Seed germination of high mountain Mediterranean species: altitudinal, interpopulation and interannual variability

Page 1

ORIGINAL ARTICLE

L. Gimenez-Benavides Æ A. Escudero Æ F. Perez-Garcıa

Seed germination of high mountain Mediterranean species:altitudinal, interpopulation and interannual variability

Received: 8 September 2004 / Accepted: 19 January 2005 / Published online: 17 March 2005� The Ecological Society of Japan 2005

Abstract The germination response of 20 species fromhigh altitude Mediterranean climates, most of them rareendemics, was studied. Our main goal was to model thegermination response of a complete set of Iberian highmountain species. The effect of temperature and otherparameters, such as spatial and temporal short gradi-ents, on germination were also evaluated. Some seedfeatures (mass and size) were also related to the germi-nation response. Finally, we tested the effect of cold-wetstratification pretreatment when germination was lowunder natural conditions. Seeds were collected at fourlocations from 1,900 to 2,400 m a.s.l. in the Sierra deGuadarrama (Spanish Central Range) over two con-secutive growing seasons (2001–2002) and submitted todifferent temperatures and a constant photoperiod of16 h light/8 h darkness. Most plants readily germinatewithout treatment, reaching an optimum at relativelyhigh temperatures in contrast to lowland Mediterraneanspecies. Seeds seem to be physiologically prepared forrapid germination even though these plants usually facevery intense summer droughts after ripening and dis-persal. Germination was also highly variable amongaltitudes, populations and years, but results wereinconsistent among species. Such flexibility could beinterpreted as an efficient survival strategy for speciesgrowing under unpredictable environments, such as theMediterranean climate. Finally cold-wet stratificationincreased germination capacity in five of nine dormantspecies, as widely reported for many arctic, boreal andalpine species. In conclusion, high mountain Mediter-

ranean species do not differ from alpine species exceptthat a relatively high number of species are ready togerminate without any treatment.

Keywords Alpine plants Æ Cold-wet stratification ÆDormancy Æ Mediterranean climate

Introduction

High levels of endemic species are a significant feature ofhigh mountain Mediterranean climates (Vare et al.2003). Global warming evidence in the mountains ofcentral Spain (Gavilan et al. 2001; Sanz-Elorza et al.2003) suggests that orophilous species are at serious riskfrom the advance of lowland plants as widely predicted(Grabherr et al. 1994; Penuelas et al. 2002). Surprisingly,there is almost no basic information for conservation ofthese plants despite the fact that Mediterranean moun-tains are considered one of the most threatened systemsin Spain and the European Union (Gomez-Campo 1987;European Community 1992).

Seed storage is considered the best way to storeand maintain large pools of genetic diversity in plants(Thompson et al. 1981; Bonner 1990). Reinforcementof wild populations using stored seed material may bea valuable tool for conserving endemic and threatenedspecies (Schemske et al. 1994; Perez-Garcıa et al. 1995;Cerabolini et al. 2004). On the other hand, restorationof alpine environments has been carried out in Europeand North America in recent decades (Brown andJohnston 1979; Urbanska and Schutz 1986). Detailedinformation about native alpine plants as potentialspecies for restoring ski runs, mine sites and othermountain habitats damaged by humans has graduallyacquired great importance (Urbanska 1986; Chambers1997), and specific germination requirements constitutevery valuable information for successful propagation(McDonough 1969, 1970; Acharya 1989; Chambers1989; Chambers et al. 1987). Despite an ambitious

L. Gimenez-Benavides (&) Æ A. EscuderoBiodiversity and Conservation Group,University of Rey Juan Carlos-ESCET,Tulipan s/n, 28933 Mostoles, Madrid, SpainE-mail: [email protected]

F. Perez-GarcıaDepartamento de Biologıa Vegetal, E.U.I.T. Agrıcola,University of Politecnica de Madrid, Ciudad Universitaria s/n,28040 Madrid, Spain

Ecol Res (2005) 20: 433–444DOI 10.1007/s11284-005-0059-4

Page 2

attempt to remove and restore an old ski resort in ourstudy area (Sanchez-Herrera 2000) there is still acomplete lack of knowledge about reproductive fea-tures of key species.

Previous studies on alpine flora around the worldsuggest the absence of a specific alpine germinationstrategy (Korner 1999). In addition, germinationbehaviour may vary greatly within a single species fromone population to another, from year to year and amongindividuals (Urbanska and Schutz 1986). There is atendency for better germination under high (Mooneyand Billings 1961; Billings and Mooney 1968) andalternating temperatures (Amen 1966; McDonough1969, 1970; Bliss 1971). Winter dormancy seems to be acommon requirement for some alpine species (Korner1999). Both innate and enforced seed dormancy havebeen successfully broken by subjecting seeds to a widevariety of treatments: gibberellins imbibition (McDon-ought 1969, 1970), scarification (Pelton 1956), stratifi-cation in multiple forms (dry-hot, dry-cold, wet-cold)(Pelton 1956; Cavieres and Arroyo 2000), or combina-tions. Among them, wet-cold stratification seems toproduce the best results in alpine plants (Baskin andBaskin 1998). The extrapolation of these germinationrequirements of alpine plants to Mediterranean envi-ronments should be conducted with caution becausehigh mountain Mediterranean plants face specific con-straints such as the development of an intense waterdeficit during the short period in which temperatures arehigh enough to enable growth (e.g., Sierra Nevada site insouthern Spain, Callaway et al. 2002).

The main objectives of this work were (1) to modelthe germination response of Iberian high mountainspecies, and compare with lowland Mediterraneanspecies and with plants from arctic and alpine environ-ments; (2) to analyse effects of altitude, populationvariability and year of collection upon seed germination,(3) to evaluate the effectiveness of the cold-wet stratifi-cation treatment for the induction of germination inspecies with high levels of dormancy, and (4) to provideuseful information on species-specific germination con-ditions for conservation purposes.

