Page  92

92 THE MICHIGAN BOTANIST Vol. 54

THE INITIAL EFFECTS OF COMMUNITY VARIABLES ON SAND PRAIRIE RESTORATION: SPECIES ESTABLISHMENT AND COMMUNITY RESPONSES

Robert C. Roos 6140 Cottonwood Drive, Suite A Fitchburg, Wisconsin 53719 robb.roos@cardno.com

Todd A. Aschenbach Grand Valley State University 1 Campus Drive Allendale, Michigan 49401 aschenbt@gvsu.edu

ABSTRACT

The tallgrass prairie was one of the most wide-ranging and diverse ecosystems in North America. This diverse ecosystem comprises a mosaic of prairie types that, in northern Lower Michigan, was historically dominated by dry sand prairie. As a result of fire suppression, silvicultural and agricul- tural activities, and degradation by invasive species, only approximately 4% of the original extent of sand prairie remains intact in the state. Despite the important ecological role and increasing scarcity of sand prairie, restoration and management of this ecosystem has been severely understudied. In order to gain a better understanding of this ecosystem and of potential restoration techniques that might influence its community variables, a sand prairie restoration experiment was established in the pine–oak barrens of northern Lower Michigan to analyze how different seeding treatments affect cer- tain community variables, including vegetative cover, species richness, diversity, and floristic quality. The seeding treatments were varied with respect to seeding concentrations (1,000 seeds/m2 and 10,000 seeds/m2); and the inclusion of grasses and/or forbs with diverse ecological characteristics, such as early flowering, late flowering, and nitrogen fixers (i.e., legumes). Measurements of the com- munity variables were taken during each of the first three growing seasons following seeding. In gen- eral, treatments that included a high concentration of grass and/or an early season forb component had the greatest overall positive impact on plant community development. These treatments resulted in significantly greater diversity and higher floristic quality, as well as a lower percentage of non-na- tive or invasive cover than other treatments. The benefit of high concentrations of grasses and early season forbs may play a critical role in initial species establishment of a sand prairie restoration due to the facilitative and competitive advantages they may provide in these harsh environments. How- ever, it remains to be seen if these initially successful communities will have continued success over longer periods of time.

KEYWORDS: Sand prairie, Community, Establishment, Restoration.

INTRODUCTION

The tallgrass prairie was one of the most wide-ranging and diverse ecosys- tems in North America. Today, only approximately 0.1% of its original extent re- mains, making tallgrass prairie one of North America’s most endangered ecosys- tems (Samson and Knopf 1994). The extensive loss and rapid decline of this

Page  93 2015 THE MICHIGAN BOTANIST 93

ecosystem can be attributed primarily to European settlers. The dark, rich, tree- free, and easily manipulated Mollic soils of the tallgrass prairie made it a prime target for agricultural use (Howe 1994). This ecosystem has also experienced many other assaults as a result of human development. Fire, which was once commonplace in the Midwestern United States, became actively managed and suppressed (Anderson 1990). Habitat fragmentation, the introduction of invasive non-native species, and the increased establishment of invasive native species also promoted habitat degradation (Noss et al. 1995; Cully et al. 2003).

The reduction of tallgrass prairie and its associated biodiversity has resulted in a substantial loss of ecosystem function. Higher levels of plant diversity are associated with greater ecosystem productivity (Tilman and Downing 1994; Hector et al. 1999; Tilman 1999), nutrient use efficiency (Risser 1988; Tilman 1997), resistance to invasion by exotic species (Tilman and Downing 1994; Kennedy et al. 2002; Pokorny et al. 2005), and resistance to environmental change (Ives et al. 2000). According to the insurance hypothesis of Yachi and Loreau (1999), high levels of biodiversity insures ecosystems against declines in their functioning, because the presence of many species guarantees that some will remain functioning even if others fail (Ives et al. 2000). Ecosystem functions that are associated with tallgrass prairie include carbon sequestration, water fil- tration, erosion control, soil enhancement, and nutrient cycling (Raison 1979; Seastedt and Knapp 1993; Wedin and Tilman 1990). Since diversity is a benefi- cial and necessary attribute that drives a fully-functioning, healthy ecosystem, it is useful to examine the factors that impact diversity.

The tallgrass prairie ecosystem comprises a mosaic of prairie types that in- clude xeric, mesic, and wet components that sometimes transition into barrens and savannah ecosystems. Tallgrass prairie historically extended eastward through Indiana and into areas of Michigan, Ohio, and Kentucky. This region is referred to as the prairie peninsula. These fingers of grassland followed areas where climate fluctuated enough to support a mosaic of prairie community types, including oak-pine barrens and oak savanna (Transeau 1935; Anderson 1990). In northern Lower Michigan, north of the tension zone (approximately 43ºN latitude), grassland was predominantly dry sand prairie (McCann 1991; Kost 2004). As a component of these open, upland mosaics, sand prairie was a primary component of roughly 5,000 hectares of northern Lower Michigan’s nat- ural landscape in the early to mid-1800s (Comer et al. 1995). This dry grassland is considered the driest ecosystem east of the Mississippi River (Schaetzl and Anderson 2005). Plant species of the sand prairies are similar to those of a xeric tallgrass prairie, but due to water, heat, and nutrient stresses, the sand prairie vegetation is typically shorter in stature and separated by patches of bare ground. The combination of wildfires, dry and well-drained soils, and the harsh frosts that are associated with northern Michigan have historically maintained these ecosystems (Kost et al. 2007).

As a result of fire suppression, silvicultural and agricultural activities, and degradation by invasive species, only approximately 4% of the original extent of sand prairie remains intact in the state (Hauser 1953; Albert and Comer 2008). Consequently, sand prairie is considered one of Michigan’s most endangered

Page  94 94 THE MICHIGAN BOTANIST Vol. 54

ecosystems. Today, less than 200 hectares of high quality sand prairie still exist in Michigan (Kost 2004).

The loss and degradation of the sand prairie has had negative consequences for species that are associated with these ecosystems. Over 25 plant and 30 ani- mal species are dependent on these dry grasslands for either all or part of their lives (Kost 2004). This includes the federally endangered Lycaeides melissa samuelis (Karner blue butterfly), a species that depends on Lupinus perennis L. Grassland birds have shown a decline that is greater than that of any other group of North American species (Knopf 1994). Dendroica kirtlandii (Kirtland’s war- bler) is a species that depends solely on the matrix of sand prairie and pine bar- rens comprised of Pinus banksiana Lamb. in northern Lower Michigan for sur- vival (Kost et al. 2007). Common northern dry sand prairie species, that nevertheless are state-listed due to the degradation of this community type, in- clude Festuca altaica Trin. Ex Ledeb., Cirsium hillii (Canby) Fernald, and Agoseris glauca (Pursh) Raf..

Despite the important ecological role and increasing scarcity of sand prairie, restoration and management of this ecosystem has been severely understudied. Published research on sand prairies within the eastern prairie peninsula, includ- ing the portion in Michigan, is sparse. Instead, the majority of sand prairie liter- ature has focused on prairie regions to the west of the Great Lakes (Gleason 1910; Plumb-Mentjes and Center 1990; Cole and Taylor 1995; Bowles et al. 2003). Studies that have addressed the Michigan sand prairies have focused on descriptive analyses (Hauser 1953; Albert 1995; Comer et al. 1995; Kost 2004), or on comparative assessments with other community types, such as dunes and jack-pine barrens (Houseman and Anderson 2002; Emery et al. 2013). To date there have been few, if any, studies that have focused on the restoration of Michi- gan’s true sand prairies.

Restoration projects in general tend to place an emphasis on many aspects of community structure and ecosystem processes. Although certain components of restoration projects have been successful, they have yet to achieve the goal of creating a historic, natural community (Martin et al. 2005). Current methods of restoring prairie communities are based on a weak scientific rationale that is not consistent with the history of how grasslands formed, and in fact may threaten biodiversity (Howe 1994). Although plant community restoration has the poten- tial to help re-establish lost diversity and ecosystem function (Foster et al. 2007), many plant community restoration attempts have not fully re-established the di- versity and function found in remnant prairie communities (Sluis 2002, Polley et al. 2005).

Sub-optimal results of previous restoration attempts have led ecologists to ex- amine different theories of plant community succession and species coexistence in order to identify more successful approaches to restoration (Cairns and Heck- man 1996). Community assembly theory is one approach that asks how species arriving at a site form an initial community (Belyea and Lancaster 1999). This approach integrates aspects of succession and species coexistence in an effort to examine how species introductions, biotic interactions (e.g., competition), and abiotic conditions (e.g., soil nutrients) influence community development (Lock- wood 1997; Belyea and Lancaster 1999; Young et al. 2001).

Page  95 2015 THE MICHIGAN BOTANIST 95

The influence of environmental factors on plant communities is complex, and multiple factors influence plant diversity and community development (Grace 1999). It is well established that the environment influences plant community succession (Tilman 1988; Howe 1995), the frequency and intensity of distur- bance (Collins et al. 1995; Suding 1999; Collins 2000), competitive intraspecific and interspecific interactions and the plant’s ability to respond (Grime 1974; Goldberg and Barton 1992; Smith et al. 1999), the availability of nutrients (Rai- son 1979; Ojima et al. 1994), and the ability of plants to be productive both veg- etatively and reproductively (Zimmerman and Kucera 1977; Gough et al. 1994; Tilman et al. 1996). A better understanding of how all of these factors influence community development will allow for more practical approaches to community restoration.

As one approach to understanding that question, we established a restoration experiment on a degraded site previously occupied by sand prairie in northern Lower Michigan in 2009. We introduced various combinations of native plants belonging to four functional groups—legumes, early flowering forbs, late flow- ering forbs, and warm-season grasses—at two different seeding concentrations in an attempt to see how these initial seeding treatments affect plant community development over time. Here we present the results obtained after two complete growing seasons and compare how the different seeding treatments have affected vegetative cover, species richness, diversity, and floristic quality in the harsh en- vironment of a Michigan sand prairie.

