Pk

C

T

R

A

erfft

1sni

f

Z

dSPwnbS(

d

1d

Basic and Applied Ecology 12 (2011) 215–226

ost-dispersal impact on seed fate by livestock trampling – A gap ofnowledge

. Faust, C. Eichberg1, C. Storm, A. Schwabe∗

echnische Universität Darmstadt, Department of Biology, Vegetation Ecology, Schnittspahnstr. 4, D-64287 Darmstadt, Germany

eceived 13 October 2010; received in revised form 11 February 2011; accepted 14 February 2011

bstract

Sheep grazing is an important management tool in threatened sandy grassland of the temperate zone. Besides direct grazingffects, previous studies have shown benefits of seed dispersal, but little is known about post-dispersal processes. We studied theole of sheep trampling for the post-dispersal fate of seeds embedded in sheep and rabbit dung and hypothesized a positive impactor the development of seedlings as a consequence of cracking the dung pellets. Sheep and rabbit dung samples were collectedrom species-rich sandy grasslands, and their seed potential was assessed in a climate room. In a factorial field experiment weested the effects of trampling and dung type on seedling emergence and fruiting success.

Seedling emergence in the field was only 5% (sheep dung) or 7% (rabbit dung) of the potential without trampling but 18 or4% with trampling. Plots with trampled sheep or rabbit dung both showed significantly more seedlings (3.6- or 2.1-fold), morepecies (2.4- or 1.9-fold), more fruiting individuals (3.9- or 2.6-fold) and more fruiting species (2.1- or 1.9-fold) compared toon-trampled dung plots. Both target as well as non-target species profited from trampling, but the proportion of target speciess clearly increased by trampling and graminoid competitors did not reach fruiting stage.

Sheep play a multifaceted role in dispersal processes: after endozoochoric transport they act as a sort of ‘gardener’ not onlyor sheep-dispersed seeds, but also for those dispersed by rabbits.

usammenfassung

Schafbeweidung ist ein wichtiger Beitrag zum Management bedrohter Sandrasen der gemäßigten Zone. Abgesehen vonirekten Beweidungseffekten, wie z. B. die Entstehung von Lücken durch extensive Viehbeweidung, konnten vorangegangenetudien die Bedeutung von Schafen für die Samenausbreitung (Endo-, Epizoochorie) zeigen, aber dennoch ist wenig überrozesse nach der Ausbreitung bekannt wie beispielsweise Effekte durch Trampeln. Wir untersuchten die Rolle der Hufein-irkung (,,trampling“) von Schafen auf das Schicksal von Samen, die in Schaf- und Kaninchendung eingeschlossenen waren,

ach deren Ausbreitung und stellten die Hypothese auf, dass die Etablierung von Keimlingen begünstigt wird als Folge des Auf-rechens der Dungpellets durch Schafhufe. Schaf- und Kaninchendung wurde in gefährdeten Sandrasen gesammelt und derenamenpotential in einer Klimakammer getestet. In einem faktoriellen Freilandexperiment wurden die Effekte des Trampelnsmittels Schafhuf-Replikaten) und des Dungtyps auf das Keimlingsaufkommen und den Fruchterfolg festgestellt.

Das Keimlingsaufkommen im Freiland entsprach für beide Dungtypen zu nur 5 (Schafdung) bzw. 7% (Kaninchendung)em Potential ohne Trampeln, mit Trampeln lagen die Werte bei 18 bzw. 14%. Flächen mit Trampel-Wirkung auf Schaf- oder

∗Corresponding author. Tel.: +49 6151163302; fax: +49 6151164630.E-mail address: schwabe@bio.tu-darmstadt.de (A. Schwabe).

1 Present address: Faculty of Geography and Geosciences, Geobotany, University of Trier, Behringstr. 21, 54296 Trier, Germany.

439-1791/$ – see front matter © 2011 Gesellschaft für Ökologie. Published by Elsevier GmbH. All rights reserved.oi:10.1016/j.baae.2011.02.009

2

KIpu

“w©

K

I

vgoqCS2rsV&mpeei&

lptsipellsedgpdaa&tslba

16 C. Faust et al. / Basic and Applied Ecology 12 (2011) 215–226

aninchendung zeigten signifikant mehr Keimlinge (3,6- oder 2,1-fach), mehr Arten (2,4- oder 1,9-fach) und mehr fruchtendendividuen (3,9- oder 2,6-fach) und fruchtende Arten (2,1- oder 1,9-fach) verglichen mit den Flächen ohne diesen Effekt. Esrofitierten nicht nur Zielarten durch das Trampeln, auch Nicht-Zielarten, aber der Anteil der Zielarten erhöhte sich deutlich,nd konkurrenzstarke Graminoide erreichten nicht das Fruchtstadium.

Schafe spielen eine facettenreiche Rolle in Ausbreitungsprozessen: nach dem endozoochoren Transport haben sie eine ArtGärtnerfunktion” nicht nur für Schaf-ausgebreitete Samen, sondern auch für die Samen, die durch Kaninchen ausgebreiteterden.2011 Gesellschaft für Ökologie. Published by Elsevier GmbH. All rights reserved.

rient-p

sAstdZir2

zqgttpia

M

S

emaoH4f(oocstgAe

eywords: Dung; Endozoochory; Facilitation; Fragmentation; Nut

ntroduction

Seed dispersal and re-colonization processes of plant indi-iduals are decisive factors for population dynamics inrassland communities (Foster & Tilman 2003). In manypen grazed habitats, herbivorous mammals transport largeuantities of seeds (e.g., Malo & Suárez 1995a, 1995b;osyns, Claerbout, Lamoot, & Hoffman 2005; Eichberg,torm, & Schwabe 2007; Bakker, Galvez Bravo, & Mouissie008). A growing number of studies have focused on theole of herbivores to overcome seed limitations in a con-ervation and restoration context (e.g., Couvreur, Christiaen,erheyen, & Hermy 2004; Mouissie 2004; Wessels-de WitSchwabe 2010). By dispersing seeds over long distances,

oving livestock provides survival opportunities for smalllant populations in fragmented landscapes (e.g., Cosynst al. 2005; Mouissie, Vos, Verhagen, & Bakker 2005; Ozingat al. 2009), but to assess the effectiveness of seed dispersal its crucial to study post-dispersal processes (Eichberg, Storm,