Materials and methods

Plant material collection and descriptionof the field localities

We have selected 20 species, most of them endemics ofthe Iberian Peninsula and with small and isolatedpopulations above the timberline. Species cover a widevariety of functional types from perennial grasses tocreeping chamaephytes and hemicryptophytes, andalso communities from rock to short pasture habitats(see Table 1). Seeds were collected directly from mo-ther plants during the fruiting season (from July toSeptember) over 2 consecutive years (2001 and 2002).

We collected from small areas, where all the availableseeds were harvested. When possible we collectedmaterial from different populations and years (2001and 2002).

Five sets of 10 or 20 seeds per species and per pop-ulation (enough to reach the accuracy of the precisionbalance, COBOS AX120, d=0.1 mg) were weighed andseed mass calculated. Seed maximum and minimumdiameter were measured from digital images (n=25)taken from the weighed seeds and were analysed usingOlympus Micro Image Version 4.0.

Collection sites were located in the highest portion ofSierra de Guadarrama, a SW-NE running mountainrange located in the north of the Madrid province(40�N, 3�W) and belonging to the Spanish CentralRange. Rainfall data for specific locations are unavail-able, but annual rainfall in the closest weather station(Navacerrada Pass, 1,890 m) is about 1,400 mm, with avery pronounced summer drought when less that 10% ofannual precipitation occurs. A complete synopsis of thecryoromediterranean belt vegetation is given in Rivas-Martınez et al. (1999).

Locality 1 was located in the vicinity of Pico Penalara(2,428 m), the highest peak of Sierra de Guadarrama.The dominant vegetation in the summit flat areas andcrests, where snow cover only remains 120–140 days peryear (Palacios et al. 2003), is a discontinuous cryophilicpasture dominated by Festuca curvifolia (associationHieracio myriadeni-Festucetum curvifoliae Rivas-Martı-nez 1963). This community is particularly rich in en-demics, such as Hieracium vahlii subspp. myriadenum,Minuartia recurva subspp. bigerrensis and Armeriacaespitosa, as well as in arctic and alpine relicts, such asAgrostis rupestris and Phyteuma hemisphaericum (seeGavilan et al. 2002; Escudero et al. 2005). Rockwallsteps and blockfields are colonised by scattered com-munities dominated by Allium schoenoprassum, Saxi-fraga willkommiana, Veronica fruticans subspp.cantabrica, and Silene boryi (Allietum latirifolii andSaxifragetum willkommianae associations). Annualaverage air temperature is 3.8�C, whereas the meanmonthly temperature ranges from �5.2�C in January to13.6�C in July.

Locality 2 was located in the summit of Dos Her-manas (2,269 m), which is located 3 km away. Due to itslower altitude, grass-dominated communities are inter-spersed with an open shrub formation dominated byCytisus oromediterraneus and Juniperus communis sub-spp. alpina (Senecioni carpetani-Cytisetum oromediter-ranei). Average snowcover duration is slightly shorter(100–120 days per year). Rockfall and blockfield habi-tats are also present. Annual average air temperature is4.9�C, whereas the mean monthly temperature rangesfrom �3.7�C in January to 14.3�C in July.

Locality 3 was in Cerro de Valdemartın (2,279 m), amountain located in an adjacent mountain range run-ning parallel and 5 km away, and separated by theheadwaters of the Lozoya river valley (see Fig. 1).Vegetation, snow cover duration and temperature

434

Page 3

Table

1Plantlife

form

,altitudinalrange,

andmeanseed

mass

andsize

for20alpinespeciescollectedontheGuadarramamountain

range.

Localities

andyears

ofcollectedplant

materialare

marked

foreach

species.

Hm

hem

icryptophyte,Cscushionchamaephyte,Hm

Caecaespitoushem

icryptophyte,A

annual,G

geophyte.SM

supra-,

OM

oro-,

COM

cryoromediterraneanbelt

Code

Species

Family

Life

form

Habitattype

Altitudinal

range

Meanseed

mass

mg(SD)

Meanmaxim

um

seed

diameter

mm

(SD)

Meanminim

um

seed

diameter

mm

(SD)

Loc.

1(2,428m)

Loc.

2(2,269m)

Loc.

3(2,279m)

Loc.

4(1,950m)

20012002200120022001200220012002

Alli

Allium

schoenoprassum

aLiliaceae

GRockwall

OM-C

OM

1.558(0.133)2.755(0.279)

1.784(0.244)

.x

..

..

.Arm

Arm

eria

caespitosa

aPlumbaginaceae

Cs

Dry

grassland

SM-C

OM

1.184(0.067)4.700(0.734)

2.020(0.341)

.x

.x

.x

..

Bis

Biscutellalaevigata

subspp.

gredensis

Cruciferae

Hm

Blockf./D

rygrass.

OM-C

OM

1.120(0.061)2.555(0.199)

1.883(0.144)

x.

..

x.

..

Fesc

Faestuca

curvifolia

Poaceae

Hm CaeDry

grassland

SM-C

OM

1.032(0.150)4.558(0.359)

1.244(0.150)

.x

.x

.x

..

Hie

Hieracium

vahliisubspp.

myriadenum

Asteraceae

Hm

Dry

grassland

OM-C

OM

0.203(0.017)1.908(0.164)

0.678(0.108)

x.

.x

x.

..

Jas

Jasionecrispasubspp.

centralisb

Campanulaceae

Cs

Dry

grassland

OM-C

OM

0.057(0.011)0.951(0.094)

0.443(0.083)

.x

.x

..

xx

Jur

Jurinea

humilis

Asteraceae

Hm

Dry

grassland

SM-C

OM

4.708(0.276)5.422(0.580)

3.064(0.462)

.x

.x

.x

..

Luz

Luzula

hispanicab

Juncaceae

Hm

Dry

grassland

COM

0.292(0.019)1.248(0.103)

0.802(0.126)

..

..

x.

..

Min

Minuartia

recurvasubspp.

bigerrensisa

Caryophyllaceae

Cs

Dry

grassland

COM

0.252(0.016)1.179(0.120)

0.876(0.094)

.x

..

xx

..

Mur

Murbeckiellaboryib

Cruciferae

ARockwall

SM-C

OM

0.106(0.007)1.235(0.162)

0.625(0.094)

..

