Am. Midi. Nat. 163:186
Transcription
Am. Midi. Nat. 163:186
Am. Midi. Nat. 163:186-196 Variation in Endophyte Symbiosis, Herbivory and Drought Tolerance of Ammophila breviiigulata Populations in the Great Lakes Region SARAH M. EMERY' Department of Biohgy, University of Louisville. Louisville, Kentucky 40292 DESIREE THOMPSON Departvient of Biology, Kalamazoo College, Kalavuuoo, MicJiigan 49006 AND jENNIEERA. RUDGERS Department of Ecology and Evolutionary Biology, Rice University, Houston. Texas 77036 T.—Ecologi-sts havi- romt- to appreciate thai plant genonpe can afTcct the success of restoration efforts, biii the role of plant .s^inbioiits has received less attention. Epichloë^.ype endophyiic fungi (EF) grow in ihf aboveground tissues of most plants and wcur systemically in 2()-30% of all grass species. Despite lhe poieniial for EF to alter plant competitive hierarchies and to be lost under improper seed storage, they have yet to be considered as an important factor in plant restoration ecotog). We surveyed EF infettion freqtiency in 42 native and restored populations of the dominant Great Lakes dtinc grass AmmojthiUi breviliguUita. We also conducted a greenhotise experiment to compart- efTects of herbivoty and drought on an uninfecled Michigan population of .\. hreinliguluta and the commercially available 'C^pe' variety, which is commonly planted for restoration. Surveys revealed low levels of EF infection in natural populations in the Great Lakes region. 'Gajw' nursery stock was 100% infected. In the greenhouse, the Michigan plants were more sensitive to grasshopper herbivory and drought than the 'Cape' plants. Our results suggest that the variety of A. breviligiilata used in dune restorations possibly could alter plant and itisect community dynamics due to differences in EF stattis. though further tests under field conditions aie needed. INTRODUCTION in rereni years, restoration has shifted from a 'trial and ertor' practice lo a scientific discipline informed by ecological Lheory (Motitalvo et ai, 1997; l*almer et ai, 1997; Young et ai, 2001). For instance, an awareness of population genetics helped restoration ecoiogists evaltiatt' the telative importance of gene flow and genetic diversity in restotation elTorts (Montalvo et ai, 1997; McKay et ai, 2Ü05). A commtmity ecology perspective shifted many restoration efforts from a single species focus to one that considers interspecific interactions, even across trophic levels (Palmer e.t ai, 1997; Gratton and Denno, 2Ü06). For example, 'target' plant species ofteti need associated mtititalisms (specifically pollinators, seed dispersers and mycorrhizal ftmgi) for sticcessful growth and reprocUtction in lestorations (Johnsoti et ai, 1997; Bevcr H ai, 2003; Yuttng et ai, 2005). Organisms that fonn very close associations with their hosts (as parasites or tnuttialisLs). may have particularly large etfects on the success of a restoration hy altering host plant competitive ' Conesponditig author: Bi<)log>' Department, University of Louisville. 139 Life Sciences Building, Louisville, Kentucky 40292; e-mail: sarah.emery@louis\ille.edu; Telephone: (502) 852-5940; FAX: (502) 852-0725 186 2010 E M E R Y E T A L . : A M M O P H I U S BREVltJGUlJ^TA PoPUIATtON 187 ability and tolerance to the harsh biotic and abiotic conditions often louiid in degraded liabitats. However, these interspecific relatiniisliips arc often overlooked in restorations because oí their cryptic nature (Clay, 2001; Bever el ai, 2003). Systemic endophytic ftmgi (EF) are one class of symbionLs whose role has gone tinnoticed in restoration efforts. Systemic EF are specialized organisms that grow in ititt-rtelhilar spaces of itlioveground plant tissue. In general, the prevalence of EF in plants is nnknowti due to their cryptic natttre; however, EpicJiloe^ypt systetnic EF are estitnated to foiin relatíotiships with 20:i()% of all grass species, especially with cool season (C3) .species in stibfamily Pooideae (Wiiite, 1987; Clay and Schaidl, 2002; Schardl et ai. 2004). These EpichloëA.yY>e endophytes (hereafter referred to simply a.s EF) can increa.se drought tolerance and nutrient uptake in some platit hosts (Malinowski and Belesky, 2000; Malinowski et ai, 2005), as well as confer resistance to herbivores through ihe production of bioactive alkaloids (Btish el ai, 1997; Clay and Schardl, '1{)()2). Ftirtlier, EF associations can increase competidve ability and slow succession in sume svstenis (Rtidgers etai, 200.'), 2007). Depending on etivironmcnlal conditions, EF may also act :is parasites (Cheplick