METHODS

Study Area

The study site is located at the historic Chittenden Nursery in the Manistee-Huron National For- est in Manistee County, Michigan. The nursery site is approximately 23 hectares and consists of 13 adjacent 1-to 2-hectare open fields. The location chosen is an appropriate study site, because the area was relatively homogenous inthat it was level, tree-free, and dominated by invasive native and non-native species. The rectangular study area covers an approximate 500 m2 portion of one of these fields. The site location is depicted in Figure 1.

The study site was historically part of the oak-pine barrens ecosystem which included pockets of sand prairie (Albert and Comer 2008). In 1934, the area was converted into the Chittenden Nursery, a tree nursery for the United States Forest Service (USFS). The tree nursery was shut down in the 1970s and has since been used rather insignificantly for USFS housing, conferences, and training ac- tivities (e.g., wildfire training, prescribed burns, all-terrain vehicle instruction).

Climatic factors at the time of the study were consistent with historical averages. Summer tem- perature highs during the survey years averaged 23.1°C, and average monthly rainfall was 17.5 cm. Site soils are mapped as Plainfield sands with a very deep water table (United States Department of Agriculture 2012). The weather and soil conditions are consistent with oak-pine barren and sand prairie ecosystems, and the vegetation is characteristic of a degraded ecosystem (Albert and Comer 2008).

Experimental Design

Prior to the initiation of our experiment, the entire study area was mowed and foliar herbicide containing glyphosate was applied to it in 2007 and again in 2008 in an attempt to reduce the abun- dance of invasive species (e.g., Centaurea stoebe L.).

In March 2009, a total of 228 1 m2 treatment plot boundaries were established. All plots were separated by a 0.5 m border along all four sides to avoid edge effects and to prevent trampling while taking measurements. Typical sand prairie species used in local U.S. Forest Service restorations were

Page  96 96 THE MICHIGAN BOTANIST Vol. 54

FIGURE 1. Chittenden Nursery, Manistee County, Michigan. Location of the approxi-

mately 500 m2 rectangular study site within the nursery and a diagram showing the lay-

out of the 228 1 m2 study plots separated by 0.5 m buffers.

chosen to be planted based on their specific guild and historic presence in Michigan sand prairies. Three species per guild were selected. Selection criteria reflect a balance of conservative versus less- conservative native species. Species were further selected to include a variety of ecological charac- teristics, including both late vs. early season flowering species, both C4 and C3 photosynthetic path- ways (among the grasses), and, with respect to soil nutrient relationships, both nitrogen fixers (legumes) and nitrogen extractors. The plants selected for seeding were placed into one of five groups—legumes, early season forbs, late season forbs, grasses, and background species. The species in each of these five groups are listed in Table 1. Twelve of the treatments consisted of one of the first three functional groups (legumes, early season forbs, late season forbs), grasses, and background species, for a maximum of 13 species per treatment. Background species were selected and added to the seed mixes in order to mimic a more traditional species rich and diverse seed mixture similar to those used in sand prairie restoration activities.

Each species other than the background species was seeded at either a high concentration (10,000 seeds/m2) or a low concentration (1,000 seeds/m2). The background species were seeded at a density of 500 seeds/m2. Seeding densities were equally divided for each species so that an equal number of seeds per species were represented in the mix. Seeds were obtained from Michigan Wildflower Farm, Portland, Michigan. All seeds are of local Michigan genotypes.

In January 2009, seeds were counted and weighed to determine the number of seeds per gram for each species. The appropriate number of seeds were mixed with clean, moist sand and placed in a freezer for a minimum of two weeks to simulate cold-moist stratification before being sown into the field. In March 2009, all seed mixes, except those containing legumes, were sown evenly by hand into treatment plots. Due to the unavailability of seed in the spring, legume seed treatments (includ- ing their grass and background species components) were sown in December 2009. All seeded group comparisons and control plots were randomly assigned. Each treatment was replicated 12 times for a

Page  97 2015 THE MICHIGAN BOTANIST 97

TABLE 1. Species in each of the five seeding groups and associated coefficient of conservatism

(CC) values. Seeding Group Species CC Value Legumes Lupinus perennis L. Desmodium canadense (L.) DC. Lespedeza capitata Michx. 7 3 5 Early Season Forbs Penstemon hirsutus (L.) Willd. Asclepias tuberosa L. Anemone virginiana L. 5 5 3 Late Season Forbs Symphyotrichum laeve (L.) Á. Löve & D. Löve Solidago nemoralis Aiton Solidago speciosa Nutt. 5 2 5 Grasses Andropogon gerardii Vitman Schizachyrium scoparium (Michx.) Nash Sorghastrum nutans (L.) Nash 5 5 6 Background Species Coreopsis lanceolata L. Euphorbia corollata L. Liatris aspera Michx. Monarda fistulosa L. Oenothera biennis L. Rudbeckia hirta L. Verbena stricta Vent. 8 4 4 2 2 1 4

total of 228 plots. The composition by functional group of each of the seeding treatments is given in Table 2. Treatments that are listed as having 24 replicates in Table 2 are those that received separate March and December seedings of 12 replicates each.

Baseline Data Collection

In July 2009, before the growth of any of the seed treatments, visual estimates were made of the percentage cover of vegetation for each plot. Measurements of the percentage of vegetative cover were broken down into seven different groupings in order to further explore the differences in each treatment group. These groupings were: (i) all species encountered (“total”), (ii) all non-native species encountered (“non-native”), (iii) all native species encountered (“native”), (iv) all non-seeded species encountered, both native and non-native (“resident”), (v) all seeded species encountered (“planted”), (vi) all grass species encountered (“grass”), and (vii) all forb species encountered (“forb”).

Initial estimates of percentage cover for each plot were averaged across the entire 228 plot study site. Mean species richness and diversity values were derived from the percentage cover data of each plot and were averaged across the entire study site. Baseline measurement data, representative of pre- restoration conditions, are provided in Table 3.

Floristic quality index (FQI) values of the different seed mixes were calculated (Table 4). These values may provide insight on the extent that restoration plot FQI may be predetermined by seed mix design. Species nomenclature follows The Taxonomic Name Resolution Service (2014), which is an online tool for automated standardization of plant names as described by Boyle et al. (2013).

In late August 2009, a floristic quality assessment (FQA) was completed for the one-half hectare area that surrounds the study site. This area was not affected by the site-preparation herbicide treat- ments. The surrounding area included a mix of disturbed, developed land, mesic-mixed forest, and xeric open fields. This FQA provides baseline data for species that occur in the immediate area adja- cent to the study site and may provide information for future measures (e.g., seed bank germination

Page  98 98 THE MICHIGAN BOTANIST Vol. 54

TABLE 2. Composition by functional groups of the 15 treatment groups. N indicates the number of replicates for each treatment. In addition to these treatments, there were 12 control plots that received no treatment.

Treatment Group N Early Season Flowering Forbs 1. Early Season Forbs (high); Grasses (low); Background Species 12 2. Early Season Forbs (low); Grasses (high); Background Species 12 3. Early Season Forbs (low); Grasses (low); Background Species 12 4. Early Season Forbs (low); Background Species 12 Late Season Flowering Forbs 5. Late Season Forbs (high); Grasses (low); Background Species 12 6. Late Season Forbs (low); Grasses (high); Background Species 12 7. Late Season Forbs (low); Grasses (low); Background Species 12 8. Late Season Forbs (low); Background Species 12 Legumes 9. Legumes (high); Grasses (low); Background Species 12 10. Legumes (low); Grasses (high); Background Species 12 11. Legumes (low); Grasses (low); Background Species 12 12. Legumes (low); Background Species 12 Grasses and Background Species Only 13. Grasses (high); Background Species 24 14. Grasses (low); Background Species 24 Background Species Only 15. Background Species 24

TABLE 3. Mean pre-restoration, baseline data collec- tion values for percentage cover, species richness, FQI, and H’ of the entire 228-plot study site in July 2009.

Variable Value Cover (%) Native 42.22 Non-Native 57.78 Resident 99.97 Planted 0.03 Grass 0.75 Forb 99.25 Total 100.00 Species Richness Total 6.98 Native 2.08 Non-Native 4.90 Resident 6.92 Planted 0.06 Grass 0.37 Forb 6.61 Diversity Floristic Quality Index (FQI) 0.99 Shannon Diversity Index (H’) 1.11

and colonization). Plants that could not be readily identified were trimmed with scissors at the base and placed into plas- tic bags. The bags were labeled with key plant and environmental characteristics

(e.g. soil conditions, exposure to shade/sun). The plants were transported on ice back to Grand Valley State Uni- versity in Allendale, Michigan, where they could further be identified. The list of species collected is provided in Table 5. It is important to note that only species that were identifiable in the fall were included in the list. Many native and non-native plants, especially early season forbs and grasses, may have been excluded in this initial sampling due to their lack of key reproductive structures. Data Collection

Data on the percentage cover of each species and on species richness were collected from the study plots in mid-July each year from 2009 through

Page  99 2015 THE MICHIGAN BOTANIST 99

TABLE 4. Floristic quality index (FQI) of each of the seed mixes.

Seed Mix FQI Early Season Flowering Forbs Early Season Flowering Forbs; Grasses; Background Species Early Season Flowering Forbs; Background Species 14.98 12.02 Late Season Flowering Forbs Late Season Flowering Forbs; Grasses; Background Species Late Season Flowering Forbs; Background Species 14.70 11.70 Legumes Legumes; Grasses; Background Species Legumes; Background Species 15.53 12.65 Grasses and Background Species Only Grasses; Background Species 12.97 Background Species Only Background Species 9.45 Entire Seeded Site 18.58

2011. These data were used to calculate the floristic quality index (FQI) and the Shannon diversity index (H’) for each plot. If the species identification of any plant was not possible due to the plant’s early life stage, it was flagged for later identification. Plants that were ultimately unidentified were not incorporated into the data analysis of percentage cover.