..

x.

..

Ran

Ranunculusollissiponensis

subspp.alpinusb

Ranunculaceae

Hm

Dry

grassland

SM-C

OM

0.744(0.067)2.860(0.270)

2.254(0.226)

..

..

x.

x.

Sax

Saxifragapentadactylis

subspp.wilkommianab

Saxifragaceae

Cs

Rockwall

M-C

OM

0.052(0.006)0.811(0.092)

0.491(0.094)

..

..

x.

..

Sbor

Sileneboryisubspp.

penyalarensis

Caryophyllaceae

Hm

Blockf./R

ockw.OM-C

OM

0.986(0.058)1.678(0.147)

1.283(0.136)

.x

.x

..

x.

Silc

Sileneciliata

subspp.

elegans

Caryophyllaceae

Cs

Dry

grassland

OM-C

OM

0.594(0.061)1.531(0.496)

1.083(0.362)

..

..

x.

x.

Sem

pSem

pervivum

vicentei

subspp.pauib

Crassulaceae

Hm

Blockf./D

rygrass.

OM-C

OM

0.053(0.010)0.954(0.126)

0.501(0.088)

..

..

x.

..

Senb

Senecio

boissierib

Asteraceae

Hm

Dry

grassland

COM

1.372(0.148)2.454(1.687)

0.605(0.455)

..

..

.x

..

Senp

Senecio

pyrenaicussubspp.

carpetanusa

Asteraceae

Hm

Dry

grassland

OM-C

OM

2.396(0.122)3.917(1.688)

1.096(0.554)

.x

..

x.

..

Sol

Solidagovirgaureasubspp.

fallit-tirones

bAsteraceae

Hm

Blockfield

SM-C

OM

0.814(0.011)3.503(1.447)

0.948(0.392)

.x

..

..

..

Thy

Thymuspraecoxsubspp.

penyalarensisa

Labiatae

Ch

Dry

grassland

COM

0.122(0.016)1.138(0.119)

0.837(0.102)

.x

.x

.x

..

Ver

Veronicafruticanssubspp.

cantabrica

bScrophulariaceaeHm

Blockfield

COM

0.096(0.040)0.853(0.107)

0.618(0.081)

.x

..

..

..

aEndem

icsofcentralmountain

range

bEndem

icsoftheIberianPeninsula

435

Page 4

regime are similar to locality 2. For further analysespopulations of localities 2 and 3 were assigned to asimilar altitude.

Finally, the Locality 4 was in Hoya de Penalara, alate Pleistocene glacial cirque situated at 1,950 m, in thetimberline zone. Seeds were collected from a depositedmoraine, covered by dispersed stunted pines (Pinus syl-vestris) in a Cytisus-Juniperus shrub matrix. Despite itslower altitude snow cover duration lasts 120–140 daysper year due to its leeward orientation that protects fromprevailing westerly winds (Palacios et al. 2003). Annual

Fig. 1 Final germination percentages for the 20 species at differenttemperature regimes. Vertical lines show standard deviations.Average germination throughout temperatures and locations arein parentheses. Superscript numbers indicate temperature require-ment for optimal germination (see Table 2)

436

Page 5

average air temperature is 6.9�C, whereas the meanmonthly temperature ranges from �1.9�C in January to15.9�C in July.

Germination test

Ripe fruits were collected directly from mother plants inrelatively small areas (50–100 m2). Seeds were cleanedand stored in plastic vials at room temperature anddarkness for several months (4–5) until trials were car-ried out. Germination responses were obtained under a16-h light/8 h-dark photoperiod at three constant tem-perature regimes (10, 15 and 20�C). Seeds of Silene cil-iata, S. boryi subspp. penyalarensis, Minuartia recurvasubspp. bigerrensis, Jasione crispa subspp. centralis andHieracium vahlii subspp. myriadenum collected in 2001were also submitted to 25�C and alternating 25/15�Ctemperatures because the number of seeds collected wassufficient. Four replicates of 25 seeds per treatment wereplaced on two layers of filter paper in 5 or 8 cm diameterpetri dishes, depending on seed size. Filter papers werekept soaked along the whole experimental period(35 days) and each 3 or 4 days seeds showing radicleemergence were counted and thereafter removed fromthe petri dishes. Dish location in the chamber was reg-ularly changed. Tests were carried out simultaneously ingermination chambers (Selecta Hotcold GL, Barcelona,Spain) equipped with six cool-white fluorescent lighttubes (Philips 18 W ‘TL’D standard type, wavelength400–650 nm) providing a photon flux density ofapproximately 19 lmol m�2 s�1. Chamber temperatureswere also monitored with a portable thermometer(Oregon Scientific MTR102) in order to avoid bias dueto failures in chamber control.

Stratification pretreatment

A cold-wet stratification treatment was carried out withthose species in which final germination percentage waslower than 60% (Allium schoenoprassum, Biscutellalaevigata subspp. gredensis, Jurinea humilis, Luzulahispanica, Ranunculus ollissiponensis subspp. alpinus,Saxifraga pentadactylis subspp. wilkommiana, Semper-vivum vicentei subspp. paui, Senecio pyrenaicus subspp.carpetanus, Silene ciliata subspp. elegans, Solidagovirgaurea subspp. fallit-tirones and Veronica fruticanssubspp. cantabrica). In order to ensure humid and darkconditions, batches from one population per taxon wereplaced between two double layers of filter paper in 8 cmpetri dishes and wetted with distilled water. Dishes werewrapped in aluminium foil and stored in a refrigerator at4�C for 3 months. Germination tests were conducted asdescribed above.

Seed viability was performed with a tetrazolium test(Hendry and Grime 1993) for those species that did notgerminate after stratification. Moistened seeds, sliced intwo with a scalpel, were incubated in 10 ml of 0.1%

tetrazolium chloride for 24 h at laboratory tempera-ture.