et ai, 1989; Saikkonen et ai, 1998), thotigh few esj^^riments have manipulated EF in more than a handful of grass species (Faeth et ai, 2004; Saikkonen et ai, 200fi). Based on the evidence to date, it set'in.s probable that EF could have strong eflects on plant competitive hieraj chies, ecosystem processes and community assembly in restorations as well as natural plant communities (Clay and Holah, 1999; Rtidgers and Clay, 2005; Rtidgera et ai. 2007; Rtidgers and Clay, in press). For this study, we stin'eyed EF infection in populations of Ammnpkil/i hrexdligulnta. the dominant native grass of Atlantic coast and Great Lakes dunes in the U.S. Beciuise sand dunes are suhject to blow outs, mining efforts and other disttirbaiices. restoration of dunes habitats of various sizes (from several sqtiare meters to several hectares) has become a common pracdce (.Albert. 2000; Palmgren, 2000), and A, breviligulata is widely phinied in restorations tliroughotit ihese areas. A. bmñligulata had variable FT infet tion by the eiidophyte EpichUië typhina in East Coast populations (White et ai, 1992), althoiigh nothing is known al)out EF presence in Great I^kes populations. For this sttidy, we surveyed EF infection in A. hreuiligulata in 42 sites throughout Michigan and Indiana, as well as from six nurseries that supply plants for dune lestorations. Because A. hreidligiiUitn is clonal and rarely reproduces by seed (Fant et ai, 2008), we generally expected either 100% or 0% infection of populations. We also condticted a greenhonse experiment to compaie the perlormiince of a widely available commercial variety that was 100% infected with EF (E+) to plants propagated from a native Michigan population free nf entlo[)hyte infection ( E - ) . We mani[)iilated both drought and herbivoiy. the tw(} priniaiy benefits ilifiught to be conferred by endophyte symbiosis in grasses (e.g.. Clay and Schardl, 2002). Based on the effects of EF infection in Lolimn arundinacnim (tall fescue), we predicted that herbivores wottld prefer to feed on the E— local Michigan plants. We also expected that E+ nursery plants would tolerate drought i)e(ter than die local Michigan plants. Differences in the peribrmance of 1<K al and nursery |)(>pulations may have important implications for successful restoration eilorts. METHODS GREAT LAKES DUNES SURVEY During the summer of 2005 we conducted a .survey of EF infection in AmmophiUi hrndligulata from natural populadons in five Michigan state parks located ahîng the shore of Lake Michigan (i.^.. Petoskey, Saugatuck, Van Buren, Warren Dunes and Wilderness State I'arks). In Jul. 2006, we conducted a much larger .survey of 18 natural populations »ï A. heviligulata in Indiana and Michigan (Table 1, Fig. 1). As part of a larger study, we also 188 163(1) T H E AMKEUCI\N Mit)LANi) NATURALIST I'.MMS. 1.—EF infection status of Ammophila breviligitUiin in natural antl restored sites in Michigan and Indiana Site lndia[ia Indiana Indiana Indiana Indiana Indiana Indiana Dunes National Lakeshore, Site 1 Dunes National Lakeshore, Site 2 Dunes National Liikeshore, Site 3 Dunes National lakfsliore. Site 4 Dunes Nalional Lakeshore, Site 5 Dunes State Park, Site 1 Dunes State Park, Site 2 Slct-pinti Sleeping Sleeping Sleeping Bear Bear Bear Bear Dunes National Dunes National Dunes National Dunes National Uikeshore, Site 1 Lakeshore, Site 2 l*ik<rshore. Site 3 Lakeshore, Site 4 Naitiral site» (infcctod/tota!) lillcr, sectl Restortd siles ( infected/t(ital) tiller, seed 0/3, n/a 0/5, n/a 0/5, n/a 0/5. n/a 0/5, n/a 0/5, n/a 0/5, n/a 0/5, n/a 0/5, n/a 0/5, n/a 0/5.n/a 0/5, 0/7 0/5, n/a 0/5. n/a 0/5,1/8 0/5, 1/3 0/5, 0/32 0/5, 0/51 0/5, 0/5, 0/5. 0/5. n/a n/a 0/36 0/33 Petoskey State Park. MI Saugaiurk State Park, MI \';ui Buren State Park. MI Warren Dunes State Park, Ml Wi!denie.ss State Park. Ml (irand Mere State Park, Ml 0/6, 0/9, 0/8, 0/8, 1/5, 0/9, Pictured Rocks National Lakeshore (Grand Sable Dunes) Escaiiaba, Ml Marquette, MI, Site I Marquetic, Ml, Site 2 Sdioohmfl County, MI US Highway 2, Ml 1/5. 0/40 0/5, n/a 1/5,0/20 0/5, 0/28 0/5. n/a 0/5, 0/36 'tlapt;' Stock, (lape Farms (Lewes, DE) Nursery Stock (Ml genoti'pe). Natural Garden (St. Charles, U.) 'Cape' Stock, Vans Pines Nursery (West Olive, Mí) 'Vans' Stock, Vans Pines Nursery (We,st Olive, MI) 'Cape' Stork, Chtirch's Nuiser\' (Cape May, Nj) ' Stock, Peat and .Sons Nursen,' (Westhampton, H\) 20/20 n/a 0/20 49/49 0/40 20/20 20/20 n/a n/a n/a n/a n/a n/a n/a n/a 0/9 n/a n/a n/a n/a n/a 5/20, 0/9 0/5, 0/5. 0/5, 0/5, 0/5. 