The percentage cover of each species, of bare-ground, and of litter percent were visually esti- mated in 2009 and 2010. In order to provide more objective quantitative data, percentage cover was determined in 2011 using the point-intercept method. This method involves dropping a metal pin at 50 points along an evenly spaced grid that covered the entire plot. Any vegetative part of a plant that touched the pin was counted as an occurrence of that plant at that particular point. Any piece of lit- ter or bare-ground that touched the pin was also accounted for. The total number of occurrences of an individual species in the plot was then divided by the total number of occurrences of all species in the plot to provide the relative percentage cover of that species. Any species noted visually as present in a plot that was not encountered by the point-intercept method was marked as being a trace amount occurrence (0.0625% cover) of the plot. In 2009, only the 144 non-legume plots were evaluated. In 2010 and 2011, all 228 plots were evaluated. In an effort to maintain consistent data collection ef- forts, estimates of percentage cover were calibrated between the same two researchers throughout the course of the multi-year study. This was a necessary approach, as cover estimates are subjective be- tween individuals.

Species richness was calculated by counting the number of species that occurred in each plot. Those species for which identification was not possible due to the plant’s early life stage were in- cluded as part of these calculations if they were identifiable to the extent that they could be deter- mined to be different from species otherwise noted in the plot. A list was kept of all species identi- fied, and all unknowns were described in detail.

Floristic quality index (FQI) measures the overall quality of an area (Swink and Wilhelm 1994) based on coefficient of conservatism (CC) values provided by the Michigan DNR (Hermann et al. 1996). Assigning CC values to individual plant species provides an approach to ranking the quality of plant communities (Swink and Wilhelm 1994). These values range from 0 to 10 and are assigned only to native species. A plant with a low value would represent a species that is common, can per- sist in highly degraded areas, and is likely not indicative of a high quality remnant community, for example, Solidago canadensis L., for which CC = 0, and Rudbeckia hirta L., for which CC=1. Con- versely, a species with a high value represents a species that would likely be found only in a place in- dicative of an intact remnant of a natural ecosystem, for example, Lithospermum canescens (Michx.)

Page  100 100 THE MICHIGAN BOTANIST Vol. 54

TABLE 5. List of species collected at the study site prior to the growth of any of the seed treatments. Non-native species are indicated by an asterisk (*). The coefficient of conservatism (CC) value for the native species are taken from Hermann et al. (1996).

Species CC Value Species CC Value

Acer rubrum L. 1 Ambrosia artemisiifolia L. 0 Andropogon gerardii Vitman 5 Asclepias syriaca L. 1 Bromus inermis Leyss.* Centaurea stoebe L.* Cichorium intybus L.* Cirsium vulgare* (Savi) Ten.* Clinopodium vulgare L. 3 Comptonia peregrina (L.) J.M. Coult. 6 Conyza canadensis (L.) Cronquist 0 Daucus carota L.* Elaeagnus umbellata Thunb.* Elymus repens (L.) Gould.* Erigeron strigosus Muhl. Ex Willd. 4 Fragaria virginiana Mill. 2 Holosteum umbellatum L.* Hypericum perforatum L.* Juniperus virginiana L. 3 Lepidium virginicum L. 0 Leucanthemum vulgare Lam.* Lupinus perennis L. 7 Melilotus alba (L.) Lam.* Melilotus officinalis (L.) Lam.* Monarda fistulosa L. 2 Monarda punctata L. 4 Oenothera biennis L. 2 Panicum capillare L. 1 Pinus strobus L. 3

Plantago rugelii Decne. 0 Populus tremuloides Michx. 1 Potentilla argentea L.* - Potentilla recta L.* - Pseudognaphalium obtusifolium (L.)

Hilliard & B.L. Burtt 2 Pteridium aquilinum (L.) Kuhn 0 Robinia viscosa Vent.* - Rudbeckia hirta L. 1 Sassafras albidum (Nutt.) Nees 5 Schizachyrium scoparium (Michx.) Nash 5 Solidago canadensis L. 1 Sorghastrum nutans (L.) Nash 6 Symphyotrichum laeve (L.) Á. Löve &

D. Löve 5 Symphyotrichum ontarionis (Wiegand) G.L. Nesom 6 Tragopogon pratensis L.* - Trifolium arvense L.* - Triosteum perfoliatum L. 5 Verbascum blattaria L.* - Verbascum thapsus L.* - Verbena hastata L. 4 Vitis aestivalis Michx. 6 Lehm.), for which CC=10, and Ceanothus americanus L., for which CC = 9 (Swink and Wilhelm 1994, Hermann et al. 1996). FQI is calculated by Equation 1, where C is the average of CC values from native plants species in the sampled community, and N is the total number of native species in the community. Non-native plant species are not included in this calculation.

Equation 1: FQI = C√N

According to Hermann et al. (1996), an FQI greater than 50 indicates an area of high conser- vatism, and an area with an FQI greater than 35 is considered to be floristically high quality in Michi- gan.

Diversity was expressed by the Shannon diversity index (H. ) (Shannon 1948), which equals (–1) times the sum for all native species in the sampled community of the product of the relative cover (pi) of each species and its natural logarithm, as shown in Equation 2.

N

H. = .. pi ln pi

i=1

Statistical Analysis

Measurements of the percentage of vegetative cover and of species richness were broken down into the previously mentioned seven different groupings: total, non-native, native, resident, planted, grass, and forb groupings.

Treatments were compared by both forb and grass groupings for percentage cover, species rich-

Page  101 2015 THE MICHIGAN BOTANIST 101

ness, and diversity. Grass groupings include control plots, background species only (treatment 15 in Table 2), forbs, including legumes with no grass (treatments 4, 8, and 12), high grass (treatments 2, 6, 10, 13), and low grass (treatments 1, 3, 5, 7, 9, 11, and 14). Forb groupings include control plots, background forb species only (treatments 13–15), early season flowering forbs (treatments 1-3), and late season flowering forbs (treatments 5–7), legumes (treatments 9-11), and forb treatments con- taining either early season flowering forbs, late season flowering forbs, or legumes that contained no grass (treatments 4, 8, and 12).

Exploratory analysis for percentage cover of bare-ground and litter values revealed that they were normally distributed. In order to produce normality in the other variables, a variety of transforma- tions were used. The percentage cover values for the native, planted, and grass groupings were square root transformed in order to correct for their positive skew. The percentage cover values for the non- native, resident, and forb groupings were transformed by formula, NormalValue(X)= √K–X. (where K is a constant from which each score is subtracted so that the smallest score is 1), to correct for their negative skew (Tabachnick and Fidell 2007). Once the data were normalized, a one-way analysis of variance (ANOVA) was used to test whether there were significant differences in per- centage cover between treatment groups in 2011. In order to further compare the differences among multiple comparison groups, Tukey post-hoc tests were run between all treatment groups. Compar- isons were considered significantly different at p . 0.05 after taking into account the Bonferroni cor- rection. All statistical analyses were performed using PASW 18 (SPSS Inc. 2011). Data are reported as non-transformed values.

Exploratory analysis for species richness and diversity values revealed that they were normally distributed. A one-way ANOVA was used to test whether there were significant differences among treatment groups in species richness and diversity in 2011. In order to further compare the differ- ences among multiple comparison groups, Tukey post-hoc tests were run between all treatment groups. Comparisons were considered significantly different at p . 0.05 after taking into account the Bonferroni correction. All statistical analyses were performed using PASW 18 (SPSS Inc. 2011).

An independent sample T-test was used to determine if there were significant differences in mea- sured variables between spring seeded treatments and fall seeded legume treatments. All treatments were statistically similar.

RESULTS

Baseline Floristic Quality Assessment

A total of 50 different plant species were encountered during the baseline floristic quality assessment of the surrounding areas of the study site in 2009, in- cluding 31 native and 19 non-native species (Table 5). The average coefficient of conservatism (CC) value was 2.94, resulting in a calculated floristic quality index (FQI) of 16.37, which indicates that the area was not considered a floristi- cally high quality site (Hermann et al. 1996). A list of resident species encoun- tered within the study plots by year is provided in Table 6.

Baseline vegetative cover data indicate that the most prevalent species were Conyzacanadensis(L.) Cronquist (mean = 41%), CentaureastoebeL. (mean = 23%), and PotentillaargenteaL. (mean = 22%). The most prevalent grass species were BromusinermisLeyss. (mean = 1%) and Agrostishyemalis(Wal- ter) Britton, Sterns & Poggenb. (mean = 1%).

Comparison among Grass Treatment Groups after ThreeYears

PercentageCover

The data reported in this section are summarized in Table 7. Significant dif- ferences (p<0.050) were found among treatment groups in the percentage cover

Page  102 TABLE 6. Resident species encountered within the study site in each of the years of the study. Non-native species are indicated by an asterisk (*).

2009 2010 2011

Acer rubrum L. Acer rubrum L. Asclepias syriaca L.

Agrostis hyemalis (Walter) Brit. , Sterns & Pogg. Agrostis hyemalis (Walter) Brit. , Sterns & Pogg. Acer rubrum L.

Arenaria serpyllifolia L. * Arenaria serpyllifolia L. * Agrostis hyemalis (Walter) Brit. Sterns Pogg.

Asclepias syriaca L. Asclepias syriaca L. Arenaria serpyllifolia L.

Berteroa incana (L.) DC. * Berteroa incana (L. ) DC. * Berteroa incana (L. DC.

Bromus inermis Leyss. * Bromus inermis Leyss. * Bromus inermis Leyss.

Centaurea stoebe L. * Centaurea stoebe L. * Centaurea stoebe L.

Clinopodium vulgare L. Clinopodium vulgare L. Clinopodium vulgare L.

Conyza canadensis (L. ) Cronquist Conyza canadensis (L. ) Cronquist Conyza canadensis (L. Cronquist

Danthonia spicata (L. ) Roem. & Schult. Danthonia spicata (L. ) Roem. & Schult. Danthonia spicata (L. Roem. Schult.

Elymus repens (L. ) Gould* Elymus repens (L. ) Gould* Elymus repens (L. Gould*

Erigeron annuus (L. ) Desf. Erigeron annuus (L. ) Desf. Fragaria virginiana Mill.

Fragaria virginiana Mill. Fragaria virginiana Mill. Galium aparine L.

Gnaphalium obtusifolium L. Gnaphalium obtusifolium L. Gnaphalium obtusifolium L.