Data analysis

The effect of population of origin (locality), altitude andyear upon final germination percentage was modelled byfitting Generalised Linear Models (GLMs) (McCullaghand Nelder 1989). Germination data follows a binomialdistribution (probability ranging from 0 to 1), so the bestway to achieve linearity is the use of a generalised linearmodel with logit link function and binomial error dis-tribution, setting the variance to ‘‘mean (1-mean)’’(Venables and Ripley 1998; Schutz and Rave 1999).Experimental parameters (temperature, population, alti-tude and year) were included as explanatory variables(fixed factors). Because data are overdispersed in somecases we used the quasilikelihood approach to overcomepossible difficulties (Guisan et al. 2002). Significantterms were identified using a stepwise addition of termsto the null model, based on the magnitude of the Cpstatistic at each step (Spector 1994), until no additionalterms improved the model. Regression coefficients weretested for significance by t-tests. v2 tests were also con-ducted to evaluate whether or not selected predictorsexplain a significant fraction of the deviance (Guisanet al. 2002). Statistical analysis was performed usingS-Plus 2000 (MathSoft 1999).

Seeds from different localities and years were notavailable for all species, so species-specific models wereperformed depending on the available material. Forexample, models for species with only one populationwere performed including only temperature as a fixedfactor (e.g., Allium schoenoprassum, Sempervivum vicen-tei subspp. paui, Senecio pyrenaicus subspp. carpetanus,Solidago virgaurea subspp. fallit-tirones and Murbecki-ella boryi). Species with populations located at differentaltitudes were analysed setting temperature and altitudeas factors (12 species). When more than one populationper altitude was collected (populations 2 and 3, both at2,200 m), population could also be added into the model(e.g., Armeria caespitosa, Festuca curvifolia, Jurineahumilis, Silene boryi subspp. penyalarensis and Thymuspraecox subspp. penyalarensis). Finally, for those specieswhere at least one population was collected both years,variable year was also fitted into the model (the case ofJasione crispa subspp. centralis and Minuartia recurvasubspp. bigerrensis). Detailed information about popu-lation of origin, altitude, and year of collection of seedsis showed in Table 1. The influence of stratificationtreatment and collecting year on final germination per-centage was tested by common analysis of variance.Germination data were previously acrsine transformed.

The Kaplan–Meier method was performed to modelthe germination curves. This method (traditionally usedin survival analysis) allows the use of germination dataconsisting of elapsed time from the beginning of theexperiment to seed germination (non-censored data), but

437

Page 6

also right-censored data (i.e., seeds that did not germi-nate before the end of the experiment). Pairwise shapecomparisons for modelled germination curves were tes-ted by non-parametric Log-Rank tests (Pyke andThompson 1986).

Results

Germination without pretreatment

Since the number of species is high, we have grouped thespecies accordingly to their ability to germinate (bothfinal germination percentage and germination rate).Mean germination percentage is showed in Fig. 1. Onaverage throughout temperatures and locations, barelyone-third (seven species) presented more than 75%germination, whilst four species (20%) failed to achieve25% of seeds germinated. We also found seven specieswith germinability around 60% (range 54–68%). Fi-nally, only two species (10%) were unable to germinatewithout pretreatment. In relation to germination rate,most plants were ready to germinate rapidly (seeexamples in Fig. 2). Germination curves showed aminimum period to initiate germination of 3–6 days,except Biscutella laevigata subspp. gredensis andRanunculus ollissiponensis subspp. alpinus that required9 days. No significant relationships were found betweenseed features (independently size and weight) and finalgermination (Kendal’s tau = �0.09 and 0.05, respec-tively). Seeds were very small in most cases (see Table 1).

Response to temperature

Differences between species can be summarised intothree groups (see Fig. 1 and Table 2 for species-specificresponses). Group 1 clusters the majority of the species,which are able to germinate equally at the three tem-peratures tested (temperature factor was not selected inthe corresponding model). Neither significant difference(P<0.0001) was found on germination curves amongtemperatures, except in Solidago virgaurea subspp. fallit-tirones and Thymus praecox subspp. penyalarensis(Fig. 2). Group 2 consists of species that better germi-nate at warmer temperatures. At lower temperatures,germination rates were significantly slower (see Sileneboryi in Fig. 2). Group 3, which is represented only bytwo species (Jurinea humilis and Silene ciliata) is char-acterised by optimum germination at relatively low(15�C) temperatures. Nevertheless, the latter species canreach substantially higher values at alternating 25/15�C.

Altitudinal, interpopulation and interannual variability

Results of the generalised linear models show the influ-ence of both altitude and population variability on thegermination behaviour of several species (Table 2).

Seeds collected at higher altitudes reached higher ger-mination percentage than those from lower elevations in3 of 12 species (Jasione crispa subspp. centralis, Min-uartia recurva subspp. bigerrensis and Thymus praecoxsubspp. penyalarensis). In contrast, 4 species respondedbetter at lower elevations (Armeria caespitosa, Biscutellalaevigata subspp. gredensis, Ranunculus ollissiponensissubspp. alpinus and Jurinea humilis). Population alsoexplained a slight but significant fraction of variation in3 of 5 species (Festuca curvifolia, Thymus praecox sub-spp. penyalarensis and Jurinea humilis).

Seeds from a given locality in both years were onlyavailable for two species, Minuartia recurva subspp.bigerrensis from locality 3 and Jasione crispa subspp.centralis from locality 4 (Fig. 3). Minuartia recurvasubspp. bigerrensis showed higher germination values in2002. Variable year was included in the final model(Table 2).

Germination after cold-wet stratification

It should be noted that germination tests could not beproperly carried out for several species below 60% ger-mination. We lost seed batches of Jurinea humilis andSolidago virgaurea subspp. fallit-tirones due to fungalcontamination during stratification. On the other hand,Sempervivum vicentei subspp. paui and Ranunculusollissiponensis subspp. alpinus surprisingly achieved highgermination values during the stratification period,leaving insufficient seeds to conduct temperature tests.Nevertheless, a positive effect of treatment in the case ofSempervivum vicentei subspp. paui was evident (germi-nation reached 78% in contrast to 2% on average beforetreatment). However, similar germination percentageswere obtained for Ranunculus ollissiponensis subspp.alpinus (76% before and 75% during the treatment).