0/5, 0/29 0/33 0/35 0/23 n/a 0/24 n/a n/\\ sui-veyed adjacent dutie restoration projects where A. breviü^ilata had been planted at sotne point in the past 25 y. The specific planting liistoiy (plant source, exact yeitr of planting) of most of these restorations was unkiiowti. At each site, we collected 25 tillers from a 10m X 10m area (one every 2 m). For some sites we were also able to collect seeds from individuals, thotigh in general Howering and seed set arc veiy low in this species. i-roiii 2()();V20()7, we ptirchased AmmofjliiUi hreviligukita planLs Irom five nurseries in New York, New Jersey, Delaware. Michigan and Illinois lo test for EF infection from these suppliers. Becatise of the ver)' low seed production by A. hreviligiilata, ntirserie.s propagate tins species cionally, and sell entire tillers for planting in restorations. Most of the.se nurseries canied the 'Cape' variety of A. frreviligidaia, which is a vigorotis strain originally developed from a New Jersey poptilation by the United StaLes Department of Agriculture. Nattiral Resources Conservation Center, Cape May Plant Materials Center in Cape May, New 2010 EMERYETAL.: AMMOPHUA IIRUVILICUIATA PoftiLAnoN 189 Jersey in the early 1970s (Soil Conservation Service, 1977). The Illinois and Michigan nurseries also carried plant.s ihat originated from natural Michigan populations. To quantify EF infection in tillers, wf took thin sections of inner leaf sheath tissue from 550 tillers per jjopulaiion (depending on the number of tillers that suiwed transport) and applied 0.1% aniline blue plus lactic acid stain, which allowed for visual identification of fungal h>phae present in the leaf tissue under 200X magnification {Clark el ai, 1983). For seeds, we soaked seeds in 5% sodium hydroxide solution for 24 h, then dissected (»ut the aleurone layer from the seed and stained this layer with the same aniliue blue plus lactic add stain used for staining tillers (Clark el ai, 1983). This technique has the limitation of detecting false negatives if EF hyphal densities are low, but it has been widely used to score t.F presence and is ihf cheapest and fastest method currently available. Also, while it is possible that detection of EF could vary with time uf sampling, field sampling occurred during mid-growing season, when EF growth should be most pie\'alent (ju et ai, 2006). tiREENHOL'SF. EXPERIMENI In the summer of 2006, we established a greenhouse experiiuent where we filled 20 plastic pots {23 cm diameter, 30.5 cm deep) with 15:1 mixtmes of screened and washed sand {Quikrete Inc., Atlanta, GA) and local field soil {collected near the W.K. Kellogg Biological Station in southwest Michigan). We planted one tiller from a Michigan population {collected from Sleeping Bear Dunes National Lakeshore, Empire Ml; 44.82N, 86.06\V) and one tiller ofthe 'Cape' variety {from Vans Pines Nursery, MI) in each pot. We measured the number of leaves and the length of each leaf for each individual, and plants were allowed to grow in the greenhouse for one week before the experiment began. We set maximum greenhouse temperatures al 37 C and provided (i b of supplcnu-ntal high pressure sodium light daily to mimic dune conditions. Treatments were imposed in a 2 X 2 factorial design, with a drought (1>+ or D - ) aud an herbivory treatment (H+ or H—). plus interactions of the two treatments. We constructed (iicin Tall cages around each pot using wooden dowels and bridal veil {O.dlcm mesh). For tlie H+ treatment, we used Trivwrotropis mcnitima (seaside grasshopper) collected ironi Cirand Mere State Park in southwest Michigan {approximately 42.00 N, 86.50 W). Prior to ciïllection, we observed T. maritima feeding on Ainwoj)hila bmnltgttlaia individuals along tbe foredime region. This species is also kno^vn to be an herbivore of A. brei'iligidata located on the Atlantic Coast {Barimo and Young, 2002). We added one grasshopper lo each H+ pot, allowing them access to plants for three days at a time to prevent the grasshoppers from destroying the plants as well as from feeding on plants they normally would not choose. For ihp D+ treatment we watered the plants imtil pots were soaking wet once a week, while non tirought (D-) pots were similarly watered every other day. In the H+D+ treatment, plants were exposed to both drought Lind herbivoi"y. Plants in the H —D— treatment were caged as A control. Each treatment combination was replicated five times and pots were randomized and rotated on greenhouse benches. After 60 d, we again measured lhe number of leaves and the length of each leaf for eacb plant ill the experiment. To quantify iudividual plant