Study Guides, Notes, & Quizzes

A fungal endosymbiont affects host plant recruitment through seed- and litter-mediated mechanisms: Leaf endophytes modify host recruitment

Description
1.  Many grass species are associated with maternally transmitted fungal endophytes. Increasing evidence shows that endophytes enhance host plant success under varied conditions, yet studies have rarely considered alternative mechanisms whereby these
Published
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Transcript
  Afungalendosymbiontaffectshostplantrecruitmentthroughseed-andlitter-mediatedmechanisms MarinaOmacini* ,1 ,EnriqueJ.Chaneton 1 ,LowellBush 2 andClaudioM.Ghersa 1 1 IFEVA-CONICET and Facultad de Agronomı´ a, Universidad de Buenos Aires, Av. San Martı´ n 4453, C1417DSE Buenos Aires, Argentina; and   2 Department of Plant and Soil Sciences, University of Kentucky, Lexington, Kentucky 40546 0091,USA Summary 1.  Many grass species are associated with maternally transmitted fungal endophytes. Increasingevidence shows that endophytes enhance host plant success under varied conditions, yet studieshave rarely considered alternative mechanisms whereby these mutualistic symbionts may affectregeneration from seed. 2.  We performed a microcosm experiment to evaluate whether infection with  Neotyphodiumoccultans  affects recruitment in the annual grass  Lolium multiflorum  either directly, by infectingthe seeds, or indirectly, by altering the suitability of recruitment microsites through the littershed by host plants. Endophyte effects on establishment were tested for different litter depthsand watering regimes under natural herbivory by leaf-cutting ants. 3.  Seed infection increased seedling emergence through the litter as well as final recruitment,irrespective of microsite conditions. However, litter produced by infected plants delayed emer-gence and decreased density of both infected and non-infected grass populations. 4.  Individual plant biomass did not change with seed infection but was increased under deep lit-ter from endophyte-infected plants. Although seed infection did not protect establishing plantsfrom leaf-cutting ants, herbivory was reduced in the presence of deep litter shed by infectedplants. 5.  We conclude that fungal endophytes may affect host plant recruitment across subsequentgenerations not only by infecting the seeds but also through the host’s dead remains. While theformer effect entailed an advantage to infected plants, litter-mediated effects did not discriminateby infection status, and generally promoted the establishment of fewer and larger plants. Thushidden foliar symbionts may play an underappreciated role in maintaining host species domi-nance through the litter produced by prior patch occupants. Key-words:  after-life effects, endophyte, herbivory, litter,  Lolium multiflorum,  seedling emer-gence, symbiotic interactions Introduction Plant-inhabiting micro-organisms play important, but oftenoverlooked, roles in terrestrial communities (Clay &Schardl 2002; van der Heijden 2004; Omacini  et al.  2005).There has recently been a renewed interest for determiningthe ecological impacts of microbial symbionts, includingfungal endophytes that live concealed within the host plantwithout causing apparent symptoms (Clay & Schardl 2002;Omacini  et al.  2005; Saikkonen  et al.  2006; Rudgers & Clay2008). The symbiosis between cool-season grasses andendophytic fungi of the genus  Neotyphodium  (Ascomycetes:Clavicipitaceae) is widespread in both natural and agricul-tural ecosystems (Roberts  et al.  2005). The fungus growssystemically in above-ground tissues and is transmittedexclusively through the host seeds. Asexual endophytes andtheir host grasses usually establish mutualistic relations;infection may increase plant fitness, contributing to