AmericanJournalofBotany87(11):1679–1692. 2000. DIFFERENTIAL EFFECTS OF FOUR ABIOTIC FACTORS ON THE GERMINATION OF SALT MARSH ANNUALS1 GREGORY B. NOE2 AND JOY B. ZEDLER3 PacificEstuarineResearchLaboratory,SanDiegoStateUniversity,SanDiego,California92182-1870USA Interspecificdifferencesinresponsivenesstotemperature,photoperiod,soilsalinity,andsoilmoistureconfirmthehypothesisthat abioticfactorsdifferentiallyaffectthegerminationofsaltmarshplants.Ingrowthchamberexperiments,fourofeightannualspecies respondedtosmalldifferencesintemperatureorphotoperiod.Increasingsoilsalinitydecreasedthefinalproportionofseedsgerminating andslowedgerminationforeachofthesevenspeciestested.Highersoilmoistureincreasedtheproportiongerminatingoffivespecies andgerminationspeedofallsevenspecies.Salinityandmoistureinteractedtoaffecttheproportiongerminatingoffivespeciesand germinationspeedofallsevenspecies.Althoughtheabioticfactorwiththelargesteffectongerminationvariedamongspecies,more species responded to, and the magnitudes of the responses were larger for, soil salinity than for the other abiotic factors. These germinationtestspartiallyexplainedinterspecificdifferencesinthetimingofgerminationinthefield.PatternsofHutchinsiaprocum- bens,Lythrumhyssopifolium,Parapholisincurva,andpossiblyLastheniaglabratassp.coulterigerminationinresponsetoanonsea- sonal rainfall could be explained by their response to salinity, temperature, or photoperiod. Fine-scale differences in the timing of establishmentwithinthetypicalgerminationwindowandspatialdistributionsalongsalinityandmoisturegradientswerenotexplained. Keywords: germination;moisture;photoperiod;salinity;saltmarshannualplants;temperature. A common goal of ecologists is to predict patternsof plant literature on the influence of temperature or photoperiod on establishment in the field using germination experiments. thegerminationoffreshwaterwetlandspecies(Thompsonand However, germination studies are frequently performed in Grime, 1983; Galinato and van der Valk, 1986; Leck, 1989, ways that make it difficult to extrapolate results to the field 1996).Finally,thechoiceofmeaningfultreatmentsisalsocrit- (Baskin and Baskin, 1998). This may be due to testing non- ical; factors should be tested at levels that can be related to meaningful factors or testing too few factors. There is an un- differences in the field at the time of germination. There is a derstanding in ecology that factors interact and that testing paucity of studies that explicitly test whether multiple inter- multiple factors may be necessary to predict the dynamics of acting abioticfactors, testedatlevelsfoundinthefieldduring plant communities (Chapin et al., 1987; Mooney, Winner,and germination, are necessary to explain observedpatternsofes- Pell, 1991; Bazzaz and Wayne, 1994; Bazzaz, 1996). For ex- tablishment in plant communities. ample,thecombinationofresponsestomultiplefactorsisnec- ThesaltmarshesofsouthernCaliforniaofferanopportunity essary to predict the establishmentandcompositionofthean- to test our ability to predict field seedlingestablishmentbased nualcommunityinoldfields(Bazzaz,1996).Inwetlands,sev- on responses to abiotic factors. The upper intertidal marsh of eral authors suggest that testing multiple abiotic factors may southern California includes up to 20 annual plant species be necessary to predict plant establishment (Galinato and van (Noe, 1999). Germination is highly punctuated due to the re- der Valk, 1986; Weiher and Keddy, 1995; Leck, 1996). How- gion’s mediterranean-type seasonality; annual plants typically ever, salt marsh studies have emphasized the effect of one germinate during the wet, mild winter season and senesce by factor,salinity,ongermination(Kingsburyetal.,1976;Ungar, thedry,hotsummerseason.Inaseparatefieldstudy,wemon- 1978,1996;Woodell,1985;ZedlerandBeare,1986;Callaway itored the density of seedlings (including those with just cot- etal.,1990;ShumwayandBertness,1992;KeifferandUngar, yledons; Noe, 1999). Counts were made weekly during the 1997; Kuhn and Zedler, 1997; Baskin and Baskin, 1998; Cal- 1996periodofgerminationandmonthlyin1997.Germination laway and Zedler, 1998) compared to the effectsof soilmois- typicallyoccursbetweenNovemberandMarch,butindividual ture (Baldwin, McKee, and Mendelssohn, 1996; Kuhn and species establish seedlings either early in, late in, or through- Zedler, 1997; Baskin and Baskin, 1998), or temperature or outthis‘‘germinationwindow’’(Table1).Wehaveshownthat photoperiod (Pihl, Grant, and Somers,1978;KhanandUngar, temporal variance in soil salinity, and to a lesser degree soil 1997; Baskin and Baskin, 1998), although there is a larger moisture, explains a significant portion of the timing of ger- mination (all species combined) at three southern California 1Manuscriptreceived17August1999;revisionaccepted29February2000. coastal wetlands (Noe, 1999). However, annual rainfall totals The authors thank the Earth Island Institute for an award to the Pacific and the seasonal timing of rainfall are variable (Noe, 1999), Estuarine Research Laboratory. G.B.N. received grants from the San Diego and most species’ germination does not followthesamestrat- StateUniversityJointDoctoralProgram,scholarshipsfromtheAchievement egy every year (Table 1). Additional variability wasobserved Rewards for College Scientists (ARCS) Foundation, and a fellowship from whenarare,early-seasonrainfallfromHurricaneNorainSep- theSanctuariesandReservesDivision,OfficeofOceanandCoastalResource Management, National Ocean Service, National Oceanographic and Atmo- tember 1997 stimulated the germination of selected species sphericAdministration.VickiTripolitisandMeghanFellowsassistedwiththe (Table 1; Noe, 1999). The seedlings of annual species also experiments. segregate along spatial gradients of surface soil salinity and 2Authorforreprintrequests,currentaddress:SoutheasternEnvironmental moisture (Noe, 1999). Research Center, Florida International University, University Park-OE 148, The variance in the timing of establishment of species is Miami,Florida33199USA. most likely due to temporal differences in soil salinity, soil 3Currentaddress:BotanyDepartmentandArboretum,UniversityofWis- consin-Madison,430LincolnDrive,Madison,Wisconsin53706USA. moisture,photoperiod,ortemperatureanddifferentialrespons- 1679 1680 AMERICAN JOURNAL OF BOTANY [Vol. 87 TABLE 1. The timing of germination of annual species in the upper intertidal marsh of southern California. Timing of germination and relative densityofseedlingsaresummarizedfromNoe(1999).Blanksinthetableindicatethatthespeciesdidnotoccurinthesubsetofwetlandsthat weremonitoredin the1997 season.ThenomenclaturefollowsHickman(1993). Nonseasonal Timingwithin Timingwithin Species germination 1996season 1997season Densityrank AmblyopappuspusillusHook.andArn.