performance in each treatment, we subtracted final total leaf length {stimined across all leaves) from the initial total leaf length. We also vÍ5ually estimated herbivore damage on a 0-5 scale {0 = no herbivor), 5 = 75%+ leaf consumed) for each plant in the H+ treatments. To compare growth of the two plant types, we subtracted leaf length change ofthe Michigan plant from leaf length change ofthe 'Cape' plant in each pol (to account for any pot effects). We feit that these were more insightfiil measures of plant performance than final biomass, as the plants were not necessarily of equal size when planted. We also mea.sured tillering by plants, but only four of 190 163(1) T H E AMKRIÍAN MIÜL.'\ND NATURAI.IST r\Bi,t: 2.—ANOV.A results for t'ffects of drought and lierbivor\' on growtli of the X^pe' and Michigan ]»huit types ill lhe greenhouse experiment <h ss 1 Hcrbivoiy Droniiht Ilcriiivor)' X Drought Krror 1 1 1 16 13.378.1 937.8 397.4 28,840.9 7.422 0.520 0.220 0.015 0.481 0.645 .\iiihi^in population: Ik'rbivorv Diouglil Hrrbivory X Drotight Krror 1 1 1 16 17.380.8 81H1.3 4253.« 10.648.6 26.115 12.293 6.391 < 0.001 0.003 0.022 SlTrss Sdiird' 1' 'dupe' iifirift%- iht; 40 plants in the cxperiinent sent tip an addition;tl tiller during the 60 d of the expcritncnt (two 'Cape' and two Michigan) and no was not insightful tor cotripaiisons across treatments. We examined the cfTecis of llie suess treatmeiiLs on each plain t^-pe tisitig factorial ANOVA. We also planned to tneasure the efïects of the stress treatments on differences in performance between the two gra.s.s t>^es nsing factorial Í \ N O V A . Becatise of our reialively low sample si/es, Iiowever, we suspected that the power to detect différences would be low, so we also condticted a confidetice interval test of tlie ntill hypothesis (no difference in between ihc two plant types) (StcidI and Thomas, 2001). RESULTS GRILAT LAKES DUNES SURVF.Y The majoriiy (18/23 sites) ot.AnumphiUiftfyí«7í^/flí<7populations from nattital dunes sites in Mi( liigau :ind Indiana were not etidophyte infected (lahle 1 ). The ïvw populalions where we did Hnd EF (iwo siles at Sleeping Bear Dunes Naiionul Lakfshore, Piciured Rocks Naiional Lakeshore, Wilderness State Pat k. Mi and one site near Miirquette, MI ) bad relatively low levels oí infection (20-r)0%). In contrast. 'Cape'variety stock plants were 100% infected with EFatall four nurseries wbere il was sold. Tlie iwo other ntirsei^- varieties ('Vans' and the Naitital (larden) showed no inlection by EF. We confmned tbese patterns for '(¿ipe' and 'Vans' from Vans Fines Ntirsery by plating sections of tiller on com meal maltagar (Bacon & White. I9Í14). All siunptes (15/l.'i) from "Cjipe" tissue grew cultures morphologically itleniiliahle as Epichbe, whereas 'Vans' culmres did not. Given tbese diiferences hetween the 'Cape' and native Gieai Lakes populations, we expected to find high le\'els of endoph\ic in tireal L^kes rcstoiTition sites. Surprisingly, Iiowever, only one restoration site bad A. brerilzgulata infected witb EF ((Inmd Mere, witb 25% infection. Table 1). GREEN>IOtISE EXPERIMENT Grassboppers. but not drought, reduced grciwth of* ibe 'Gape' plants in tbe expeiitncni and there were no interactions between the two Stressors (Table 2. Fig. 2A). Both gra.sshoppers and drotigbt reduced growth of tbe Micliigan plants and the combination of the IWO Stressors compounded the effect (Table 2. Fig. 2B). Wiien comparing tbe growtb of 'Cape' and Miihigan |)lani types, there were no significant effects of lierhivory or drought ireatmenis in the ANOVA (H: F,.y„ ^ 0.11, P = 0.75; D: Fi ^.o = 1.49, P - 0.24; min EMERYET AL.: .AMMOPHILA BREVÍUCULATA POPULATION 191 Marquette 2 sites_ PRNL SBDNL 4 sites IDNL 5 sites SSP VBSP GMSP QSP IDSP 2 sites FK;. I .—Map of sites for survey (if endophyte infection in populations of Ammophila breviiigulata. Site iihbreviaiions correspond to those listed in Table 1. Each site included paired restored and natural dune communides HxD: Fi_2o = 0.85, P = 0.37), though the plant types did differ in herbivoty damage (H: F|.-jo = 28.84, P < 0.001; D:F,.ao-8-37, P = 0.011; HxD: F|..^o= 8-37, P = 0.001, Fig, 2C). A retrospective power analysis using the program P;\SS {Hintze, 2007) indicated low power levels for this stitdy (0.06-0.21 for each factor). To follow tip on this analy.sis, we conducted a comparison of 95% confidence intervals for differences in plant growth in each treatment. This analysis showed that the herbivory (H+D-) and drotight (H-D+) treatments had differences in plant growth greater than zero indicating that the 'Cape' type grew more than the Michigan type in these treatments (Fig. 2D), while the H - D ;ind H+D+ treatments included zero in tlie 95% CI, indicating no significant difference in between plant types. 