IMPACTS OF ASIATIC SAND SEDGE ON NATIVE PLANTS AND ARBUSCULAR MYCORRHIZAL FUNGI IN A BARRIER DUNE

The recent expansion of the nonnative invasive Asiatic sand sedge (Carex kobomugi Ohwi) at East Beach State Park, Rhode Island, is reducing populations of the most important native, dune-building species and their associated arbuscular mycorrhizal fungi (AMF). In contrast to the native American beachgrass (Ammophila breviligulata Fern.) that is dependent upon AMF to thrive in nutrientpoor sand dunes, C. kobomugi does not form beneficial associations with the fungi. Furthermore, assessments suggest that the sedge is competitively superior in obtaining the essential nutrient phosphorous without AMF-facilitation. Analysis of data from transects of the dune system revealed significant negative correlations between distributions of C. kobomugi and A. breviligulata that are being extirpated. Percent cover of A. breviligulata was significantly reduced in areas of C. kobomugi. Other native plant species were not significantly reduced as a result of C. kobomugi expansion. Spore populations of AMF showed significant positive correlations with percent cover of A. breviligulata and significant negative correlations with percent cover of C. kobomugi. Mean spore abundance of AMF in areas of C. kobomugi was less than in areas dominated by A. breviligulata. The number of species of AMF was not significantly reduced as a result of C. kobomugi likely because of highly aggregated and infrequent distribution of some species‟ spores. Assessment of mycorrhizal inoculum potential (MIP) of soils taken from the field mirrored the spore-population data: mean root colonization of plants grown in field soil of C. kobomugi (12%-24%) was between three and five times lower than that of plants grown in field soil of A. breviligulata (55%-72%). This study was unique in quantifying the effect of an invasive species on populations of mycorrhizal fungi in a dune habitat. It was novel in assessing the reduction of native plant and fungi species by C. kobomugi in Rhode Island. The replacement of AMF-forming species on dunes by a species that does not form AMF (and support spore production by these obligately biotrophic fungi) will have serious consequences when attempts are made to re-establish native species in the sites that are eventually cleared of C. kobomugi.

Running head: Impacts of an invasive sedge in RI dunes 1  Fern.) that is dependent upon arbuscular mycorrhizal fungi (AMF) to thrive in nutrient-poor sand dunes, C. kobomugi does not form beneficial associations with the fungi. Furthermore, assessments suggest that the invasive is competitively superior in obtaining the essential nutrient phosphorous without facilitation by mycorrhizae.
Analysis of data from transects of the dune system revealed significant negative correlations between distributions of C. kobomugi and A. breviligulata that are being extirpated. Percent cover of A. breviligulata was significantly reduced in areas of C.
kobomugi. Spore populations of AMF showed significant positive correlations with percent cover of A. breviligulata and significant negative correlations with percent cover of C. kobomugi. Assessment of mycorrhizal inoculum potential (MIP) of soils taken from the field mirrored the spore-population data. This study was unique in quantifying the effect of an invasive species on populations of mycorrhizal fungi in a dune habitat. The replacement of AMF-forming species on dunes by a species that does not form AMF will have serious consequences when attempts are made to reestablish native species.
They form mutualistic relationships with the roots of a majority of plant species by the formation of specialized cells that enable the fungi to acquire carbohydrates from the host plants" roots while absorbing inorganic nutrients from the soil (Smith and Read, 1997). The increased exploitation of soil volume made possible by AMF is critical for the uptake of less mobile nutrients like phosphorous, zinc, and copper (Sorensen, Larsen, and Jakobsen, 2005). Studies by McGonigle and Fitter (1988) and Tinker, Jones, and Durall (1992) found that the uptake rate of phosphorous per unit root length in mycorrhizal plants can be 2-3 times higher than in non-mycorrhizal plants. AMF have been shown to increase drought tolerance and improve the overall structure and stability of soil (Auge, 2001;Koske, 1975). They enhance the uptake of inorganic nitrogen (Govindarajulu et al., 2005) and provide protection against harmful plant-pathogens (Cooper and Grandison, 1986;Newsham, Fitter, and Watkinson, 1994;Pozo et al., 2002). AMF have the ability to physically reduce the amount of root exudates responsible for nematode attraction, and they may possibly reduce nematode development in substrate and roots altogether (Maun, 2009). It has been demonstrated that AMF competitively exclude harmful plant-parasitic nematodes from regions of roots of A. breviligulata specifically (Hussey and Roncadori, 1982).
Although major dune-building species are well adapted to a variety of abiotic stresses including frequent variations in wind, salt spray, temperature, moisture, sand movement, and organic matter deposition, their association with AMF to acquire critical nutrients is thought to be highly essential for survival (Maun, 2009). Sand dunes typically have relatively low nutrient concentrations that can vary with daily and seasonal wind speed and direction, as well as rainfall amount (Gemma and Koske, 1997;Maun and Baye, 1989;Mosse, 1973). A major source of nutrients in dune soil comes from the deposition of algae that are distributed randomly along the dune (Maun and Baye, 1989). Nutrients deposited as a byproduct of salt spray are quickly leached from the upper soil when it rains and, as a result, even in light of continuous deposition of nutrients from these sources, average levels of soil cations are low (van der Valk, 1974). Foredune areas especially are regions of extremely low nutrient concentrations as topographic rise restricts the deposition of detritus in general (Maun and Baye, 1989). Phosphorous generally occurs at deficient or neardeficient levels in sand dune substrates yet is one of the most critical nutrients for plant species (Atkinson, 1973;Gemma, Koske and Habte, 2002;Halvorson and Koske, 1988;Koske, unpublished observations). Available plant phosphorous and potassium levels in dune soils may be as low as 0.011% and 0.008%, respectively (Maun and Baye, 1989).
Physical aspects of the dune such as the sand mobility, the rate of nutrient deposition, fresh water availability, and distance from the shoreline, contribute to overall functionality as habitat for any array of species and as a barrier to protect inland areas against storms (Maun, 2009;Maun and Baye, 1989). AMF-facilitated enhancement of dune structure further reduces the potential for degradation by constant wind erosion. Fungal hyphae of AMF enhance the geophysical structure of sand dunes by binding sand grains into larger aggregates thereby trapping organic matter and promoting microbial activity (Forster and Nicolson, 1981;Jehne and Thompson, 1981;Koske, Sutton and Sheppard, 1975;Sutton and Sheppard, 1976).
Glomalin-related soil protein (GRSP), a component believed to be primarily synthesized by AMF, likely serves a larger role in binding sand grains. The formation of GRSP by AMF also retains organic matter and moisture that can be utilized by dune plant species (Rillig, 2004;Rillig and Steinberg, 2002).