Hieracium auranticum L. * Hieracium auranticum L. * Hieracium auranticum L.

Hieracium caespitosum Dumort. * Hieracium caespitosum Dumort. * Hieracium caespitosum Dumort.

Hypericum perforatum L. * Hypericum perforatum L. * Hypericum perforatum L.

Lepidium virginicum L. Lepidium virginicum L. Lepidium virginicum L.

Leucanthemum vulgare Lam. * Leucanthemum vulgare Lam. * Leucanthemum vulgare Lam.

Oxalis stricta L. Oxalis stricta L. Monarda punctata L.

Panicum capillare L. Panicum capillare L. Oxalis stricta L.

Physalis heterophylla Nees Physalis heterophylla Nees Panicum capillare L.

Poa compressa L. * Poa compressa L. * Physalis heterophylla Nees

Potentilla argentea L. * Potentilla argentea L. * Poa compressa L.

Potentilla recta L. * Potentilla recta L.* Potentilla argentea L.

Prunus serotina Ehrh. Prunus serotina Ehrh. Potentilla recta L.

Quercus rubra L. Quercus rubra L. Prunus serotina Ehrh.

Rubus occidentalis L. Rubus occidentalis L. Quercus rubra L.

Rumex acetosella L. * Rumex acetosella L. * Rubus occidentalis L.

Taraxacum offi cinale F.H. Wigg. * Taraxacum offi cinale F.H. Wigg. * Rumex acetosella L.

Tragopogon pratensis L. * Tragopogon pratensis L. * Taraxacum offi cinale F.H. Wigg.*

Trifolium arvense L. * Trifolium arvense L. * Tragopogon pratensis L.

Trifolium pratense L. * Trifolium pratense L. * Trifolium arvense L.

Trifolium repens L. * Trifolium repens L. * Trifolium repens L.

Verbascum thapsus L. * Verbascum thapsus L. * Verbascum thapsus L.

Veronica arvensis L. * Veronica arvensis L. *

102 THE MICHIGAN BOTANIST Vol. 54

Page  103 2015 THE MICHIGAN BOTANIST 103

of the bare-ground (F4,227=5.79, p<0.001), native (F4,227=11.20, p<0.001), non- native (F4,227=11.41, p<0.001), resident (F4,227=24.09, p<0.001), planted (F4,227=23.90, p<0.001), grass (F4,227=30.39, p=0.014), and forb (F4,227=30.39, p=0.014) groupings. There were no significant differences (p>0.050) among treatment groups in litter cover (F4,227=0.70, p=0.595, mean = 17.86%).

Bare-ground cover averaged 36.03% across all grass treatment groups. High grass treatments had significantly lower bare-ground cover (mean = 31.85%, p<0.050) than all other treatment groups except the control plots (mean = 37.20%), which showed high variability.

Native cover averaged 19.34% across all grass treatment groups. High grass treatments had significantly higher native cover (mean = 35.02%, p<0.050) than all other treatment groups. Low grass treatments had significantly higher native cover (mean = 22.99%, p<0.050) than all other treatment groups except the high grass treatments. There were no significant differences in native cover among all other treatment groups (mean range: 8.01–16.94%) Comparison of mean native cover in each of these treatment groups is shown in Figure 2.

Non-native cover averaged 80.54% across all grass treatment groups. High grass treatments had significantly lower non-native cover (mean = 64.38%, p<0.050) than all other treatment groups, while low grass treatments had signif- icantly lower cover (mean = 77.01%, p<0.050) than all other treatment groups except the high grass treatment. There were no significant differences in non-na-

FIGURE 2. Grass Treatment Comparison: Mean relative percentage cover of species in the native grouping for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

Page  104 104 THE MICHIGAN BOTANIST Vol. 54

FIGURE 3. Grass Treatment Comparison: Mean relative percentage cover of species in the Non-Na- tive grouping for several treatment groups in 2011. Bars that do not share the same letter are signif- icantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

tive cover among all other treatment groups (mean range: 83.06–91.99%). Com- parison of mean non-native cover in each of these treatment groups is shown in Figure 3.

Resident cover averaged 86.45% across all grass treatment groups. Resident cover in control plots (mean = 99.83%) was significantly greater (p<0.050) than all other treatments (mean range: 68.00–91.12%). High grass treatments had sig- nificantly lower resident cover (mean = 68.00%, p<0.050) than all other treat- ments. Low grass treatments had significantly lower (mean = 82.71%, p<0.050) resident cover than the treatment with background only species and no grass treatment and the control plots.

Planted cover averaged 13.43% across all grass treatment groups. Control plots had significantly less planted cover (mean = 0.17%, p<0.050) than all other treatment groups (mean range: 8.88–31.40%). High grass treatments had signif- icantly higher planted cover (mean = 31.40%, p<0.050) than all other treatment groups. Low grass treatments had significantly higher planted cover (mean = 17.29%, p<0.050) than the background only treatments with no grass treatments (mean = 8.88%), but was not significantly different (p>0.050) from the forb treatments with no grass (mean = 9.43%).

Grass cover averaged 10.35% across all grass treatment groups. High grass treatments had significantly more grass cover (mean = 28.47%, p<0.050) than all other treatment groups. Low grass treatments had significantly higher grass

Page  105 TABLE 7. Mean percentage cover values of bare-ground and litter and of the native, non-native, resident, planted, grass, and forb groupings in the grass treat- ment groups in 2011. Percentage cover in each grouping was tested only against the treatments within the same grouping. Values within column that do not sharea superscript letter are significantly different (p≤0.05) as determined by Tukey post-hoc analysis.

TreatmentGroup Bare-ground Litter Native Non-Native Resident Planted Grass Forb

Control Plots 37.20ab(±1.56) 18.45(±1.66) 8.01a(±1.93) 91.99a(±1.93) 99.83(±0.17) 0.17(±0.17) 3.32(±1.22) 96.68(±1.22)

Background 37.04b(±1.62) 16.93(±1.35) 13.73ab (±4.59) 86.27ab (±4.59) 91.12b(±4.25) 8.88b(±4.25) 3.88(±2.35) 96.12(±2.35) Only, No Grass

Forb 38.38b(±1.09) 18.17(±1.00) 16.94ab (±3.29) 83.06ab (±3.29) 90.57ab (±2.03) 9.43ab (±2.03) 4.93(±2.23) 95.07(±2.23) Treatment, No Grass

High Grass 31.85a(±1.20) 18.65(±0.51) 35.02(±3.28) 64.38(±3.28) 68.00c(±3.17) 31.40c(±3.17) 28.47b(±3.01) 71.53a(±3.01) Low Grass 35.67b(±0.64) 17.66(±0.50) 22.99b(±2.06) 77.01b(±2.06) 82.71a(±1.92) 17.29a(±1.92) 11.14a(±1.40) 88.86b(±1.40)

TABLE 8. Mean species richness values of total, native, non-native, resident, planted, grass, and forb groupings in the grass treatment groups in 2011. Speciesrichness in each grouping was tested only against the treatments within the same grouping. Values within a column that do not share superscript letter are sig- nificantly different (p≤0.05) as determined by Tukey post-hoc analysis.

TreatmentGroup Total Native Non-Native Resident Planted Grass Forb

Control Plots 7.75b(±0.54) 2.33(±0.28) 5.25(±0.49) 7.33(±0.47) 0.25(±0.13) 0.83(±0.24) 6.92(±0.36) Background Only, No Grass 9.92bc (±0.60) 4.63b(±0.61) 5.25(±0.41) 7.08(±0.60) 2.79b(±0.44) 0.75(±0.32) 9.17a(±0.50) Forb Treatment, No Grass 10.61c(±0.49) 5.33b(±0.42) 5.28(±0.23) 7.25(±0.29) 3.36b(±0.29) 0.94(±0.19) 9.69a(±0.38) High Grass 12.88a(±0.31) 7.48a(±0.26) 5.37(±0.21) 7.25(±0.21) 5.60a(±0.19) 3.50b(±0.09) 9.40a(±0.28)

Low Grass 12.49a(±0.29) 6.96a(±0.25) 5.50(±0.15) 7.43(±0.20) 5.03a(±0.19) 2.63a(±0.13) 9.86a(±0.23)

2015 THE MICHIGAN BOTANIST 105

Page  106 106 THE MICHIGAN BOTANIST Vol. 54

cover (mean = 11.14%, p<0.050) than all other treatment groups except the high grass treatments.

Forb cover averaged 89.65% across all grass treatment groups. High grass treatments had significantly less forb cover (mean = 71.53%, p<0.050) than all other treatment groups. Low grass treatments had significantly less forb cover (mean = 88.86%, p<0.050) than all other treatment groups except the high grass treatments.

Species Richness

The data reported in this section are summarized in Table 8. There were sig- nificant differences (p<0.050) in species richness among treatment groups in the total (F4,227=16.09, p<0.001), native (F4,227=19.78, p<0.001), planted (F4,227=40.16, p<0.001), grass (F4,227=51.68, p<0.001), and forb (F4,227=5.32, p<0.001) groupings. There were no significant differences (p>0.050) in species richness among grass treatment groups in the non-native (F4,227=0.26, p=0.905, mean = 5.33 species) and resident (F4,227=0.21, p=0.932, mean = 7.27 species) groupings.

The total species richness averaged 10.73 species across all grass treatment groups. It averaged significantly lower for the control plots (mean = 7.75 species, p<0.050) than for all forb treatments with no grass (mean = 10.61 species), high grass treatments (mean = 12.88 species), and low grass treatments (mean = 12.49 species). High grass and low grass treatments exhibited signifi- cantly higher total species richness than all other grass treatment groups (p<0.050). Comparison of mean total species richness in each of these treatment groups is shown in Figure 4.

Species richness of the native grouping averaged 5.35 species across all grass treatment groups. The control plots had significantly lower native species rich- ness (mean = 2.33 species, p<0.050) than all other treatment groups (mean range: 4.63–7.48 species). High grass and low grass treatments had significantly higher native species richness (mean = 7.48 and 6.96 species, respectively, p<0.050) than all other treatment groups. Comparison of mean native species richness in each of these treatment groups is shown in Figure 5.