192 T H E AMERICAN 150 150 rowth (cm) A B 100 100- 50 50- 2 0 Cape' 163(1) NATtJRAt.isT c 1- -1 (0 -*- s -50h- -50 -100 -100 H-D- H+D- H-D+ H+D+ H-D- Treatment Q. > (0 CD • > H+D+ 150 D ^ 3- 100- to JZ "i H+D- H-D+ Treatment c > H- • 5020S Ü -50|u 1 H-D- I I I 1 H+D- H-D+ Treatment H+D+ I I o r -100 H-D- H+D- H-D+ H+D+ Treatment FH.. 2.—Leaf Icngih change (growth) of 'Cape" (A) and Michigan (B) planes in c^ach stress treatment. Positive values indicate growth over time, while negative values indicate senescence or leaf destniction {herbivory) over lime. (C) Difference in visually estimated herbivore damage between Michigan and 'Cape' planLs in each stress treatnieni {on relative scale from 0-5, where 0 = no herbivory and 3 = 7.'i%+ leaf consumed). Positive values indicate more herbivore damage on Michigan plants than the 'Capt-' |ii.iiH.s grown wilh ihem. {D) Avei-age differences in leaf length change {growth) between 'Cape' and Mil higiin plaTiLs in eacb stress tieaimenL. Po.silive values indicate more growth in 'Cape' than Michigan, while negative \-ahies indicate more growth in Michigan than 'Cape'. A difFerence of zero {dashed line) indicates no difference in g;rowth between 'Cape' and Michigan plants. Error bars represent one + / - S E {A-Cl) or 95% Q s {D) DISCUSSION GREAT LARES .\MMOPHIIJ\ AN1Í L. We showed that tnany populations of Ammophila bretñligulata sampled in Indiana and Michigan had no evidence of EF infection. However, EF are not absent from this area. At least five sites had A. fneviligiilata populations with low to moderate levels of EF infection (<ñO%, Table 1). Interestingly, these five sites were all located in the northern lower peninsula and upper peninsula of Michigan. Given the potential benefits of EF for drought 2010 EMERVETAL.: AniMOPHitA BREVILÍCULATA POPULATION 193 tolerance (Malinowski and Belesky, 2000; Malinowski at ai, 2005), we would have expected higher frequencies of EF in sotithcrn poptilaiions with liigher rates of evapo transpirât i on, as lia-s been repotted for other EF infected grass species (Lewis el al., 1997; Leyxonas and Raynal, 2001). More extensive surveys spanning the geographic range of A. breviligiilata would lielp generate hypotheses about cHmadc factors that could influence EF distribution within thi,s species. We expected ihat populations would lypicalty he 100% infected with EF or 100% free of EF for several reasons. First, EF arc rarely horizontally transmitted in this system (White el nl, 1992). Second. Ammofúila breoiligiilala reproduces almost exclu,sively by clonal growth, resiiliing in ver\' tew genotypes in a given poptilalioii. For example, Fatil et ai (2008) found only 1-3 genotypes m '~ and 1-13 genotypes per population in Illinois and Wisconsin siies. Third, a study by White et al. (1992) along the New Jersey coast showed a simitar infection pairetn, althotigh suffered from small sample sizes. Wliite et ai (1992) also foutid an increase in poptilations widi EF presence tht ough time, possibly reflecting Uie inct eased use uf the 'Cape' r^riety iti tesiorations along the coast afier 1971, Tliis expecied pattern does not seem lo hold for Great I^kes poptilaiions. wliich hosted either no. or relatively low levels of infection. While il is possible our method of scoring turned up false negatives, all Cape' plants were 100% infected, suggesiing that scoring w;is reliable. Temporal and spatial variation in the cosLs and benefits oí Epichl/ti' (CVAV and SchardI, 2002; Morse W ai, 2002), as well as the potential for imperfect vertical transmi.s.sion to seeds (Ravel et ai, 1997) could explain the maintenance of low levels of symbiosis in some natural poptilations. Surprisingly, only one of the restoratirjus we examined showed any presence of EF itifectinn. despite a histoiy of .Ammofihila lnniU:i}>^i.t/ita plantings over the past ^0 y, when [tresutnably the only commercial variety available would have been the '(^ape' variety. It is possible that A. breviiigulata conld lose its infection over Ume, especially if the EF are deirimental to the plants in a new habitat, or if seeds or tillers lose tlie infection (Clay el ai, 1989; Faeth, 2002; Sachs and Simms, 2006). Great l^kes dunes may be less osmotically stressful for plants than ocean dunes dne to lack of salt water, and so EF may not be as beneficial iji this system. However, persistent