thelong-term maintenance of the symbiosis (Clay & Schardl2002; Gundel  et al.  2008). Nevertheless, uninfected plantsdo persist in natural populations. Infected plants normallyproduce non-infected as well as infected seeds (i.e. verticaltransmission is imperfect). In addition, environmental con-ditions can modify the outcome of the symbiosis, dilutingpotential advantages of harbouring endophytic fungi (Faeth2002; Saikkonen  et al.  2006; Krauss  et al.  2007; Rudgers &Swafford 2009). *Correspondence author. E-mail: omacini@agro.uba.ar   2009 The Authors. Journal compilation    2009 British Ecological Society Functional Ecology  2009,  23 ,1148–1156 doi: 10.1111/j.1365-2435.2009.01582.x  Endophyte-inducedbenefitshavebeengenerallyassociatedwith increased host tolerance to multiple biotic and abioticstresses (Malinowski & Belesky 2000; Clay & Schardl 2002).The symbiosis produces different types of alkaloids, whichcan protect theplant against vertebrateorinvertebrateherbi-vores (Bush  et al.  1997; Wilkinson  et al.  2000). Grassesinfected with fungal endophytes may also become strongercompetitors than non-infected conspecifics and co-occurringplant species in the absence of herbivores (Malinowski  et al. 1999; Clay  et al.  2005; Omacini  et al.  2006). Many studieshave focused on how endophytes modulate the host’s abilityto grow in different environments. Evidence shows that themagnitude and direction of plant responses to endophyteinfection vary depending on resource availability (Malinow-ski et al. 1998;Cheplick2007;Marks&Clay2007;Kannadan & Rudgers 2008). Much less attention has been given to pat-terns of seedling establishment of endophyte-infected andnon-infectedconspecificsunderdifferentmicrositeconditions(Faeth & Hamilton 2006). Differential responses to environ-mental factors during plant regeneration may help to under-stand the persistence of infected and non-infected plantswithin a local population. Endophyte-mediated effects onrecruitment would be most critical for maintaining infectionfrequenciesinannualgrassspecies(Gundel et al. 2008).Endophyte infection may affect host seedling recruitmentin the next generation through direct and indirect mecha-nisms. Direct mechanisms comprise the effects of endophytepresence on seed physiology and seedling performance (Clay1987; Faeth & Hamilton 2006; Gundel  et al.  2006). Existingresults on seed behaviour and establishment of endophyte-infected grasses are surprisingly scarce and variable. Endo-phyte-free and infected seeds may show no differences inresponse to varying soil water potentials or temperatures(Neil  et al.  2003; Faeth  et al.  2004). Yet, some studies foundthat seeds from infected plants have higher germination ratesthan seeds from uninfected conspecifics (Clay 1987; Novas et al. 2003).Incontrast,othersreportedreducedgerminationand seedling success for infected seeds (Hamilton & Faeth2005),althoughsuchnegativeendophyteeffectsmaybeover-ridden by enhanced seedling survival in later life stages (VilaAiub  et al.  2005). Therefore, it is far from clear whetherhereditaryendophyticfungigenerallyincreaseordecreasetheperformanceofhostgrassspeciesattheseed   ⁄   seedlingstages.Fungalendophytesmayalsoaffectregenerationinanindi-rect way, by altering microsite conditions for recruitmentthrough the host’s dead remains. While the role of litter inseedling establishment has been studied in many systems(Xiong& Nilsson1999;Olson & Wallander 2002; Hovstad&Ohlson 2008), its ecological impacts have not been related tothe presence of foliar endophytes in the donor plant. Littermay alter physical (light, temperature, humidity) and chemi-cal conditions for seeds, modifying both the timing of germi-nation and the rates of seedling emergence (Facelli & Pickett1991; Olson & Wallander 2002; Hovstad & Ohlson 2008). In