(Asteraceae) slight prolonged early 4 Cordylanthusmaritimusssp.maritimusBenth.(Scrophulariaceae) no late late 12 Cotula coronopifoliaL.(Asteraceae) yes early 9 Hutchinsiaprocumbens(L.)Desv.(Brassicaceae) no prolonged early 5 JuncusbufoniusL.(Juncaceae) no late prolonged 3 Lastheniaglabratassp.coulteri(A.Gray)Ornd.(Asteraceae) yes early prolonged 6 LoliummultiflorumLam.(Poaceae) yes early 11 LythrumhyssopifoliumL.(Lythraceae) yes early 8 MesembryanthemumnodiflorumL.(Aizoaceae) slight prolonged early 2 ParapholisincurvaC.E.Hubb.(Poaceae) no prolonged prolonged 1 Polypogon monspeliensis(L.)Desf.(Poaceae) yes early 7 Sonchus oleraceusL.(Asteraceae) slight late early 13 Spergulariamarina(L.)Griseb.(Caryophyllaceae) slight late 10 es of species to these abiotic factors. In the upper intertidal vemberandMarchwerechosentorepresentthebeginningand marsh of southern California, soil moisture and salinity fluc- end of the germination window. We also chosearangeofsoil tuate in response to the amount of daily rainfall during the salinity and moisture levels that is similar to conditions when germination window (Noe, 1999). We expect that soilsalinity germination occurs in the field. willhavethelargesteffectonspeciesbecauselevelsofsalinity We tested ten annual species of the upper intertidal marsh duringtheshortgerminationwindowarestressfulandvariable in southern California that accounted for 88% of allseedlings while differences in temperature and photoperiod are small. inthe1996germinationwindow(Noe,1999).Responseswere Additionally, there is a large literature on the influence of sa- assessed with both the final proportion of seeds germinating linity on the germination of salt marsh plants. Some annual andthespeedofgermination.Thefinalproportiongerminating species in the upper intertidal marsh of southern California (‘‘proportion’’ hereafter) is a useful measure of establishment likely have a proportion of seeds that are dormant and entera and potential community compositionandrelativeabundance. seed bank (G. B. Noe and J. B. Zedler, personalobservation); Speed of germination can be an important determinant of in- however, it is unlikely that seeds are buried in this rarely in- traspecific and interspecific interactions (Harper, 1977;Grace, undated community with low litter production and little soil 1987; Bazzaz, 1996). Because of the potential importance of disturbance. Therefore, the species may not respond to tem- germination speed, we created an index of the speed of ger- perature or light like other species with buried seed banks mination that factors out the proportion of seeds germinating (e.g., Thompson and Grime, 1979, 1983; Grime et al., 1981). and is therefore independent ofviability,ascomparedtoTim- For seeds in mediterranean-type climates,coolertemperatures son’s(cid:83)n(Timson,1965)andotherindexes(BrownandMayer, could signal the onset of the wet season and favorable con- 1988). ditions for seedlings. Experiments determining the effect of salinity on germina- MATERIALS AND METHODS tion are typically conducted at very high moisture levels for thedurationof theexperiment.Suchtestsmaynotberelevant Seed collection, storage, and treatment—Seeds of ten species were col- tofieldconditionsbecauseupperintertidalsaltmarshsoilsare lected after plant senescence in 1997 from plants that were growing within rarely saturated for more than a few days at a time. Salinity thesaltmarsh(Table2).Of20annualplantspeciesthatoccurinthesystem, can have both osmotic and toxic effects on halophyte seeds these ten species produced sufficient seed to permit collection and were se- (Waisel, 1972; Ungar, 1978; Baskin and Baskin, 1998). The lectedtoincludearangeofnativeandexotic,commonandrareannualspe- interaction of soil salinity and moistureongerminationwould cies. Seedswere collected from areas of sparse perennial canopy and along have important implications for predicting the distribution of the edges of salt pannes (dominant perennial plant species in both habitats salt marsh plants. It would be more appropriate to conduct areSalicorniasubterminalisandMonanthochloe¨littoralis),areaswheremost such laboratory experiments at several moisture levels and to seedlingsarefoundintheupperintertidalmarshofsouthernCalifornia.Seeds use a factorial experimental design in order to identify the were collected from three locations in San Diego County, California, USA: interaction. thenorthernarm(OneontaSlough)ofTijuanaRiverNationalEstuarineRe- search Reserve; Sweetwater Marsh National Wildlife Refuge; and Los Pen˜- The goal of this study was to explain differences in the asquitosLagoon.Alltenspecieshavesmallseeds((cid:59)0.2–5mm3). establishmentofspeciesasdescribedinthreecoastalwetlands The seeds were stored dry and under ambient room temperatures (16(cid:56)– in southern California (Table 1; Noe, 1999). To address this 37(cid:56)C) and a light regime that was typical of summer photoperiod. To our goal, we (1) determined the effect of soil salinity, soil mois- knowledge there is no evidence for recalcitrance among these species. We ture, temperature, and photoperiod on germination, (2) com- believethatthestorageconditionsusedinthisstudyaresimilartowhatseeds pared the magnitude of the effects of each abiotic factor on experienceintheupperintertidalmarshofsouthernCaliforniaintheperiod germination, and (3) assessed the interaction of soil salinity afterdispersalinlatespringorearlysummerandpriortogerminationinthe and moisture on germination. We tested ecologically relevant winter.Thelengthofstoragevariedfrom5to20modependingonthedate treatments that could be related to field conditions when the ofseedmaturationandcollectioninthefield(Table2)andthetimingofthe annual species germinate in the upper intertidal marsh of experiments(Table3). southern California. The average climatic conditions of No- Seeds were only used if they had intact seed coats and appeared to have November 2000] NOE AND ZEDLER—ABIOTIC EFFECTS ON GERMINATION OF SALT MARSH ANNUALS 1681 TABLE2. Speciesusedinthephotoperiod-temperatureexperimentandsalinity-moistureexperiment,informationontheircollection,andthestarting date of the salinity-moisture experiment runs. Wetland: LPL(cid:53)Los Pen˜asquitos Lagoon, SW(cid:53)Sweetwater Marsh National Wildlife Refuge, TE(cid:53)TijuanaRiverNationalEstuarineResearchReserve. Photo- period temper- Salinity- Species Date(s)ofseedcollection Seedprocessing Wetland ature moisture Amblyopappuspusillus 6-2-97 none SW 8-22-98 Cordylanthusmaritimusssp.maritimus 8-97 none TE 10-26-98 Cotula coronopifolia 4-3-97 none LPL X Hutchinsiaprocumbens 3-24-97,3-31-97 none SW X 7-14-98 Lastheniaglabratassp.coulteri 3-31-97,4-3-97 none LPL,SW