low levels of EF infection in natural populations may indicate at least occasional benefits of this symbiosis. Further, because restoration records were often poorly kept, there is a need to verify the identily of these plants before we can confidently claim that EF infections are losl in ihese populations. HERBIVORY. OROUGHT AND AMMOPHILA In the greenhouse experiment, the '{¡ape' type of Ammophila breinligula.ta had less lierbivory damage than the Michigan type. This could he a result of local adaptation of the ^grasshopper (Hanks and f^enno, 1994; Mopper and Strattss, 1998), as only gritsshoppers ( ollected from Michigan were tised in these experiments. However, the Michigati plants used in the greenhouse experiment were collected approximately 350ktn noith ofthe site where the giasshoppcrs were collected, in an area where Trivierotrofm imtrithrui does not occur (Schokens et al., 2005). Thus, we think it tmlikely that local adaptation by the heibivore was the cause of differences in herbivore resistance between the A. breviligulata populations. EF infection may explain the reduced herbivore damage on 'Cape' plants, though we cannot be certain oí' this Ix'cause of the confounding oí plant lype widi EF infectit)n suittis. However, EF infection is widely knnwti to reduce herbivory in several agriciilltiral giass species {e.g., Lolium perenne, 1 .olium arundinareum) due to the production of peiamitie and loline alkaloids (reviewed by Schard! et ai, 2004; Rndgers and Clay, 2005). While the efVects of EF on insect herbivores of nadve grasses have received less attetition and can X'ary with 194 T H E AAÍERICAN MIIHJVND NATtiRAiJírr 163(1) environmental conditions (Saikkouen et ai, 1998), EF often deicr herbivores in other .sysit-ms that have been rigorously tested {reviewed by Clay and SchardI, 2002; Schardl et ai, 2004; Rtidgeis and Clay, 2005). EF have also been shown lo incrt-íLse drought tolerance of host plants by increasing osmotic adjustment, increasing stomatal resistance, moderating rates of photosynthesis, increasing leal rolling and increasing growth of root bairs (Malinowski et ni, 1999; Bultman and Bell, 2003; Schardl *•/ al, 2004; Malinowski et ai, 2005). In our greenhouse experiment, 'Cape' plants maintained positive growth under drotight stress, while Michigan plants lost leaf length due to leaf senescence. Again, tins may be explained by differences in local adaptation to osmotic stress between the two Ammophila lirei'iligulata poptilation.s. as we cannot separate plant genotype from endophyte presence in this sttidy. Osmotic stress is expected to be greater along the Atlantic coast due to salt spray and increased soil salinity (Boyd and Barbour. 1986). However, drotight and high evapotransporation rates are common in young dunes along the Great Lakes as well {Lichter, 1998). While the 'Cape' variety was a better performer than the Michigan plants under drought and herhivory stress, the combination of both stipsses eliminated significant differences between 'Cape' and Micbigan plant growth, While other studies have proposed that the plant benefits of EF infection shotild be maximized ai extreme environmental conditions {Saikkonen et ai, 19*18; Farih and Fagan, 20{)2), we know of only one sttidy that has specifically examined the interaction between drought and berbivur)' on EF infected plants. Bultman and Bell (2003) found tliat drougbt generally increased EF infected plant resistance to fall armyworms {Spodoptera frugiperda). but not to aphids. The costs of maintaining a symbioiil in high stress environments may cancel any benL-fits conferred by the endophyte (Johnson et ai, 1997) CtlNOLUStONS Our greenhouse study demonstrates the importance of considering both plant source and symhionts in planning restoration efforts. Though we are unable to explicitly separate effects of plani type fr((m effects of EE infection in this study, studies in other .systems have sliown that EF infected plants can slow successional dynamics (Rudgers etai, 2007). Nursery stock plants of'Cape' are 100% infected with EE, and tolerate drought and herbivoiy bcttt-r than native Michigan ¡liants. Uliile the 'Cape' variety is widely available for restoration efforts diroughout lhe Eastern US, managers should consider the possible effects of cryptic symbionts such as endopbytes, as well as plant genotype, when selecting stock populations for restorations. Further work is needed to confirm field performance of this variety, to uiultTstaiu! lhe iniplicaiions of altering EF infection lrf(]ueiKÍfs in dune restorations, as wfU as to add lo our knowledge of mechani.sms regulating biodiversity of dune systems. Acknmoledgments.