Species richness of the planted grouping averaged 3.41 species across all grass treatment groups. All treatment groups had significantly higher planted species richness (mean range: 2.79–5.60 species, p<0.050) than the control plots (mean = 0.25 species). High grass and low grass treatments had significantly higher planted species richness (mean = 5.60 and 5.03 species, respectively, p<0.050) than all other treatment groups.

Species richness of the grass grouping averaged 1.73 species across all grass treatment groups. High grass treatments had significantly higher grass species richness (mean = 3.50 species, p<0.050) than all other treatments. Low grass treatments had significantly higher grass species richness (mean = 2.63 species, p<0.050) than all other treatments.

Species richness of the forb groupings averaged 9.01 species across all grass treatment groups. All treatment groups had significantly higher forb species richness (mean range: 9.17–9.86 species, p<0.050) than the control plots (mean = 6.92 species).

Page  107 2015 THE MICHIGAN BOTANIST 107

FIGURE 4. Grass Treatment Comparison: Mean species richness encountered in the Total grouping for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

FIGURE 5. Grass Treatment Comparison: Mean species richness encountered in the Native grouping for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

Page  108 108 THE MICHIGAN BOTANIST Vol. 54

TABLE 9. Comparison of the mean floristic quality index (FQI) and Shannon diversity index (H’) values of the grass treatment groups and selected non-grass treatment groups in 2011. Each different measure was tested independently. Values within a column that do not share a superscript letter are significantly different (p≤0.05) as determined by Tukey post-hoc analysis.

Treatment Group FQI H. Control Plots 1.37(±0.56) 1.26b(±0.09) Background Only, No Grass 6.56a(±0.45) 1.40ab(±0.06) Forb Treatment, No Grass 6.71a(±0.39) 1.34b(±0.06) High Grass 10.37c(±0.17) 1.61a(±0.04) Low Grass 9.35b(±0.18) 1.49ab(±0.04)

Floristic Quality and Diversity

There were significant differences among treatment groups in FQI values (F4,227=87.83, p<0.001). The mean FQI of grass treatment groups was 6.87. The FQI of the control plots in 2011 was significantly lower (mean = 1.37) than that of all other treatments (mean range: 6.56–10.37). The high grass treatments had a significantly higher FQI (mean = 10.37) than all other treatment groups while low grass treatments had a significantly higher FQI (mean = 9.35) than all other treatments except the high grass treatments (Table 9; Figure 6).

FIGURE 6. Grass Treatment Comparison: Mean FQI for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

Page  109 2015 THE MICHIGAN BOTANIST 109

FIGURE 7. Grass Treatment Comparison: Mean H’ for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

There were also significant differences among treatment groups in the Shan- non diversity index (H’) (F4,227=4.43, p=0.002) (Table 9). The mean H’ of the grass treatment groups was 1.42. The mean H. of the high grass treatments was significantly higher (mean = 1.61) than for all other treatments except the back- ground species only with no grass treatment (mean = 1.40) and the low grass treatment (mean = 1.49) (Table 9; Figure 7).

Comparison among Forb Treatment Groups after Three Years

Percentage Cover

The data reported in this section are summarized in Table 10. Significant dif- ferences (p<0.050) were found among treatment groups in the percentage cover of the bare-ground (F4,227=2.99, p=0.020), native (F4,227=8.75, p<0.001), non- native (F4,227=8.33, p<0.001), resident (F4,227=17.72, p<0.001), planted (F4,227=18.51, p<0.001), grass (F4,227=3.21, p=0.014), and forb (F4,227=3.21, p=0.014) groupings. There were no significant differences (p>0.050) among treatment groups in litter cover (F4,227=0.16, p=0.959, mean = 18.04%).

Bare-ground cover averaged 35.59% among forb treatment groups. The late season flowering forb treatment had significantly less (p<0.050) bare-ground cover than the legume treatments (mean = 33.40% and 37.96%, respectively).

Native cover averaged 21.24% among forb treatment groups. Treatments of early season flowering forbs and late season flowering forbs showed significantly

Page  110 110 THE MICHIGAN BOTANIST Vol. 54

FIGURE 8. Forb Treatment Comparison: Mean relative percentage cover of species in the Native grouping for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

higher percent native cover (mean = 34.10% and 27.19%, respectively, p<0.050) than the control plots (mean = 8.01%) and legume treatments (mean = 14.68%). Comparison of mean native cover in each of these treatment groups is shown in Figure 8.

Non-native cover averaged 78.61% among forb treatment groups. The control plots and legume treatments showed significantly higher non-native cover (mean = 91.99% and 84.57%, respectively, p<0.050) than early season and late season flowering forb treatments (mean = 65.90%, 72.81%, respectively). Comparison of mean non-native cover in each of these treatment groups is shown in Figure 9.

Resident cover averaged 84.39% among forb treatment groups. Resident cover in control plots (mean = 99.83%) was significantly greater (p<0.050) than in all other treatments (mean range: 71.16%–90.96%). Early season flowering forbs and late season flowering forbs (mean = 71.16% and 76.46%, respectively) had significantly lower (p<0.050) resident cover than all other treatments.

Planted cover averaged 15.46% among forb treatment groups. Control plots had significantly less planted cover (mean = 0.17%, p<0.050) than all other treatment groups (mean range: 8.29%–28.84%). Early season flowering forb and late season flowering forb treatments (mean = 28.84% and 23.54%, respectively) had significantly more (p<0.050) planted cover than all other treatment groups.

Grass cover averaged 11.93% among forb treatment groups. All treatment groups had significantly higher grass relative cover (mean range:

Page  111 2015 THE MICHIGAN BOTANIST 111

FIGURE 9. Forb Treatment Comparison: Mean relative percentage cover of species in the Non-Na- tive grouping for several treatment groups in 2011. Bars that do not share the same letter are signif- icantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

8.30%–17.48%, p<0.050) than the control plots (mean = 3.32%). Early season flowering forbs and late season flowering forbs had significantly higher grass cover (mean = 17.48% and 16.18%, respectively, p<0.050) than all other treat- ment groups except the background only treatments (mean = 14.36%.

Forb cover averaged 88.07% among forb treatment groups. All treatment groups had significantly lower (p<0.050) forb (mean range: 82.52%–91.70%) relative cover than the control plots (mean = 96.68%). Early season flowering forbs and late season flowering forbs had significantly lower forb cover (mean = 82.52% and 83.82%, respectively) than all other treatment groups except the background only treatments (mean = 85.64%.

Species Richness

The data reported in this section are summarized in Table 11. There were sig- nificant differences (p<0.050) in species among treatment groups in the total (F4,227=9.99, p<0.001), native (F4,227=15.77, p<0.001), planted (F4,227=26.28, p<0.001), grass (F4,227=3.98, p=0.004), and forb (F4,227=9.96, p<0.001) group- ings. There were no significant differences (p>0.050) in species richness among forb treatment groups in the non-native (F4,227=0.65, p=0.629, mean = 5.35 species) and resident (F4,227=0.38, p=0.825, mean = 7.29 species) groupings.

The total species richness averaged 11.20 species among forb treatment groups. It averaged significantly less for the control plots (mean = 7.75 species,

Page  112 TABLE 10. Mean percentage cover values of bare-ground and litter and of the native, non-native, resident, planted, grass, and forb groupings in the grass treat- ment groups in 2011. Percentage cover in each grouping was tested only against the treatments within the same grouping. Values within column that do not sharea superscript letter are significantly different (p≤0.05) as determined by Tukey post-hoc analysis.

TreatmentGroup Bare-ground Litter Native Non-Native Resident Planted Grass Forb

Control Plots 37.20ab(±1.56) 18.45(±1.66) 8.01a(±1.93) 91.99a(±1.93) 99.83(±0.17) 0.17(±0.17) 3.32(±1.22) 96.68(±1.22) Background Forbs Only 35.41ab(±0.77) 17.97(±0.50) 22.24ab (±2.39) 77.76ab (±2.39) 83.55a(±2.26) 16.45a(±2.26) 14.36ab (±2.08) 85.64ab (±2.08)

Legumes 37.96a(±1.01) 17.69(±0.80) 14.68a(±2.39) 84.57a(±2.39) 90.96a(±1.58) 8.29a(±1.58) 8.30a(±1.80) 91.70a(±1.80) Early Season 33.99ab(±1.22) 17.74(±0.68) 34.10c(±3.80) 65.90c(±3.80) 71.16b(±3.65) 28.84b(±3.65) 17.48bc (±3.14) 82.52bc (±3.14) Late Season 33.40b(±1.13) 18.33(±0.75) 27.19bc (±3.31) 72.81bc (±3.31) 76.46b(±3.12) 23.54b(±3.12) 16.18bc (±3.15) 83.82bc (±3.15)

TABLE 11. Mean species richness values of total, native, non-native, resident, planted, grass, and forb groupings in the forb treatment groups in 2011. Speciesrichness in each grouping was tested only against the treatments within the same grouping. Values within a column that do not share superscript letter are sig- nificantly different (p≤0.05) as determined by Tukey post-hoc analysis.

TreatmentGroup Total Native Non-Native Resident Planted Grass Forb

Control Plots 7.75(±0.54) 2.33(±0.28) 5.25(±0.49) 7.33(±0.47) 0.25(±0.13) 0.83a(±0.24) 6.92(±0.36) Background Forbs Only 11.53a(±0.29) 5.93a(±0.25) 5.58(±0.17) 7.51(±0.21) 4.00a(±0.19) 2.46b(±0.19) 9.08b(±0.21)

Legumes 11.83ab (±0.49) 6.46a(±0.43) 5.35(±0.23) 7.21(±0.29) 4.60a(±0.32) 2.06ab (±0.22) 9.77ab (±0.38) Early Season 13.25b(±0.44) 8.06b(±0.37) 5.15(±0.19) 7.29(±0.27) 5.92b(±0.29) 2.56b(±0.22) 10.71a(±0.33) Late Season 11.63a(±0.35) 6.17a(±0.30) 5.42(±0.23) 7.13(±0.26) 4.46a(±0.20) 2.40b(±0.19) 9.23b(±0.26)

112 THE MICHIGAN BOTANIST Vol. 54

Page  113 2015 THE MICHIGAN BOTANIST 113

FIGURE 10. Forb Treatment Comparison: Mean species richness encountered in the Total grouping for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

p<0.050) than for all other treatments. The early season flowering forb treatment had significantly higher total species richness (mean = 13.25 species, p<0.050) than all other treatments except the legumes (mean = 11.83 species). Compari- son of mean total species richness in each of these treatment groups is shown in Figure 10.