Schwabe 2005; Ramos, Robles, & Castro 2006).As a model to study post-dispersal processes we used

ivestock dung with an endozoochorously dispersed seedotential. Our hypothesis is that on pastures one impor-ant factor for the post-dispersal fate of dung-embeddedeeds is trampling, which has a high-frequency disturbancempact on pastures (Hobbs 2006). Previous studies on tram-ling effects in pastures mainly focused on the gap-creatingffect of hooves (e.g., Stammel & Kiehl 2004). It seemedikely that trampling also affects the germination and estab-ishment of dung-embedded seeds, but evidence for this istill lacking. We hypothesized that the post-dispersal fate ofndozoochorous seeds could be improved by cracking theung pellets. Sheep dung pellets which are deposited in dryrasslands during the spring–summer season (main grazingeriod) dry out quickly and can persist over years withoutisintegrating (Eichberg et al. 2007). Hooves of ungulateslso crack dung deposited by other herbivore species, suchs rabbits, as was observed in our area (Faust, Süss, Storm,

Schwabe 2011). Especially in the case of sheep it is likelyhat trampling of dung has a significant impact on the grazedystem, for the following reasons: first, sheep are kept in

arge flocks, generating a high trampling density. Second,esides concentrating their dung unintentionally in specialreas (e.g., resting places) sheep scatter it while grazing. Con-

tsa

oor grassland; Rabbit; Sheep; Seedling emergence; Target species

equently the grazed area will be affected to a large extent.s an example of a grazed semi-natural grassland type, we

tudied threatened sheep-grazed inland sand ecosystems in aemperate region. The high phytodiversity of these systemsepends on continuous disturbance dynamics (Süss, Storm,ehm, & Schwabe 2004; Eichberg et al. 2007). Previous stud-

es have shown a high endozoochorous potential of sheep andabbits (Pakeman, Engelen, & Attwood 1999; Eichberg et al.007; Wessels & Schwabe 2008).

We focused on the following questions: (1) Which endo-oochorously dispersed plant species are contained in whatuantities in sheep and rabbit dung collected from sandyrasslands? (2) What proportion of this seed potential is ableo establish itself in the field after 1 or 2 years? (3) Is sheeprampling on dung pellets of sheep and rabbits a facilitativeost-dispersal process for dung-embedded seeds? (4) Whats the proportion of target species contained in the dung andmong the emerging/fruiting individuals?

ethods

tudy sites

Our study was conducted on two sites of inland sandcosystems in the northern upper Rhine valley (Hesse, Ger-any), which were part of one large area ca. 70 years

go. Both sites are characterised by a mainly calcare-us and nutrient-poor soil and protected by the EU 92/43abitat Directive. The area ‘Griesheimer Düne’ (8◦39′E,9◦53′N; 45 ha; hereafter ‘GD’) served as a source areaor the dung. The area ‘Ehemaliger August-Euler-Flugplatz’8◦35′E, 49◦51′N; 71 ha; ‘AEF’) was used for the installationf experimental plots. The studied vegetation of GD consistedf a well-developed species-rich grassland (Allio-Stipetumapillatae), whereas the vegetation of AEF was made up ofpecies-poor grasslands (consolidated Armerio-Festucetumrachyphyllae stands) which were dominated by competitiverasses (Cynodon dactylon, Poa angustifolia). The topsoil ofEF is mainly slightly acidic but in the case of disturbances,

.g., by burrowing rabbits, calcareous material is transferred

o the topsoil. The study region is managed by an extensiveheep grazing regime: short but intensive grazing periodsre followed by long periods for re-growth of plants (pad-

pplied E

d1a2

D

mdsawvrtTlptrtdwu

E

pVf1(ssw(dwwfoH

F

E

eemesae

ctDitP(esbBtgsOata

htpewt

cp

V

idwdbrttthft

d(

D

C. Faust et al. / Basic and A

ock size: <1–2 ha, flock size: 500–800 animals, duration:–9 days). The annual precipitation is 629 mm and the meannnual temperature is 11.1 ◦C (data from Frankfurt airport,000–2009, Deutscher Wetterdienst).

ung collection

In August 2007, dung samples were taken from four tamedale Rhoen sheep, which grazed for 2 weeks in a 2-ha pad-

ock in the GD area. Sheep dung was collected daily from theoil surface. Rabbit dung was collected in the same area anddditionally up to 150 m away because rabbit dung densitiesere too low in the paddock. The similarity of the adjacentegetation and the vegetation of the paddock was proved byelevés. These relevés were used also as a reference to con-extualize actual vegetation and endozoochorous seed pools.o allow comparability of the two dung types, only fresh pel-

ets were collected. Rabbit pellet age was estimated by theellets’ colour, surface structure and degree of decomposi-ion. During the sampling period seeds of most species wereipe. Before grazing, all vascular plant species occurring inhe paddock (presence/absence) and their phenological con-ition (vegetative/flowering/fruiting) were recorded. Dungas washed with tap water, dried at 40 ◦C and stored dryntil utilization.

xperiment in climate room

To test the potential seed availability of the dung sam-les we used the sample concentration method of Ter Heerdt,erweij, Bekker, and Bakker (1996), which was adapted

or dung samples by Wessels and Schwabe (2008). In total,920 g (dry weight) crumbled and concentrated sheep dungDS) was divided into 48 samples of 40 g dry weight each andpread out in trays. The same was done with the rabbit dungamples (DR). All trays were established in a climate roomith a day/night air temperature of 20/5 ◦C and 16 h light

PPFD 300–800 �mol m−2 s−1) per day. Water was providedaily up to field capacity. The samples were exposed for 16eeks: two 6-week intervals of exposure in the climate roomere interrupted by a 4-week interval of cold stratification

rom 0 to 4 ◦C. Seedling emergence was recorded through-ut with the help of determination literature (Muller 1978;anf 1999).

ield experiment

xperimental designIn November 2007, we started a 2-year factorial field

xperiment at the AEF site. Three 14 m × 14 m sheep + rabbitxclosures – established in 1999 – were used as an experi-ental area. Within each exclosure two 5 m × 5 m plots were

stablished, one for preliminary vegetation analyses (whichhowed a high degree of similarity of the three exclosures)nd the other for the establishment of plots for the tramplingxperiment.

g2fs

cology 12 (2011) 215–226 217

The following two factors were tested: dung (D0,ontrol; DS, sheep; DR, rabbit) and trampling (T0, con-rol; T1, trampling). Thus, six treatments (D0T0/D0T1/

ST0/DST1/DRT0/DRT1) were applied, eight-fold replicatedn each case, and randomly distributed within the plot area inhe three exclosures (completely randomised block design).lot size was 15 cm × 15 cm according to Eichberg et al.2007). Before treatment, the soil surface of each plot wasxperimentally disturbed by spades and hand rakes in order toimulate disturbed areas which are more likely to be colonisedy new plant individuals (Bullock et al. 2002; Cosyns,ossuyt, Hoffmann, Vervaet, & Lens 2006). We removed

he complete above-ground vegetation and as much below-round plant material as possible without losing too muchoil. The total disturbed area of each plot was 25 cm × 25 cm.n each dung plot, 40 g dung pellets were distributed asclosed single-layer (in total 1920 g dry weight per dung

ype). Plot-surrounding vegetation was clipped in spring andutumn.