We found a significant increase in final germinationin response to cold-wet stratification in five of ninespecies. Best effects were found in Veronica fruticanssubspp. cantabrica and Luzula hispanica, in whichabsolute dormancy of fresh seeds (total absence of ger-mination) was broken after treatment (Fig. 4). Veronicafruticans subspp. cantabrica showed the highest germi-nation value (96%) at the low temperature (10�C),decreasing progressively at warmer temperatures. On theother hand, Luzula hispanica germinated only at 10�C,reaching 50%. Optimal results were also obtained forA. schoenoprassum and Silene ciliata subspp. elegans.The former exhibited maximum germination at warmertemperatures (98% both at 15� and 20�C, compared to43% at 10�C) while Silene ciliata subspp. elegansreached 96% at 15�C, decreasing at extreme tempera-tures. Germination curves always followed these ten-dencies (see Fig. 5). While seed stratification inSaxifraga pentadactylis subspp. wilkommiana causedlittle increase of final germination at 10 and 15�C (butdecreasing radically at 20�C), germination curvesshowed a high germination rate at unmodified condi-

438

Page 7

tions. That is, Saxifraga pentadactylis subspp.wilkommiana germinate slightly better, but slower, afterstratification. Finally, Biscutella laevigata subspp.gredensis and Senecio pyrenaicus subspp. carpetanuswere unable to germinate after stratification. A seed

viability test assessed after the experiment revealed thatseeds of both species were unviable.

Discussion

Germination without pretreatment

Surprisingly most species germinated without pretreat-ment despite the fact that germination assays were car-ried out 4–5 months after seed collection (July–August).

Fig. 2 Modelled germination curves by the Kaplan–Meier methodfor fresh seeds. These four cases exemplify contrasting germinationbehaviours obtained for the 20 species. Superscript numbersindicate temperature requirement for optimal germination (seeTable 2). n.s. non-significant differences between curves(P>0.0001)

439

Page 8

Previous results suggest that alpine plants are unable togerminate during the current season (autumn), so theyget into a dormancy process to avoid harsh winterconditions and germinate after snowmelt in early sum-mer (Urbanska and Schutz 1986; Korner 1999). Thisfact has been interpreted by different authors as a mat-uration period to complete development of embryos(Urbanska and Schutz 1986), a physical restriction ofseed coats that needs scarification by freeze-thaw cycleson soil (Urbanska et al. 1979; Zuur-Isler 1982), or

merely an environmental constraint rather than aphysiological requirement (Billings and Mooney 1968).Our results show that in most high mountain Mediter-ranean plants storage periods of 4–5 months may besufficient to induce germination, even under room tem-perature conditions. Since immediately post-harvestgermination tests were not carried out with these species,assumptions about prewinter germination capabilityshould be taken with care. In any case, seeds seem to bephysiologically prepared to germinate after very short

Table 2 Results of the reduced Generalized Linear Models for finalgermination percent of fresh seeds. Selection of variables temper-ature, altitude, population (locality) and year was carried out bymeans of a forward stepwise procedure. Species are groupedaccordingly to temperature requirement for optimal germination.Coef. (SE) coefficient value (standard error), t t-test for signifi-

cance of the coefficients, Prob. (t) P value of t-test, Res. df residualdegrees of freedom, Res. Dev. residual deviance, Adj. D2 adjustedD2 following Guisan and Zimmermann (2000), F v2 tests used toevaluate if selected predictors explain a significant fraction of thedeviance, Prob. (F) P value of v2 tests

Species Variable Coef. (SE) t Prob. (t) Res.df

Res.Dev.

Adjus. D2 F Prob. (F)

Group 1: No temperature requirementMur Null 11 0.3411

Temp. – – – 10 0.3190 0.9681 0.5684 0.4683Sax Null 11 0.4734

temp. – – – 10 0.4503 0.0488 0.5055 0.4933Semp Null 11 0.6114

Temp. – – – 10 0.5147 0.1581 1.8059 0.2087Senb Null 11 0.1846

Temp. – – – 10 0.1052 0.4303 2.5361 0.1553Sol Null 11 0.8298

Temp. – – – 10 0.6917 0.9308 2.0797 0.1798Arm Null 35 3.1354

Alt. �1.056 (0.247) �4.2656 0.0001 34 2.1074 0.3279 18.2298 0.0001Bis Null 20 3.1957

Alt. �0.847 (0.439) �1.9311 0.0677 19 2.6607 0.1674 4.0113 0.0597Ran Null 23 1.8630

Alt. �0.6568 (0.2363) �2.7798 0.0104 22 1.3971 0.2501 7.9172 0.0101Fesc Null 35 8.7073

Pobl. 1.0685 (0.2431) 4.3943 0.0001 34 5.3761 0.3826 22.9723 0.0000Thy Null 35 6.1207

Alt. 2.4313 (0.4459) 5.4520 0.0000 34 3.9351 0.3382 25.3992 0.0000Pobl. 0.8255 (0.2494) 3.3101 0.0021 33 2.9699 0.5005 11.2168 0.0020

Group 2: High temperature requirement (20�C)Alli Null 11 1.7839

Temp. 1.5189 (0.2501) 6.0731 0.0001 10 0.3427 0.8079 53.0164 0.0000Senp Null 11 2.7624

Temp. 1.1500 (0.1873) 6.1399 0.0001 10 0.5392 0.8048 44.1109 0.0000Hie Null 43 6.9432

Temp. 0.3594 (0.1576) 2.2795 0.0275 42 6.1359 0.1163 5.6993 0.0215Sbor Null 35 26.3755

Temp. 2.4412 (0.3886) 6.2808 0.0000 34 7.1796 0.7278 71.0145 0.0000Jas Null 47 28.8022

Temp. 2.2152 (0.2023) 10.9472 0.0000 46 6.2726 0.7775 196.4671 0.0000Alt. 0.5369 (0.1621) 3.3122 0.0002 45 4.9575 0.8241 11.4680 0.0015

