be explained by the M. albus treatment level in either year (Table 2; Fig. 3c), nor could it be explained by the number of M. albus inflorescences at the site (Table 3).
Fig. 3

R. groenlandicum conspecific pollen loads on stigmas (a), percent flowers pollinated (b), percent flowers setting fruit (c), and number of seeds per fruit (d) in sites without M. albus (Control), Mel 40 and Mel 120 sites during the summers of 2011 and 2012. Bars are mean ± SE of site-level averages for each treatment in each year. The Mel 120 treatment was only conducted in 2012. *p < 0.05, +p < 0.1 (for differences between treatment means). For abbreviations, see Figs. 2 and 3

Relative importance of M. albus and other environmental factors in predicting reproduction

V. vitis-idaea

The number of M. albus inflorescences, canopy cover, flower richness, and mean temperature were identified as important in explaining the variation among sites in percent V. vitis-idaea flowers pollinated (Table 4). M. albus had a lower cumulative parameter weight compared to the other three variables, all three of which were positively related to  % flowers pollinated (Table 4). M. albus inflorescence number was important in explaining % V. vitis-idaea flowers pollinated in black spruce sites, but not in mixed deciduous and white spruce sites (Appendix 4). The number of M. albus inflorescences was not identified as being as important as the other environmental variables for any of the other three V. vitis-idaea response variables across all sites (Table 4) or in the two different habitat types.
Table 4

Modeled Akaike’s information criterion (AIC) average parameter estimates (b) and relative variable importance (cumulative parameter weights; \(\sum \, \omega_{i}\)) for candidate variables explaining differences in total pollen deposited on stigmas, % flowers well pollinated (≥10 pollen grains), % flowers setting fruit, and seeds per fruit (total number seeds for V. vitis-idaea and seed mass for R. groenlandicum) for focal species across all sites

Species

Explanatory variables

Response variables

Conspecific pollen

% Flowers pollinated

% Fruit set

Seeds or seed mass per fruit

\(\overline{b}\)

\(\sum \, \omega_{i}\)

\(\overline{b}\)

\(\sum \, \omega_{i}\)

\(\overline{b}\)

\(\sum \, \omega_{i}\)

\(\overline{b}\)

\(\sum \, \omega_{i}\)

V. vitis-idaea

No. M. albus inflorescences

0.23

−8 × 10−5

0.58

0.19

0.26

 

Canopy cover (%)

0.12

0.67

0.003

0.86

0.28

0.09

0.86

 

Shrub cover (%)

0.42

0.33

0.30

0.35

 

No. V. vitis-idaea flowers

0.26

0.39

2.25

0.88

0.34

 

No. All flowers

0.26

0.40

−2.30

0.90

0.30

 

Flower richness

0.41

0.09

0.68

16.45

0.84

3.40

0.72

 

Average temperature

0.24

0.06

0.67

0.28

0.40

 

Hours of rain

0.24

0.22

−0.13

0.86

0.23

R. groenlandicum

No. M. albus inflorescences

−0.01

0.68

−0.01

0.71

0.54

0.19

 

Canopy cover (%)

0.24

0.25

−0.31

0.99

7 × 10−6

0.83

 

No. R. groenlandicum inflorescence

0.22

0.23

0.21

0.28

 

No. All flowers

0.24

0.24

0.23

0.20

 

Flower richness

0.30

0.23

−5.93

0.71

0.20

 

Average temperature

0.32

0.36

0.31

3 × 10−4

0.91

 

Hours of rain

0.29

0.24

−0.04

0.56

0.40

Values in italic indicate well-supported variables \(\sum \, \omega_{i}\) > 0.55) and average parameter values for these variables were taken over models with a difference in AIC adjusted for small sample size <2

R. groenlandicum

The number of M. albus inflorescences at a site outweighed the importance of all the other vegetation and weather variables in explaining conspecific pollen loads and the percent pollinated R. groenlandicum flowers (Table 4). R. groenlandicum fruit set, however, was better explained by canopy cover, flower richness, and hours of rain, which were all negatively related to fruit set (Table 4). The R. groenlandicum seed mass per fruit was best explained by the percent canopy cover and average temperature at the sites, both of which had positive relationships with the seed mass per fruit (Table 4). The number of M. albus inflorescences was not identified as important in explaining the response variables when we divided the sites by habitat type.