Experimental trampling was done with 12 replicas of sheepooves, which were screwed on a wooden cube, to obtain arampling stamp of approximately plot size. Dung was tram-led in a manner of stamping for as long as it took to crackvery pellet. Plots without dung but with trampling treatmentere trampled with the same intensity. Hooves were cleaned

horoughly between the trampling procedures.To prevent disturbance by dung beetles we installed wire

ages (mesh size 6 mm) from autumn till spring above eachlot, and plastic lawn borders around each plot.

egetation samplingOn each plot, all soil- and dung-borne plants were recorded

n spring and autumn of 2008 and 2009 as presence–absenceata (‘Field data set A’). In addition, dung-borne seedlingsere counted (‘Field data set B’). Seedlings that emerged andied within the periods between counts could not be recorded,ut germination events between the counts were relativelyare. For a correct detection as ‘dung-borne’ the origin ofhe seedlings was observed very carefully: on plots withoutrampling treatment dung-borne seedlings should be rooted inhe intact pellets, on plots with trampling treatment seedlingsad to root in the layer of the cracked and clumped pellets. Toollow the individual fate of dung-borne seedlings we markedhem with sticks.

The data of both years were pooled within each of theata sets. Nomenclature follows Wisskirchen and Haeupler1998).

ata analysis

The qualitative (TSRqual) and quantitative ratio of tar-

et species (TSRquant; Eichberg, Storm, Stroh, & Schwabe010) was calculated to elucidate whether target species areacilitated due to endozoochory and trampling or not. Targetpecies are considered as species with main occurrence in the

2 pplied Ecology 12 (2011) 215–226

cgrow

(U2ABetefiwo

tFFSsaswa

R

Ss

asfb

C

sa

stisVsss

0

1

2

3

4

5

6

7

8

9

DRT1DRT0DST1DST0D0T1D0T0

Nu

mb

er o

f sp

eci

es

Treatment

D0T0 D0T1 DST0 DST1 DRT0 DRT1

aab

bccc

a

F-valueDen dfNum dfFactor p<0.000112.591382Dung

0.00378.731381Trampling

0.13532.031382Dung × trampling

Fig. 1. Mean number of dung- and soil-borne species(presence–absence data) per plot (No. of plots per treatment:24; plot size: 225 cm2) in the field experiment and the results ofthe mixed-linear-model analysis of influences of the independentvariables dung and trampling and/or interaction effects. Meanssharing the same letter are not significantly different (p > 0.05).Num df: degrees of freedom Numerator, Den df: degrees of freedomDi

n11ni

Tcl

pstsssdv

F

D

ht

18 C. Faust et al. / Basic and A

lasses Festuco-Brometea (‘steppes, rocky steppes and sandyrasslands of the sub-continental temperate and sub-borealegions’), or Koelerio-Corynephoretea (‘pioneer vegetationn primitive soils and rocky outcrops in regions with mildinter climate’) (according to ETC/BD 2008).The following data were analysed by mixed linear models

SAS 9.2, PROC GLIMMIX, SAS Institute Inc., Cary, NC,SA; Littell, Miliken, Stroup, Wolfinger, & Schabenberger006): (1) climate room data (DS/DR), (2) field data set

(D0T0/D0T1/DST0/DST1/DRT0/DRT1), (3) field data set(DST0/DST1/DRT0/DRT1). For the latter two models, the

xclosure was included as random effect. The influence onhe following dependent variables was assessed: number ofmerging seedlings, number of emerging species, number ofruiting individuals, number of fruiting species, both TSRndices and all single species. Tukey-adjusted post hoc testsere carried out to test for effects of treatments (‘slice’ optionf PROC GLIMMIX).

For the calculation of the degrees of freedom, we selectedhe Kenward-Roger approximation (Schaalje, McBride, &ellingham 2002; Arnau, Bono, & Vallejo 2009). Vallejo,ernández, Herrero, and Conejo (2004) and Jacqmin-Gadda,ibillot, Proust, Molina, and Thiébaut (2007) were able tohow that mixed linear models using this method are robustgainst deviation from normal distributions. Nevertheless, thetudentised residuals and conditional studentised residualsere examined for normality by means of graphical display;nearly Gaussian distribution could be ascertained.

esults

pecies pool of the actual vegetation of theource area (Table 1)

In total, 59 vascular plant taxa were present in the sourcerea (GD), 37 of which were in fruiting condition (17 targetpecies, 20 non-target species; TSRqual: 0.46). Among theruiting species most were short-lived: 18 annuals (49%), 5iennials (13%) and 14 perennials (38%).

limate-room data

In sheep dung we found a total of 26 species and 101eedlings per 100 g air-dry dung, in rabbit dung 31 speciesnd 86 seedlings per 100 g air-dry dung emerging.

As shown in Tables 1 and 2 the spectra of species fromheep and rabbit dung were to a high degree complemen-ary. Nine species showed a significantly higher seed contentn sheep than rabbit dung; among them the three dominantpecies were tall-growing (>50 cm): Oenothera biennis s.l.,

erbascum phlomoides and Amaranthus retroflexus. Eightpecies were significantly more abundant in rabbit than inheep dung, all of which were small- or medium-growingpecies (≤50 cm). Short-lived species represented the domi-

wtai

enominator, p: level of significance. Error bars = 95% confidencenterval, CI.

ant life form in both dung types: sheep dung: 58% annuals,2% biennials and 30% perennials, rabbit dung: 61% annuals,3% biennials, and 26% perennials. Furthermore, predomi-antly small-seeded species (size groups 1 or 2) were foundn both dung types.

The mean TSRqual was in both dung types nearly identical.he mean TSRquant, however, showed slightly but signifi-antly higher values in rabbit dung, although on a very lowevel (0.10 vs. 0.13).