Ammophila breviligulata and Carex kobomugi
Ammophila breviligulata is a cool-season deciduous perennial that propagates primarily via extensive rhizomes (Maun and Baye, 1989). It is a dominant pioneer species in dunes, with a habitat range in eastern coastal North America, from 35° N to 53° N, as well as in the Great Lakes region (Halvorson and Koske, 1988;Koske and Halvorson, 1981;Maun and Baye, 1989;Watkinson, 1988).
A functioning population of AMF is essential for this species to survive in sand dune environments (Francis and Read, 1994;Gange, Brown and Sinclair, 1993;Gemma and Koske, 1997;Little and Maun, 1996;Maun, 2009;Maun and Baye, 1989;Maun and Lapierre, 1984;Miller, 1987;. In evaluating individual plants" roots throughout the species" distribution on the Atlantic coast, it has been reported that as much as 80% of roots of individual A. breviligulata can be colonized by AMF hyphae (100% being total root coverage) (Koske, 1987).
Rhizomes of A. breviligulata grow horizontally seaward as much as 2 cm per day, and extend some overall distance of 1 to 2 m from the foredune and into the berm of the beach (Brodhead and Godfrey, 1977;Koske, personal communication). Koske and Halvorson (1981) showed that other dominant dune species, including Solidago sempervirens L., Lathyrus japonicus Willd., and Myrica pensylvanica Mirbel, were also colonized by AMF. Sylvia (1989) reported that transplants of sea oat (Uniola paniculata L.), another predominant sand dune and mycorrhizae-forming species in the warmer dunes of the U.S. Atlantic coast and Gulf of Mexico, grew 219% larger shoots and 53% more tillers when grown in soil containing AMF.
As a pioneer species, A. breviligulata plays a substantial role in the development and maintenance of sand dunes. The leaves slow on-shore winds, allowing sand to accrete, contributing to increased dune height. A. breviligulata is especially important because it frequently inhabits the foredune, an area critical in the early formation of dunes and highly susceptible to erosion and disturbance events (Maun, 2009;Maun and Baye, 1989;Maun and Lapierre, 1984). Additionally, A.
breviligulata responds positively to increased deposition of sand, making it further well-suited for dune initiation, growth and stabilization of dunes over time (Maun and Baye, 1989;Olson, 1958). Without such consistent sand accretion, A.
breviligulata suffers from decreases in shoot weight, height and density. Though past studies have suggested that improved plant vigor due to sand burial was associated with a release from plant-pathogenic nematodes, Little and Maun (1996) found that plant vigor and the reduced impact from plant-parasitic biota in the substrate was more a result of enhanced tolerance and resistance provided by AMF that become more available with increased sand deposition.
Carex kobomugi is a robust primary colonizer of sand dunes that spreads rapidly by numerous rhizomes (Miyata and Haramoto, 1986;Miyata and Haramoto, 1987;Ishikawa et al., 1993 Park, New Jersey (Small, 1954). Though the introduction pathway to New Jersey is uncertain, it was suggested that the nonnative invasive was introduced via packing materials (Halsey, 2002;Small, 1954). However, due to the relative scarcity of the plant in its native range, using C. kobomugi in packing material in the past appears unlikely. Rather, seeds or rhizomes were likely transported in dry ballast (Wootton, 2007).  (Standley et al., 1983;Belcher et al., 1984;United States Department of Agriculture, 1983, 1984, thus expanding its range (Wootton et al., 2005). In the 1930s, C. kobomugi was deliberately planted in areas of southeastern Virginia for dune stabilization purposes, later escaping into non-target dune communities (Virginia Department of Conservation and Recreation, 2011). In the early 1990s, increased awareness of the proliferation of introduced species in the United States essentially halted the planting of C. kobomugi as a dune stabilizer, though fugitive populations continued to expand (Wootton, 2002).
Disturbance events that negatively affect native plants, such as naturallyoccurring dune erosion or increased anthropogenic activity, likely further contributed to the increased spread of C. kobomugi. In areas in New Jersey, Wootton et al. (2005) reported exponential increases in overall population as high as 780% over the last 20 years. The largest and oldest stand of the invasive in North America at Island Beach State Park, New Jersey increased from an area of 2,000 m 2 to 90,032 m 2 between 1939 and 2005 (Belcher et al., 1984;Wootton et al., 2005). Its current range along the North American east coast extends from Massachusetts to North Carolina.
At present the species is listed as invasive in Rhode Island, Connecticut, Maryland, New Jersey, and Virginia (Enser, 2005;Enser, 2006;MacLachlan, personal communication;Shisler, Wargo and Jordan, 1987 The populations of C. kobomugi first discovered at East Beach in 1981 (Champlin, 1994;Enser, 2006), have expanded significantly in the last 30 years. In 1983, C. kobomugi was reported as a single main population covering approximately 170 m 2 (Standley, 1983). After 2005, six distinct patches of C. kobomugi were documented in foredune areas, encompassing a total area estimated as 8,000 and 12,000 m 2 (Enser, 2005;Johnson, personal observation;MacLachlan, personal communication). In addition to dense foredune populations, C. kobomugi occurs in sparser populations beyond the crest of the dune, near backdune roads and clearings surrounding Ninigret Pond. In these areas C. kobomugi is more interspersed with other naturally occurring backdune species (Johnson, personal observation).
C. kobomugi typically propagates through vegetative means, as sexual reproduction yields seeds with low germination rates and high seedling mortality (Nobuhara and Miyazaki, 1974;Sasaki, 1987;Yamamoto, 1964). Plants are able to regrow fully from the vegetative remains or rhizomes left after manual removal (Lea and McLaughlin, 2002). Rhizomes can extend horizontally outward 0.5 m to 1.2 m from an individual, depending on degree of dune maturation (Ishikawa and Kachi, 1998). A survey in New Jersey by Small (1954) indicated that rhizomes of C.
kobomugi are deeper, have shorter internodes, and root more profusely than A.
breviligulata. Roots of C. kobomugi capable of producing new shoots occur at a depth of up to 60 cm (Ishikawa and Kachi, 1998;Nobuhara, 1967;Park, 1982;Wootton et al., 2003), as compared to roots of A. breviligulata that can extend vertically for 150 cm (Maun and Baye, 1989). In New Jersey, rhizomes of C. kobomugi have been documented at lower depths than those of A. breviligulata, though viable shoots typically do not emerge from these rhizomes below a depth of 60 cm (Ishikawa and Kachi, 1998;Park, 1982;Small, 1954;). Rhizomes of C. kobomugi extend laterally from 50 to 250 cm depending on dune maturation (i.e., sand accumulation) and available resources (Miyata and Haramoto, 1987;Nobuhara, 1967). Although individual plants of C. kobomugi are short (10-30 cm) in comparison to those of A.
Staining and analysis of root samples from 6 individuals of C. kobomugi collected from multiple locations at East Beach indicate that it is non-mycorrhizal (Johnson, personal observation). A number of publications have indicated that other species in the Cyperaceae are facultative or non-mycorrhizal as well (Brundrett, 1991;Gerdemann, 1968;Miller, 2005;Newman and Reddell, 1987;Tester, Smith, and Smith, 1987). The long and fine root systems characteristic of sedges is thought to be an alternative to a dependence on AMF for nutrient acquisition and uptake (Brundrett and Kendrick, 1988;Miller et al. 1999). In general, the mycorrhizal status of species depends on genetic characteristics and on a series of environmental factors, such as the presence of AMF inoculum or soil moisture. In the assessment of mycorrhiza-forming ability of plants, careful measures must be taken to avoid misidentification of structures in roots formed by a variety of non-mycorrhizal soilborne fungi whose appearance resembles that of the hyphae, arbuscules, and vesicles of AMF (Johnson, personal observation; Koske, personal communication).
Although a large number of studies on the interactions between invasive and native plant species have been published, evaluating differences in phenology and attributes, (Baker, 1974(Baker, , 1986Rejmanek and Richardson, 1996), very few assess the effect of plant invasions on AMF (Vogelsang and Bever, 2009). As AMF are especially important to dependent host-plant species in obtaining nutrients and tolerating a variety of physical stresses (e.g., drought, salinity, excessive sunlight) in dune systems (Koske et al., 2004;Maun, 2009;Smith and Read, 1997), a reduction in the population of AMF in soil by an invasive species could have a significant ecological impact, and may be linked to the overall success of a non-mycorrhizal, nonnative invasive such as C. kobomugi.
The degraded mutualist hypothesis, as proposed by Vogelsang and Bever, (2009), describes a situation in which natives have been completely replaced by a competitively-dominant invasive, and are unable to re-colonize due to an absence of associated AMF. As non-mycorrhizal invasive plant species are less dependent on interactions with AMF to establish in novel areas (Allen and Allen, 1980;Pendleton and Smith, 1983;Reeves et al., 1979;Vogelsang and Bever, 2009), ecological dominance could likely shift from a mycorrhizal-dependent native to an invasive in a relatively short period of time. Furthermore, the longevity of the mycorrhizal communities themselves is put in jeopardy, as exotic species, even if facultatively mycorrhizal, are generally poorer hosts for the AMF mutualism than native, mycorrhizal-dependent species (Vogelsand and Bever, 2009).
Studies have shown that a change in the composition of AMF belowground alters plant carbon exudation and nutrient uptake, suggesting that the alteration of AMF has the capacity to dramatically influence the composition of aboveground vegetation (Cavagnaro et al., 2005;Schwab, Leonard, and Menge, 1984). In California, Hawkes et al. (2006) found that exotic grasses (Avena barbata Link and Bromus hordeaceus L.) reduce species richness of mycorrhizal fungi, causing dramatic shifts in a mycorrhizal community of native grasses and lupines (Nasella pulchra (Hitchc.) Barkworth and Lupinus bicolor Lind). Vogelsang and Bever (2009) demonstrated that soils associated with a nonnative invasive thistle (Carduus pycnocephalus L.) did not promote growth of AMF as compared to soils conditioned with native herb species (Gnaphalium californicum (D.C.) Anderb.).
In addition to altering composition of host-plants, the direct, chemical disruption of AMF has also recently been identified as a means by which an invasive plant species achieves a competitive advantage in an established ecosystem. The best known example of this is a study by Stinson et al. (2006) demonstrating that a nonnative invasive species" active suppression of AMF fungi was connected to its ability to invade and successfully supplant native species in Northeastern U.S.
forests. Garlic mustard (Alliaria petiolata (M. Bieb.) Cavara and Grande) significantly reduced growth and vigor of tree seedlings by killing mycorrhizal fungi in the soil (Stinson et al., 2006). Chemical compounds isolated from the invasive"s root tissues, including allyl isothiocyanate, benzyl isothiocyanate, and glucotropaeolin, had allelopathic effects on native plants in the absence of AMF.
Exposing AMF spores to these extracts severely reduced germination rates (Stinson et al., 2006).
Like the non-mycorrhizal garlic mustard it is possible that C. kobomugi produces allelopathic compounds to suppress AMF and established native species. Though Li, Henry, and Seeram (2009)