Discussion

The existing body of literature addressing the effects of invasive plants on native plant pollination and reproduction has documented a diversity of competitive (Chittka and Schürkens 2001; Brown et al. 2002; Kandori et al. 2009; Flanagan et al. 2010), facilitative (Nielsen et al. 2008; Tepedino et al. 2008; Da Silva et al. 2013) and neutral effects (Bartomeus et al. 2008a). Our study confirms the complexity of invasive plant impacts on pollination within a plant community. We found that addition of M. albus increased pollinator diversity in our sites and tended to increase pollinator visitation rates to native V. vitis-idaea and R. groenlandicum. We saw a facilitative effect of M. albus on the seed production of V. vitis-idaea in 2012 and no strong effect on R. groenlandicum reproduction. However, there was a weak competitive effect on R. groenlandicum pollen loads and percent flowers pollinated at the highest M. albus densities.

Along with other multi-species studies and plant-pollinator network studies (Moragues and Traveset 2005; Jakobsson et al. 2009; Bartomeus et al. 2008b; Albrecht et al. 2014), we suggest that the identity of shared pollinators between specific invasive and native pairs or identity of new pollinators attracted to the invaded sites are likely determining factors in the reproductive responses of native plants. Our study also indicates that site environmental conditions and resource limitations to plant reproduction further complicate the generalizations that can be made from the existing corpus of experimental studies on invasive plant impact on native plant reproduction. We discuss here how both pollinator identity and site environmental conditions may have influenced our results.

M. albus effect on pollination and reproduction of two boreal shrubs

The addition of M. albus did little to change pollen loads or pollination rates of V.vitis-idaea. Seed production per berry, however, did increase in the presence of the highest M. albus abundance level. A higher seed set without evidence of higher pollen loads suggests that there is an increase in the proportion of outcross pollen being delivered by pollinators when higher densities of M. albus are introduced. This explanation is consistent with previous hand-pollination experiments in which cross-pollinated V. vitis-idaea plants produced more seeds than self-pollinated plants (Jacquemart and Thompson 1996; Fröborg 1996; Jacquemart 1997), but did not have higher fruit set (Jacquemart and Thompson 1996; Jacquemart 1997).

The shift in the pollinator community composition in the presence of M. albus provides a mechanism through which the outcrossing rates may change. Densely flowered clonal plants like V. vitis-idaea are subject to high levels of within-genet pollen transfer due to the foraging strategies of bumblebees in particular (Jacquemart and Thompson 1996). Other pollinating guilds such as butterflies tend to take longer flights between plants than Bombus spp. (Proctor et al. 1996) thereby increasing outcross potential. Butterflies, syrphid flies, other flies, and wasps were observed visiting the native focal species only at sites where M. albus was added. Other studies have documented changes in pollinator behaviors (e.g., changes in distance traveled between plants) as a result of non-native plant invasions (Ghazoul 2004), which could also explain a possible shift in outcrossing rates.

The modest decline in conspecific pollen loads and pollination rates of R. groenlandicum flowers with greater numbers of M. albus flowers is suggestive of a greater overlap in pollinator community and a shift by more effective R. groenlandicum pollinators to visiting M. albus when it is at high densities. Indeed a high proportion of pollinator guilds are shared between R. groenlandicum and M. albus (plant-pollinator networks are explored in Schneller et al., in prep). Despite a potential for reduced conspecific pollen flow, we did not observe an associated decline in fruit or seed set. Resource limitation is likely to play a major role in limiting sexual reproduction in these boreal communities (Grainger and Turkington 2013) and pollination rates beyond a minimum threshold may not result in changes in fruit and seed set.