With reference to the species pool of the sampled sheepaddock we found 50% (9 species) of the fruiting annualpecies in sheep dung again (biennials: 60%, 3 species), inhe case of rabbit dung 56% (10 species; biennials: 80%, 4pecies). The percentages of fruiting perennials were 60% (3pecies, sheep dung) or 14% (2 species, rabbit dung). Fivepecies emerged from sheep dung and 12 species from rabbitung samples which were neither recorded in fruiting nor inegetative condition in the paddock.

ield experiment

ung- and soil-borne individualsSpecies numbers were lowest on plots without dung and

ighest on plots where either dung type was combined withrampling treatment (Fig. 1). Species numbers on DST0 plotsere only slightly higher than on plots without dung. Statis-

ical analyses revealed a significant positive impact of dungs well as trampling on the species number. No significantnteraction dung × trampling occurred, but, by trend, the data

C.Faustetal./B

asicand

Applied

Ecology

12(2011)

215–226219

Table 1. Synoptic table showing all plant species and fates across the different stages of generative reproduction including fruiting in the actual vegetation, endozoochorous seed dispersaland establishment. G = species group (t: target, n: non-target); L = life cycle (a: annual, b: biennial, p: perennial); H = growth height (based upon observations in the field: 1: <20 cm, 2:20–50 cm, 3: >50 cm, w: woody species); S = seed size group (1: <1.0 mm, 2: 1.0–2.0 mm, 3: 2.1–3.0 mm, 4: 3.1–4.0 mm, 5: >4.0 mm). Data on L and S according to Klotz, Kühn, andDurka (2002). Dots indicate the frequency of occurrence in plots/samples: � = <5%; �� = 5–50%; ��� = >50%; © = species occurred in one plot/sample only; + = species was presenton the studied sheep paddock in fruiting condition.

G L H S Seed potential(climate room)

Establishment(field experiment)

Actual vegetation(sheep paddock)

DS DR Seedling emergence Fruiting success

DST0 DST1 DRT0 DRT1 DST0 DST1 DRT0 DRT1 Fruiting species

Present in sheep and rabbit dung, successful emergence and fruitingOenothera biennis s.l. n b 3 2 ��� ��� ��� ��� �� ��� �� �� ��

Potentilla argentea agg. t p 1 1 ��� ��� � �� �� �� �� �� � +Arenaria serpyllifolia agg. t a 1 1 �� ��� �� �� �� �� �� �� © �� +Verbascum phlomoides n b 3 1 ��� �� �� �� ��� �� �� �� �� +Conyza canadensis n a 2 2 �� �� �� �� �� �� © � �� �� +Sedum acre t p 1 1 �� �� �� �� �� © © �� +Vicia lathyroides t a 1 2 © � �� �� �� © © ��

Only present in sheep dung, successful emergence and fruitingChenopodium album agg. n a 2 2 �� �� �� � �� © © ©Trifolium campestre t a 1 2 � �� © +Only present in rabbit dung, successful emergence and fruitingRumex acetosella s.l. t a 1 2 ��� � � � �

Alyssum alyssoides t a 1 2 �� �� © © +Cerastium semidecandrum t a 1 1 �� © ©Petrorhagia prolifera t a 2 2 �� �� © +Centaurea stoebe s.l. t b 3 3 © © © +Erodium cicutarium agg. t a 1 3 © © © +Present in either sheep or rabbit dung, successful emergencePoa angustifolia n p 2 2 � ©Crepis capillaris n a 2 3 © © +

220C

.Faustetal./Basic

andA

ppliedE

cology12

(2011)215–226

Table 1 (Continued)

G L H S Seed potential(climate room)

Establishment(field experiment)

Actual vegetation(sheep paddock)

DS DR Seedling emergence Fruiting success

DST0 DST1 DRT0 DRT1 DST0 DST1 DRT0 DRT1 Fruiting species

Present in sheep and rabbit dung, no establishmentFragaria × ananassa n a 1 2 © ���

Echium vulgare t b 2 3 �� �� +Portulaca oleracea n a 1 1 �� �� +Urtica dioica s.l. n p 3 2 �� ��

Solanum nigrum n a 2 2 � �� +Psyllium arenarium n a 1 3 � � +Medicago minima t a 1 2 � © +Robinia pseudoacacia n p w 5 © ©Only present in sheep dung, no establishmentAmaranthus retroflexus n a 3 2 ��� +Chenopodium strictum n a 2 2 ���

Diplotaxis tenuifolia n p 2 2 �� +Herniaria glabra t a 1 1 ��

Saponaria officinalis n p 3 2 ��

Cynodon dactylon n p 2 1 �

Geranium molle n a 1 2 © +Myosotis ramosissima t a 1 2 ©Only present in rabbit dung, no establishmentAgrostis capillaris n p 2 1 ���

Veronica praecox t a 1 2 ��

Silene conica t a 1 1 �

Spergula arvensis n a 2 2 �

Thymus pulegioides n p 1 1 �

Agrostis vinealis t p 2 2 ©Eragrostis minor n a 1 1 ©Koeleria macrantha t p 2 3 ©Stellaria media n a 1 2 ©Neither found in sheep nor rabbit dung, successful seedling emergenceHelianthemum nummularium ssp. obscurum t p 1 2 �� �� +Veronica arvensis n a 1 1 ��

Vicia angustifolia t a 1 2 � ©Silene otites t p 3 1 ©Total number of species 26 31 7 10 13 16 3 8 9 14 37a

aThis value includes further 16 species which were only present in the actual vegetation and are listed in Appendix A: Table* 1.

C. Faust et al. / Basic and Applied Ecology 12 (2011) 215–226 221

Table 2. Climate room data. Mean content of viable seeds per 100 g air-dried faeces (with ±95% confidence interval, CI, in parentheses)and the results of the mixed-linear-model analysis of significant influences of the dung type calculated for seed abundances. No. of samplesof each dung type: 48; total dry weight of faeces of each dung type: 1920 g. Fr: frequency of occurrence in samples (%). TSR: target speciesratio. Num df: degrees of freedom Numerator, Den df: degrees of freedom Denominator, p: level of significance.