—Thanks lo Megan Rua for lab assistance and the W.K. Kc^llogg Biological Station for grecnhoiLSc space. Many thanks to the government agencies, park managers and htcal landowners for access to the- sites for surveys. This research was made possihle through the National Park.s Ecological Research Fellowship Program, a pannershtp between the National Purk Service, the Ecological Society of /Vmcrica and the National Park Foundation. It is ftiiided through a generous grant from lhe Andrew W. Mellon Fouiidaiion. LrrERATtiRE CtTEll .\L[ILK[. D . 2000. Borne of the Wind: An introduction to the ecology of Michigan sand dunes. Michigan Natural Features Inventory, East Lansing Michigan, 63 p. 2010 EMERYET .y,.: AMMOPHiiABREVn.iGuiATA POPLOATION 195 J. F. wii 1). R. YouMi. 2002. Grasshopper (Orthoptera: Acddidacj-plant-cnvironraentaJ iiitcr;Ktioiis in relation to zonation on an Atlantic coa.st barrier island. Envir. Entom., 31:1158-1167. llt:\'KR. [. D., P. A. Sr:nm-T7. R. M. MILLER. L. GADKS ANU |, D . JASTROW. 2003. Inoaihition with prairi(! mycorrhizal fungi may improve restoration of tiative ¡irairie plant diversity. Ecol. fteU., 21:311-312. IStiYD. R. S. AND M. G. BARBÍJOR. 1986. Relative Salt Tolerance of Cakile-Edenttila (Bra.'wicaceae) frntn Lacustrine and Marine Beaches. Am. J. Hot., 73:2S6-241. ikii.TMAN, T. L. AN!) G. D. Rui.. 2003. Imcratlion bciwcen iiingal endophytes and environ men ul streswors influences plant resistance lo insccw. Oikus, 103:182—190. BUSH, L . P., H. II. WILKINSON AND C. L. SCHARDL. 1997. Bioprotective alkaloids of grass-fungal endophytr symbioses. Plant Phys., 114:1-7. CHEPLICK, G. P., K. CtAV AND S. M^RÜS. 1989. Interactions between infection by etidophytir ftmgi and nutrient limitation in tlit- grasses Lolium perenne and Fesiuca anindinacea. Nni- Phytol., 111:89-98. CL^RK, E . M.. J . F. WHITE .XNI» R. M . PARIIRSON. 1983. Improved histocheinical techniques for the deiectiiin of Anrtmmiuin rtienofthialum in tall fescue and nu-tliods ol in vitro culture of the fungii.s. /. Microh. Mrtb.. 1:149-155. (j,\v, K. 2001. Symbiosis aiid the regulation of communities. Am. Zooi, 41:810-824. , G. P. GtiKfUCK A.ND S. MARKS. 1989. Impact of the Fungus Balansia-Henningsiana on PanicumAgrostoides.- Frequency nf Infection. Plant-Growth and Reproduction, and Resistance to Pests. Oecologia. 8 0 : 3 7 4 - : Î « 0 . AND |. Hoi,\n. 1999. Fungal endophyte symbiosis and plant diversily in successional fields. Science, 285:1742-1744. — • ANn C. ScuARiii.. 2002. Evolutioiiaiy origins and ecological consequences of endophyte symbiosis witli glasses. Am. Nat, 160:S99-S127, ['.•VETH, S. H. 2002. Are endoph>tic fungi defensive plant mutualists? Oikos, 98:25-3fi. .\Nt) W. F. FAGAN. 2002. Fungal endophytes: Common host plant symbionts but nnrommon mtittialisLs. /iJíí^r. Ompar. Itiiii, 42:;i(iO-.^fi8. , M. L. HK.[ANI)HR .KNVI K. T . S.MKKONÜN. 2004. /Vsexual Neotyphodium endophyte.s in a native grass reduce conipeiitive abiÜLies. Eroi /Wi., 7:304-313. l\.vj. ]. B.. R. M. HoL.M.s^^i'ROM, E. .SiRJüN, ). R. [.Tn:RSf)N .\ND S. M.\sr. 2008. Genetic .structure of tbreatened native ptjputations and propaguics used for restoration, in a clonat sfwcies, AmmophiUi breviligulata {American beachgrass). Restor. Ecoi. 16:594-603. , C. ANn R. F. DENNO. 2006. Arthropod food web restoration following removal of an invasive wetland plant. Ecoi App., 16:622-631. L. M. .MMi) R. F. DENNO. 1994. Local adaptation in ihe armored scale insecl Pinidaulacaspis pgntagima (Homoptera, Diaspididae). Ecohgs. 75:2301-2310. i, J. 2007. N<;SS. P.\SS. and GKSS. NCSS. Kaysville, UT. [DHN.'SON. N . C.,]. H. (ÍRAIIAM AND F. .\. SMfni. 1997. Functioning of mycorrhi7.al associations along the niutiialisni'parasitUni continuum. Neui Phytd., 155:575-586. I'josr, R. E. 1995. Acremonititn in fescue and ryegnis.s: Boon or bane? A review.y. Anim. Sei.. 73:881—888. [il, H. J.p N. S. Hiu., I. ABWITT AND K. T. INCÍR\M. 2006. Temperature influences on endophyte growth in tall fescue. C.rof> .Sri.. 46:404-412. I.EW1S, G. c , f ^ R.-\\r.i. W. NAFFA.\, C. ASTIKRAN» G. CHARMET. 1997. Occurrence of Afrnnitwmwi endophytes in wild poptilation.s of Eolium sjtp. in European counlHes and a relationship between level of infection an<l climate in France. Annals App. Bioi. 130:227-238. S. C. ANn G. RAW.^:. 2001. Presence otNeotyphodiurtk^like endopbytes in European grasses. Annals Apfi. Bioi. 139:1 U)-127. R, J. 1998. Primary succes.sion and forest development on coastal Lake Michigan sand dunes. Ecol. Monog., 68:487-510. MALLNIIWSKI, D . P. AND D . P. BELESKY. 