reported that several members of the genus
Carex produce stilbenes and other bioactive polyphenols that are capable of such activity, it is unclear whether the C. kobomugi specifically uses exudates competitively. In stem-density experiments, Burkitt and Wootton (2010) found that native plants of diverse functional type (i.e., annuals, perennials, dicots, or monocots) were all equally negatively affected when interacting with C. kobomugi.
They suggested that it may actively replace established, vigorously growing species, rather than solely colonizing recently disturbed areas. These authors further hypothesized that allelopathic chemicals could be the primary means of the invasive to outcompete natives, especially in consideration of its relatively uniform effect for such a wide functional range of species. Preliminary research found that germination of spores of Gigaspora gigantea treated with root-exudates from C. kobomugi was not significantly reduced (Johnson, personal observation; see Appendix).
Even without the use of direct allelopathic suppression, the replacement of breviligulata, a mycorrhizae-dependent species and primary dune-building native plant, the population of associated AMF also will decline, thus affecting the ability of future host-plants to acquire nutrients, and successfully colonize the area (Koske and Gemma, 1997;Miller, 1979;Reeves et al., 1979). Leaf tissues from C. kobomugi and native mycorrhiza-forming plant species, and available phosphorous in soils were analyzed to examine differences in phosphorous acquisition ability and further examine competitive interactions between the target plant species. These field phosphate measurements supplemented the main objective of investigating plant competition and the impact on mycorrhizal fungi.
The primary objective of this study was to address whether the rapidly expanding C. kobomugi reduced the dominant sand-dune building plant, A.
breviligulata, as well as critical mycorrhizal fungi. This was tested by surveying percent cover of plant species and by sampling populations of AMF in soil along transects at foredune sites of C. kobomugi. Significant negative interactions between A. breviligulata, spore populations of AMF, and C. kobomugi supported the hypothesis that C. kobomugi effectively replaces and reduces native plant and fungi species. A secondary hypothesis was that C. kobomugi reduced future mycorrhizal inoculum potential of areas after it had invaded. This was investigated by collecting field soils from areas of C. kobomugi and A. breviligulata, and conducting growth trials using corn as an indicator species. A significant decrease in root colonization of reduces AMF populations in soils. The objective of the field plant leaf tissue phosphate assessment was to examine the ability of the invasive to acquire phosphorous, specifically without the AMF interaction that is essential for many native plants. Soil phosphorous was assessed to document baseline available phosphorous in dunes for comparison with leaf tissue phosphate of target plant species.

Study Site
East Beach (approx. 41º N, 71º W) borders the Ninigret Conservation Area in The dune extends for the length of the beach, and at its crest, the maximum overwash and beach sand deposits are approximately 3 m above mean sea level (Urish, 1982). Vegetation on the dunes at East Beach is typical of Mid-and North Atlantic coastlines of the U.S. (Godfrey, 1977;Koske and Halvorson, 1981 (Stuckey, 1976;Urish, 1982). Populations of Pinus thunbergii Parl. occur commonly in both the foredune and backdune areas of the beach. C. kobomugi occurs at five distinct locations along the foredune ( Figure 2) and in a number of areas in the backdune (Enser, 2005; Johnson, personal observation). Each patch-site of C. kobomugi generally consists of a dense stand of the invasive interspersed with other plant species in low densities, and surrounded by common native species assemblages. Sites were designated E#1 through E#5, beginning with the southwestern-most patch. Study sites out of the five total foredune patch-sites of C. kobomugi were selected for sampling based on the quality of transition zones between different vegetation types (i.e., distinct areas of dense C.
kobomugi, areas of a near-even ratio of cover between both C. kobomugi and A.
breviligulata, and areas of dense A. breviligulata). In 2009 the study surveyed three distinct foredune patch-populations (E#2, E#3 and E#5) for spore abundance and richness of AMF, using one of the three patches (E#3) for percent cover assessment.
Based on findings from the 2009 assessments, in 2010 the study was modified to intensely sample one patch-site (E#3) for spore abundance and richness of AMF as well as percent cover of plant species.
Permission to access and conduct surveys on the dunes for this study was obtained from the Rhode Island Department of Environmental Management Division of Parks and Recreation.