X 10-26-98 Lythrumhyssopifolium 6-3-97 none LPL X Mesembryanthemumnodiflorum 7-21-97 fruitsoaked15 m SW X 9-23-98 Parapholisincurva 5-5-97,5-12-97,5-14-97 seedleftin floret LPL,SW,TE X 9-23-98 Polypogon monspeliensis 5-14-97 seedremovedfromfloret LPL X Spergulariamarina 4-3-97 none LPL X 8-22-98 embryos.TheseedsofthegrassPolypogonmonspeliensiswereremovedfrom blocked by growthchamber shelf.Thenumberof germinatedseedsin each theflorettoensurethepresenceofaseed.Parapholisincurvaseedswereleft petridishwascountedevery3dfor30d.Germinationwasdefinedasroot withinthefloretbecausetheirpresencecouldbedeterminedwithoutremoval radicle or cotyledon emergence. The species in this study germinate with and because they germinate from within the floret in the field (G. B. Noe, differentspeeds,althoughtherankofspecies’speedofgerminationisaffected personalobservation).Mesembryanthemumnodiflorumfruitsweresoakedin bytheexperimentaltreatments(seeResults).Thelocationofthepetridishes deionized water for 15 min to facilitate seed removal. Seeds of the federal withinablockwasrerandomizedevery6d.Thetemperatureandphotoperiod andstate-listedendangeredCordylanthusmaritimusssp.maritimuswerecol- treatmentsoccurredconsecutively(Table3).Todeterminetheeffectofseed lectedpursuanttoUnitedStatesDepartmentoftheInteriorEndangeredSpe- aging,thefirsttreatmentwasrepeatedattheendoftheplannedtemperature ciesActrecoverypermitnumberPRT823806andCaliforniaDepartmentof andphotoperiodcomparisons. FishandGameResearchPermit96–01-RP.Preliminaryattemptstotransplant seedlingsfailed.Therefore,Cordylanthusmaritimusssp.maritimusseedlings Soil salinity and moisture—Seven upper intertidal marsh annual species werenotreturnedtothefield. weretestedfortheirresponsetoconstantmoistureandsalinitylevels(Table 2). Three moisture treatments, high, medium, and low, were fully crossed Temperatureandphotoperiod—Eightupperintertidalmarshannualspecies withfoursalinitytreatments,34,17,8,and0ppt.Themoistureandsalinity (Table 2) were testedfor differencesin germinationbetweenthemeancon- treatmentlevelswerechosentorepresentarangeofconditionsfoundinsouth- stant temperatures, mean diurnal fluctuating temperatures, and photoperiods ernCaliforniahighsaltmarshduringperiodsofgermination(Noe,1999).In ofNovemberandMarch(Table3).Fourcomparisonsweretested:November ordertobreakanytemperature-relateddormancy,seedsofeachspecieswere vs. March constant mean daily temperature, November constant vs. diurnal coldtreatedat5(cid:56)Cfor15dpriortothestartoftheexperiment.Theduration fluctuating mean temperature, March constant vs. diurnal fluctuating mean andtemperatureofthecoldtreatmentwerechosentosimulateconditionsin temperature,and15Novembervs.15Marchphotoperiod.Thethreetemper- coastalSanDiegoduringthewinter;minimumtemperaturerarelyreaches5(cid:56)C aturecomparisonswererunatMarchphotoperiod,andthephotoperiodcom- forlongperiodsoftime.Speciesweretestedtwoatatimeafterthefirstrun parisonwasconductedatNovemberconstantmeantemperature.Theduration withonespecies(Table2). ofdiurnaltemperaturestreatmentscorrespondedwiththelengthoflightand The experimental unit consisted of 25 seeds of a species placed in a mi- dark in the March photoperiod. Temperatures for coastalSanDiegoCounty crocosm,locatedinsideatemperature-andlight-controlledPercival(cid:116)growth were determined from the National Weather Service’s 30-yr average daily chamber, as above. A microcosm consisted of a soil-filled plastic cup that temperature normals for Lindbergh Field (Table 3; National Climatic Data rested on a wood block inside an outer plastic cup (Fig. 1). The inner cup Center, Comparative Climatic Data). Daylengths (sunrise to sunset) on 15 wasfilledwith250mLofmineralsoil(48%sand,41%silt,and11%clay) November and 15 March for San Diego were obtained from the Nautical that had been passed through a 2-mm sieve. To create the three moisture AlmanacOffice(1965). treatments,theoutercuphadaholeatoneofthreedifferentheightstoreg- Theexperimentalunitconsistedof25seedsofaspeciesplacedinapetri ulate the depth of water relative to the soil surface. Water of one of four dish (6.0-cm diameter) inside a temperature- and light-controlled Percival(cid:116) differentsalinitylevelswasaddedtotheoutercuptocreatethedifferentsoil growthchamber.Theseedswereevenlyplacedonaglassfiberfilter(4.7-cm salinitytreatments.Seedswereevenlyplacedonflatareasofthesoilsurface diameter)thatrestedonathinstyrofoamwaferfloatingon9mLofdeionized to avoid differences in microtopography. Species with nonspherical seeds water.Theedgeofthefilterwasincontactwiththewatertokeepthefilter wereplacedwiththeirlongitudinalaxisflatonthesoilsurface.Theoutercup uniformly moist but not waterlogged. The experiment used a randomized ofeachmicrocosmwascoveredwithapetridishlidtolimitevaporationand completeblockdesignwithfourreplicatespertreatment,withthepetridishes maintainconstantsoilsalinityandmoisture.Thesalinityandmoisturetrials TABLE3. Treatmentsinthe temperature,photoperiod,andseedaging tests. Experimental Treatment Temperature((cid:56)C) Photoperiod(h:min) Startingdate order Novemberconstanttemperature/Marchphotoperiod 16.7 11:57 light,12:03dark 12-3-97 1 Novemberdiurnaltemperaturefluctuation/Marchphotoperiod 21.1/12.2 11:57 light,12:03dark 1-20-98 2 Marchconstanttemperature/Marchphotoperiod 15.3 11:57 light,12:03dark 4-14-98 4 Marchdiurnaltemperaturefluctuation/Marchphotoperiod 19.1/11.6 11:57 light,12:03dark 5-22-98 5 Novemberconstanttemperature/Novemberphotoperiod 16.7 10:29 light,13:31dark 2-26-98 3 Novemberconstanttemperature/Marchphotoperiod(seedage test) 16.7 11:57 light,12:03dark 7-8-98 6 1682 AMERICAN JOURNAL OF BOTANY [Vol. 87 drymass(Gardner,1986).Reverseosmosiswater(salinity(cid:53)0ppt)wasadded tothesamedriedsoilsampleuntilthesaturationpointwasreached(Richards, 1954).Thesaturatedsoilsamplewasthenaddedtoa10-mLsyringeloaded withfilterpaperandadropofwaterwasexpressedontoatemperature-com- pensatedsalinityrefractometer(PacificEstuarineResearchLaboratory,1990). Allpastesweremixedbyoneperson(G.B.Noe).Saturatedsoilpasteextracts underestimatefieldsoilsalinityconcentrations,exceptwhensoilsaresaturat- ed. Instead of estimating the salt concentrations of soils, the salinity of sat- urated soil paste extractsisa measureof thesaltmassin soils.Soilsinthe low-moisturetreatmentsweretoodrytomeasurethesalinityoftheinterstitial water. The experiment used a randomized complete block design with four rep- licatespertreatment,withthemicrocosmsblockedbygrowthchambershelf. Thenumberofgerminatedseedsineachmicrocosmwascountedevery3d for 30 d. The location of the microcosms within a block was rerandomized every6d. Statistical analyses—The response variables for all experiments were the proportion of seeds germinating at the end of