Species richness of the native grouping averaged 5.79 species among forb treatment groups. The control plots had significantly lower native species rich- ness (mean = 2.33 species, p<0.050) than all other treatment groups. The early season flowering forb treatment group had significantly higher native species richness (mean = 8.06 species, p<0.050) than all other treatment groups (mean range: 5.93–6.46 species). Comparison of mean native species richness in each of these treatment groups is shown in Figure 11.

Species richness of the planted grouping averaged 3.85 species among forb treatment groups. All treatment groups had significantly higher (p<0.050) planted species richness than the control plots (mean = 0.25 species). The early season flowering forbs had significantly greater planted species richness (mean = 5.92 species, p<0.050) than all other treatment groups (mean range: 4.00–4.60 species).

Species richness of the grass grouping averaged 2.06 species among forb treatment groups. All treatments except for legumes (mean = 2.06 species) ex- hibited significantly higher grass species richness (mean range: 2.40–2.56 species, p<0.050) than the control plots (mean = 0.83 species).

Page  114 114 THE MICHIGAN BOTANIST Vol. 54

FIGURE 11. Forb Treatment Comparison: Mean species richness encountered in the Native group- ing for several treatment groups in 2011. Bars that do not share the same letter are significantly dif- ferent (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

Species richness of the forb grouping averaged 9.14 species among forb treat- ment groups. The control plots had significantly lower forb species richness (mean = 6.92 species, p<0.050) than all forb treatment groups (mean range: 9.08–10.71 species). The early season flowering forbs had significantly higher (mean = 10.71 species, p<0.050) forb species richness than all other treatment groups except for legumes (mean = 9.77 species).

Floristic Quality and Diversity

There were significant differences among treatment groups in FQI values (F4,227=37.96, p<0.001). The mean FQI of forb treatment groups was 7.43. The FQI of the control plots in 2011 was significantly lower (mean = 1.37, p<0.050) than that of all forb treatments (mean range: 8.39–9.99). The early season flow- ering forb treatments had a significantly higher FQI (mean = 9.99, p<0.050) than all other treatments except the legumes (mean = 8.89) (Table 12; Figure 12).

There were also significant differences among treatment groups in the Shan- non diversity index (H.)(F4,227=4.86, p=0.001) (Table 12). The mean H. of forb treatment groups was 1.44. The mean H. of the control plots and legume treat- ments (mean = 1.26, 1.33, respectively) was significantly lower (p<0.050) than for the early season flowering forb treatments (mean = 1.62) (Table 12; Figure 13).

Page  115 2015 THE MICHIGAN BOTANIST 115

FIGURE 12. Forb Treatment Comparison: Mean FQI for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

TABLE 12. Comparison of the mean floristic quality index (FQI) and Shannon diversity index (H.) values of the forb treatment groups in 2011. Each different measure was tested independently. Values within a column that do not share a superscript letter are signifi- cantly different (p≤0.05) as determined by Tukey post-hoc analysis.

Treatment Group FQI H. Control Plots Background Forbs Only 1.37(±0.56) 8.39b(±0.23) 1.26a(±0.09) 1.50ab(±0.04) Legumes Early Season Late Season 8.89ab(±0.40) 9.99a(±0.31) 8.51b(±0.27) 1.33a(±0.06) 1.62b(±0.05) 1.50ab(±0.05)

DISCUSSION

Comparison Among Grass Treatment Groups

The percentage cover of resident species was significantly lower in all grass seeded treatments (i.e., no grass, low grass, and high grass), which was likely due to a change in competitive balance (Connell 1983, Fargione and Tilman 2006). The percentage cover of the planted grouping, the species richness of the native and planted groupings, and the FQI were all significantly higher in all the

Page  116 116 THE MICHIGAN BOTANIST Vol. 54

FIGURE 13. Forb Treatment Comparison: Mean H’ for several treatment groups in 2011. Bars that do not share the same letter are significantly different (p≤.05) as determined by Tukey post-hoc analysis. Error bars represent one standard error about the mean.

seeded grass plots than in the control plotsm which is due, in part, to an increase in forb species richness in the grass treatments. The baseline data (Table 3) indi- cates that the community was heavily forb-dominated in 2009. However, native forb species included in the grass treatments further increased the species rich- ness of forbs. Although some grass treatments included forbs, the grass-only treatments also had significantly greater forb species richness than the control plots. Forb species may be benefiting from seeded treatments, because as seeded species grow, they may facilitate the growth of other seeded species around them (Callaway and Walker 1997, Peltzer and Kochy 2001). This may also apply to non-seeded species in the surrounding area. Species currently on or near the site, as well as those species whose seeds are currently dormant in the seed bank, may realize more favorable conditions for germination from the establishment of seeded species, and thereby increase the species richness of forbs.

Comparison Among Forb Treatment Groups

The percentage cover of resident and forb species was significantly higher in the control plots than in the seeded forb treatments. Conversely, the percentage cover of the planted grouping, the species richness of the total, native, planted, and forb groupings, and the FQI were all significantly higher in all seeded forb treatments than in the control plots. Resident cover may be higher in control plots due to reduced competition for available open niche space resulting from

Page  117 2015 THE MICHIGAN BOTANIST 117

the lack of seed additions in these plots. In contrast to the grass-only treatments, forb cover was higher in treatments that included the addition of forbs. Similarly, planted cover was higher in seeded plots than in the control plots due to the ad- dition of native seed. The species richness of the total, native, and planted group- ings and the FQI all increased with the addition of seeded forbs.

Community Responses to Grass Seeding and Forb Seeding Efforts

Treatments with either high or low concentrations of grasses were not signif- icantly different from each other in any measure of species richness other than grass species richness. High grass treatments had significantly higher grass species richness than both low and no grass treatments. Both low and high grass treatments had significantly more total and planted species richness than treat- ments with no grass.

High grass treatments were more effective in covering more ground and hence had lower bare-ground cover than all other treatments except the highly variable control plots. High grass treatments also had a significantly higher FQI. Diversity (H.) did not differ significantly between high and low grass treatments. However, these treatments were significantly more diverse than all other treatments except for the background only species with no grass treat- ment.

Although the cover was dominated by grasses, the high grass plots displayed significantly higher native and planted cover than any other treatment. This may be attributed to the inclusion of native species within these planting groups. However, high grass only plots consisting of no planted forb species still exhib- ited significantly higher native and planted cover than treatments that included forbs and low grass concentrations. High grass treatments also had significantly lower non-native and resident cover than all other treatments.

Native and planted covers were significantly greater in the early season forb and late season forb treatments than in all other forb treatments. Conversely, non- native and resident covers were significantly less in early season and late season forb treatments than in all other treatments.

Early season forb treatments had significantly higher native and planted species richness than all other treatments. Although early season forb treatments had a significantly higher FQI than most other treatments, the FQI of legume treatments were not significantly different. Early season forb treatments also were significantly more diverse (H’) than most other treatments except the back- ground only and late season forb treatments.

Facilitation

These results suggest that seed additions may facilitate plant community de- velopment in the relatively hot and dry conditions of a sand prairie during the first few years of community development. The increased growth of native species within the grass treatments—especially within high grass seed treat- ments—as compared with treatments that did not contain grass suggests that seeded grasses may provide suitable microclimatic conditions for native seed

Page  118 118 THE MICHIGAN BOTANIST Vol. 54

growth, such as shade for reduced heat stress and moisture capture (Plumb-Men- tjes 1990; Smith and Huston 1990; Peltzer and Kochy 2001). Similarly, the growth of early season forbs may create greater cover and biomass early in the year which could result in greater shade and moisture capture for late season forb and grass species (Henderson et al. 1988; Brooker et al. 2007; Callaway and Walker 1997).

Similar results indicating a facilitative effect from seed additions have been shown in an experiment where bunchgrasses protected rare plant species during dry years (Greenlee and Callaway 1996) and in other studies where adult plants provided a facilitative response during restoration in arid ecosystems (Maestre et al. 2001; Barchuk et al. 2005). Studies have also attempted to model light and moisture capture of species in dry, harsh environments such as sand prairies. Holmgren et al. (1997) and Smith and Huston (1990) studied the role facilitation plays in community development by modeling the positive and negative effects that “nurse plant” or mature plant canopy cover plays on the establishment of new plants in dry communities. This model shows that the increased availability of water due to facilitation outweighs the detrimental effect of shade (decreased light and photosynthesis) in harsh conditions. Therefore, facilitative effects may outweigh competitive effects of neighboring plants in such environments.

Competition

These results also indicate that competition resulting from seed additions may play a key role in promoting initial community development. Individual resident species respond in different ways to species additions based on their ability to compete for limited resources (Grime 1974, Goldberg and Barton 1992, Smith et al. 1999). However, the overall decrease in resident species growth within the grass treatments—especially within high grass seed treatments—suggests that resident species are at a competitive disadvantage compared to seeded species. Conversely, the overall increased growth of seeded species in the grass treat- ments suggests that species that are native to the sand prairie may be better able to compete for the limited resources than other early successional, weedy, or non-native resident species such as Centaurea stoebe L.

Similar results indicating a competitive effect from grass seed additions have been shown by Jordan et al. (2008), where they found that native grasses were less affected by invasive species. That study found that grasses are relatively in- sensitive to altered soil biota from invasive plants, and in turn, may shift the competitive balance of restoration efforts on an ecosystem. Similar studies sug- gest that grasses in prairie restoration enhance overall diversity and reduce exotic species cover (Middleton et al. 2009; Carter and Blair 2012). These studies also indicate that native grasses in restoration may represent an effective management strategy to reduce exotic plant density. Although grass seeding in restorations has been shown to be beneficial, long-term studies have shown that grasses can dom- inate and exclude other native species over time (Schramm 1990).