Vegetation Percent Cover Analysis
The vegetation percent cover method was used to document populations of C.
kobomugi and A. breviligulata, and those areas of transition between the species.
Three transects were made on the foredune, parallel to the shore length. These transects represented three elevation zones, denoted low, nearest the front of the dune, mid, and high, nearest the crest of the dune. Generally they were spaced 6-7 m apart.
The purpose of establishing elevation transects was to investigate whether elevation on the dune had any relevant effect on vegetative cover or spore abundance. Overall patch site measurements were a combination of assessments made by Enser (2005) and personal measurements using tape measures and marker objects. In 2009, transects were established from areas of dense C. kobomugi to areas of dense A. breviligulata.
In 2010, transects were made from areas of dense A. breviligulata to areas of C.
kobomugi and back into areas dominated by A. breviligulata ( Figure 3). kobomugi produces leaves that are typically curled and yellow-green compared to leaves produced by A. breviligulata that are straighter and darker-green in color (Ishikawa and Kachi, 1998; Johnson, personal observation; Maun and Baye, 1989).

Spore Abundance and Species Richness of AMF (2009)
Soil samples were collected in December of 2009 in accordance with seasonal peaks of sporulation for anticipated AMF species in the dune (Gemma and Koske, 1988;Gemma, Koske, and Carreiro, 1989;Lee and Koske, 1994). A stratified random sampling design was chosen based on sampling strategies suggested by Gemma, Koske and Carreiro (1989) and Tews and Koske (1986 (Gemma, Koske, and Carreiro, 1989;Ishikawa and Kachi, 1998;Maun and Baye, 1989;Small, 1954). Collection holes were dug using a small shovel and 200 ml of substrate were retrieved with a plastic collection cup scraped along all sides of the hole to ensure a representative sample. The shovel and collection cup were wiped clean between collections to remove macro debris such as sand grains or root material. Based on the spores" size and sampling techniques used in the past this method was considered sufficient to reduce spore contamination between collection samples (Koske, personal communication). Sand was placed into sealable, polypropylene plastic bags (volume approximately 600 ml) and kept in refrigerated storage (5° C) until spore extraction and processing.

Spore Abundance and Species Richness of AMF (2010)
Soil sampling took place in November 2010, in accordance with the highest yearly occurring abundance peaks of sporulation by AMF (Gemma and Koske, 1988;Lee and Koske, 1994 (Jansa, Wiemken, and Frossard, 2006).

Spore Extraction
For efficiency in identifying and counting, spores were extracted from substrate with as minimal detritus and sand as possible. This was accomplished by wet sieving and decanting in combination with a sucrose centrifugation-extraction method (Gerdemann and Nicolson, 1963;Walker, Mize, and McNabb, 1982 to re-suspend the pellet and then centrifuged at near 3000 rpm (50 Hz) for approximately 1 minute. The solution was filtered through a fine-mesh sieve (53 µm) onto a filter paper (5.5 cm diameter, medium/fast qualitative crystalline retention) and suspended in a ceramic filter container connected to a Buchner funnel vacuum filtration apparatus. The contents in the ceramic filter were rinsed with deionized water repeatedly until the filter paper was removed and placed in a plastic dish for spore abundance and richness analysis.

Spore Identification and Analysis
Spore analysis was used to identify and quantify AMF spores retrieved from field sampling. Dead, parasitized, or spores filled with atypical contents were not used in assessments, as they are considered unable to produce typical mycorrhizal structures (INVAM, 2009). Only spores with a healthy, viable appearance as according to descriptions outlined by Lee and Koske (1994) and INVAM (2009) were used in the spore counts. Spores on filter paper were sorted by appearance morphology, collected using a metal inoculating needle with PVLG (8.33 g polyvinyl alcohol, 50 ml lactic acid, 5 ml glycerol, 50 ml water) and placed onto glass microscope slides. Spores were crushed by pressing the inoculating needle firmly on a cover slip. Crushing spores served to facilitate the recognition of differences in sporewall structure, and was necessary for species identification. Characteristics such as color, shape, size, wall composition, and hyphal attachments were also used to further identify species under a compound microscope (40X-1000X magnification). Species identification was supplemented by descriptions by Schenck and Pérez (1990) and Koske (personal communication). A number of typical dune-inhabiting species of AMF were anticipated (Gemma and Koske, 1988;Gemma, Koske, and Carreiro, 1989;Halvorson and Koske, 1988;Koske, personal communication). The number of each species" spores on an individual filter paper was counted and recorded. Spores of individual species were totaled and combined to represent the total spore abundance of AMF per 200 ml of substrate from a particular collection point. The number of species encountered in a sample was considered overall AMF species richness for that sample.

Mycorrhizal Inoculum Potential (MIP)
The objective of growth assays was to determine if areas of the dune with a history of C. kobomugi had reduced mycorrhizal inoculum potential (MIP) as compared to native (A. breviligulata-established) soils. The successful colonization by AMF of indicator plants" roots was used as an indicator of AMF functionality and availability in soil. The final MIP configuration (i.e., growing medium, watering schedule, fertilizer amount, time of harvest, and staining methodology) was determined by conducting preliminary trials and modifying previous MIP growth studies Gemma and Koske, 1988;Giovannetti and Mosse, 1980;Tarbell and Koske, 2007).
Field soils were obtained from areas of highest plant density under the assumption that these areas had a relatively longer and more-concentrated exposure to the desired study-species. Soil was sampled from typical root zones of plant species (20-30 cm) and overlying surface sand was excluded as much as possible. Five ca. 1-L samples were taken at random locations throughout dense vegetation areas. Soils were collected from the field in polypropylene plastic bags and kept in refrigerated storage (5° C) for approximately 5 weeks to break dormancy and to promote germination of spores (Gemma and Koske, 1988;Gemma, Koske and Carreiro, 1989 MIP studies consisted of the following aspects and methods. Corn (Zea mays L., Jubilee hybrid, W. Atlee Burpee and Co., Warminster, PA) was used in growth assays as it is the standard species for measuring the mycorrhiza-forming potential of soils (Gerdemann and Trappe, 1974;Moorman and Reeves, 1979;Reeves et al., 1979).
Corn seeds were planted at a depth of approximately 5 cm in Cone-tainers, using a small wooden stick to prime planting holes. Plants were kept in a temperature- together with enough deionized water to produce 1000 ml) dissolved in 3800 ml of deionized water. The pH of the fertilizer solution was checked before use, as pH levels below 6.0 are generally not sufficient for plants to access critical nutrients (Buckman and Brady, 1960). Fertilizer was adjusted to between pH 6.2 and 6.5 by adding 0.5 ml of 2.5% KOH to each 500 ml of fertilizer solution (40X stock in deionized water).
Fertilizer was kept in refrigerated storage (5° C).
Plants were harvested at 6 weeks (approximately 42 days), a duration considered sufficient to allow development of AMF root colonization Tarbell and Koske, 2007). For assessment of root colonization, root systems for each plant were washed, cleared and stained . In general, sampled roots were approximately 15-20 cm in length. Roots were washed for 3 minutes with a hard stream of water to remove large debris. After washing, the top 5 cm (near base of plant) was discarded, and the bottom 5 cm (near base of Cone-tainer) was discarded. The middle 5 cm was used for assessment, as this is typically where AMF colonize (Koske, personal communication). Root colonization and staining was carried out in accordance with a combination of standard methods (Giovannetti and Mosse, 1980;Koske and Gemma, 1989;Moorman and Reeves, 1979) (see Appendix for detailed procedure). Stained root matter was placed in an 8 cm x 8 cm square Petri dish, with a 1.5 cm x 1.5 cm grid of 25 total intersections. Modifications were made to the intersect method outlined by Giovannetti and Mosse (1980)  was not achieved as this number did not account for the uneven distribution of plants on the dune -some locations across the transect did not have representative plants, as they were not naturally occurring at collection points (i.e., high density areas of C. kobomugi at times lacked other target species).
The most expanded, newly emerged leaf was sampled on each plant, as it was presumed that this would contain the highest tissue phosphate concentration (Gemma, Koske, and Habte, 2002;Koske, personal communication). Leaf tissue phosphorous was determined by adapting the molybdate blue/ascorbic acid procedures from Aziz and Habte (1987), Habte, Huang and Fox (1987) and Habte and Manjunath, (1987).
Leaves were placed in labeled plastic scintillation vials or polypropylene plastic bags for transport back to the lab. For L. japonicus, the youngest, fully expanded pinnule was used. For S. sempervirens, 1.5 cm of the leaf tip was discarded, and the 1.5 cm following was used. A. breviligulata leaves were cut 2 cm from the base and tip, and a middle 3 cm piece was used. The same process was applied to C. kobomugi leaves as deionized water in a 100 ml volumetric flask, resulting in a solution of 2 µg P/ml. One ml of this 2 µg P/ml was then added to 9 ml deionized water and 2.5 ml of Reagent B.
Two test tubes were made as phosphate standards. The absorbance of phosphate standards was roughly 150-200 nm (using a Turner SP-830 spectrophotometer), equivalent to approximately 0.16 µg P/ml. Five milliliters of test tube solutions were pipetted into spectrophotometer cubettes for absorbance assessment. The spectrophotometer was zeroed between each sample using a deionized water blank.
The outside surfaces of cuvettes were wiped clean with Kim Wipes between samples to ensure that no particulate matter interfered with absorbance readings. Cuvettes themselves were thoroughly rinsed with deionized water between samples. If absorbance of a sample was over 600 nm, the sample was diluted and re-measured. A dilution consisted of 1 ml of the sample added to 9 ml of deionized water in a clean test tube. This was then vortexed and re-measured. Calculations for tissue phosphate were corrected for the effect of the dilution.