the experiment (day 30) and thespeedofgermination.Thespeedofgerminationwasexpressedas((cid:83)n)/ t (n t), where n is the cumulative proportion germinating at each sampling f t time,n isthecumulativeproportiongerminatingattheendoftheexperiment, f and t is the number of sampling times. When no germination occurs (n (cid:53) f 0),theindexvalueisdefinedtobezero.Theindexrangesfromzerotoone, increasingasgerminationoccursearlierintheexperiment.Sincecumulative germination was sampled ten times in this experiment, a 0.1-change in the indexcorrespondstoadifferenceinthe timingofgerminationbyonesam- plingtime,3d,fortheaverageseed. Fourtemperatureandphotoperiodcomparisonsweretested:Novembervs. Marchconstantmeantemperature,Novemberconstantvs.diurnalfluctuating meantemperature,Marchconstantvs.diurnalmeanfluctuatingtemperature, Novembervs.Marchphotoperiod.Seedagewasalsoevaluatedasapotential factorinfluencinggerminationbycomparingtheNovemberconstanttemper- ature and March photoperiod treatment at the beginning of the experiment and 7 mo later (Table 3). For each species, the final proportion of seeds germinatingandthespeedofgerminationwereeachanalyzedwithananalysis Fig.1. Thedesignofthe soilmoistureand salinityexperimentalmicro- ofvariance(ANOVA)withrandomizedblockingandphotoperiod/temperature cosms. Holes were placed at 0.5, 3.5, and 7.0 cm from the bottomofa 9.5 treatmentasthemainfactor.Proportiondatawerearcsinesquare-roottrans- cmtall296-mLplasticinnercup.Theinnercupwasfilledto1cmfromthe topofthecupwithmineralsoilandrestedona1.9cmtallwoodblockinside formed to improve normality and homogeneity of variance in the residuals a473-mLoutercup.Theoutercuphadaholeat1,5,or9cmfromtopof (Zar,1996).Significantdifferences(P(cid:44)0.05)foreachofthefivecompari- the cup for high, medium, and low moisture treatments, respectively. The sonsweretestedwithTukey’sHonestlySignificantDifference(HSD)tests. mediummoisturetreatmentisshown;wateristhelightershade. Theeffectsofsoilsalinityandmoistureonthefinalproportiongerminating FigureAbbreviations:Ap,Amblyopappuspusillus;Cm,Cordylanthusmar- and germination speed of each species were each analyzed with a two-way itimusssp.maritimus;Cc,Cotulacoronopifolia;Hp,Hutchinsiaprocumbens; analysis of variance (ANOVA) with randomized blocking and moistureand Lg,Lastheniaglabrataspp.couleri;Lh,Lythrumhyssopifolium;Mn,Mesem- salinitytreatmentsasthemainfactors.Proportiondatawerearcsinesquare- bryanthemum nodiflorum; Pi, Parapholis incurva; Pm, Polypogon monspe- root transformed to improve normality and homogeneity of variance in the liensis;Sm,Spergulariamarina;SW,SweetwaterMarsh;TE,TijuanaEstuary; residuals(Zar,1996).Allsignificant(P(cid:44)0.05)maineffectsweretestedfor LPL,LosPen˜asquitosLagoon. differences between treatmentlevels with Tukey’s HSDtests.However,dif- ferencesamongtreatmentlevelsofindividualfactorsarenotreportedifthere occurredatNovemberphotoperiodanddiurnalmeantemperaturefluctuation wasasignificantinteractionofsoilsalinityandmoisture.Allstatisticalanal- (Table3). yseswereperformedusingSYSTATsoftware(SYSTAT,1992). As suggested by preliminary trials, seawater diluted to 25, 12, and 5 ppt anddeionizedwaterwere addedtotheoutercuptothelevelof thehole to Relative effects of salinity, moisture, temperature, and photoperiod—To create target surface soil salinity of 34, 17, 8, and 0 ppt, respectively. The compare the magnitude of the effects of the different abiotic variables and differentsalinitywaterswereaddedatotalofthreetimesonthefirstday,and seed aging on the proportion and speed of germination, we calculated the once on the second day, to equilibrate water levels in the outer and inner rangeinthemeansofbothresponsevariablesforeachspeciesinresponseto cups.Onthethirdday,thewaterintheoutercupwasreplacedwithdeionized each of the four abiotic factor tests (temperature, photoperiod, salinity, and watertomaintainconstantsoilsalinity.Seedswereaddedtothemicrocosms moisture) as well as the seed aging test. The temperature effect for each onthefourthday. specieswascalculatedasthelargestrangeamongthethreecomparisons(No- Surfacesoilmoistureandsalinityineachrunwerequantifiedinnonrepli- vembervs.Marchconstant,Novemberconstantvs.diurnalfluctuating,March catedseedlessmicrososms.A1.3-cm2diametersoilcorewastakentoadepth constantvs.diurnalfluctuatingtemperatures).Threeoftheeightspeciestested of1cmineachseedlessmicrocosmwhentheseedswereaddedtotheseeded inthetemperature,photoperiod,andagingtrialsinthisstudywerenottested microcosms(day0)andondays3,6,9,15,and30.A1-cmdeepsoilcore fortheirresponsetosoilsalinityandmoisture.Inordertocompareallofthe wastakenonday30ineachoftheseededmicrocosms. abioticfactorsonalleightspecies,theresponseofCotulacoronopifolia,Lyth- Soil moisture was determined gravimetrically. The soil sample was dried rumhyssopifolium,andPolypogonmonspeliensistodifferentsoilsalinityand at60(cid:56)Cfor24h.Soilmoisturewascalculatedaschangeinmassdividedby moisturelevelswasobtainedinaseparategreenhouse-basedmicrocosmex- November 2000] NOE AND ZEDLER—ABIOTIC EFFECTS ON GERMINATION OF SALT MARSH ANNUALS 1683 Fig.2. Species responsestothetemperature,photoperiod,andagingcomparisons.(a)Novembervs. March constanttemperature,(b)Novemberconstant vs.diurnalfluctuatingtemperature,(c)Marchconstantvs.diurnalfluctuatingtemperature,(d)Novembervs.Marchphotoperiod,(e)seedagingcomparison.* (cid:53)significant(P(cid:44)0.05)differencebetweentreatments. perimentthatisreportedinNoe(1999).Theconstantmoisturelevelsinthe Temperature and photoperiod—Few species responded to soil-basedmicrocosmsofthegreenhousestudy(35–45%soilmoisture)were the temperature or photoperiod treatments (Fig. 2a-d). Differ- similar to moisture contents in the soils of this growth chamber study. The ent constant temperatures had a significant effect on the pro- four constant salinitytreatments inthe greenhousestudy were2, 7, 15, and portion germinating of Lythrum hyssopifolium (P (cid:44) 0.001), 31ppt,withthe31-ppttreatmenthavinghighersalinitythanthehighestsa- with a higher proportion of seeds germinating at constantNo- linitytreatmentinthisstudy(seeResults). vember (0.72; 16.7(cid:56)C) than constant March (0.31; 15.3(cid:56)C) temperature. This species’ response to a 1.4(cid:56)C difference in RESULTS temperature indicates high sensitivity to temperature during Seed age—With one exception, the consecutive arrange- germination, although speed was not affected. A higher pro- ment of the temperature and photoperiod treatments had no portion of Lasthenia glabrata ssp. coulteri (P (cid:53) 0.029) and meaningfuleffectonspecies’germination.Theproportionger- Parapholis incurva (P (cid:53) 0.004) seeds germinated at constant minating of Cotula coronopifolia decreased 59% after an