These results also indicate that forb additions—particularly a benefit of early season forb treatments and a lack of response from legume treatments— may play a key role in influencing community competitive balance (Howe

Page  119 2015 THE MICHIGAN BOTANIST 119

1994; Dukes 2001). Additions of early season forb species may provide an ini- tial, early-in-the-year increase in species cover and richness that may in turn directly affect the competitive balance of a community by providing greater re- sistance or buffer capability to further invasions by invasive species, which in turn may show increased diversity from other native seeded species later in the growing season (Naeem and Wright 2003; Kennedy et al. 2002; Dukes 2001). Contrary to early season forb treatments, legume treatments, although ex- pected to be highly competitive in a nutrient-poor sand prairie ecosystem due to their ability to fix nitrogen, had little effect on plant community composi- tion. Three competitive reasons may explain why we did not see significant es- tablishment of seeded legume species. The dominant and widespread resident legume species Trifolium arvense L. may have adverse impacts on seeded legumes. It may suppress seed additions either via interspecific competition, being temporally competitive due to its annual life cycle and producing large amounts of seeds both in the spring and summer, or possibly because nitrogen fixing niche space is already being consumed by Trifolium arvense L. (Dukes 2001; Fargione et al. 2003).

Similar experimental results have suggested that seeding of early season forbs within a restoration may provide positive competitive effects to a community. Martin et al. (2005) suggested that restoration seeding efforts that contain early season forbs may be more diverse because these species are better able to co- exist with other, later growing seeded species because they come to occupy an early season niche. Thus, early season forbs will ultimately increase diversity, species richness, and other important community variables. Similarly, a study by Foster and Tilman (2003) suggests that early season forb seeding (as part of a complete restoration seed mix) presents an opportunity for transient coexistence through competition-colonization trade-offs, thus allowing many species to co- exist at the community scale. Additionally, Seabloom et al. (2003) proposed that a single seeding of native forbs (including early season forb species), even in the presence of high densities of exotic species, may be sufficient to create viable populations of native species in areas that are currently dominated by exotic species.

Similar results showing a lack of legume establishment after seed additions have been found. In a twenty-five year study of prairie establishment following restoration, Schramm (1990) found that legumes did not become a major com- ponent of restorations until later phases of restoration succession. However, after legumes became more established in later years of restoration, it was found that they had greater staying power. This could be due to a lack of com- petitive ability during their early establishment periods (Schramm 1990). This may also be due to interspecific competition for limiting resources among legumes with neighboring plants belonging to the same functional group (Nemec et al. 2013). Most studies that have shown minimal impacts of legumes (species cover and richness) on restorations have been in response to herbivory. Restoration attempts that were not excluded by fencing and had populations of deer or voles saw high mortality or reduced fruiting in legume species due to the grazing preference of these animals for these species (Diaz et al. 2003; Howe et al. 2002).

Page  120 120 THE MICHIGAN BOTANIST Vol. 54

CONCLUSION

In general, a comparison between seeded treatments and non-seeded control treatments indicates that our efforts have been more successful in the restoration of native sand prairie than would have resulted from succession alone. Restora- tion attempts displayed a significant decrease in invasive, resident species rich- ness and increased diversity compared to succession. Treatments that included a high concentration of grass and/or an early season forb component had the great- est overall positive impact on plant community development. These treatments exhibited significantly greater diversity and higher floristic quality than most other treatments, and they also displayed less non-native or invasive cover than most other treatments. High concentrations of grasses and early season forbs may therefore play a critical role in initial species establishment of a sand prairie restoration due to the facilitative and competitive advantages they may provide in these harsh environments. However, it remains to be seen if these initially suc- cessful communities will have continued success over longer periods of time.

ACKNOWLEDGMENTS

This article is based on the graduate research and associated master’s program thesis completed as part of the first author’s graduate requirements at Grand Valley State University, Allendale, Michi- gan. The thesis upon which this article is based can be found at Roos (2014). He thanks his Gradu- ate Committee for their support and advice: Todd A. Aschenbach (the second author of this article and chair of the committee), Tim Evans, of Grand Valley State University, and David Warners, of Calvin College, Grand Rapids, Michigan.

LITERATURE CITED

Albert, D. A. (1995). Regional landscape ecosystems of Michigan, Minnesota, and Wisconsin: A working map and classification. Gen. Tech. Rep. NC-178. St. Paul: USDA, Forest Service, North Central Forest Experiment Station, St. Paul, Minnesota. Available at http://nr.fs.fed.us/pubs/242 (Version 03UN1998).

Albert, D. A., and P. J. Comer. (2008). Atlas of early Michigan’s forests, grasslands, and wetlands. Michigan State University Press, East Lansing.

Anderson, R. C. (1990). The historic role of fire in the North American grassland. Pp. 8–18 in Fire in North American tallgrass prairies. S. L. Collins and L. L. Wallace, editors. University of Okla- homa Press, Norman.

Barchuck, A. H., A. Valiente-Banuet, and M. P. Diaz. (2005). Effect of shrubs and seasonal variabil- ity of rainfall on the establishment of Aspidosperma quebrachoblanco in two edaphically contrast- ing environments. Austral Ecology 30: 695–705.

Belyea, L. R., and J. Lancaster. (1999). Assembly rules within a contingent ecology. Oikos 86: 402–416.

Bowles, M. L., M. D. Jones, and J. L. McBride. (2003). Twenty-year changes in burned and unburned sand prairie remnants in northwestern Illinois and implications for management. The American Midland Naturalist 149: 35–45.

Boyle, B., et al. 2013. The Taxonomic Name Resolution Service: An Online Tool for Automated Standardization of Plant Names. BMC Bioinformatics 14:16.

Brooker, R. W., F. T. Maestre, R. M. Callaway, C. L. Lortie, L. A. Cavieres, G. Kunstler, P. Liancourt, et al. (2007). Facilitation in plant communities: The past, the present, and the future. Journal of Ecology 96: 18–34.

Cairns, J., and J. R. Heckman. (1996). Restoration ecology: The state of an emerging field. Annual Review of Energy and the Environment 21: 167–189.

Page  121 2015 THE MICHIGAN BOTANIST 121

Callaway, R. M. and L. R. Walker. (1997). Competition and facilitation: A synthetic approach to in- teractions in plant communities. Ecology 78: 1958–1965.

Carter, D. L. and J. M. Blair. (2012). High richness and dense seeding enhance grassland restoration establishment but have little effect on drought response. Ecological Applications 22: 1308–1319.

Cole, K. L., and R. S. Taylor. (1995). Past and current trends of change in a dune prairie/oak savanna reconstructed through a multiple-scale history. Journal of Vegetation Science 6: 399–410.

Collins, S. L., S. M. Glenn, and D. J. Gibson. 1995. Experimental analysis of intermediate distur- bance and initial floristic composition: Decoupling cause and effect. Ecology 76: 486–492.

Collins, S. L. 2000. Disturbance frequency and community stability in native tallgrass prairie. The American Naturalist 155: 311–325.

Comer, P. J., D. A. Albert, H. A. Wells, B. L. Hart, J. B. Raab, D. L. Price, D. M. Kashian, R. A. et al. (1995). Michigan’s native landscape as interpreted from the General Land Office surveys 1816–1856. Michigan Natural Features Inventory, Lansing. Report n. 1995–07.

Connell, J. H. (1983). On the prevalence and relative importance of interspecific competition: Evi- dence from field experiments. The American Naturalist 122: 661–696.

Cully, A. C, J. F. Cully, and R. D. Hiebert. (2003). Invasion of exotic plant species in tallgrass prairie fragments. Conservation Biology 17: 990–998.

Diaz, S., A. J. Symstad, F. S. Chapin III, D. A. Wardle, and L. F. Huenneke. (2003). Functional di- versity revealed by removal experiments. Trends in Ecology and Evolution 18: 140–146.

Dukes, J. S. (2001). Biodiversity and invisibility in grassland microcosms. Oecologia 126: 563–568.

Emery, S. M., P. J. Doran, J. T. Legge, M. Kleitch, and S. Howard. (2013). Aboveground and below- ground impacts following removal of the invasive species Baby’s Breath (Gypsophila paniculata) on Lake Michigan Sand Dunes. Restoration Ecology 21: 506–514.

Fargione, J., C. S. Brown, and D. Tilman. (2003). Community assembly and invasion: An experi- mental test of neutral versus niche processes. Proceedings of the National Academy of Sciences

100: 8916–8920. Fargione, J., and D. Tilman. (2006). Plant species traits and capacity for resource reduction predict yield and abundance under competition in nitrogen-limited grassland. Functional Ecology 20: 533–540.

Foster, B. L., C. Murphy, K. Keller, T. Aschenbach, E. Questad, and K. Kindscher. (2007). Restora- tion of prairie community structure and ecosystem function in an abandoned hayfield: A sowing experiment. Restoration Ecology 15: 652–661.

Foster, B. L. and D. Tilman. (2003). Seed limitation and the regulation of community structure in oak savanna grassland. Journal of Ecology 91: 999–1007.

Gleason, H. A. (1910). The vegetation of the inland sand deposits of Illinois. Bulletin of the Illinois State Laboratory of Natural History 9: 23–174.

Goldberg, D. E. and A. M. Barton. (1992). Patterns and consequences of interspecific competition in natural communities: A review of field experiments with plants. The American Naturalist 139: 771–801.

Gough, L., J. B. Grace, and K. L. Taylor. (1994). The relationship between species richness and com- munity biomass: The importance of environmental variables. Oikos 70: 271–279.

Grace, J. B. (1999). The factors controlling species density in herbaceous plant communities: An as- sessment. Perspectives in Plant Ecology, Evolution, and Systematics 2: 1–28.

Greenlee, J. T. and R. M. Callaway. (1996). Abiotic stress and the relative importance of interference and facilitation in montane bunchgrass communities in western Montana. The American Natural- ist 148: 386–396.