The magnitude of the M. albus effect on V. vitis-idaea fruit set and R. groenlandicum seed mass per fruit was greater in 2012 compared to 2011. In 2011 it rained for almost twice as many hours as in 2012 (173 vs. 97 h during the experimental period), which likely allowed for a greater amount of time for pollinator activity in the second year (Kuchko 1988; Tuell and Isaacs 2010). It may also have affected which types of pollinators are active (Corbett et al. 1993). Other studies have found it difficult to disentangle the role of variations in weather in V. vitis-idaea fruit set (Jacquemart 1997; Krebs et al. 2009), with factors like late spring frosts having a potential effect on both flowers and insect populations. Similarly, we cannot determine whether the warmer conditions in the second year of our study reduced pollinator limitation or resources limitation for V. vitis-idaea fruit set and R. groenlandicum seed production. V. vitis-idaea fruit set was negatively related to the number of heterospecific flowers and positively related to the number of conspecific flowers, which is consistent with pollen limitation. Further, seed production for V. vitis-idaea increased only under the highest level of M. albus addition, suggesting that this variable is pollinator limited in a warm year.

Relative importance of M. albus and other environmental variables in explaining reproduction

M. albus inflorescence number was more important for explaining R. groenlandicum pollination than were the other biotic and abiotic variables we measured. The importance of M. albus did not persist in subsequent R. groenlandicum reproduction. Similarly, the number of M. albus inflorescences was only important in explaining the percent flowers pollinated for V. vitis-idaea, while fruit set was more influenced by the environmental conditions. This finding is consistent with Totland et al. (2006), who documented non-native Phacelia tanacetifolia affecting pollinator visitation to a native boreal plant species (Melampyrum pratense), but not reproductive success. Totland et al. (2006) similarly attributed this finding to the high resource limitations on reproduction in boreal forest habitats.

Our finding that M. albus had far greater importance in explaining pollination rates of V. vitis-idaea in black spruce sites than in mixed deciduous-spruce sites further supports the important role of resource availability in mediating relationships between invasive and native boreal plant reproduction. Black spruce sites tend to have lower soil temperatures (Viereck et al. 1992) and lower densities of native flowers than the mixed deciduous-spruce sites. The reduced floral resources and the lower temperatures for ground-nesting pollinators likely limit the number of pollinators in black spruce sites, and the addition of M. albus may act as a distraction for the few pollinators present.

Our study sets the stage for further investigation of the pollination of native plants in the face of accelerating rates of invasion in boreal forest ecosystems.

Author contribution statement

K. V. S., L. C. S., C. P. H. M. and M. L. C. conceived, designed, and performed the experiments. K. V. S. analyzed the data and wrote the manuscript with assistance from L. C. S. (Pollinator activity sections in Materials and methods and Results); C. P. H. M. and M. L. C. provided analytical and editorial advice.

Notes

Acknowledgments

Funding for this project was provided by grants from the US Department of Agriculture NIFA (ALKR-2009-04931) and National Science Foundation IGERT (grant no. 0654441). We thank our technicians (S. Decina, P. Hurtt, M. Kain, J. Malthot, L. Medinger, K. Moeller, L. Ponchione, T. Saunders) and volunteers (J. Conn, J. Martin, B. Spellman, D. Uliassi, E. Uliassi, L. Uliassi, K. Schnaars Uvino, J. Villano, and T. Villano) for assistance in field and lab work, M. Wright for greenhouse support, and the Bonanza Creek Long-Term Ecological Research Program for providing access to sites. Thoughtful comments from L. Conner, A. D. McGuire, and D. Wagner greatly helped us improve this manuscript.