ad- Novembertemperature(0.87and0.95,respectively)compared ditional7moofdrystorage(P(cid:44)0.001)(Fig.2e).Afteraging to diurnal fluctuating November temperatures (0.62 and 0.63, for 7 mo, the germination speed index of Polypogon monspe- respectively).Inaddition,Parapholisincurva(P(cid:53)0.002)and liensis increased 0.03 (P (cid:53) 0.007) (Fig. 2e). This change in Spergularia marina (P (cid:53) 0.008) germinated fasteratNovem- germination speed corresponds to a difference of about one ber constant temperature (0.88 and 0.99, respectively) com- day, a shorter time than the sampling interval (3 d). pared to November diurnal fluctuating temperatures(0.81and 1684 AMERICAN JOURNAL OF BOTANY [Vol. 87 affected by salinity (P (cid:44) 0.001) and moisture treatments (P (cid:53) 0.006). The germination speed of Amblyopappus pusillus also responded to salinity (P (cid:44) 0.001) and moisture (P (cid:44) 0.001). However, salinity and moisture treatments interacted for both the proportion (P (cid:44) 0.001) and speed ofgermination (P (cid:44) 0.001) (Fig. 4). The proportion of Amblyopappus pus- illus seeds germinating and speed of germination declined drastically at the highest salinity and low moisture (Fig. 4). Although salinity and moisture interacted, the range in ger- mination among the treatments of both soil salinity and soil moisture is a measure of the responsiveness of a species and allows broad comparison between these two factors. In this experiment, the proportion of Amblyopappus pusillus germi- nating responded to salinity treatments more than moisture treatments (Table 4). The range in the proportion germinating inresponsetothesalinitytreatmentswasthreefoldgreaterthan the range in the moisture treatments. The range in the germi- nationspeedindexvalueswassimilarinthesalinityandmois- ture treatments (Table 5). The number of Cordylanthus maritimus ssp. maritimus seeds germinating differed among both salinity (P (cid:44) 0.001) and moisture treatments (P (cid:44) 0.001). Similarly, the speed of Cordylanthus maritimus ssp. maritimus germination differed among salinity (P (cid:44) 0.001) and moisture levels (P (cid:44) 0.001). Soil salinity and moisture treatments interacted for both the proportion germinating (P (cid:53) 0.014) and the speed of germi- Fig. 3. Mean ((cid:54)1 SE) surface (top 1 cm) soil moisture and salinity of nation(P(cid:53)0.003),withsalinitytolerancemuchgreaterunder treatmentsinseedlessandseededmicrocosms. high moisture than low moisture conditions (Fig. 4). Thepro- portionofCordylanthusmaritimusssp.maritimusgerminating 0.92, respectively). However, the proportion and the speed of was affected strongly by soil moisture, with a 0.40 difference germination of Parapholis incurva and Spergularia marina between the low and high moisture treatments. Moisture was did not differ between the March diurnal temperature fluctu- relativelymoreimportantthansoilsalinity,withalargerrange ation and the March constant temperature. Seasonal photope- in proportion germinating among moisture treatments (0.40) riod differences affected the germination of two species, with than salinity treatments (0.25; Table4).However,therangein MarchphotoperiodresultinginahigherproportionofLythrum germination speed was larger among soil salinity treatments hyssopifolium germinating (0.72; P (cid:53) 0.009) and faster Par- (0.25) than soil moisture treatments (0.16; Table 5). apholis incurva germination (0.88; P (cid:53) 0.009) than the No- The proportion of Hutchinsia procumbens germinating dif- vemberphotoperiod(0.41and0.82,respectively).Eightofthe feredamongbothsalinity(P(cid:44)0.001)andmoisturetreatments nine significant differences among the temperature, photope- (P (cid:53) 0.001). In addition, germination slowed with increasing riod,andseedagecomparisonshadP(cid:44)0.001,indicatingthat salinity(P(cid:44)0.001)anddecreasingmoisture(P(cid:44)0.001).For Type I errors are unlikely despite the large number of statis- both the proportion and speed of germination (P (cid:53) 0.001 and tical tests that were performed. P (cid:53) 0.048, respectively), moisture and salinity treatmentsin- teractedwithdifferencesamongmoisturetreatmentsbecoming Soil salinity and moisture—Soil moisture levels in the much more apparent at high salinity (Fig. 4). The proportion seedless microcosms were relatively constant through time of Hutchinsia procumbens germinating was affected more by (Fig. 3). Final soil moisture content in the top 1 cm of soil of salinity than moisture, with a fivefold larger range in the sa- the seeded microcosms was 37.1, 45.5, and 50.5% in thelow, linity treatments compared to the moisture treatments (Table medium,andhighmoisturetreatments,respectively.Thethree 4). Speed ofgerminationhadanearlytwofoldhigherrangein highest salinity treatments fluctuated and decreased slightly the salinity treatments than the moisture treatments (Table5). during the first week but still exhibited differences among The proportion of Lasthenia glabrata ssp. coulteri germi- treatments. These decreases in salinity may be due to the fre- nating responded to salinity (P (cid:44) 0.001) and moisture treat- quent removal of soil from the seedless microcosms for sam- ments (P (cid:44) 0.001). The speed of germination also differed pling; removal of the surface soil decreased the elevation of among salinity (P (cid:44) 0.001) and moisture (P (cid:44) 0.001) levels. soil in the inner cup relative to the level of the low salinity Salinity interacted with moisturefor boththe proportion(P(cid:53) water in the outer bath. Final soil salinity was lower than in- 0.012)andspeed(P(cid:44)0.001)ofLastheniaglabratassp.coul- tended. Soil salinity on day 30 in the seeded microcosmswas teri germination; there were larger differences in germination 1.8, 6.8, 13.1, and 22.9 ppt in the 0, 8, 17, and 34 ppt target among moisture treatments at higher salinity levels (Fig. 4). treatments, respectively. Hereafter, the salinity treatmentswill Lastheniaglabrata ssp.coulteriwasmoreresponsivetosalin- be referred to as 2, 7, 13, and 23 ppt. itythanmoisture.Therewasmorethanathreefoldlargerrange When a species responded to the soil moisture or salinity in proportion germinating among the salinity treatments than treatments, the proportion germinating or germination speed the moisture treatments (Table 4). In addition, germination decreasedwithdecreasingmoistureorincreasingsalinity(Fig. speed varied more inresponsetosalinitythanmoisture(Table 4). The proportion of Amblyopappuspusillusgerminatingwas 5). November 2000] NOE AND ZEDLER—ABIOTIC EFFECTS ON GERMINATION OF SALT MARSH ANNUALS 1685 Fig.4. Interactionsofsoilmoistureandsalinitytreatmentsonthemean((cid:54)1SE)finalproportiongerminatingandgerminationspeedindexofeachspecies. An‘‘I*’’indicatesasignificant(P(cid:44)0.05)interactionbetweensoilmoistureandsalinity. The