addition, microenvironmental changes generated by litterdeposition may indirectly suppress or favour certain plantspeciesbyalteringtheoutcomeofcompetition,orthesuscep-tibility of seeds and seedlings to various pathogens or herbi-vores (Facelli 1994; Moles & Westoby 2004; Finkes  et al. 2006).Inarecentwork,Antunes et al. (2008)showedthatlit-terfromendophyte-infectedplantsreducedmycorrhizalcolo-nizationofanon-endophyticspecies,presumablybyleachingof endophyte-derived allelochemicals. Hence, there is amplepotential for endophytes to drive transgenerational, litter-mediated effects involving changes in litter quality (QL),microhabitatconditionsandconsumerpressure.In this study, we examine how endophyte infectionaffects plant establishment through seed-mediated (direct)and litter-mediated (indirect) mechanisms. We tested thehypothesis that endophyte presence in the seed enhanceshost recruitment, especially under stressful microsite condi-tions created by deep litter layers and low soil moisturelevels. Moreover, we hypothesized that the litter shed bythe previous generation of infected plants exerts a negativeeffect on current seedling recruitment. We expected litter-mediated effects of prior patch occupants to become moreintense with higher litter quantities and be most negativefor endophyte-free plants. To examine these hypotheses,we conducted a factorial, microcosm   ⁄   glasshouse experi-ment, in which seeds of the annual grass  Lolium multiflo-rum  with contrasting levels of endophyte infection weresown into different microsites created by the amount andsrcin of litter (whether from infected or uninfected plants)and watering regime (WA). We allowed for an additional(biotic) stress by exposing the microcosms to natural her-bivory by leaf-cutting ants. In this way, we were able toexamine endophyte effects on seedling establishment andherbivory rates over a wide range of microhabitats. Weexpected that benefits conferred by endophytes would bepartly associated with changes in host plant chemistryderived from increased alkaloid contents in the shoots(Bush  et al.  1997). Materialsand methods EXPERIMENTAL DESIGN Lolium multiflorum  Lam. (Italian ryegrass) is a cool-season annualspecies srcinary from the Mediterranean zone, which has becomewidely naturalized throughout the world (Beddows 1973; Roberts et al.  2005). Seeds of ryegrass populations naturally infected withthe endophyte  Neotyphodium occultans  (SI+) were collected fromold fields in the Inland Pampa, Argentina, where the host speciesis a major component of plant communities undergoing post-agri-cultural succession (Omacini  et al.  1995; Fig. 1). A subset of theseeds was treated with the fungicide triadimenol (5 mg activeingredient per g seed) to obtain endophyte-free seeds (SI ) ). Plantmonocultures from both seed types were separately grown in out-door plots for two consecutive annual cycles (April–December).After two generations, we collected seeds and harvested all theabove-ground dead material (litter) produced by endophyte-infected and non-infected monocultures. Infection levels were 98%and 9% for SI+ and SI )  seeds respectively (after microscopicexamination of 50 seeds for each type using aniline blue stain).The litter was harvested at the end of the summer (March),   2009 The Authors. Journal compilation    2009 British Ecological Society,  Functional Ecology ,  23 , 1148–1156 Leaf endophytes modify host recruitment  1149  air-dried for 21 days and stored until the start of the experiment(  40 days later).On 31 May 2002, we established a glasshouse experiment com-prising 30 plots (hereafter, ‘microcosms’) arranged in five blocks(Fig. 2). The microcosms were made of wooden boxes(50 cm  ·  50 cm, 15 cm deep) filled with soil from a native grasslanddominated by a mix of tussock grasses ( L. multiflorum  was not pres-ent) (for details, see Omacini  et al.  2004). Coarse plant debris weremanually removed but the srcinal soil fauna was left intact. Eachmicrocosm was randomly assigned to one of three levels of litterquantity (QT) and to one of two watering treatments (WA). In eachmicrocosm we marked two 16 cm  ·  34 cm subplots, which wererandomly designated to receive litter from either endophyte-infected(QL+) or –non-infected (QL ) )  L. multiflorum  plants (QL treat-ment). Litter was added to each microcosm at one of three quantities(QT): 125, 250 or 500 g m ) 2 , representing the range of   L. multiflo-rum  litter found in pampean old fields after 2–10 years of succession(Chaneton  et al.  2001). These litter quantities reduced photosynthet-ically active radiation at the soil level to 25%, 8% and 3% of ambient light respectively.The two moisture treatments were applied by regularly wateringthe microcosms through the litter layer to field capacity (  3000 mL)and allowing themto drain freely until the volumetric soil water con-tent reached 70% (WA high: watered once a week) or 40% of fieldcapacity(WAlow:wateredeverythirdweek).Soilmoisturewasmea-sured to 6 cm depth every third day using a ThetaProbe sensor(Delta-T Devices, Cambridge, UK). During the experiment, soilwater content was significantly affected by the WA (Fig. 3), but noQT or quality effects were detected. In this experiment, watering andQThadnosignificanteffectsonlittermassloss(14–20%,seeOmacini et al. 2004).On 16 June, 100 SI )  and 100 SI+ seeds were separately sown intwo 10 cm  ·  10 cm microsites delimited within each subplot of themicrocosms (Fig. 2). Seeds were put in contact with the soilbeneath the litter layer. As a result, the full experiment comprisedfour main factors arranged in a split-plot blocked design, with litterQT and WA treatments crossed at the main plot level, and litterQL and SI treatments crossed at the subplot level. Prior to theexperiment, the germination potential of SI+ and SI )  seeds wastested by incubating 100 seeds ( n  = 4) at 20–30   C, with 9 h of light (ISTA 1996). No significant difference in germination wasdetected between SI+ (97 %) and SI )  (98%) seed batches( t  = 0 Æ 67,  P  = 0 Æ 52, d.f. = 6). PLANT MEASUREMENTS Every 2–3 days, the number of   L. multiflorum  plants emergingthroughthelitterwascountedineachmicrocosm.Toavoidremovingthe litter, we recorded dead seedlings only when they passed throughthe litterlayer;seedlings thatgerminated and diedunderneaththe lit-ter were not counted. Probit analysis was used to model seedlingemergencedynamicsforeachsownmicrosite(Finney1971).Thispro-cedure allows one to calculate the slope and  x -intercept of the emer-gence curve; the slope equals the rate of emergence while the x -interceptmeasuresthetimeelapsedtothestartofemergence.Basedon these parameters, we calculated the time to 50% emergence ( E  50 ).On 20 August (12 weeks after sowing), the shoots of all established WA-highWA-low125 250 500QT:QL+QL–SI+SI–SI–SI+ Fig. 2.  Schematic of the experimental design showing all the treat-mentsforoneblock.Eachblockincludedsixmicrocosms(mainplots)receivingdifferentwateringregimes(WA,-high:wateredonceaweek,or -low: every third week) and litter quantities (QT, 125, 250 or500 g m ) 2 ).Eachmicrocosmwassplitintotwosubplotscoveredwithlitter produced by endophyte-infected (QL+) or non-infected (QL ) )plants. Each subplot was sown with endophyte-infected (SI+) andnon-infected (SI ) )  Lolium multiflorum  seeds in two separate micro-sites(100 cm 2 ).Thefullexperimentcomprisedfivecompleteblocks. Fig. 1.  Lolium multiflorum  seedlings emerging through the litterdeposited by previous generations of this annual grass in pampeanoldfields(PhotographbyM.Omacini). 