Species Sheep dung Fr Rabbit dung Fr Num df Den df F-Value p

Significantly higher content in sheep dungAmaranthus retroflexus 5.21 (1.33) 79 0 0 1 94 59.10 <0.0001Chenopodium album 1.25 (0.65) 33 0 0 1 94 14.10 0.0003Chenopodium strictum 2.34 (0.57) 67 0 0 1 94 64.35 <0.0001Echium vulgare 0.68 (0.43) 21 0.16 (0.17) 6 1 94 4.83 0.0305Herniaria glabra 0.36 (0.25) 15 0 0 1 94 8.02 0.0056Oenothera biennis s.l. 66.20 (6.47) 100 13.80 (1.91) 100 1 94 231.75 <0.0001Portulaca oleracea 1.46 (0.56) 44 0.16 (0.17) 6 1 94 18.84 <0.0001Sedum acre 1.61 (0.64) 44 0.52 (0.29) 21 1 94 9.21 0.0031Verbascum phlomoides 12.08 (1.91) 100 2.76 (0.56) 42 1 94 110.47 <0.0001Significantly higher content in rabbit dungAgrostis capillaris 0 0 36.25 (5.71) 98 1 94 109.20 <0.0001Alyssum alyssoides 0 0 0.73 (0.36) 27 1 94 16.10 0.0001Arenaria serpyllifolia agg. 2.08 (0.87) 46 5.36 (1.29) 83 1 94 17.15 <0.0001Cerastium semidecandrum 0 0 0.73 (0.38) 25 1 94 13.79 0.0003Conyza canadensis 0.21 (0.20) 8 0.89 (0.40) 31 1 94 8.87 0.0037Fragaria × ananassa 0.16 (0.31) 2 11.56 (3.13) 85 1 94 50.38 <0.0001Rumex acetosella s.l. 0 0 3.23 (0.95) 65 1 94 44.85 <0.0001Solanum nigrum 0.10 (0.14) 4 1.46 (0.94) 31 1 94 7.73 0.0065

Total no. of speciesa 26 31Mean no. of seeds per 100 ga 100.52 (8.22) 85.88 (8.63) 1 94 102.20 <0.0001

Mean TSRquala 0.31 (0.04) 0.32 (0.03) 1 94 0.11 0.7373

M a 3 (0.0

able 2.

so

E

cnnsr

ineogtmhw(

p96

ta8rsao

(ldcsf(

F

dc1

ean TSRquant 0.10 (0.02) 0.1

aValues include non-significant species which are listed in Appendix A: T

how a more pronounced trampling effect on sheep dung thann rabbit dung and no effect on plots without dung.

mergence success of dung-embedded seedsThe effect of dung type was less pronounced than in the

limate-room data. Only O. biennis s.l. and Helianthemumummularium ssp. obscurum emerged in significantly higherumbers or exclusively from sheep dung. Dung type exertedignificant effects on both TSR indices, with higher values inabbit dung.

The effect of trampling, however, was considerable andncreased seedling emergence and species number sig-ificantly: Trampling of sheep dung enhanced seedlingmergence 3.6-fold and species number 2.4-fold. In the casef rabbit dung the factors were 2.1-fold (seedling emer-ence) and 1.9-fold (mean species number). Trampling ledo a significant enhancement of TSRqual. This effect wasore pronounced on plots with sheep dung (TSRqual 2.2-fold

igher) than with rabbit dung (1.4-fold higher). The TSRquantas not significantly influenced, but the tendency is the same

sheep dung 1.7-fold, rabbit dung 1.4-fold higher).

Significant effects of trampling occurred for Arenaria ser-

yllifolia agg. (increase by factor 3.5 on sheep dung and.4 on rabbit dung) and O. biennis s.l. (factors 3.5 and.4). Trampling of sheep dung slightly favoured eight fur-

i(22

2) 1 94 6.91 0.0100

her species and none was suppressed. The favoured speciesre 50% annuals, 20% biennials, 30% perennials; altogether0% are small- or medium-growing species. In the case ofabbit dung, 10 further species benefited. Mainly short-livedpecies benefited from trampling on rabbit dung (83% annu-ls, 8% biennials). Again, the highest percentages were small-r medium-growers (83%).

Compared to the number of seeds contained in the dungclimate room data), the emergence success in the field wasow: sheep dung: 5% without and 18% with trampling, rabbitung: 7% without and 14% with trampling. As found for thelimate room data, the emergence success was highest formall-seeded species. Twenty-four species did not emergerom dung samples in the field experiment; most of them16) occurred in small numbers only in the climate room.

ruiting success of dung-borne seedlingsThe fruiting success was not significantly influenced by

ung type but increased by trampling (Tables 3 and 4): inase of sheep dung, the number of fruiting individuals per00 g of dung was 3.9-fold increased and the number of fruit-

ng species 2.1-fold. The factors for rabbit dung are 2.6-foldfruiting individuals) and 1.9-fold (fruiting species). Thus,1% of the seedlings on DST0 plots had fruiting success,2% on DST1 plots, 25% on DRT0 plots and 31% on DRT1

222 C. Faust et al. / Basic and Applied Ecology 12 (2011) 215–226

Table 3. Results of the mixed-linear-model analysis of significant influences of the independent variables dung and trampling and/or interactioneffects on (A) the number of seedlings emerging and (B) the number of individuals fruiting in the field experiment. Significant effects areshown in bold print. TSR: target species ratio. Num df: degrees of freedom Numerator, Den df: degrees of freedom Denominator, p: levelof significance. A dash indicates that no test was done, because no individuals fruited successfully. The following species were tested butdid not reveal significant results: Alyssum alyssoides, Centaurea stoebe s.l., Cerastium semidecandrum, Chenopodium album agg., Conyzacanadensis, Crepis capillaris, Erodium cicutarium agg., Petrorhagia prolifera, Poa angustifolia, Potentilla argentea agg., Rumex acetosellas.l., Sedum acre, Silene otites, Trifolium campestre, Verbascum phlomoides, Veronica arvensis, Vicia angustifolia, Vicia lathyroides.

Variables Num df Den df No. of seedlings No. of species TSRqual TSRquant

F-Value p F-Value p F-Value p F-Value p

A. Seedling emergenceDung 1 92 0.45 0.5028 0.07 0.7980 4.48 0.0371 6.68 0.0114Trampling 1 92 21.14 <0.0001 22.86 <0.0001 5.73 0.0187 2.92 0.0911Dung × trampling 1 92 1.08 0.3009 0.53 0.4668 0.18 0.6762 0.00 0.9606B. Fruiting successDung 1 92 0.01 0.9409 1.29 0.2588 6.71 0.0212 7.84 0.0062Trampling 1 92 14.38 0.0003 8.43 0.0046 2.07 0.1537 2.18 0.1431Dung × trampling 1 92 0.27 0.6039 0.02 0.8998 0.33 0.5664 0.62 0.4336

Individual species Num df Den df Arenariaserpyllifolia agg.

Helianthemum nummulariumssp. obscurum

Oenothera biennis s.l.