interaction of salinity and moisture on the proportion moisturetreatments,althoughtherewasamaximumdifference of Mesembryanthemum nodiflorum germinating was nearly of only 0.12 among treatments (Table 4). However, salinity statistically significant (P (cid:53) 0.052). The proportiongerminat- had a large effect on the speed of germination (Table 5). Me- ing responded to soil salinity (P (cid:53) 0.035) and was highest at sembryanthemum nodiflorum germination speed index values 13 ppt and lowest at 23 ppt. Soil moisture had no effect on were halved from 0.94 at 2 ppt to 0.47 at 23 ppt. theproportionofMesembryanthemumnodiflorumgerminating The effects of soil moisture and salinity on the proportion (P (cid:53) 0.429). The speed of germination responded to salinity of Parapholis incurva germinating did not interact (P (cid:53) (P (cid:53) 0.001) and moisture (P (cid:53) 0.001) and the interaction of 0.790) (Fig. 4). The proportion of Parapholis incurva germi- salinity and moisture (P (cid:53) 0.002). At high salinity, germina- nating decreased at the 23 ppt treatment, but was similarat2, tion was much slower at low and medium moisture than at 7,and13ppt(P(cid:53)0.010)anddidnotrespondtothemoisture high moisture, whereas germination speed did not differ treatments(P(cid:53)0.254).TheproportionofParapholisincurva among the moisture treatments at low salinity (Fig. 4). Dif- germinating was higher than 0.90 foralltreatments.However, ferences in the proportion of Mesembryanthemum nodiflorum thespeedofgerminationatdifferentsalinitylevelsdiddepend germinatingwerelargeramongthesalinitytreatmentsthanthe on moisturelevels(P(cid:44)0.001).Germinationspeedresponded 1686 AMERICAN JOURNAL OF BOTANY [Vol. 87 TABLE4. Therangeinthefinalproportiongerminatingofeachspecies TABLE 5. The range in the germination speed index of each species among the treatment levels of each abiotic factor and seed age. among the treatment levels of each abiotic factor and seed age. Boldfacednumbersindicatenoteworthylargeeffectsofafactoron Boldfacednumbersindicatenoteworthylargeeffectsofafactoron a species. a species. Photo- Temper- Photo- Temper- Species period Age ature Salinity Moisture Species period Age ature Salinity Moisture Amblyopappuspusillus — — — 0.30 0.10 Amblyopappuspusillus — — — 0.38 0.34 Cordylanthusmaritimusssp. — — — 0.25 0.40 Cordylanthusmaritimusssp. — — — 0.25 0.16 maritimus maritimus Cotula coronopifolia 0.12 0.59 0.28 0.35a 0.10a Cotula coronopifolia 0.01 0.04 0.02 0.57a 0.10a Hutchinsiaprocumbens 0.01 0.02 0.18 0.67 0.12 Hutchinsiaprocumbens 0.01 0.02 0.03 0.54 0.28 Lastheniaglabratassp.coulteri 0.06 0.01 0.25 0.72 0.22 Lastheniaglabratassp.coulteri 0.00 0.01 0.01 0.21 0.15 Lythrumhyssopifolium 0.31 0.02 0.41 0.55a 0.12a Lythrumhyssopifolium 0.01 0.01 0.03 0.76a 0.12a Mesembryanthemumnodiflorum 0.05 0.05 0.04 0.12 0.05 Mesembryanthemumnodiflorum 0.02 0.02 0.05 0.47 0.21 Parapholisincurva 0.13 0.02 0.32 0.08 0.04 Parapholisincurva 0.05 0.04 0.07 0.24 0.16 Polypogon monspeliensis 0.03 0.03 0.04 0.21a 0.08a Polypogon monspeliensis 0.00 0.03 0.02 0.22a 0.01a Spergulariamarina 0.15 0.19 0.23 0.28 0.19 Spergulariamarina 0.01 0.00 0.07 0.24 0.08 aRangecalculatedfromdatainNoe(1999). aRangecalculatedfromdatainNoe(1999). to both salinity (P (cid:44) 0.001) and moisture (P (cid:44) 0.001), al- dylanthus maritimus ssp. maritimus germinating was strongly affected by soil moisture and was the only species tobemore though differences among salinity treatments were only ap- responsive to soil moisture than soil salinity (Table 4). How- parent at low moisture (Fig. 4). Salinity had alargereffecton ever, Cordylanthus maritimus ssp. maritimus germination theproportiongerminatingthanmoisture,butbothfactorshad speed varied more in response to soil salinity treatmentsthan smalldifferencesamongtreatments(Table4).Salinityaffected moisture treatments (Table 5). the speed of germination slightly more than moisture (Table 5). DISCUSSION Bothsalinity(P(cid:53)0.001)andmoisture(P(cid:53)0.006)affected the proportion of Spergularia marina germinating. Germina- We attempted to explain patterns of field germination by tionspeedalsorespondedtosalinity(P(cid:53)0.001)andmoisture testing multiple factors, interactions among factors, and field- (P (cid:44) 0.001). Salinity and moisture effects interacted for both relevant levels of factors on the proportion and speed of ger- the proportion and speed of germination (both P (cid:44) 0.001). mination of half of the annual species in southern California When seeds experienced high and medium moisture levels salt marshes. Other studies of wetland plants have found that they did not respond to the different salinity treatments (Fig. germination can be affected by several abiotic factors (Leck, 4). In contrast, no seeds germinated at low moisture and high 1989).Leck(1996)testedtheeffectsofdifferentsoilmoisture, salinity. The range in Spergularia marina proportion germi- light, and temperature regimes on germination and concluded nating and germination speed index values was similar be- that each species responds to the abiotic environment in a tween the salinity and moisture treatments (Tables 4, 5). unique manner. Similarly, Galinato and van der Valk (1987) found that each of the species they tested was affecteddiffer- Relative effects of five factors—Of temperature, photope- ently by temperature, salinity, and light. Interactions among riod, seed aging, soil salinity, and soil moisture, the speed of abiotic factors also occur (Ungar, 1978); Khan and Ungar germination responded most strongly to soil salinity for all (1997) determined thattemperature,light,andsalinityinteract eight species (Table 5). The proportion germinating of four to effect germination of halophytes. However, most other ex- (Hutchinsiaprocumbens,Lastheniaglabratassp.coulteri,Me- perimentsonthegerminationofsaltmarshplantshavefocused sembryanthemum nodiflorum, and Polypogon monspeliensis) solely on theeffectof soilsalinity.Wewilldiscusstheeffects of the eight species had a greater range among salinity treat- of each of the abiotic factors tested, the interactionofsalinity ments compared to moisture, photoperiod, temperature,orag- and moisture, the differential response of the proportion ger- ing treatments (Table 4). Two species (Lythrumhyssopifolium minating vs. germination speed, and the ability of these ex- andSpergulariamarina)hadsimilarrangesintheirproportion periments to explain patterns of field germination. germinating in both the soil salinity and temperature treat- ments. Temperature had the largest effect on Parapholis in- Effects of each factor—Species