Supplementary material

442_2015_3364_MOESM1_ESM.docx (863 kb)
Supplementary material 1 (DOCX 863 kb)

References

  1. AKEPIC—Alaska Exotic Plant Information Clearing House (2014) AKEPIC mapping project inventory field data. University of Alaska Anchorage Alaska Natural Heritage Program and USDA Forest Service, Anchorage, Alaska. Retrieved 9 September 2014 from http://akweeds.uaa.alaska.edu/
  2. Albrecht M, Padrón B, Bartomeus I, Traveset A (2014) Consequences of plant invasions on compartmentalization and species roles in plant–pollinator networks. Proc R Soc B 281:20140773. doi:10.1098/rspb.2014.0773 PubMedCentralCrossRefPubMedGoogle Scholar
  3. Arnold TW (2010) Uninformative parameters and model selection using Akaike’s information criterion. J Wildl Manage 74:1175–1178CrossRefGoogle Scholar
  4. Bartomeus I, Bosch J, Vila M (2008a) High invasive pollen transfer, yet low deposition on native stigmas in a Carpobrotus-invaded community. Ann Bot 102:417–424PubMedCentralCrossRefPubMedGoogle Scholar
  5. Bartomeus I, Vila M, Santamaria L (2008b) Contrasting effects of invasive plants in plant–pollinator networks. Oecologia 155:761–770CrossRefPubMedGoogle Scholar
  6. Bartomeus I, Vila M, Steffan-Dewenter I (2010) Combined effects of Impatiens glandulifera invasion and landscape structure on native plant pollination. J Ecol 98:440–450CrossRefGoogle Scholar
  7. Brower JE, Zar JH (1984) Field and laboratory methods for general ecology, 2nd edn. Brown, DubuqueGoogle Scholar
  8. Brown BJ, Mitchell RJ, Graham SA (2002) Competition for pollination between an invasive species (purple loosestrife) and a native congener. Ecology 83:2328–2336CrossRefGoogle Scholar
  9. Burnham KP, Anderson DR (2002) Model selection and multimodel inference: a practical information-theoretic approach, 2nd edn. Springer, NewyorkGoogle Scholar
  10. Campbell DR (1985) Pollinator sharing and seed set of Stellaria pubera: competition for pollination. Ecology 66:544–563CrossRefGoogle Scholar
  11. Carlson ML, Shephard M (2007) Is the spread of non-native plants in Alaska accelerating? In: Harrington TB, Reichard SH (ed) Meeting the challenge: invasive plants in Pacific Northwest ecosystems. Gen Tech Rep PNW-GTR-694. US Department of Agriculture, Forest Service, Pacific Northwest Research Station, Portland, OR, pp 111–127Google Scholar
  12. Carlson ML, Lapina IV, Shephard M, Conn JS, Densmore R, Spencer P, Heys J, Riley J, Nielsen J (2008) Invasiveness ranking system for non-native plants of Alaska. Technical paper R10-TP-143. USDA Forest Service, Alaska Region, AnchorageGoogle Scholar
  13. Chittka L, Schürkens S (2001) Successful invasion of a floral market. Nature 411:653CrossRefPubMedGoogle Scholar
  14. Coe HS, Martin JN (1920) Sweet clover seed. Bulletin 844. US Department of Agriculture, WashingtonGoogle Scholar
  15. Conn JS, Stockdale CA, Morgan JC (2008a) Characterizing pathways of invasive plant spread to Alaska: propagules from container-grown ornamentals. Inv Plant Sci Mgmt 1:331–336CrossRefGoogle Scholar
  16. Conn JS, Beattie KL, Shephard ML, Carlson ML, Lapina I, Hebert M, Gronquist R, Densmore R, Rasy M (2008b) Alaska Melilotus invasions: distribution, origin, and susceptibility of plant communities. Arct Antarct Alp Res 40:298–308CrossRefGoogle Scholar
  17. Corbet SA, Fussell M, Ake R, Fraser A, Gunson C, Savage A, Smith K (1993) Temperature and the pollinating activity of social bees. Ecol Entomol 18:17–30CrossRefGoogle Scholar