F-Value p F-Value p F-Value p

A. Seedling emergenceDung 1 92 2.06 0.1547 4.95 0.0285 5.88 0.0172Trampling 1 92 5.46 0.0216 0.42 0.5170 13.00 0.0005Dung × trampling 1 92 1.71 0.1937 0.42 0.5170 0.47 0.4934B. Fruiting successDung 1 92 0.35 0.5534 – – 9.44 0.0028T –D –

psbm

cs

ca(

D

S

ftEqwa

m7sbsiTdttstwFiaas

E

rampling 1 92 1.93 0.1685 –ung × trampling 1 92 0.98 0.3241 –

lots. The percentage of short-lived species which fruiteduccessfully was very high (range of the four treatment com-inations: 75–100%) as well as the percentage of small- oredium-growing species (66–85%).Both TSR indices of the fruiting individuals were signifi-

antly higher in rabbit than in sheep dung. Trampling did notignificantly affect the TSR indices (Tables 3 and 4).

The fruiting success of only one plant species was signifi-antly affected by the treatments: O. biennis s.l. was morebundant on sheep dung plots and favoured by tramplingTables 3 and 4).

iscussion

eed availability

The mean seed density in the climate-room experimentound in sheep dung (101 seeds 100 g−1 dry matter) was inhe range of the results found in other studies (Mouissie 2004;

ichberg et al. 2007; Kuiters & Huiskes 2010). Referring touestion 1, the number of taxa (26) found in sheep faecesas similar to the results of Eichberg et al. (2007; 28 taxa)

nd Mitlacher, Poschlod, Rosén, and Bakker (2002; 27) butnc

7.72 0.00661.05 0.3086

uch lower than in the study by Kuiters and Huiskes (2010;2). The number of species contained in sheep dung was onlylightly lower and the density only slightly higher than in rab-it dung. More interesting are the differences in composition:heep dung contained more tall-growing species, whereasn rabbit dung small- or medium-growing species prevailed.his is obviously a reflection of the food selection in depen-ence on body height. Since many target species are small,he TSR indices of rabbit dung were also higher. In addition,he home range size of rabbits varies based on population den-ity and food supply (Myers & Poole 1963). We suppose thathis is also an explanation for the occurrence of those specieshich were found in dung samples but not on the paddock.or example, the abundance of Fragaria × ananassa seeds

n rabbit dung was very high. Strawberries were cultivateddjacent to the sampling area. In the case of sheep dung wessume that sheep may have eaten seed-containing plant oroil material from the litter layer.

stablishment success in the field

Differences between sheep and rabbit dung were less pro-ounced regarding seedling emergence, but there was a clearontrast between the climate room and the field experiment

C. Faust et al. / Basic and Applied Ecology 12 (2011) 215–226 223

Table 4. Mean number of emerging seedlings (A) and fruiting individuals (B) per 100 g air dry faeces (with ±95% confidence interval, CI,in parentheses) in the field experiment within 2 years. No. of plots of each treatment type: 24; total dry weight of faeces of each plot type:960 g. Fr: frequency of occurrence in plots (%). Means sharing the same letter are not significantly different (p > 0.05).

Plot type DST0 Fr DST1 Fr DRT0 Fr DRT1 Fr

A. Seedling emergenceAlyssum alyssoides 0 0 0 0 0.52 (0.83) 8 0.10 (0.20) 4Arenaria serpyllifolia agg. 0.21a (0.28) 8 0.73a (0.81) 13 0.21ab (0.28) 8 1.98b (1.22) 38Centaurea stoebe s.l. 0 0 0 0 0.10 (0.20) 4 0 0Cerastium semidecandrum 0 0 0 0 0 0 0.10 (0.20) 4Chenopodium album agg. 0.21 (0.28) 8 0.73 (0.62) 21 0.21 (0.41) 4 0.31 (0.45) 8Conyza canadensis 0.31a (0.45) 8 0.94ab (1.01) 17 0.52ab (0.59) 13 1.98b (1.50) 33Crepis capillaris 0 0 0 0 0 0 0.10 (0.20) 4Erodium cicutarium agg. 0 0 0 0 0 0 0.10 (0.20) 4Helianthemum nummularium ssp. obscurum 0.42 (0.48) 13 0.83 (0.76) 21 0 0 0 0Oenothera biennis s.l. 2.71a (1.06) 58 9.48b (2.96) 92 0.52a (0.59) 13 3.33a (1.46) 67Petrorhagia prolifera 0 0 0 0 0 0 0.21 (0.28) 8Poa angustifolia 0 0 0 0 0.10 (0.20) 4 0 0Potentilla argentea agg. 0.73a (1.43) 4 2.81b (2.01) 42 1.46ab (1.02) 29 1.25ab (0.78) 33Rumex acetosella s.l. 0 0 0.42 (0.82) 4 0 0 0.83 (1.13) 8Sedum acre 0 0 0.52 (0.51) 17 0.21 (0.28) 8 0.31 (0.45) 8Silene otites 0 0 0 0 0 0 0.10 (0.20) 4Trifolium campestre 0 0 0 0 0 0 0.21 (0.28) 8Verbascum phlomoides 0.42 (0.48) 13 1.46 (0.97) 38 1.46 (1.02) 33 0.73 (0.69) 17Veronica arvensis 0 0 0 0 0.21 (0.28) 8 0 0Vicia angustifolia 0 0 0 0 0.21 (0.41) 4 0 0Vicia lathyroides 0 0 0.21 (0.28) 8 0.21 (0.28) 8 0.63 (0.53) 21

Number of species per plot type 7 10 13 16Mean no. of seedlings per 100 g 5.00a (2.13) 18.13c (5.63) 5.94ab (1.97) 12.29b (3.18)Mean no. of species per treatment 1.13a (0.36) 2.71c (0.51) 1.46ab (0.44) 2.71bc (0.63)

Mean TSRqual 0.16a (0.13) 0.35b (0.10) 0.33ab (0.17) 0.46b (0.13)Mean TSRquant 0.16a (0.13) 0.27ab (0.08) 0.33ab (0.17) 0.45b (0.13)