Available Soil Phosphorous (2009)
Available soil phosphorous was determined based on the molybdate-blue procedure by Fox and Kamprath (1970). Soil from the spore counts in 2009 was used, from one site only (i.e., 3 transects). Approximately 1-2 g of soil per sample was dried in an oven at 70° C for 2 days in small aluminum dishes. After this, soil samples were ashed for 4 hours in a muffle furnace at 500° C and allowed approximately 2 hours to cool. The values 0.020 mg P/L and 0.200 mg P/L were concentrations of interest because these have previously been found to be the critical levels necessary to classify the mycorrhizal dependency of plant species (Habte and Manjunath, 1987). Findings from soil assessments in 2009 were compared with leaf phosphate assessments from 2010. As phosphorous inputs from the atmosphere, leaf litter and weathering of materials are considered very small, coupled with the small amount lost to leaching each year, it was assumed that available soil phosphorous in soils did not differ substantially between years (Maun, 2009).

Statistical Analysis
Statview (Ver. 5.0.1, SAS Institute Inc., Cary, NC) was used to analyze all data in this study. Initially, distributions of vegetation across transects were examined using bivariate line representations to visually interpret trends. Spore population and percent cover data were investigated with linear and curvilinear regression analyses, as it was hypothesized that there were direct causal relationships between them. Where indicated in results, data for spore abundance were base-10 log-transformed (log (x+k) where k=1) to account for non-normal distribution in the field, and data from vegetation percent cover were square-root transformed to account for extreme outliers in data. MIP data were subject to one way analysis of variance (

Vegetation Percent Cover
There were strong inverse associations between aboveground percent cover of C.
kobomugi and of A. breviligulata: high cover values of one species across the transects were associated with low cover values of the other species. In general, the distributions of both species were inversely related to one another (Figures 8-11).
There were no significant differences in percent cover of C. kobomugi and A.
breviligulata between dune elevation transects. In 2009, percent cover of C. kobomugi and A. breviligulata across each transect followed linear trends. C. kobomugi occurred at higher densities towards the beginning of transects (0 m) and lower densities towards the ends (50 m). In 2010, trends in percent cover of both species were curvilinear, as the transect length was increased from a radius of the site (50 m) to cover the entire length of the vegetation gradient (120 m). The highest densities of C.
kobomugi were detected in the middle portion of transects (between 20 m and 100 m), whereas the highest percent cover of A. breviligulata was detected at either end.
Investigation of percent cover data combined from 2009 and 2010 with linear regression analysis found that as C. kobomugi increased in percent cover along the dune, A. breviligulata decreased significantly (P < 0.0001) (Figure 12).
Native plant species outside the scope of this study, such as L. japonicus, R.
rugosa, S. sempervirens, and M. pensylvanica were also present across transects, but had far lower percent cover values, and were not significantly reduced by C. kobomugi (data not shown).

Spore Abundance and Species Richness of AMF
The relationship between percent cover, meters and spore abundance of AMF was investigated using linear and curvilinear regression analysis. Transects made in 2010 for surveying vegetation were standardized at 4 m intervals and did not precisely mirror the intervals selected for spore abundance surveys. But, after reviewing data from 2009 assessments, it was assumed in 2010 that the difference between vegetation cover and AMF distribution across a 4 m distance allowed for the coupling of both standardized and randomized data in this case.
In general, both 2009 and 2010 data suggest that increased percent cover of C.
kobomugi is associated with decreased mean spore abundance of AMF and increased percent cover of A. breviligulata is associated with increased mean spore abundance of AMF. Linear regression analysis found that as percent cover of C. kobomugi increased, mean spore abundance of AMF decreased significantly (P < 0.0001) ( Figure 13). Conversely, as aboveground percent cover of A. breviligulata increased, the mean spore abundance of AMF increased significantly (P < 0.0001) ( Figure 14).
In the 2009 transects, made from areas of dense C. kobomugi to A. breviligulata, mean spore abundance of AMF increased significantly with increasing meters across a transect (Figure 15 and 16). In the 2010 transects, sampling an entire patch site rather than one radius, the relationship between mean spore abundance of AMF and meters was also significant, and followed curvilinear trends (Figure 17 and 18 Sanders. Though each species of AMF was analyzed individually in reference to percent cover of C. kobomugi and A. breviligulata, no one species declined more than another as a result (data not shown). Species richness of AMF data was investigated using linear regression analysis. Richness was defined as the mean number of species encountered from collected samples. Linear regression detected a statisticallysignificant trend between both percent cover of C. kobomugi and percent cover of A. breviligulata and species richness, but this was not biologically significant ( Figure   19).

MIP Growth Experiment
In both MIP trials, corn plants grown in soil from the root zone of A.
breviligulata had significantly greater root colonization than plants grown in soil from the root zone of C. kobomugi (P < 0.0001 for each) (Figure 20 and 21). Plants demonstrated higher root colonization in soil of A. breviligulata (55%-72%) than plants grown in soil of C. kobomugi (12%-24%) ( Table 1). Mean root colonization of plants grown in the 50/50 combination soil of A. breviligulata and C. kobomugi was 48%, significantly less than the two other treatments (P < 0.004) (MIP 3).

Soil Phosphate
Differences in soil solution phosphate (mg P/L) across transects were not significant ( Figure 22). Mean available soil phosphate detected at East Beach (0.029 mg P/L) was similar to values previously reported for unfertilized field soils (Fox and Kamprath, 1970;Habte and Manjunath, 1987;Gemma, Koske and Habte, 2002). Soil solution phosphate is reported as available phosphate for plants as determined by Fox and Kamprath (1970) and Habte and Manjunath (1987).