  18. Da Silva EM, King VM, Russell-Mercier JL, Sargent RD (2013) Evidence for pollen limitation of a native plant in invaded communities. Oecologia 172:469–476CrossRefPubMedGoogle Scholar
  19. Davis AN (2002) Pollination biology of the lingonberry, Vaccinium vitis-idaea subsp. minus L. Master’s thesis. University of Alaska Fairbanks, Fairbanks, AlaskaGoogle Scholar
  20. Davis AN, Holloway PS, Kruse JJ (2003) Insect visitors and potential pollinators of lingonberries, Vaccinium vitis-idaea subsp. minus, in sub-arctic Alaska. Acta Hortic 626:441–446Google Scholar
  21. Dlusski GM, Glazunova KP, Perfilieva KS (2005) Mechanisms that limit pollinator range in Ericaceae. Z Obshch Biol 66:224–238Google Scholar
  22. Eckardt N (1987) Element stewardship abstract for Melilotus alba—sweetclover or white sweetclover, Melilotus officinalis—yellow sweetclover. The Nature Conservancy, MinneapolisGoogle Scholar
  23. Flanagan RJ, Mitchell RJ, Karron JD (2010) Increased relative abundance of an invasive competitor for pollination, Lythrum salicaria, reduces seed number in Mimulus ringens. Oecologia 164:445–454CrossRefPubMedGoogle Scholar
  24. Fröborg H (1996) Pollination and seed production in five boreal species of Vaccinium and Andromeda (Ericaceae). Can J Bot 74:1363–1368CrossRefGoogle Scholar
  25. Garibaldi A (1999) Medicinal flora of the Alaska natives. University of Alaska Anchorage Press, AnchorageGoogle Scholar
  26. Gathmann A, Tscharntke T (2002) Foraging ranges of solitary bees. J Anim Ecol 71:757–764CrossRefGoogle Scholar
  27. Ghazoul J (2004) Alien abduction: disruption of native plant pollinator interactions by invasive species. Biotropica 36:156–164Google Scholar
  28. Hall IV, Beil CE (1970) Seed germination, pollination, and growth of Vaccinium vitis-idaea var. minus Lodd. Can J Plant Sci 50:731–732CrossRefGoogle Scholar
  29. Holloway PS (2006) Managing wild bog blueberry, lingonberry, cloudberry, and crowberry stands in Alaska. Report for University of Alaska Fairbanks and Natural Resource Conservation Service, FairbanksGoogle Scholar
  30. Hultén E (1968) Flora of Alaska and neighboring territories. Stanford University Press, StanfordGoogle Scholar
  31. Irwin DL (1945) Forty-seven years of experimental work with grasses and legumes in Alaska. Univ Alaska Agric Exp Stn Bull 12:47Google Scholar
  32. Jacquemart AL (1997) Pollen limitation of three sympatric species of Vaccinium (Ericaceae) in the Upper Ardennes, Belgium. Plant Syst Evol 207:159–172CrossRefGoogle Scholar
  33. Jacquemart AL, Thompson JD (1996) Floral and pollination biology of three sympatric Vaccinium (Ericaceae) species in the upper Ardennes, Belgium. Can J Bot 74:210–221CrossRefGoogle Scholar
  34. Jakobsson A, Padrón B, Traveset A (2009) Competition for pollinators between invasive and native plants: effects of spatial scale of investigation (note). Ecoscience 16:138–141CrossRefGoogle Scholar
  35. Kandori I, Hirao T, Matsunaga S, Kurosaki T (2009) An invasive dandelion unilaterally reduces the reproduction of a native congener through competition for pollination. Oecologia 159:559–569CrossRefPubMedGoogle Scholar
  36. Kearns CA, Inouye DW (1993) Techniques for pollination biologists. University Press of Colorado, BoulderGoogle Scholar