Min Null 35 6.5573Year 1.0425 (0.3301) 3.1578 0.0004 34 4.6897 0.2415 29.0067 0.0000Temp. 1.0546 (0.2156) 4.8909 0.0000 33 2.8502 0.5390 28.5711 0.0000Alt. 1.9933 (0.6910) 2.8844 0.0066 32 2.0757 0.6643 12.0287 0.0015

Group 3: Medium temperature requirement (15�C)Silc Null 19 7.7464

Temp. 0.6694 (0.1939) 3.4511 0.0025 18 4.5027 0.4187 14.3020 0.0014Jur Null 34 10.5831

Alt. �2.4180 (0.6211) �3.8929 0.0004 33 7.6814 0.2288 19.7693 0.0001Pobl. 0.7377 (0.3477) 2.1213 0.0411 31 5.0465 0.4934 4.5900 0.0401Temp. 0.6445 (0.1791) 3.5972 0.0010 32 5.7202 0.4257 13.3613 0.0009

440

Page 9

periods even though these plants usually suffer a veryintense summer drought along or immediately afterripening and dispersal (Gavilan et al. 2001).

Response to temperature

As a general rule alpine plants do not differ from low-land plants in their temperature requirements for ger-mination (Korner 1999), so most of them preferrelatively high temperatures (Mooney and Billings 1961;

Billings and Mooney 1968). This also seems to be thecase for Mediterranean orophyllous plants. Most speciesshow temperature indifference (Group 1, 50% of cases)or warm-preference (Group 2, 30%) for germination,but no cold-adaptation. Surprisingly, this behaviourdiffers from that of typical Mediterranean plants, forwhich germination at low temperatures is a widely ex-tended trait (Bell et al. 1993; Escudero et al. 1997;Thanos et al. 1995). Logically, this must be a conse-quence of harsh mixture of conditions of high Medi-terranean habitats. Alternating conditions are alsoreported to offer very good results for many alpinespecies (Amen 1966; Bliss 1971), and we have also foundsuch a type of response in the two tested Silene species(although only five species were tested with this tem-perature regime).

Fig. 3 Interannual variability(2001–2002) in finalgermination percentage forJasione crispa subspp. centralisand Minuartia recurva subspp.bigerrensis. Vertical linesstandard deviations. * P £ 0.05,n.s. non-significant differences

Fig. 4 Final germination percentage before and after submittingseeds to cold-wet stratification treatment. Vertical lines standarddeviations. ** P £ 0.001, * P £ 0.005, n.s. non-significant differ-ences

441

Page 10

Seeds readily germinate in almost all cases (Figs. 2, 5)as similarly reported for many alpine species around theworld (Baskin and Baskin 1998). Since alpine ecosystemsare typically subjected to short growing seasons, thisstrategy could be claimed to ensure enough time forplant establishment (Chambers et al. 1987; Angosto andMatilla 1994; Korner 1999). Under harsh Mediterraneanclimatic conditions, where topsoil remains moistened foronly few weeks after snowmelt, the number of days re-quired for germination may be of special importance.

Altitudinal, interpopulation and interannual variability

Alpine plants show a very great variability in the ger-mination behaviour which sometimes has been attrib-uted to environmental conditions but not alwaysassociated with habitat features (Bliss 1971; Urbanskaand Schutz 1986; Chambers 1989). Our results clearlyagree with these generalities. Variation in germination ofdifferent species had been observed in relation to altitudeby several authors (Holm 1994; Vera 1997). However,we have not found a consistent pattern in germinationrelated to altitude variation. Among those tested, somespecies showed better results at higher (25%), and othersat lower altitudes (33%). Population variation contrib-utes to the explanation of germination behaviour inthree of five species. The year variable accounts for the

greatest proportion of variability in one of two speciestested. The variability of germination characteristicscould be interpreted as one of the most important sur-vival strategies for species growing under unpredictableenvironmental conditions (Gutterman 1994; Kigel 1995)and will reduce the risk of seedlings being subjected topoor growing conditions due to the establishment ofintense competition hierarchies.

Germination after cold-wet stratification

We found stratification enhancement in seven Mediter-ranean mountain species (including S. vicentei subspp.paui, which emerged in profusion during treatment),whereas two species presented similar germination per-centages and two others lost viability. Cold stratificationis known to improve germination percentage in manyalpine and non-alpine species of eastern Europe andNorth America (Bell and Bliss 1980; Marchand andRoach 1980; Bewley and Black 1982; Baskin and Baskin1998). Cavieres and Arroyo (2000) also proved effectiveinduction of Phacelia secunda seed germination bymeans of cold-wet stratification in the Mediterranean-type climate of the Andes in Chile, with longer treatmentperiods required for populations from higher elevations.Additionally, cold stratification may level out inter-population differences in germination patterns (Schutzand Milberg 1997; Milberg and Andersson 1998).Unfortunately, we tested just one locality per taxon, sowe cannot establish altitudinal and/or population com-parisons.

Fig. 5 Modelled germination curves by the Kaplan–Meier methodbefore and after submitting seeds to cold-wet stratification. n.s.non-significant differences between curves (P>0.0001)

442

Page 11

Seed coat structure has previously pointed out asa possible cause of seed dormancy in alpine plants(Urbanska et al. 1979; Zuur-Isler 1982). Luzula hispanicaseeds are covered by a hard waxy layer that turnsgelatinous when wetted. Amen (1965) reported the samekind of structure for L. spicata, and proved its role in thetotal dormancy of the species. Following his recom-mendation we tried a manual scarification on the radiclearea, obtaining a similar germination percentage to thatreached by cold-wet stratification.