Field Plant Leaf Tissue Phosphate
Analysis of variance of leaf tissue phosphate taken from field plants in 2010 indicated significant differences among some species. C. kobomugi had significantly

Effects of Invasive Plants on Belowground Biota and Processes
As most plant invasion research focuses on trends in aboveground distribution, a substantial gap in knowledge is left as to the effects on the composition of soil biota (Levine et al., 2003). Few studies examining the effect of invasive plants on belowground soil communities have been conducted and these have been among a very limited number of geographic regions (Wolfe and Klironomos, 2005). Recent research suggests that interactions with soil biota have the capacity to greatly influence community processes, as well as survival of native plants (Bever, 2003;Reinhart and Callaway, 2006;Vogelsang and Bever, 2009). After an invasive plant has established in a novel habitat, it has the ability to completely re-shape a soil community, changing a number of critical soil functions in an ecosystem, including mineralization of nutrients, aeration, and moisture retention (Bever et al., 1996;Wolfe and Klironomos, 2005).

Some invasive plants modify soil composition by releasing organic compounds
and secondary metabolites. Li, Henry, and Seeram (2009) demonstrated that several members of the genus Carex produce stilbenes and other bioactive polyphenols that are potentially fungitoxic or allelopathic. Garlic mustard, a rapidly-expanding invasive in North American forests, kills AMF in soil by releasing root exudates (Stinson et al., 2006). Invasive plants are able to transform the soil community by introducing novel nutrient acquisition abilities to a system. In volcanic sites on Hawaii, the invasive evergreen shrub Morella faya (Ait.) Wilbur and associated nitrogen-fixing bacterialsymbionts alter nitrogen cycling in soils, raising available soil nitrogen levels and potentially allowing other invasives to become established in previously nitrogenlimited areas (Vitousek and Walker, 1984). Invasives also have the capacity to directly change physical properties of the substrate, as seen in the hyperaccumulation of sodium in rangelands of the United States by the noxious weed, saltlover (Halogeton glomeratus (M. Bieb.) C.A. Mey.) (Duda et al., 2003;Wolfe and Klironomos, 2005).
It has been documented that by altering aboveground host-species density, invasive plants can substantially reduce a diverse AMF community (Vogelsang and Bever, 2009). The strong interdependence between AMF and host-plants likely contributes to this effect (Richardson et al., 2000;Hawkes et al., 2006). In plant communities dependent on mycorrhizal interactions, such as sand dunes, the alteration or reduction of AMF could promote the expansion of an invasive species that is nonmycorrhizal. Even in the event that nonnative invasives are at least facultatively mycorrhizal, they still may reduce densities of AMF by contributing fewer roots relative to those from appropriate, native plant-hosts (Vogelsang and Bever, 2009 (Gemma, Koske and Carreiro, 1989) and the variation in spore abundance across a dune varies with species (Gemma and Koske, 1988), vigor of the plant host (Koske and Halvorson, 1981), as well as the maturity of the dune (Puppi and Riess, 1987). In general, the composition and availability of appropriate host plants play a critical role in enabling mycorrhizal interactions -if native species are replaced by a poorer host, or a plant that may suppress the interaction, the community becomes in danger of becoming not only dominated by a non-native, but also a place where natives have difficulty recolonizing. If these mycorrhizal-dependent host species are destroyed, future plant establishment and continued longevity of the area made possible by mycorrhizal inoculum in soil, is effectively put in jeopardy (Gemma and Koske, 1997;Miller, 1979;Reeves et al., 1979).
Spore abundance of AMF is directly related to the overall mycorrhizal inoculum potential of sand dune soils. The colonization of plant roots at the beginning of the growing season is facilitated by a viable and seasonally-sporulating AMF (Lee and Koske, 1994). In sand dune habitats, growth conditions for host-plants include the availability of soil nutrients, the extent of wind and salt spray deposition, seasonal and daily temperature and moisture changes as well as sand movement and deposition. By affecting the vigor of the host plants, these same physical factors influence spore populations of AMF and soil inoculum potential values in the dune (Koske and Halvorson, 1981). In the dune soil AMF are dispersed primarily by the growth of roots and rhizomes and as rhizomes grow, spawning new plants and roots in novel areas, they subsequently establish new AMF populations (Gemma, 1987;. In general, the fluctuation of spores and the emergence of new host plants is an interdependent cycle -increased plant growth in response to the AMF mutualism leads to a greater number of AMF-plant interactions, resulting in greater overall AMF-growth and subsequently a greater number of spores. A higher number of spores is further supplemented by increased hyphal expansion, and an increase in colonized root fragments, resulting in a substantially higher inoculum potential for soils. Hence, whereas a reduction in spore abundance due to an invasive may indicate a more recent, localized effect, a reduction of overall inoculum potential may be interpreted as a long-term suppression of AMF and host-plants.

Reduction of A. breviligulata and AMF by C. kobomugi
A number of studies address the spread and mitigation of C. kobomugi in North America but none evaluate its potential to disrupt AMF (Burkitt and Wootton, 2010;Enser, 2006;Lea and McLaughlin, 2002;Wootton, 2002;Wootton, 2007;Wootton et al., 2003;Wootton et al., 2005;).The present study was the first to document a reduction of AMF populations by C. kobomugi, a quickly-expanding and relatively recent invasive to Rhode Island dune habitats. As AMF are essential to the growth and survival of many native species in sand dunes the great decline in spore populations and soil inoculum potential resulting from the invasion by C. kobomugi has serious implications for the dune, including its vegetation, value for recreation, and ability to protect the coast.
By replacing A. breviligulata, the dominant host for AMF in East Beach dunes, C. kobomugi appears to indirectly reduce spore populations of AMF and soil inoculum potential. The species richness of AMF was not reduced significantly as a result of increasing C. kobomugi, and this was possibly because either species of AMF did not occur regularly enough to be detected in sampling. It is also likely that the typical distribution of AMF in general is non-normal, as spores develop in aggregated patches relative to spreading clusters of host plant rhizomes. Certain species of AMF were so rarely encountered in samples that correlations along transects, much less across Carex species (Li, Henry, and Seeram, 2009). However, preliminary studies adapted from Koske (1981), using sand-plant microcosms and assessing the ability of spores of G. gigantea to germinate in soil collected from the root zones of plants of C.
kobomugi, did not indicate that the invasive had a direct negative effect on germination, number of germ tubes produced per spore, or length of hyphae formed per spore (see Appendix; Johnson, personal observation). Although not investigated in this study, it is possible that a reduction in the ability of AMF to form successful associations with native species was reduced by root exudates that inhibited postgermination stages (e.g., contact of hyphae with roots, synthesis of the mycorrhiza, and growth of new hyphae in the soil origination from the roots).

Findings of the MIP growth experiments demonstrated that plants grown in
soils with a history of C. kobomugi had significantly lower root colonization than those plants grown in soil mixes or field soil collected from areas of A. breviligulata.