  37. Kevan PG, Tikhmenev EA, Usui M (1993) Insects and plants in the pollination ecology of the boreal zone. Ecol Res 8:247–267CrossRefGoogle Scholar
  38. Krebs CJ, Boonstra R, Cowcill K, Kenney AJ (2009) Climatic determinants of berry crops in the boreal forest of the southwestern Yukon. Botany 87:401–408CrossRefGoogle Scholar
  39. Kuchko AA (1988) Bilberry and cranberry yields and the factors controlling them in the forests of Karelia USSR. Acta Bot Fenn 136:23–25Google Scholar
  40. Malacalza NH, Caccavari MA, Fagúndez G, Lupano CE (2005) Unifloral honeys of the province of Buenos Aires, Argentine. J Sci Food Agric 85:1389–1396CrossRefGoogle Scholar
  41. Molina-Montenegro MA, Badano EI, Cavieres LA (2008) Positive interactions among plant species for pollinator service: assessing the magnet species concept with invasive species. Oikos 117:1833–1839CrossRefGoogle Scholar
  42. Moragues E, Traveset A (2005) Effect of Carpobrotus spp. on the pollination success of native plant species of the Balearic Islands. Biol Conserv 122:611–619CrossRefGoogle Scholar
  43. Morales CL, Traveset A (2008) Interspecific pollen transfer: magnitude, prevalence and consequences for plant fitness. Crit Rev Plant Sci 27:221–238CrossRefGoogle Scholar
  44. Muñoz AA, Cavieres LA (2008) The presence of a showy invasive plant disrupts pollinator service and reproductive output in native alpine species only at high densities. J Ecol 96:459–467CrossRefGoogle Scholar
  45. Myneni RB, Keeling CD, Tucker CJ, Asrar G, Nemani RR (1997) Increased plant growth in the northern high latitudes from 1981 to 1991. Nature 386:698–702CrossRefGoogle Scholar
  46. Nelson JL, Zavaleta E, Chapin FS III (2008) Boreal fire effects on subsistence resources: landscape diversity as a critical component of rural livelihoods in Alaska and adjacent Canada. Ecosystems 11:156–171CrossRefGoogle Scholar
  47. Nielsen C, Heimes C, Kollmann J (2008) Little evidence for negative effects of an invasive alien plant on pollinator services. Biol Invasions 10:1353–1363CrossRefGoogle Scholar
  48. Peterson SF (1989) Beekeeping under the northern lights. Am Bee J 129:33–35Google Scholar
  49. Proctor M, Yeo P, Lack A (1996) The natural history of pollination. Timber Press, PortlandGoogle Scholar
  50. Quiner M (2005) Ranchers get the blues: area citizens consider branching out into berries. Peninsula Clarion, 11 October 2005, http://www.peninsulaclarion.com/stories/101105/
  51. R Development Core Team (2012) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. ISBN. 3-900051-07-0, http://www.R-project.org/
  52. Royer F, Dickinson R (1999) Weeds of the Northern US and Canada. The University of Alberta Press, EdmontonGoogle Scholar
  53. Sanderson LA, McLaughlin JA, Antunes PM (2012) The last great forest: a review of the status of invasive species in the North American boreal forest. Forestry 85:329–339CrossRefGoogle Scholar
  54. Serreze MC, Walsh JE, Chapin FS III, Osterkamp T, Dyurgerov M, Oechel WC, Romanovsky V, Morison J, Zhang T, Barry RG (2000) Observational evidence of recent change in the northern high latitude environment. Clim Change 46:159–207CrossRefGoogle Scholar
  55. Shugart HH, Leemans R, Bonan GB (1992) A systems analysis of the boreal forest. Cambridge University Press, CambridgeCrossRefGoogle Scholar
  56. Spellman BT, Wurtz T (2011) Invasive white sweetclover (Melilotus officinalis) impacts native recruitment along rivers in interior Alaska. Biol Invasions 13:1779–1790CrossRefGoogle Scholar