0 25 50 75 010203040 125250500125250500 WA-high WA-low Time from seeding (day)    W  a   t  e  r  c  o  n   t  e  n   t   (   %   ) Fig. 3.  Soilvolumetricwatercontentintheexperimentalmicrocosmsunder different watering (WA) and litter quantity (QT) treatments.WA-high: watered once a week (solid lines), WA-low: watered everythirdweek(dottedlines).Litterwasaddedat125,250and500 g m ) 2 .Eachpointshowsthemeanof10values.   2009 The Authors. Journal compilation    2009 British Ecological Society,  Functional Ecology ,  23 , 1148–1156 1150  M. Omacini   et al.  plants were individually harvested and oven-dried at 80   C for 48 htodeterminefinalshootdrymassperplant.Leaf-cutting ants ( Acromirmex  sp., Formicidae: Attini) wereallowed to colonize the experimental microcosms to determinewhether endophyte infection could modify plant–herbivore interac-tions through its influence on host plant attributes (direct effect)and   ⁄   or microhabitat conditions associated with QL (indirect effect).Ant damage on seedlings became apparent early in the experimentand occurredinall the microcosms. InAugust, just beforeendingtheexperiment,weassessedthefrequencyofherbivoryineachsubplotbycounting the number of damaged seedlings (i.e. plants having at leastoneleafcutcross-sectionally). CHEMICAL ANALYSES Nitrogen (N) and pyrrolizidine alkaloid (lolines) tissue contents wereanalysed at the end of the experiment. Ten plants were randomlytakenfromeachsubplot;however,aswedidnothavetheresourcestoinclude all the treatments, subplots with intermediate litter quantitieswere excluded. Nitrogen content ( n  = 5, total 80) was determinedusing a flow injection autoanalyser (Alpkem Corporation, Wilson-ville, OR). Due to logistic constraints, loline alkaloid concentrationswere only evaluated for plants emerging through the QL )  treatment( n  = 5,total 40).For eachsubplot, 10plants were harvested, lyophi-lized and ground before extraction and separation of loline alkaloidsbycapillarygaschromatography, usingamodificationofthemethodproposed by Yates  et al.  (1990). For each sample, 100 mg of tissuewas extracted with 1 mL of CH 2 Cl 2  and 50  l L of a 40% MeOH and5% NH 4 OH solution. Extracts were assayed to determine the pres-ence of the most frequently detected pyrrolizidine alkaloids in endo-phyte-infected grasses, including  N  -formylloline,  N  -acetylloline, N  -acetyl norloline,  N  -norloline and loline (TePaske  et al.  1993; Bush et al. 1997). STATISTICAL ANALYSIS Final plant density, number of days to first emergence ( x -intercept), E  50 ,andemergencerate(slope)werejointlyanalysedusingasplit-plotmultivariate analysis of variance ( MANOVA ) model with blocks. WAand QT entered the model as main-plot factors, while QL and seedinfection (SI) status were included as subplot factors. When  MANOVA showed significant results, we used univariate  ANOVA s to determinewhich of the response variables was most affected by the treatments(Scheiner 2001). In these analyses, four-way treatment interactionswere pooled into the residual error. Cumulative mortality, final seed-ling biomass and frequency of plants damaged by ants were analysedusing univariate split-plot  ANOVA  with blocks. Alkaloid concentra-tions were analysed using  ANOVA  with blocks including WA and QTas main effects (QL was not considered in these analyses). SI statuswas not included because alkaloids were not detected in seedlingsemerged from SI )  subplots. The Levene test was used to check forvarianceheterogeneity(test P  < 0 Æ 05);accordingly,finalplantdensi-ties were square root-transformed before analysis. For clarity, meanvaluesarepresentedinthesrcinalscale. Results SEEDLING EMERGENCE AND GROWTH Between 53% and 79% of sown  L. multiflorum  seeds gener-ated seedlings that emerged through the litter and becameestablished in the different treatments (Fig. 4). Of the 8101seedlings