B. Fruiting successAlyssum alyssoides 0 0 0 0 0 0 0.10 (0.20) 4Arenaria serpyllifolia agg. 0.21 (0.28) 8 0.31 (0.45) 8 0.10 (0.20) 4 0.73 (0.86) 13Centaurea stoebe s.l. 0 0 0 0 0.10 (0.20) 4 0 0Cerastium semidecandrum 0 0 0 0 0 0 0.10 (0.20) 4Chenopodium album agg. 0 0 0.10 (0.20) 4 0.10 (0.20) 4 0.10 (0.20) 4Conyza canadensis 0.10 (0.20) 4 0.21 (0.41) 4 0.21 (0.28) 8 0.52 (0.51) 17Crepis capillaris 0 0 0 0 0 0 0 0Erodium cicutarium agg. 0 0 0 0 0 0 0.10 (0.20) 4Helianthemum nummularium ssp. obscurum 0 0 0 0 0 0 0 0Oenothera biennis s.l. 0.73a (0.46) 29 2.08b (1.20) 38 0 0 0.63a (0.53) 21Petrorhagia prolifera 0 0 0 0 0 0 0.10 (0.20) 4Poa angustifolia 0 0 0 0 0 0 0 0Potentilla argentea agg. 0 0 0.63 (0.90) 8 0.21 (0.28) 8 0.21 (0.41) 4Rumex acetosella s.l. 0 0 0 0 0 0 0.31 (0.61) 4Sedum acre 0 0 0.10 (0.20) 4 0.10 (0.20) 4 0.21 (0.28) 8Silene otites 0 0 0 0 0 0 0 0Trifolium campestre 0 0 0 0 0 0 0.10 (0.20) 4Verbascum phlomoides 0 0 0.52 (0.51) 17 0.42 (0.38) 17 0.31 (0.34) 13Veronica arvensis 0 0 0 0 0 0 0 0Vicia angustifolia 0 0 0 0 0.10 (0.20) 4 0 0Vicia lathyroides 0 0 0.10 (0.20) 4 0.10 (0.20) 4 0.31 (0.45) 8

Number of fruiting species per plot type 3 8 9 14Mean no. of fruiting individuals per 100 g 1.04a (0.50) 4.06c (1.93) 1.46ab (0.72) 3.85bc (1.67)Mean no. of fruiting species per treatment 0.42a (0.20) 0.88ab (0.85) 0.58ab (0.29) 1.08b (0.42)

Mean TSRqual 0.08a (0.11) 0.15a (0.13) 0.23ab (0.16) 0.38b (0.17)Mean TSRquant 0.08a (0.11) 0.13a (0.12) 0.23ab (0.16) 0.39b (0.17)

2 pplied E

(wrTsrtehAemsowihg

T

tcfosbtertrsSgsFWesal

hewcssirt

p

ptbtnpsdtet&is

ebsferiiy

E

ddhatih

ccttomita

etTtt

24 C. Faust et al. / Basic and A

question 2). Emergence success in the field was much lower:ithout trampling only 5 or 7% of the potential in the climate

oom, but with trampling the values were 3.6- or 2-fold higher.he trampling effect can be considered, to some degree, asimilar to the crumbled dung which was used in the climateoom. It is self-evident that the optimised growing condi-ions in a climate room are beneficial for germination andstablishment of plants while natural conditions are mostlyarsh, as was also found in previous studies (Welch 1985;uman, Call, & Wiedmeier 1998; Mouissie 2004; Eichberg

t al. 2007). Nevertheless, it was remarkable that H. num-ularium ssp. obscurum exclusively emerged in the field on

heep dung but never in the climate room. The seed sizeften plays a crucial role concerning seed dispersal. Speciesith bigger seeds neither emerged in the climate room nor

n the field. We assume that they do not survive digestion byerbivores, especially ruminants. The emergence success oframinoid competitors was very low.

he impact of trampling

Plots with dung and particularly those with tramplingreatment in addition harboured most plant species. Thisonspicuously indicates the importance of trampling as aacilitative post-dispersal process (question 3). Tramplingn dung led to a significant increase in mean dung-borneeedling and species numbers. A number of single speciesenefited from trampling. The impact of trampling was par-icularly obvious in the case of sheep-dung pellets (3.6-foldnhancement of the number of seedlings). But also trampledabbit-dung pellets showed about twice as many seedlings ashe non-trampled ones. A positive effect of trampling on plantegeneration has been shown also for non-dung-embeddedeeds lying on the ground (Wessels-de Wit & Schwabe 2010).mall-scale disturbances, especially those performed by dig-ing rabbits, can create bare soil patches, which are crucial foreedling establishment in various grassland types (Jentsch,riedrich, Beyschlag, & Nezadal 2002; Bakker & Olff 2003).e found that trampling on dung, placed in bare soil patches,

nhanced seedling establishment additionally, which empha-izes on the one hand the importance of gaps in dry grasslandsnd on the other hand the potential of post-dispersal processesike trampling.

The proportion of annual species emerging from dung wasigh but with trampling even higher. Some species nevermerged on plots without trampling. Seedling emergenceith trampling treatment might even be stronger since we

annot exclude a potential underestimation of dung-borneeedlings on plots with trampling treatment, because someeeds may have been separated from the dung substrate dur-ng the trampling process. After trampling the dung formed a

elatively compact layer on the soil and only seedlings fromhis layer were included in our records.

We infer that the most important impact of trampling is,articularly in the case of sheep dung, the cracking of the hard

sn

s

cology 12 (2011) 215–226

ellets’ surface, and for both dung types the amplification ofhe dung’s surface. Hence, dung-embedded seeds obtain aetter light and water supply after trampling. Additionally,he dung layer probably promotes dung-borne seedlings byutrient pulses. For larger dung-embedded seeds a furtherositive effect of trampling may be the incorporation into theoil (Eichberg et al. 2005). A germination experiment con-ucted by Wessels-de Wit and Schwabe (2010) revealed thathe relatively large seeds of Cynoglossum officinale almostxclusively germinated when buried. Large seeds are believedo require less light for germination than smaller ones (Baskin

Baskin 1998). However, this effect seems less importantn our experiment, since these populations comprise mostlymall-seeded species.

The fruiting success of dung-borne seedlings in the fieldxperiment was high and, again, plants profited significantlyy preceding trampling. Particularly annual and biennialpecies, which had a high emergence success, reached theirruiting stages within the 2-year census period. Eichbergt al. (2007) found only 0.4% of sheep-dung-borne seedlingseaching fruit ripeness, which is in clear contrast to our find-ngs, but their study was conducted with untrampled dung,n pioneer plant community stages and in the extremely dryear 2003.

ffects on target species ratios

TSRqual of the seed pools contained in sheep and rabbitung were reduced as compared to the index of the actual pad-ock vegetation. The TSRquant was slightly but significantlyigher for rabbit than sheep dung. These results are probablygain caused by dietary preferences and differences betweenhe two animal species. The preference of sheep for tall grow-ng species, which are often ruderals (Artemisietea species),as been shown by Stroh, Storm, Zehm, and Schwabe (2002).

Trampling, however, enhanced the relative emergence suc-ess of target species markedly (question 4), especially in thease of sheep dung. As a result, both TSR indices of both dungypes are much enhanced by trampling and exceed those ofhe dung seed potential. This is due to the higher emergencef habitat-typical species (6 species in the field, 9 in the cli-ate room) than that of non-target species (4 in the field, 17

n the climate room). Hence, the harsh environmental condi-ions are beneficial for the relative success of target speciess a group (environmental filter).