Colonization of growth assay plants is determined by the viability of soil inoculum in
terms of the abundance of spores and infective propoagules, such as previouslycolonized root fragments or pieces of hyphae. Successful colonization by AMF is also controlled by abiotic factors, such as light, nutrients, soil pH and moisture (Tarbell and Koske, 2007). Stinson et al. (2006) used MIP assays to evaluate whether an invasive specifically caused decline of AMF in native soils. In that investigation, a significantly lower colonization by AMF in soils conditioned by Alliaria petiolata suggested that not only was the invasive non-mycorrhizal in nature, but also that it reduced native plant performance by interfering with mycorrhizal associations. In the present study, the reduced colonization in MIP studies suggests that C. kobomugi effectively reduces inoculum potential in areas of the foredune, having consequences for the rereestablishment of mycorrhizae-dependent native plants. As C. kobomugi does not form beneficial associations with mycorrhizal fungi, the plant is not growth-limited by populations of AMF, making it potentially far more pliant in its expansion in dune habitats as compared to AMF-dependent natives (Johnson, personal observation).
C. kobomugi was found to have a significantly higher mean leaf tissue phosphate concentration as compared to native plant species (A. breviligulata, L. japonicus and S. sempervirens). For native mycorrhizae-forming plants, phosphorous acquisition serves as a primary indicator of the functionality of AMF in soil (Fox and Kamprath, 1970). As phosphorous is deficient in sand dunes, most plant species that grow there require the AMF association to obtain it in any significant quantity (Gemma, Koske, and Habte, 2002;Habte and Manjunath, 1987;Koske, personal communication) however, the higher leaf phosphate concentrations detected in C.
kobomugi suggests that it is highly capable of obtaining phosphorous from the soil, apparently independent of interactions with AMF. An alternative explanation, one that could be further investigated using seasonal sampling of phosphate, is that phosphorous use and storage among these plants differs greatly. The leaf tissue phosphate study was conducted multiple times over a two year period, the most recent experiment"s results reported in this study. It may be that C. kobomugi sequesters phosphate differently than native plants, especially in a phenologic sense, though this study was not meant to specifically focus on this mechanisminstead, sampling of leaf tissue phosphate was meant to supplement spore counts, and to determine in a broad sense the relative amounts of phosphate among species. These data were particularly useful as the nutrient is highly deficient in sand dunes, and natives are largely dependent on AMF to obtain it in any substantial quantity. Furthermore, available soil phosphate levels did not differ between C. kobomugi-dominated areas and A. breviligulata-dominated areas, suggesting the invasive does not interfere with AMF obtaining the nutrient for mycorrhizal species.
Increased phosphorous absorption could be accomplished by using finer roots, organized in a relatively more exploratory and denser pattern, as compared to other plants. The difference between phalanx-type and guerrilla-type rhizome spread could contribute to both plants" acquisition abilities in general. Guerrilla-type plants, such as C. kobomugi, are opportunistic and pioneering in their root development, growing quickly to search for and acquire nutrients (De Kroon and Knops, 1990). Conversely, phalanx-type plants, such as A. breviligulata, tend to maintain a fixed position and slowly colonize over time (Watkinson, 1988). If C. kobomugi can acquire critical nutrients in greater concentrations than AMF-dependent species, especially in deficient soils, it could likely gain a competitive advantage. As C. kobomugi does not require AMF to acquire phosphorous and does not form a mycorrhizal association with AMF, the AMF populations dies off in C. kobomugi dominated areas. Thus, serious consequences may exist for re-planted native mycorrhiza-dependent species after the invasive has been mitigated and removed.

Effects of Replacement and Reduction of A. breviligulata
In its native range, C. kobomugi is distributed in a more seaward orientation and direction (e.g., areas of the dune associated with relatively harsher conditions) as compared to other native vegetation, due to its ability to withstand relatively harsher abiotic conditions (Ishikawa, Furukawa, and Oikawa, 1995). In Rhode Island, Enser (2006) suggested that this ability allows it to persist and expand into areas where lesstolerant natives cannot. At East Beach it appears that C. kobomugi shares a similar tolerance for salinity and desiccation associated with sand movement, and soil-water content stresses (Barbour, 1978;Ishikawa and Kachi, 1998;Kachi and Hirose, 1979;Maruyama and Miura, 1981;Nobuhara, 1967) as A. breviligulata and other dune species (Bertness and Ellison, 1987;Maun, 2009;Snow and Vince, 1984). In other words, the seaward expansion of C. kobomugi appears to be as limited by similar abiotic stresses, including anthropogenic activity, as populations of A. breviligulata.
The effect of existing environmental heterogeneity on the distribution of natives and C. kobomugi in New Jersey has recently been explored by Burkitt and Wootton (2010), but further investigations in Rhode Island have yet to reveal how dune structure and specific conditions dictate the distribution of these plant species.
In general, the replacement of A. breviligulata by C. kobomugi may have considerable impact on the structure of sand dunes. It has been suggested that in contrast to tall, native plants that buffer the dune from the strong forces of wind and salt spray, the low-growing C. kobomugi makes the dunes vulnerable to shifting sands and blowouts, and in effect, areas of secondary invasion (Virginia Department of Conservation and Recreation, 2011). The replacement of A. breviligulata by C.
kobomugi could contribute to erosion and loss of protection of inshore areas because of its shorter roots (Hayes, 2009;Lea and McLaughlin, 2002). Consistently lower dunes may not be as effective as habitat or barrier, especially in light of the fact that Ninigret Pond exists within such close proximity to the shoreline.
Conversely, some evidence tends to suggest that C. kobomugi is as an effective dune stabilizer as A. breviligulata or even perhaps more effective, as periodic dieback of A. breviligulata, due to marasmius blight, coastal storms, erosional phases, or washover drift lines, make the species vulnerable to replacement (Wootton et al., 2005). At East Beach it appears that the fine and extensive rooting of C. kobomugi binds sand into cohesive sod-like pieces (Johnson, personal observation). Photographs taken of roots in areas of dense C. kobomugi and dense A. breviligulata depict substantial differences in ability to bind sand solely with roots (Figures 24 and 25).
Testable and quantifiable comparisons of dune-forming abilities between C. kobomugi and A. breviligulata need to be made in order to determine which is better at maintaining the geomorphology of dunes. Measurements of the maximum depth of live roots frequently are used to identify which species are the best dune stabilizers (Hayes, 2009;Lea and McLaughlin, 2002;Wootton et al., 2005), but other factors, including the aggregation of sand grains by hyphae of AMF, may be of equal or greater importance. Furthermore, an essential characteristic of dune-building and stabilizing species is the ability to tolerate anoxia associated with sand burial. In general, this differential tolerance by certain species may be one of the principal causes of plant species zonation on coastal foredunes (Maun, 2004). As A.
breviligulata is associated with sand deposition in mobile and early fixed dunes  cm per year), its tolerance for sand burial and makes it a well-suited dune stabilizer (Maun and Baye, 1989;Maun, 2009). Typically, burial depths of 5-20 cm maintain and enhance shoot density, percent cover, as well as belowground biomass for native dune grasses (Seliskar, 1994;Maun, 2009). Maun (2004) reported that A. breviligulata and C. kobomugi are vigorous in areas with average sand deposition of 17 to 28 cm/year, an amount typical for the first 40 m of the foredune. After that point (in the next landward 20 m of the dune) sand deposition decreased by 3-5 cm resulting in a significant decline of vigor and distribution of both species. Conversely, Wootton et al. (2005) reported that burying C. kobomugi in sand, a technique that has been shown increase vigor in A. breviligulata (Maun and Baye, 1989;Seliskar, 1994), causes extensive mortality (United States Department of Agriculture, 1983;Disraeli 1984).
The possible causes for decline in plant vigor from excessive sand burial are unclear, and appear to be caused by a number of interactions, including increased desiccation, physiological deterioration of plant functions, and increased soil microorganism activity (Maun, 2004).