counted across all experimental microcosms, only229 plants passed the litter layer but died thereafter (<3%).No treatment effects on seedling mortality were detected inthis experiment (split-plot  ANOVA , all effects  P  > 0 Æ 10).Hence, final plant recruitment was largely determined byearlyemergencepatterns.Seed infection, QL and WA all significantly affected seed-ling recruitment dynamics ( MANOVA , Table 1). QT did notinfluence seedling emergence, and there were no significanttreatmentinteractions(all P  > 0 Æ 10,seeTable 1).Univariate ANOVA  showed that SI increased final recruitment by 4–15%across microsite treatments ( F  4,74  = 14 Æ 08,  P  = 0 Æ 0004;Fig. 4), but endophyte infection did not alter the timing andrate of seedling emergence ( x -intercept:  F  4,74  = 0 Æ 02, WA-high0 25 50 750255075100 WA-low0 25 50 750255075100SI–SI+0 25 50 750255075100    S  e  e   d   l   i  n  g  p  e  r  m   i  c  r  o  s   i   t  e   (   #   ) 0 25 50 750255075100QL+QL–Time from seeding (day) Fig. 4.  Mean cumulative number of   Loliummultiflorum  seedlings emerging throughthe litter under different watering regimes(WA-high: watered once a week, WA-low:watered every third week). Seedling emer-gence differed between endophyte-infected(SI+) and non-infected (SI ) ) seeds irrespec-tive of watering and litter treatment (upperpanels). Emergence patterns also differedbetween microcosms covered with litter frominfected (QL+) and non-infected (QL ) )plants (lower panels). In each panel, datapoints represent the mean of 30 values, afterpoolingovernonsignificanttreatments.   2009 The Authors. Journal compilation    2009 British Ecological Society,  Functional Ecology ,  23 , 1148–1156 Leaf endophytes modify host recruitment  1151  P  = 0 Æ 9; slope:  F  4,74  = 0 Æ 27,  P  = 0 Æ 61;  E  50 :  F  4,74  = 0 Æ 59, P  = 0 Æ 44; Table 2). This endophyte effect was independentof litter and watering treatments (all interactions with SI, P  > 0 Æ 10).Litterqualitysignificantlyaffectedboththetimingofemer-gence and final seedling density (Table 2, Fig. 4). The pres-ence of litter from endophyte-infected plants (QL+) delayedthe onset of emergence ( x -intercept:  F  4,74  = 4 Æ 17, P  = 0 Æ 044) and the time elapsed to reach  E  50  ( F  4,74  = 5 Æ 24, P  = 0 Æ 025), and also reduced the number of establishedplants ( F  4,74  = 4 Æ 22,  P  = 0 Æ 043). On average, we recorded5% less seedlings in QL+ than in QL )  microcosms(Table 2). Lastly, WA significantly influenced all seedlingemergence parameters (Table 2) but did not modify theobserved effects of SI or litter QL (interactions  P  > 0 Æ 10).Seedlings emergedmorerapidly inWA-highthanin WA-lowmicrocosms ( x -intercept:  F  1,20  = 60 Æ 08,  P  = 0 Æ 0001;  E  50 : F  1,20  = 90 Æ 91,  P  = 0 Æ 0001) and also attained higher finaldensitiesinWA-highmicrocosms( F  1,20  = 15 Æ 27, P  = 0 Æ 001;Fig. 4).  E  50  was reached about 12 days earlier in WA-highthan in WA-low, while final recruitment was increased by26% in WA-high plots (see Table 2). Noteworthy, the lowestnumbersofestablishedplantsoccurredinsubplotssownwithendophyte-free seeds (SI ) ) covered with the largest amount(500 g m ) 2 ) of QL+ litter, although this three-way interac-tionwasmarginallynonsignificant(seeTable 1).Seed infection and WA did not affect final shoot mass perplant ( ANOVA , SI:  F  1,72  = 0 Æ 13,  P  = 0 Æ 71, WA:  F  1,72  = 1 Æ 71, P  = 0 Æ 21). However, seedling mass varied significantlydepending on QT and quality (QT  ·  QL:  F  2,72  = 9 Æ 51, P  = 0 Æ 0002). Seedlings emerging through the highest QTattained greater final mass in subplots covered with QL+thaninthosewithQL ) litter(Fig. 5). HERBIVORY BY LEAF-CUTTING ANTS Litter quality significantly reduced the frequency of plantsdamagedbyleaf-cuttingantsatbothlowandhighlitterquan-tities (QL:  F  