It is important to note that all large-seeded specieswere unable to germinate after stratification. Seeds ofJurinea humilis and Solidago virgaurea subspp. fallit-ti-rones died during stratification due to severe fungal at-tack, whereas Biscutella laevigata subspp. gredensis andSenecio pyrenaicus subspp. carpetanus seeds lost theirviability during treatment. In contrast, all small-seededspecies improve their final germination after stratifica-tion, except Saxifraga pentadactylis subspp. wilkommi-ana. Similar results were found by Schwienbacher andErschbamer (2002), who showed that most species of theCentral Alps tested had high germination rates after ashort cold treatment, with S. oppositifolia among theexceptions. They argued that a brief cold storage mightbe insufficient to break the deep dormancy of this spe-cies. These results partially agree with the generalassumption that large heavy-weight seeds are mainlyshort-lived, whereas small low-weight seeds are adaptedto maintain persistent seed banks (Thompson et al.1993). However, results must be taken with caution be-cause contradictory results between laboratory and fieldexperiments have been reported in this sense (seeSchwienbacher and Erschbamer 2002).

Conclusions

The absence of a genuine alpine germination strategyhas been demonstrated in high mountain species(Korner 1999). High mountain Mediterranean plants donot seem to differ in germination characteristics fromother alpine plants but there is a relatively high numberof species that are ready to germinate without anytreatment. Temperature optima differ from those oftypical Mediterranean lowland plants. Germination re-sponse can vary between altitudes, sites and years, butwithout clear trends among species. There is a consid-erable proportion of species with different levels ofdormancy that can be successfully broken by cold-wetstratification. Type and duration of dormancy, as well asgenotypic versus phenotypic causes of variation, and itsimplications on recruitment from seed banks need to beclarified.

Acknowledgements We would like to thank graduate studentsMarıa Gomez and Silvia Gonzalez for assistance in the laboratory.Arantzazu Lopez de Luzuriaga reviewed an earlier draft of themanuscript. This research was financed by the Spanish Ministry ofScience and Technology (Project REN 2003-03366).

References

Acharya SN (1989) Germination response of two alpine grassesfrom the Rocky Mountains of Alberta. Can J Plant Sci69:1165–1177

Amen RD (1965) Seed dormancy in the alpine rush, Luzula spicataL. Ecology 46:361–364

Amen RD (1966) The extent and role of seed dormancy in alpineplants. Q Rev Biol 41:271–281

Angosto T, Matilla AJ (1994) Modifications in seeds of Festucaindigesta from two different altitudinal habitats. Seed SciTechnol 22:319–328

Baskin CC, Baskin JM (1998) Seed ecology, biogeography, andevolution of dormancy and germination. Academic, San Diego

Bell KL, Bliss LC (1980) Plant reproduction in high arctic envi-ronment. Arctic Alpine Res 12:1–10

Bell DT, Plummer JA, Taylor SK (1993) Seed germination ecologyin southwestern Australia. Bot Rev 59:24–73

Bewley JD, Black M (1982) Physiology and biochemistry of seedsViability, dormancy, and environmental control, vol 2.Springer, Berlin Heidelberg New York

Billings WD, Mooney HA (1968) The ecology of arctic and alpineplants. Biol Rev 43:481–529

Bliss LC (1971) Arctic and alpine plant life cycles. Annu Rev EcolSyst 2:405–438

Bonner FT (1990) Storage of seeds, potential and limitations forgermplasm conservation. For Ecol Manage 35:35–43

Brown RA, Johnston RS (1979) Revegetation of disturbed alpinerangelands. In: Johnson DA (ed) Special management needs ofalpine ecosystem Range Science Series No 5. Society for RangeManagement, Denver, Colo., pp 78–94

Callaway RM, Brooker RW, Choler P, Kikvidze Z., Lortie CJ,Michalet R, Paolini L, Pugnaire FI, Newingham B, Asche-houg ET, Armas C, Kikodze D, Cook BJ (2002) Positiveinteractions among alpine plants increase with stress. Nature417:844–848

Cavieres LA, Arroyo MTK (2000) Seed germination response tocold stratification period and thermal regime in Phacelia se-cunda (Hydrophyllaceae). Plant Ecol 149:1–8

Cerabolini B, De Andreis R, Ceriani RM, Pierce S, Raimondi B(2004) Seed germination and conservation of endangered spe-cies from the Italian Alps, Physoplexis comosa and Primulaglaulescens. Biol Conserv 117:351–356

Chambers JC (1989) Seed viability of alpine species, variabilitywithin and among years. J Range Manage 42:304–308

Chambers JC (1997) Restoring alpine ecosystem in the WesternUnited States, environmental constrains, disturbance charac-teristics, and restoration success. In: Urbanska KM et al (eds)Restoration ecology and sustainable development. CambridgeUniversity Press, Cambridge, pp 161–187

Chambers JC, McMahon JA, Brown RW (1987) Germinationcharacteristics of alpine grasses and forbs, a comparison ofearly and late serial dominants with reclamation potential.Reclam Reveg Res 6:235–249

Escudero A, Carnes LF, Perez-Garcıa F (1997) Seed germinationof gypsophytes and gypsovags in semi-arid central Spain. J AridEnviron 36:487–497

Escudero A, Gimenez-Benavides L, Iriondo JM, Rubio A (2005)Soil amelioration under vegetated patches in a high mountainMediterranean community. Arct Antarct Alp Res (in press)

European Community (1992) Council Directive 92/43/EEC on theconservation of natural habitats and of wild fauna and flora

Gavilan RG, Fernandez-Gonzalez F, Rivas-Martınez S (2001)Variaciones bioclimaticas en Madrid: un estudio sobre cambioclimatico local. In: Gomez F, Mota JF (eds) Vegetacion ycambos climaticos. Servicio de publicaciones de la Universidadde Almerıa, Almerıa, pp 243–256

Gavilan RG, Sanchez-Mata D, Escudero A, Rubio A (2002)Spatial structure and interactions in Mediterranean highmountain vegetation (Sistema Central, Spain). Isr J Plant Sci50:217–228

443

Page 12

Gomez-Campo C (1987) Libro rojo de las especies vegetales ame-nazadas de Espana peninsular e Islas Baleares. ICONA,Madrid

Grabherr G, Gottfried M, Pauli H (1994) Climate effects onmountain plants. Nature 369:448

Guisan A, Zimmermann NE (2000) Predictive habitat distributionmodels in ecology. Ecol Model 135:147–186