differed in their respon- curvaproportiongerminating(Table4).Finally,theproportion siveness to abiotic factors, as quantified by the rangesoftheir of Cotula coronopifolia germinating was affected by seedag- proportion germinating and speed of germination (Tables 4, ing more than the other factors; soil salinity and temperature 5). The differing treatments ofeach abioticfactorcorrespond- elicited larger responses in Cotula coronopifolia compared to ed to the range of conditions that the species are exposed to moisture and photoperiod (Table 4). inthefieldduringperiodsofgermination.Therefore,therange Two additional species, Amblyopappus pusillus and Cor- of the response variables among the treatments of each of the dylanthusmaritimusssp.maritimus,weretestedwithonlysoil abiotic factors is a measure of the relative importanceofeach salinity and moisture. The range in the proportion of Ambly- abiotic factor in determining germination, and thereforeinflu- opappus pusillus germinating was greatest in response to the encing population dynamics, in the field. Some species had soil salinity treatments (Table 4). Amblyopappus pusillus ger- small ranges among the salinity, moisture, temperature, and minationspeedvariedthemostamongsoilsalinitytreatments, photoperiodtreatments;othersrespondedwithlargedifferenc- although there was a large difference between the high and es in proportion germinating or germination speed among lowsoilmoisturetreatments(Table5).TheproportionofCor- treatments. November 2000] NOE AND ZEDLER—ABIOTIC EFFECTS ON GERMINATION OF SALT MARSH ANNUALS 1687 Of the eight species tested, the proportion germinating or in this study. Callaway et al. (1990) used seeds from Carpin- germination speed of four (Lythrum hyssopifolium, Lasthenia teria Marsh, farther north than the seed sources in this study, glabrata ssp. coulteri, Parapholis incurva, and Spergularia KeifferandUngar(1997)testedseedsfrominlandsaltmarsh- marina) responded to the various treatments that tested small es in Ohio, and Kingsbury et al. (1976) collected seeds from differences in either temperature or photoperiod (Fig. 2). The Los Pen˜asquitos Lagoon. Of these studies, the salt tolerance temperatures and photoperiods tested in this experiment rep- of Lasthenia glabrata ssp. coulteri in Kingsbury et al. (1976) resented conditions at the beginning and end of the typical is most similar to the results of this study. Kingsbury et al. periodofgerminationintheupperintertidalmarshofsouthern (1976) found regional differences in Lasthenia glabrata salt California.Ofthethreetemperaturecomparisons,theonewith tolerance and concluded that the salt tolerance of different the greatest number of significant differences in germination populationswasrelatedtothesoilsalinityfoundinthehabitat was November diurnal fluctuating vs. constant temperature. of each population. Beare and Zedler (1987) also found dif- Thompson and Grime (1983) found that the germination of ferences in the salt tolerance during germination among dif- many temperate wetland species respond to as small as 1(cid:56)C ferent southern California Typha domingensis populations. fluctuations in temperature compared to constant conditions. This study is not directly comparable with these other studies A possibleexplanation forwhythegerminationofthespecies because others used filter paper or sand as experimental sub- in thisstudy werelessresponsivetotemperaturethanthosein strates and collected seeds from populations that may differ Thompson and Grime (1983) is that the germination of most genetically from the seeds in this study. Because saline soils speciesiscuedtovariationsinsalinity,nottemperature,insalt arenotalwayssaturated,soil-basedstudiesofthesalttolerance marshes with a mediterranean-type climate. of germination may be more realistic and predictive of field The consecutivearrangement ofthetreatmentshadlittleef- patterns. fect on theexperiments. OnlytheproportionofCotulacoron- Microcosm moisture levels were also similar to the range opifolia germinating was greatly affected by the 7-mo period of conditionsfoundinsouthernCaliforniaupperintertidalsalt between the beginning and ending temperature and photope- marshes during periods of germination (Noe, 1999). Wetter riod experimental treatments (both November constant tem- soil resulted in more seeds germinating for five species and perature and March photoperiod). Baskin and Baskin (1998) the germination speed of all seven species increasedinwetter recommendstartinggerminationexperimentswithin7–10dof soils. No studies examining theeffect ofsoilmoisturecontent seed collection to prevent changes in germination responses on these species are available for comparison. In addition, during storage. However, up to six months commonly elapse most experiments on other wetland species test much higher between seed dispersal and germinationintheupperintertidal moisture levels (flooded or saturated conditions) than the marsh in southern California (G. B. Noe, personal observa- moisture levels tested in this study. tion). While seed storage in the laboratory differed from con- ditions in thefield, surface soilsinthefieldarealsodry((cid:59)5– Salinity and moisture interaction—Salinity and moisture 15%; Fig. 6) during this period of summer dormancy. interacted to affect the proportion germinating of five species Themicrocosmsalinitylevelstestedinthisexperimentwere and germinationspeedofallsevenspecies(Fig.4).Theinflu- similar to those when germination occurs in the upper inter- ence of salinity became more evident at low moisture, like- tidalmarshofsouthernCalifornia(Noe,1999).Increasingsoil wise, moisture effects were largest at high salinity. Salinity salinity elicited declines in the proportion germinating and and moisture effects were independent of each other for the slowed the speed of germination for each species. These ef- proportion of Mesembryanthemum nodiflorum and Parapholis fects are common among halophytes (Waisel, 1972; Ungar, incurva germinating, the two species most tolerant of high 1978). However, some species were more tolerantof saltthan salinity and low moisture. The germination of many wetland otherspeciesandthedegreeofslowedgerminationinresponse plant species is also determined by interactions between soil tosalinityvariedamongspecies.Forexample,thegermination salinity and the duration of soil saturation (Kuhn and Zedler, of the average seed of Lasthenia glabrata ssp. coulteri was 1997) or inundation (Baldwin, McKee, and Mendelssohn, delayed(cid:59)6d(0.21indexdifference),andHutchinsiaprocum- 1996). bens germination was delayed (cid:59)16 d (0.54 index difference) Theinteractioninthisexperimentmaybedueinparttothe at 23 ppt compared to 2 ppt. Such a 