Spread of C. kobomugi and Implications for Future Mitigation
The robust growth form of C. kobomugi, and its propensity to expand via extensive rhizomes, makes it particularly difficult to manage. Widespread herbicide treatments for C. kobomugi have the potential to destroy non-target species, leaving the dunes susceptible to becoming rapidly destabilized. Wootton et al. (2005) found that localized herbicide application failed to effectively eliminate C. kobomugi when applied on sites in New Jersey. Mechanical removal of C. kobomugi has great potential to disturb or even destroy non-target rhizomes of A. breviligulata and the associated beneficial mycorrhizal networks (Koske, personal communication).
McGonigle and Miller (1993)  Past studies have indicated that C. kobomugi fills in areas, primarily those spaces that become available as A. breviligulata becomes sparser in its distribution over time (Belcher et al., 1984). The growth characteristics of C. kobomugi are likely the main cause for this trend, as the plants are shorter and have a greater number of leaves than A. breviligulata. Populations of C. kobomugi anecdotally documented in backdune areas potentially compound mitigation efforts at East Beach (Figure 26 and 27).
Backdune populations were less dense relative to those of the foredune, and likely spread by seeds or small rhizome fragments, rather than advancement of a clonal rhizome network. Though the seeds of C. kobomugi have been shown to have a low germination rate and seedlings high mortality (Yamamoto, 1964;Nobuhara, 1974;Sasaki, 1987;Ishikawa et al., 1993), transport of seeds is quite plausible, given their small size. The germination of seeds of C. kobomugi is controlled by external dormancy, initiated by scarification and followed by a moist-chilling event caused by low soil temperature and warming to approximately 35° C (Ishikawa et al., 1993). In terms of these germination requirements, Wootton et al. (2005)  characteristic to backdune areas is tantamount to decreased pathogen protection, and nutrient availability associated with AMF (Hawk and Sharp, 1967;Shisler, Wargo, and Jordan, 1987), essentially making A. breviligulata grow relatively poorly there (Maun and Baye, 1989).
The relative ages of C. kobomugi patch sites could be of importance in further addressing its expansion. At East Beach, though the approximate age of the initial invasion is known (Champlin, 1994;Standley, 1983), relative ages of secondary patch sites are unknown. In a series of stem-density assessments conducted in New Jersey, Burkitt and Wootton (2010) determined that the size of secondary invasion sites was actually a poor proxy for age. Instead they proposed that larger beds may just represent areas where multiple introductions of C. kobomugi occurred in close proximity to one another and the smaller populations merged to form a single one. Ishikawa and Kachi (1998)  are also permitted on the beach itself, frequently disturbing areas of the foredune. As the intertidal beach area is important as a nutrient cycling system, both in terms of algae and bacteria, one result of frequent vehicle activity is the exposure of sensitive organisms to desiccation. Though vehicles tend to travel in a corridor of minimal beach biota (Godfrey and Godfrey, 1981) when vehicles stray into the transition areas of foredune, crushing both seedlings and fragments of A. breviligulata, they potentially reduce the population of viable plants that act as stabilizers for new dunes ( Figure 30). This straying also quickly deteriorates the sloping incline of the foredune that is necessary to prevent widespread and unchecked wind erosion. With strong enough prevailing winds, large blow-out areas can occur, inevitably threatening an entire dune system (Godfrey and Godfrey, 1981). The robust growth characteristics of C. kobomugi make it an ideal candidate for colonizing these areas, recently made available by vehicle straying and beach traffic.
As a functioning population of AMF can have overarching effects on not only individual plants but also entire plant communities (Smith and Read, 1997), the disruption of the AMF community likely has significant, far-reaching effects for the future identity and functionality of East Beach dunes. The reduction in the population of AMF and MIP values in the East Beach dunes in response to invasion of the area by C. kobomugi, a non-mycorrhizal species (Johnson, personal observation), seems to be the result of a decline in the vigor and extent of the population of mycorrhiza-forming native species (especially the dominant A. breviligulata) that are necessary to maintain a large and diverse AMF community in the soil rather than the direct suppression of AMF activity (sporulation, germination of spores, hyphal growth, etc.) by exudates from the invasive. By replacing A. breviligulata, C. kobomugi disrupts the associated AMF populations therefore critically reducing the potential for natives to persist, and subsequently re-colonize areas. These findings are of particular importance to the successful re-planting of dunes after mitigation of the invasive. Concurrent with vegetation surveys in New Jersey that indicated overall native species diversity and density was reduced in areas of C. kobomugi when compared to non-invaded areas, (Wootton et al., 2005) it appears that C. kobomugi effectively replaces natives in the plant community at East Beach. Though difficult to quantify, a monoculture of C.
kobomugi will certainly be a significant aesthetic loss and may substantially alter overarching ecosystem dynamics. Management methods are difficult to implement both due to the rapid and expansive rhizomatous growth of the invasive as well as the inherent sensitivity both of native dune plants and AMF hyphal networks. Widespread anthropogenic pressures and prolonged shoreline erosion may result in the elimination of species from dune habitats, or the introduction of others (Maun, 2004 1. Soil into 1.5 L beaker. Add 200 ml water to break up clods and wash substrate.

2.
Add water in hard stream to aerate and agitate.

4.
Wash material on fine sieve into small beaker.

5.
Wash material from beaker into large centrifuge tube (50 ml capacity). Fill with deionized water until approximately 40 ml level.

7.
Pour water off and wipe inside of centrifuge tube thoroughly to remove debris/dead spores.

8.
Add 40% sucrose solution, shake well and vortex to resuspend pellet of kaolin.

9.
Centrifuge for approximately 1 minute near maximum speed.
11. Wash material on 53 um sieve onto filter paper, in Buchner filtration system. 12. Remove filter paper from Buchner funnel and place onto plastic Petri dish for further examination.
13. Remove spores from filter paper using inoculating needle with PVLG at tip.

Appendix D: Spores of G. gigantea germination assessment (adapted and modified from Koske, 1981):
Sand was obtained from root zones of A. breviligulata and C. kobomugi at East Beach in Fall 2009. Sand was kept in refrigerated (5°C) storage for approximately 5 months to stimulate germination before spore extraction. Spores were obtained from the sand using a modified wet sieving and decanting procedure from Walker, Mize, and McNabb (1982). This extraction method was similar to that used in spore abundance assessments, but did not incorporate sucrose suspension to remove debris (essentially water aeration and two sieve-filtration onto collection dish). Only viable, healthy spores of G. gigantea were used (Lee and Koske, 1994). Sand plates were made to represent soil conditions in the field -5 glass Petri dishes, each with 40 ml of sand from root zones of A. breviligulata and C. kobomugi were made. Petri dishes had a filter paper soaked with deionized water that was placed on the surface of the sand.
Molecularporous membrane tubing (Spectrum Medical Industries, Inc., Los Angeles, California) was cut into two, 1 cm squares and placed onto the soaked filter paper.          . Plants grown in field soil of A. breviligulata had significantly higher root colonization in comparison to those in field soil of C. kobomugi (P < 0.0001) and the 50% C. kobomugi / 50% A. breviligulata soil combination (P = 0.0009). Plants grown in soil of C. kobomugi had less mean root colonization than the 50% C. kobomugi / 50% A. breviligulata soil combination (P = 0.0004).