Guisan A, Edwards TC Jr, Hastie T (2002) Generalized linear andgeneralized additive models in studies of species distributions:setting the scene. Ecol Model 157:89–100

Gutterman Y (1994) Strategies of seed dispersal and germination inplants inhabiting deserts. Bot Rev 60:373–425

Hendry GA, Grime JP (1993) Methods in comparative plantecology. Chapman and Hall, London

Holm SO (1994) Reproductive patterns of Betula pendula and B.pubescens coll. along a regional altitudinal gradient in northernSweden. Ecography 17:60–72

Kigel J (1995) Seed germination in arid and semiarid regions. In:Kigel J, Galili G (eds) Seed development and germination.Marcel Dekker, New York, pp 645–699

Korner C (1999) Alpine plant life. Springer, Berlin Heidelberg NewYork

Marchand PJ, Roach DA (1980) Reproductive strategies of pio-neering alpine species, seed production, dispersal, and germi-nation. Arct Alp Res 12:137–146

MathSoft (1999) S-Plus 2000. Guide to statistics, vol I. Seattle,Washington

McCullagh P, Nelder JA (1989) Generalized linear models. Chap-man and Hall, New York

McDonough WT (1969) Effective treatments for the induction ofgermination in mountain rangeland species. Northwest Sci43:18–22

McDonough WT (1970) Germination of 21 species collected from ahigh-elevation rangeland in Utah. Am Midl Nat 84:551–554

Milberg P, Andersson L (1998) Does cold stratification level outdifferences in seed germinability between populations. PlantEcol 134:225–234

Mooney HA, Billings WD (1961) Comparative physiologicalecology of arctic and alpine populations of Oxyria digyna. EcolMonogr 31:1–29

Palacios D, de Andres N, Luengo E (2003) Distribution andeffectiveness of nivation in Mediterranean mountain, Penalara(Spain). Geomorphology 54:157–178

Penuelas J, Filella I, Comas P (2002) Changed plant and animal lifecycles from 1952 to 2000 in the Mediterranean region. GlobalChange Biol 8:531–544

Pelton J (1956) A study of seed dormancy in eighteen species ofhigh altitude Colorado plants. Butler Univ Bot Stud 18:74–84

Perez-Garcıa F, Iriondo JM, Gonzalez-Benito ME, Carnes LF,Tapia J, Prieto C, Plaza R, Perez C (1995) Germination studiesin endemic plant species of the Iberian Peninsula. Isr J Plant Sci43:239–247

Pyke DA, Thompson JN (1986) Statistical analysis of survival andremoval rate experiments. Ecology 67:240–245

Rivas-Martınez S (1963) Estudio de la vegetacion y flora delas Sierras de Guadarrama y Gredos. An Inst Bot Cavanill21:5–330

Rivas-Martınez S, Canto P, Fernandez-Gonzalez F, Molina JA,Pizarro JM, Puente E (1999) Sinopsis of the Sierra de Gua-darrama vegetation. Itinera Geobot 13:189–260

Sanchez-Herrera F (2000) La restauracion ambiental de la Antiguaestacion de esquı de Valcotos y el programa de conservacion delParque Natural de Penalara. In: Segundas Jornadas Cientıficasdel Parque Natural de Penalara y del Valle del Paular. Direc-cion General del Medio Natural, Madrid, pp 19–25

Sanz-Elorza M, Dana ED, Gonzalez A, Sobrino E (2003) Changesin the high-mountain vegetation of the Central Iberian Penin-sula as a probable sign of global warming. Ann Bot 92:273–280

Schemske DW, Husband BC, Ruckelshaus CG, Goodwillie C,Parker IM, Bishop JG (1994) Evaluating approaches to theconservation of rare and endangered plants. Ecology 75:584–606

Schutz W, Milberg P (1997) Seed dormancy in Carex canescens,regional differences and ecological consequences. Oikos 78:420–428

Schutz W, Rave G (1999) The effect of cold stratification and lighton the seed germination of temperate sedges (Carex) fromvarious habitats and implications for regenerative strategies.Plant Ecol 144:215–230

Schwienbacher E, Erschbamer B (2002) Longevity of seeds in aglacier foreland of the Central Alps—a burial experiment. BullGeobot Inst ETH Zurich 68:63–71

Spector P (1994) An introduction to S and S-PLUS. Duxbury,Belmont

Thanos CA, Kadis CC, Skarou F (1995) Ecophysiology of germi-nation in the aromatic plants thyme, savory and oregano(Labiatae). Seed Sci Res 5:161–170

Thompson PA, Smith RD, Dickie JB, Sanderson RH, Probert RJ(1981) Collection and regeneration of populations of wildplants from seed. Biol Conserv 20:229–245

Thompson K, Band SR, Hodgson JG (1993) Seed size and shapepredict persistence in soil. Funct Ecol 7:236–241

Urbanska KM (1986) Behaviour of alpine plants and high altituderevegetation research. In:Schuster MA, Buck RH (eds) Pro-ceedings of the high altitude revegetation workshop no 7. ForthCollins, Colo., pp 215–226

Urbanska KM, Schutz M (1986) Reproduction by seed in alpineplants and revegetation research above timberline. Bot Helvet96:43–60

Urbanska KM, Schwank O, Fossati A (1979) Variation withinLotus corniculatus L. s. l. from Switzerland, II. Reproductivebehaviour of L. alpinus (DC) Schleicher. Berlin Geobot InstETH 46:62–85

Vare H, Lampinen C, Humphries C, Williams P (2003) Taxonomicdiversity of vascular plants in the European alpine areas. In:Nagy L et al (eds) Alpine biodiversity in Europe. Ecologicalstudies, vol 167. Springer, Berlin Heidelberg New York, pp133–148

Venables WN, Ripley BD (1998) Modern applied statistics withS-PLUS. Springer, Berlin Heidelberg New York

Vera ML (1997) Effect of altitude and seed size on germination andseedling survival of heathland plants in north Spain. Plant Ecol133:101–106

Zuur-Isler D (1982) Germinating behaviour and early life phase ofsome species from alpine serpentine soils. Berlin Geobot InstETH 49:76–107

444