2000. Adaptations of endophyte-infected cool-season grawes to environmenial stresses: Mechanisms of drought and nüneral stress tolerance. Crop Sd.. 40:923-940. T H E AMERICAN MtDLANu NATLIRALIST 163(1) .\Ni) |. M. FKIMIK.RS. 1999. Endophyte infection may affect Uie competitive ability of lail fcsiiie grown wilh red clover,/ Agimt. Crop Sri., 183:91-101. G. C Ltwis. 2005. Ahioiic stresses in endophytic gra.sses, p. 187-199. In: C. A. Rollens. C. V. West and D. E. Spiers (eds.). Neotyphodium in cool-season grasses. Blackweil Publishing, Ajiies, Iowa. M(xw,J. K-.C-t. CHRISTIAN. S.IIAKRISON AMI K. |. Rtct.. 200.^1. "How local is local?"-A review of practical atid coticeptiial issues in the geneLics oi restoration. Rest. Ecoi, 13:432-440. MONTALVO, A. M., S. L WIUJAMS, K. J. R I Œ , S. L. BÜCHMVJN, C. CORY. S. N. HANDEL. C. P. NABIIAN, R. pRiMAfïi APJD R. H. RoBiCHAnx. 1997. Restoration biology: a population biology perspective. liest. Ecoi, 5:277-290. R, S. .\M) S. Y. Snt\tis.s. I9!)H. Ceneric slnicture and local adaptütioii in natural insecl populations: Effects of ecolop-, life histoiy, and behavior. Chapman and I lall. Inc. New York. K. l..J.,T. A. I)AV.\NI.>S. 11. F.\t.ni. 2002. Effect of AVw/v/j/iorfiumendophyte infection on growth and le;if gas exchange of Arizona fescue tinder contrasting water aMiilabílity regimes. Envir. Exper. Bot., 48:257-268. M. A-. R. F. ABRasE /VND N . I.. PorF. 1997. Ecological theor)' and community restoration ecology. liest, EcoL, 5:291-3(K). PAI.M{;RFN. t;. 2000, Sand mitie resi(inuii»ti plan, Grand Mere State Park. Michigan Natural Features Inventory. Lansing, Michigan. Rwii. C:.. V, MiciiAi.AKis ANt) G. CiiiRMn. 1997. The effect of itiiperfect transmission on the frequency of mutnalistjc seed-bt)me endophytes in natural poptilations of grasses. Oikos, 80:18-24. Ri iKiERS. |. A. ANn K. CiAY, 200,"). Fungal endophytes in terrestrial communities and ecosystems, p. -12S-442. /fi. F..J. Dighton, P. Oudeman.s andJ. F.J. Uliite (eds,). The fungaJ community. M. Dekker, New York, New York. \Nn . In pnss. Endophyte symbiosis wilh (all fescue: How strong are the impacts on communities and ecos^'siems? I'ungal BioL Rei'. . J. Hot.Mi. .S. P. ORR .VND K. CLAV. 2007, Forest succession suppressed by an introduced planiftmgal symbiosis. Ecology, 88:18-25. . W. B. M.\rnNCit.YANn J. M. Ki isi.uw. 2005. Mutualistic fimgiis protnotes plant invasion into diverse communities. Oecologia, 144:46^-471. SACHS, I. L. AND E. L, SIMMS. 200ti. Paihwavs to tntittialism breakdown. Trends EcoL Evoi, 21:585-592, S.'.iKKtiNtN, K.. S. H. FAtvni, M. HFLWDKR AW T . J. SII.I.IVAN. 1998. Ftmgal endophytes: A continuum of interactions with hosi planls. .\nii. lirr. EcoL SyU.. 29:^lit-343. . P- LxHToNEN. M. HtiANOFR, J. KnRicHj.vA AM) S. H. FAKTU. 2006. Model S7.stems in ecology: (iis.secting ihe endopliyte-grass literaitire. Trends Plant .Sä., 11:428-433. SCHARDL, C- L., A. LEL¡CHTM.\NN AND M. J. SPIERINI;. 2004. Symbioses of grasses with seedbome fungal endophytes. Ann. Rrv. Plant. BioL, 55:315-340. Soioi.TKNS, B. G.. J. REZNIK AND J . HOUJVND. 2005. Factors affecting the distribution of the threau-ned Lake Huron loctist (Orthopiera: Aciididae). /. Orthoptera Res., 14:45-52. Suit. CkjNSKKVATtoN .StRMtjL. 1977. "Capc" Anieriran ßeachgrass: Clon.servauon plani for Mid-Atlantic s:uid dunes. Report ()-:22l-IH7. I'.S. Department of Agriculttire. SmiîL. R. .\Nn L. T[ioM,\.s. 2001. Power AnalysLs and Experimenul Design, p. 14-36. /n.S. Scheiner and J, Gure\ilch (eds.). Design and Analysis of Ecological ElxperimenLs. Oxford University Press, Oxfonl. WiirtF.. J. F., P. M. HALISKY. S. C . SLUJ. G. MORGANIONF^S ,VND G. R. FUNK. 1992. Endopliyte-hosi ;is.sociati(ms in grasses. 16, Patterns of endophyte distribution in species of the tribe Agrostideae. 4m. /. Bol., 79:472-477. . 1987. Widespread disu-ibution of endophytes in the Poaceae. Plant Disease, 71:340-342. , T, P,. J. M. CiL^SK .\.si) R. T. HI™)I.^:S-IUN. 2001. Communiiy succfs.sion and a.s.semblj': comparing. contrasting, an<l combining pitradigtns in the context of ecológica! resioration. EcoL Restor., 19:5-18. , D. .\. PPTKRSKN ANtiJ. ], GiAR\. 2005. The ecology of restoration: historical links, emerging issues and tmexplort-d realms. EcoL Ij-tl., 8:6fi2-íi73. 1 i DECEMBER 2008 Aía^:i:pTED 10 FKÍIRU,VR'Í 2009 Copyright of American Midland Naturalist is the property of University of Notre Dame / Review of Politics and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.