Grime, J. P. (1974). Control of species density in herbaceous vegetation. Journal of Environmental Management 1: 151–167.

Hauser, R. S. (1953). An ecological analysis of the isolated prairies of Newaygo County, Michigan. Ph.D. Dissertation. Michigan State College.

Hector, A., B. Schmid, C. Beierkuhnlein, M. C. Caldeira, M. Diemer, and P. G. Dimitrakopoulos. (1999). Plant diversity and productivity experiments in European grasslands. Science 286: 1123–1127.

Henderson, C. B., K. E. Petersen, and R. A. Redak. (1988). Spatial and temporal patterns in the seed bank and vegetation of a desert grassland community. Journal of Ecology 76: 717–728. Hermann, K. D., L. A. Masters, M. R. Penskar, A. A. Reznicek, G. S. Wilhelm, and W. W. Brodow- icz. (1996). Floristic quality assessment with wetland categories and computer application pro-

Page  122 122 THE MICHIGAN BOTANIST Vol. 54

grams for the state of Michigan. Michigan Department of Natural Resources, Wildlife Division, Natural Heritage Program. Lansing.

Holmgren, M., M. Scheffer, and M. A. Huston. (1997). The interplay of facilitation and competition in plant communities. Ecology 78: 1966–1975.

Houseman, G. R., and R. C. Anderson. (2002). Effects of jack pine plantation management on bar- rens flora and potential Kirtland’s warbler nest habitat. Restoration Ecology 10: 27–36.

Howe, H. F. (1994). Managing species diversity in tallgrass prairie: Assumptions and implications. Conservation Biology 8: 691–704.

Howe, H. F. (1995). Success and fire season in experimental prairie plantings. Ecology 76: 1917–1925.

Howe, H. F., J. S. Brown, and B. Zorn-Arnold. (2002). A rodent plague on prairie diversity. Ecology Letters 5: 30–36.

Ives, A. R., J. L. Klug, and K. Gross. (2000). Stability and species richness in complex communities. Ecology Letters 3: 399–411.

Jordan, N. R., D. L. Larson, and S. C. Huerd. (2008). Soil modification by invasive plants: Effects on native and invasive species of mixed-grass prairies. Biological Invasions 10: 177–190.

Kennedy, T. A., S. Naeem, K. M. Howe, J. M. Knops, D. Tilman, and P. Reich. (2002). Biodiversity as a barrier to ecological invasion. Nature 417: 636–638.

Kost, M. A. (2004). Natural community abstract for dry sand prairie. Michigan Natural Features In- ventory, Lansing.

Kost, M. A., D. A. Albert, J. G. Cohen, B. S. Slaughter, R. K. Schillo, C. R. Weber, and K. A. Chap- man. (2007). Natural communities of Michigan: Classification and description. Michigan Natural Features Inventory, Report Number 2007–21, Lansing.

Lockwood, J. L. (1997). An alternative to succession: Assembly rules offer guide to restoration ef- forts. Restoration and Management Notes 15: 45–50.

Maestre, F. T., S. Bautista, and J. Cortina. (2001). Potential for using facilitation by grasses to estab- lish shrubs on a semiarid degraded steppe. Ecological Applications 11: 1641–1655.

Martin, L. M., K. A. Moloney, and B. J. Wilsey. (2005). An assessment of grassland restoration suc- cess using species diversity components. Journal of Applied Ecology 42: 327–336.

McCann, M. T. (1991). Land, climate, and vegetation of Michigan. Pp. 15–31 in The atlas of breed- ing birds of Michigan. R. Brewer, G. A. McPeek, and R. J. Adams Jr., editors. Michigan State Uni- versity Press, East Lansing.

Middleton, E. L., J. D. Bever, and P. A. Schultz. (2009). The effect of restoration methods on the qual- ity of the restoration and resistance to invasion by exotics. Restoration Ecology 18: 181–187.

Naeem, S. and J. P. Wright. (2003). Disentangling biodiversity effects on ecosystem functioning: De- riving solutions to a seemingly insurmountable problem. Ecology Letters 6: 567–579.

Nemec, K. T., C. R. Allen, C. J. Helzer, and D. A. Wedin. (2013). Influence of richness and seeding density on invasion resistance in experimental tallgrass prairie restorations. Ecological Restoration

31: 168–185. Noss, R. F., E. T. LaRoe III, and J. M. Scott. (1995). Endangered ecosystems of the United States: A preliminary assessment of loss and degradation. Biological Report 28. United States Department of Interior, National Biological Service, Washington, D.C.

Ojima, D. S., D. S. Chimel, W. J. Parton, and C. E. Owensby. (1994). Long-and short-term effects of fire on nitrogen cycling in tallgrass prairie. Biogeochemistry 24: 67–84.

Peltzer A., and M. Kochy. (2001). Effects of grasses and woody plants in mixed-grass prairie. Ecol- ogy 89: 519–527.

Plumb-Mentjes, M. L., and G. E. Center. (1990). The dynamics of a sand prairie plant community. Pp 123–127 in Proceedings of the Seventh North American Prairie Conference. C. L. Kucera, ed- itor. Southwest Missouri State University, Springfield.

Pokorny, M. L., R. L. Sheley, C. A. Zabinsky, R. E. Engel, T. J. Svejcar, and J. J. Borkowski. (2005). Plant functional group diversity as a mechanism for invasion resistance. Restoration Ecology 13: 448–459.

Polley, H. W., J. D. Derner, and B. J. Wilsey. 2005. Patterns of plant species diversity in remnant and restored tallgrass prairies. Restoration Ecology 13: 480–487. Raison, R. J. (1979). Modification of the soil environment by vegetation fire, with particular refer- ence to nitrogen transformations: A review. Plant and Soil 51: 73–108.

Page  123 2015 THE MICHIGAN BOTANIST 123

Risser, P. G. (1988). Abiotic controls on primary productivity and nutrient cycles in North American grasslands. Ecological Studies: Analysis and Synthesis 67: 115–130.

Roos, R. C. (2014). The initial effects of community variables on sand prairie restoration: species es- tablishment and community responses. Masters Thesis, Grand Valley State University, Allendale, Michigan. Available at http://scholarworks.gvsu.edu/cgi/viewcontent.cgi?article=1735&context= theses.

Samson, F., and F. Knopf. (1994). Prairie conservation in North America. BioScience 4: 418–421.

Schaetzl, R. J, and S. Anderson. (2005). Soils: Genesis and geomorphology. Cambridge University Press, New York, N.Y.

Schramm, P. (1990). Prairie restoration: A twenty-five year perspective on establishment and man- agement. Pp. 169–178 in Proceedings of the Twelfth North American Prairie Conference. D. D. Smith and C. A. Jacobs, editors. University of Northern Iowa, Cedar Falls.

Seabloom, E. W., W. S. Harpole, O. J. Reichman, and D. Tilman. (2003). Invasion, competitive dom- inance, and resource use by exotic and native California grassland species. Proceedings of the Na- tional Academy of Sciences 100: 13384–13389.

Seastedt, T. R. and A. K. Knapp. (1993). Consequences of nonequilibrium resource availability across multiple time scales: The transient maxima hypothesis. American Naturalist 141: 621–633.

Shannon, C. E. (1948). A mathematical theory of communication. Bell System Technical Journal 27: 379–423.

Sluis, W. J. (2002). Patterns of species richness and composition in re-created grassland. Restoration Ecology 10: 677–684.

Smith, M. D., D. C. Hartnett, and G. W. T. Wilson. (1999). Interacting influence of mycorrhizal sym- biosis and competition on plant diversity in tallgrass prairie. Oecologia 121: 574–582.

Smith, T. and M. Huston. (1990). A theory of spatial and temporal dynamics of plant communities. Vegetatio 83: 49–69.

SPSS Inc. 2011. PASW Statistics 18.0. SPSS Inc., Chicago, Illinois.

Suding, K. N. 1999. Processes responsible for changes in plant species abundance following distur- bance. Ph.D. Dissertation, University of Michigan, Ann Arbor.

Swink, F., and G. Wilhelm. (1994). Plants of the Chicago region. Indiana Academy of Science, Indi- anapolis.

Tabachnick, G. G., and L. S. Fidell. (2007). Experimental designs using ANOVA. Duxbury Press, Belmont, C alifornia.

The Taxonomic Name Resolution Service. (2014). The Taxonomic Name Resolution Service. iPlant Collaborative. Version 3.2. Available at http://tnrs.iplantcollaborative.org. (Accessed June 27, 2014).

Tilman, D. (1988). Plant strategies and the dynamics and structure of plant communities. Princeton University Press, Princeton, New Jersey.

Tilman, D., and J. A. Downing. (1994). Biodiversity and stability in grasslands. Nature 367: 363–365.

Tilman, D., D. Wedin, and J. Knops. (1996). Productivity and sustainability influenced by biodiver- sity in grassland ecosystems. Nature 379: 718–720.

Tilman, D. (1997). Community invasibility, recruitment limitation, and grassland biodiversity. Ecol- ogy 78: 81–92.

Tilman, D. (1999). The ecological consequences of changes in biodiversity: A search for general principles. Ecology 80: 1455–1474.

Transeau, E. N. (1935). The prairie peninsula. Ecology 16: 423–437.

United States Department of Agriculture. (2011). Web Soil Survey, Natural Resources Conservation Survey. Available at http://www.websoilsurvey.nrcs.usda.gov. (Accessed March 11, 2011).

Wedin, D. A., and D. Tilman. (1990). Species effects on nitrogen cycling: A test with perennial grasses. Oecologia 84: 433–441.

Yachi, S., and M. Loreau. (1999). Biodiversity and ecosystem productivity in a fluctuating environ- ment: The insurance hypothesis. Proceedings of the National Academy of Sciences 96: 1463–1468.

Young, T. P., J. M. Chase, and R. T. Huddleston. (2001). Community assembly and succession: Com- paring, contrasting, and combining paradigms in the context of ecological restoration. Ecological Restoration 19: 5–18.

Zimmerman, U. D., and C. L. Kucera. (1977). Effects of composition changes on productivity and biomass relationships in tallgrass prairie. American Midland Naturalist 97: 465–469.