1,72  = 6 Æ 40,  P  = 0 Æ 01, QL  ·  QT:  F  2,72  = 4 Æ 65, P  = 0 Æ 01).Onaverage,11%(SE = 4 Æ 6)oftheplantsemerg-ing through QL+ litter were damaged, while 30%(SE = 10 Æ 2) of the plants were damaged in QL )  litter(Fig. 6). SI and WA did not affect herbivory rates( P  < 0 Æ 60). To explore whether the severity of ant herbivoryaltered biomass patterns at the subplot level, we performedanalysisofcovarianceonfinalseedlingmassincludingthefre-quencyofdamagedplantsasthecovariate.Thisanalysisindi-cated that differences associated with litter QT and QLremained significant after controlling for the effect of herbiv-ory (split-plot ANCOVA, QT  ·  QL:  F  2,71  = 6 Æ 75, P  = 0 Æ 002;covariate: F  1,71  = 7 Æ 97, P  = 0 Æ 006). PLANT TISSUE CHEMISTRY Endophyte infection did not affect plant N content ( ANOVA , P  > 0 Æ 10 for all effects). Nitrogen concentrations rangedfrom1 Æ 18%to3 Æ 55%forSI+plants,andfrom1 Æ 04to3 Æ 98%for SI )  plants. Regardless of endophyte infection, seedlings Table 1.  Results of multivariate analysis of variance ( MANOVA ) fortheeffectsofwateringregime(WA),litterquantity(QT),litterquality(QL) and seed infection status (SI) on four parameters describingdynamics of seedling emergence of   Lolium multiflorum  inexperimentalmicrocosms.Thesplit-plotdesignincludedWAandQTas main plot effects, and QL and SI as subplot effects. Responsevariables were derived from probit analysis and included final plantrecruitment (no. plants), time elapsed to the onset of emergence( x -intercept, days), time to half total emergence ( E  50 , days) andemergencerate(slope,day ) 1 )Effect d.f. Wilk’s lambda  P -levelBlock 16, 52 0 Æ 385 0 Æ 2956Watering (WA) 4, 17 0 Æ 101  0 Æ 0001 Litter quantity (QT) 8, 34 0 Æ 602 0 Æ 3119QT  ·  WA 8, 34 0 Æ 632 0 Æ 3908Litter quality (QL) 4, 71 0 Æ 841  0 Æ 0142 Seed infection (SI) 4, 71 0 Æ 816  0 Æ 0055 QL  ·  SI 4, 71 0 Æ 981 0 Æ 8525WA  ·  QL 4, 71 0 Æ 898 0 Æ 1028WA  ·  SI 4, 71 0 Æ 944 0 Æ 3851QT  ·  QL 8, 142 0 Æ 933 0 Æ 7534QT  ·  SI 8, 142 0 Æ 906 0 Æ 5154WA  ·  QT  ·  SI 4, 71 0 Æ 961 0 Æ 5762WA  ·  QT  ·  QL 8, 142 0 Æ 864 0 Æ 2232WA  ·  QT  ·  SI 8, 142 0 Æ 967 0 Æ 9655QT  ·  QL  ·  SI 8, 142 0 Æ 834 0 Æ 1079Significant effects ( P  < 0 Æ 05) are shown in bold. Table 2.  Effects of main experimental factors on four parameters describing the dynamics of seedling emergence for  Lolium multiflorum  inexperimentalmicrocosmsResponse variableWatering Litter quantity Litter quality Seed infectionWA )  WA+ QT I  QT II  QT III  QL )  QL+ SI )  SI+ R  (no. plants)  58 Æ 1  (1 Æ 7)  73 Æ 1  (1 Æ 7) 66 Æ 0 (4 Æ 0) 68 Æ 8 (3 Æ 3) 62 Æ 0 (3 Æ 2)  67 Æ 3  (1 Æ 9)  63 Æ 9  (2 Æ 1)  62 Æ 4  (2 Æ 0)  68 Æ 8  (1 Æ 9) x -Int (days)  3 Æ 2  (0 Æ 02)  2 Æ 6  (0 Æ 05) 3 Æ 0 (0 Æ 08) 2 Æ 9 (0 Æ 10) 2 Æ 9 (0 Æ 11)  2 Æ 9  (0 Æ 06)  3 Æ 0  (0 Æ 05) 2 Æ 9 (0 Æ 06) 2 Æ 9 (0 Æ 05)Slope (day ) 1 )  3 Æ 1  (0 Æ 1)  1 Æ 9  (0 Æ 1) 2 Æ 7 (0 Æ 2) 2 Æ 5 (0 Æ 3) 2 Æ 3 (0 Æ 2) 2 Æ 6 (0 Æ 1) 2 Æ 4 (0 Æ 1) 2 Æ 5 (0 Æ 1) 2 Æ 5 (0 Æ 1) E  50  (days)  26 Æ 2  (0 Æ 6)  14 Æ 4  (0 Æ 5) 21 Æ 9 (1 Æ 7) 19 Æ 6 (1 Æ 8) 19 Æ 4 (1 Æ 6)  19 Æ 7  (1 Æ 0)  20 Æ 9  (1 Æ 0) 20 Æ 5 (1 Æ 0) 20 Æ 1 (0 Æ 9)Data show means (with SE in brackets) for each level of a main factor, after pooling over the rest of the treatments. Response variables werederived from probit analysis and included final recruitment ( R ), the time elapsed to the onset of emergence ( x -intercept) and to half totalemergence ( E  50 ), and emergence rate (slope). For each response variable, values shown in bold under a given experimental factor weresignificantly different ( P  < 0 Æ 05, univariate  ANOVA ).   2009 The Authors. Journal compilation    2009 British Ecological Society,  Functional Ecology ,  23 , 1148–1156 1152  M. Omacini   et al.
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x