  57. Spellman KV, Mulder CPH, Hollingsworth TN (2014) Susceptibility of burned black spruce (Picea mariana) forests to non-native plant invasions in Interior Alaska. Biol Invasions 16:1879–1895CrossRefGoogle Scholar
  58. 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 Area J 28:44–50CrossRefGoogle Scholar
  59. Totland Ø, Nielsen A, Bjerknes AL, Ohlson M (2006) Effects of an exotic plant and habitat disturbance on pollinator visitation and reproduction in a boreal forest herb. Am J Bot 93:868–873CrossRefPubMedGoogle Scholar
  60. Tuell JK, Isaacs R (2010) Weather during bloom affects pollination and yield of highbush blueberry. J Econ Entomol 103:557–562CrossRefPubMedGoogle Scholar
  61. Turkington RA, Cavers PB, Rempel E (1978) The biology of Canadian weeds. 29. Melilotus alba Desr. and M. officinalis (L.) Lam. Can J Plant Sci 58:523–537CrossRefGoogle Scholar
  62. US Census 2010 (2010) Census 2010 data for the state of Alaska. US Census Bureau. http://www.census.gov/census2010/states/ak.html
  63. Viereck LA, Dyrness CT, Batten AR, Wenzlick KJ (1992) The Alaska vegetation classification. Gen Tech Rep PNW-GTR-286. US Department of Agriculture, Forest Service, Pacific Northwest Research Station. Portland, ORGoogle Scholar
  64. Villano KL, Mulder CPH (2008) Invasive plant spread in burned lands of interior Alaska. Technical report for National Park Service-Alaska Region and National Aeronautics and Space Administration, Fairbanks, AlaskaGoogle Scholar
  65. Walker DA, Walker MD (1991) History and pattern of disturbance in Alaskan arctic terrestrial ecosystems: a hierarchical approach to analysing landscape change. J Appl Ecol 28:244–276CrossRefGoogle Scholar
  66. Waser NM (1978) Competition for pollination and sequential flowering in two Colorado wildflowers. Ecology 59:934–944CrossRefGoogle Scholar
  67. Waser NM (1983) Competition for pollination and floral character differences among sympatric plant species: a review of evidence. In: Jones CE, Little RJ (eds) Handbook of experimental pollination biology. Van Nostrand Reinhold, New York, pp 277–293Google Scholar
  68. Westphal C, Steffan-Dewenter I, Tscharntke T (2003) Mass-flowering crops enhance pollinator densities at a landscape scale. Ecol Lett 6:961–965CrossRefGoogle Scholar
  69. Westrich P (1996) Habitat requirements of central European bees and the problems of partial habitats. In: Matheson A, Buchmann SL, O’Toole C, Westrich P, Williams H (eds) The Conservation of bees. Linnaean Society of London and the International Bee Research Association. Academic Press, London, pp 1–16Google Scholar
  70. Wheelwright NT, Dukeshire EE, Fontaine JB, Gutow SH, Moeller DA, Schuetz JG, Smith TM, Rodgers SL, Zink AG (2006) Pollinator limitation, autogamy and minimal inbreeding depression in insect-pollinated plants on a boreal island. Am Midl Nat 155:19–38CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2015

Authors and Affiliations

  • Katie V. Spellman
    • 1
  • Laura C. Schneller
    • 2
  • Christa P. H. Mulder
    • 1
  • Matthew L. Carlson
    • 2
    • 3
  1. 1.Department of Biology and Wildlife, Institute of Arctic BiologyUniversity of Alaska FairbanksFairbanksUSA
  2. 2.Department of Biological SciencesUniversity of Alaska AnchorageAnchorageUSA
  3. 3.Alaska Natural Heritage ProgramUniversity of Alaska AnchorageAnchorageUSA

Personalised recommendations