10-d difference in the method of quantifying soil salinity. The soil paste extractses- timing of germination between species could shift relative timate salt concentrations in saturated soils, and effective sa- growth rates and alter the outcome of interspecific competi- linitiesweremuchhigheratlowmoisturethanwasmeasurable tion. Grace (1987) was able to measure a competitive advan- by thistechnique.It isdifficulttoascertainwhetherthemech- tage between two species with as little as a 2-d difference in anism of the interaction in this study isdue to osmoticeffects the timing of seed sowing. However, the magnitude of the on water potential or the toxic effects of ions. However, the interspecific differences in the delay of germination in this effects of salinity on the germination of halophytes are most study isnotsufficienttoexplainthe1–3modifferencesinthe commonly osmotic (Ungar, 1978; Baskin and Baskin, 1998). timing of germination in the field (Noe, 1999). The interaction of salinity and moisture has implications for Others have examined the effect of salinity on the germi- studies determining the effect of soil salinity on germination. nation of some of the same species. In these studies, seedsof Mostexperimentstestforeffectsofsalinityongerminationby Lasthenia glabrata ssp. coulteri (Kingsbury et al., 1976; Cal- placing seeds on filter paper in water-filledpetridishes,there- lawayetal.,1990),Parapholisincurva(Callawayetal.,1990), by providing very high or saturated moisture levels. Fewer and Spergularia marina (Callaway et al., 1990; Keiffer and experiments omit the filter paper or use petri dishes with sat- Ungar, 1997) were less salt tolerant than in this study. Soil urated sand. In the upper intertidal zone, where tidal inunda- pasteextractsestimatesaltconcentrationsafterdilution;hence, tion is infrequent and of short duration, soils are often both the differences in salt tolerance between this study and other salineand dry and theinteractionofsoilsalinityandmoisture studies are exacerbated by the underestimation of soil salinity could be important. 1688 AMERICAN JOURNAL OF BOTANY [Vol. 87 Fig.5. DailyhighandlowtemperaturesatLindberghFieldanddaylength(sunrisetosunset)inSanDiego.Shadedboxesindicatethegerminationwindows duringthe1996season,nonseasonalrainfallfromHurricaneNora,and1997season. Proportion vs. speed of germination—The germination decreased during the 1996 and 1997 germination windows speed of each species, but not proportion (at 4 wk), was af- comparedtoperiodswithoutgermination(Fig.6).Soilsalinity fected by soil salinity and moisture. At high salinity and low and moisture varied during the germination window of both moisture, slowing of germination is a more general trait than years. A week after the rainfalls from Hurricane Nora in late decreasing proportion germinating among the species in this September1997,soilsalinitywashigherandsoilmoisturewas study. In the photoperiod and temperature experiments, pro- lower than during the germination windows. portion germinating was more responsive than germination speed (Fig. 2). Statistically significant differences in the pro- Explaining observed field patterns—The upper intertidal portion germinating ranged from 0.25 to 0.41 between tem- marsh of southern California has a germinationwindowof2– perature/photoperiod treatments. In comparison, significant 3 mo (Noe, 1999). In general, by testing the germination re- differences in the germination speed index ranged from 0.04 sponses of species to soil salinity, soil moisture, temperature, to 0.07 in the temperature/photoperiod trials, or a 1.2- to 2.1- and photoperiod, we can explain why certain species canger- ddifferenceintheaveragetimeofgermination.Despitesmall- minate outside this germination window. We cannot explain er differences in germination speed than proportion among the details of germination timing within this period, nor can species and abiotic factors, the time it takes for a seed to ger- we explain the spatial distributions of species. Two species minate can have large effects on interspecific interactions were restricted to the November to March germination win- (Grace, 1987). Therefore, while the magnitude of thechanges dow; that is, they did not germinate afterHurricaneNora(Ta- in germination speed could not explain the differences in the ble 1). The absence of Hutchinsia procumbens after that non- timing of germination that were observed in the field, the seasonal rain event is most likely due its low salt tolerance, slowing of some species’ germination could affect the com- as soil salinity was high at that time (Fig. 6). However, Par- petitive balance among species. The lack of concordance be- apholis incurva is the most salt-tolerant species, so itslackof tween the response of the two traits to abiotic stress suggests germination following the hurricane is best explained by its that an index of germination speed should be independent of decreasedproportiongerminatingatthehighesttemperaturein the proportion germinating, as is the index used in this study. this experiment (Fig. 2) and the high temperatures that oc- ThisiscontrarytothesuggestionofBrownandMayer(1988), curred during and after the hurricane (Fig. 5). The proportion who promote combining these aspects of germination into a ofLastheniaglabratassp.coulterigerminatingalsodecreased single, concise index. in the highest temperature treatment(Fig.2)butitgerminated following thehurricane. Ofthe twowetlandswhereLasthenia Field environmental conditions during germination— glabrata ssp.coulteriwasfound,nonseasonalgerminationoc- Temperatures in each of the 1996 and 1997 germination win- curred only at the wetland with very low salinity and high dows were similar throughout the period of germination (Fig. moisture(Noe,1999),indicatingthatthesalinityandmoisture 5).Daylengthincreasedbyaboutanhourfromthestarttoend conditions may have overridden any effect of temperature on of thegerminationwindows. Duringthegerminationpulseaf- germination. Parapholis incurva was found at the same sites ter Hurricane Nora in late September1997, high temperatures as Lasthenia glabrata ssp. coulteributdidnotgerminateafter were(cid:59)5(cid:56)Chigherandlowtemperatures10(cid:56)Chigherthandur- thenonseasonalrainfall,suggestingthattemperaturelimitation ing the typical germination windows. Photoperiod during this of Parapholis incurva was more important than salinity tol- nonseasonaleventwassimilartoconditionsinMarch,theend erance. Both the temperature and photoperiod trials correctly of the typical period of germination. predicted that Lythrum hyssopifolium would germinate in the Noe (1999) found that soil moisture increased and salinity conditions following Hurricane Nora.