The fruiting success of target species or individuals, how-ver, was not significantly influenced by trampling and, inhis development stage, the TSR indices are reduced again.his is in strong contrast to results of Eichberg et al. (2007):

hey found predominantly target species and even threatenedarget species emerging out of sheep dung, and the latter

pecies group had very great fruiting success (80%) whereason-target species never set fruits.

In conclusion, sheep trampling in extensively managedand grasslands with short grazing periods and long periods

pplied E

opSefr

A

tDrAfEfmmT

A

cd

R

A

A

B

B

B

B

C

C

C

D

E

E

E

E

F

F

H

H

J

J

K

K

L

M

C. Faust et al. / Basic and A

f plant re-growth has a significant positive impact on theost-dispersal development of many sheep-dispersed seeds.heep play a multifaceted role in dispersal processes: afterndozoochoric transport they serve as “gardeners” not onlyor sheep-dispersed seeds, but also for those dispersed byabbits.

cknowledgements

We are grateful to Reiner Stürz (Ober-Ramstadt) forhe sheep husbandry. Berend Koch (Technische Universitätarmstadt) is acknowledged for making the sheep hoove

eplicas. Parts of the study were funded by the Federalgency for Nature Conservation and the Federal Minister

or Environment, Nature Conservation and Nuclear Safety.specially, thanks to the ‘Regierungspräsidium Darmstadt’

or permission to work in the area. We thank two anony-ous reviewers for very valuable suggestions to improve theanuscript. The improvement of the English text by Dr. Annhorson (Oxford) is much appreciated.

ppendix A. Supplementary data

Supplementary data associated with this arti-le can be found, in the online version, atoi:10.1016/j.baae.2011.02.009.

eferences

rnau, J., Bono, R., & Vallejo, G. (2009). Analyzing small samplesof repeated measures data with the mixed-model adjusted F test.Communications in Statistics – Simulation and Computation, 38,1083–1103.

uman, B. S., Call, C. A., & Wiedmeier, R. D. (1998).Crested wheatgrass establishment in livestock dung depositedon degraded rangeland vegetation types in the Intermoun-tain West, USA. Arid Soil Research and Rehabilitation, 12,317–333.

akker, E. S., & Olff, H. (2003). Impact of different-sized herbivoreson recruitment opportunities for subordinate herbs in grasslands.Journal of Vegetation Science, 14, 465–474.

akker, J. P., Galvez Bravo, L., & Mouissie, A. M. (2008). Dispersalby cattle of salt-marsh and dune species into salt-marsh and dunecommunities. Plant Ecology, 197, 43–54.

askin, C. C., & Baskin, J. M. (1998). Seeds: Ecology, biogeog-raphy and evolution of dormancy and germination. San Diego:Academic Press.

ullock, J. M., Moy, I. L., Pywell, R. F., Coulson, S. J., Nolan, A. N.,& Caswell, H. (2002). Plant dispersal and colonization processesat local and landscape scales. In J. M. Bullock, R. E. Kenward,& R. S. Hails (Eds.), Dispersal ecology (pp. 279–302). Malden,

MA: Blackwell Publishing.

osyns, E., Claerbout, S., Lamoot, I., & Hoffman, M. (2005). Endo-zoochorous seed dispersal by cattle and horse in a spatiallyheterogeneous landscape. Plant Ecology, 178, 149–162.

M

cology 12 (2011) 215–226 225

osyns, E., Bossuyt, B., Hoffmann, M., Vervaet, H., & Lens,L. (2006). Seedling establishment after endozoochory in dis-turbed and undisturbed grasslands. Basic and Applied Ecology,7, 360–369.

ouvreur, M., Christiaen, B., Verheyen, K., & Hermy, M. (2004).Large herbivores as mobile links between isolated naturereserves through adhesive seed dispersal. Applied VegetationScience, 7, 229–236.

eutscher Wetterdienst, http://www.dwd.de/bvbw/appmanager/bvbw/dwdwwwDesktop.

ichberg, C., Storm, C., & Schwabe, A. (2005). Epizoochorous andpost-dispersal success in a rare plant species: Jurinea cyanoides(L.) Rchb (Asteraceae). Flora, 200, 477–489.

ichberg, C., Storm, C., & Schwabe, A. (2007). Endozoochorousdispersal, seedling emergence and fruiting success in disturbedand undisturbed successional stages of sheep-grazed inland sandecosystems. Flora, 202, 3–26.

ichberg, C., Storm, C., Stroh, M., & Schwabe, A. (2010). Is thecombination of topsoil replacement and inoculation with plantmaterial an effective tool for the restoration of threatened sandygrassland? Applied Vegetation Science, 13, 425–438.

uropean Topic Centre on Biological Diversity (ETC/BD) (2008).European Nature Information System (EUNIS) Database. Habi-tat types and habitat classifications. Paris: ETC/BD-EEA.

aust, C., Süss, K., Storm, C., & Schwabe, A. (2011). Threatenedinland sand vegetation in the temperate zone under differenttypes of abiotic and biotic disturbances during a ten-year period.Flora, doi:10.1016/j.flora.2010.09.013

oster, B. L., & Tilman, D. (2003). Seed limitation and the regulationof community structure in oak savanna grassland. Journal ofEcology, 91, 999–1007.

anf, M. (1999). Ackerunkräuter Europas mit ihren Keimlingen undSamen. München: BLV.

obbs, N. T. (2006). Large herbivores as sources of disturbance inecosystems. In K. Danell, P. Duncan, R. Bergström, & J. Pas-tor (Eds.), Large herbivore ecology, ecosystem dynamics andconservation (pp. 261–288). Cambridge: University Press.

acqmin-Gadda, H., Sibillot, S., Proust, C., Molina, J. M., &Thiébaut, R. (2007). Robustness of the linear mixed modelto misspecified error distribution. Computational Statistics andData Analysis, 51, 5142–5154.

entsch, A., Friedrich, S., Beyschlag, W., & Nezadal, W. (2002).Significance of ant and rabbit disturbances for seedling estab-lishment in dry acidic grasslands dominated by Corynephoruscanescens. Phytocoenologia, 32, 553–580.

lotz, S., Kühn, I., & Durka, W. (Eds.). (2002). BIOLFLOR –Eine Datenbank zu biologisch ökologischen Merkmalen derGefäßpflanzen in Deutschland. Schriftenreihe für Vegetations-kunde 38. Bonn: Bundesamt für Naturschutz.

uiters, A. T., & Huiskes, H. P. J. (2010). Potential of endo-zoochorous seed dispersal by sheep in calcareous grasslands:Correlations with seed traits. Applied Vegetation Science, 13,163–172.

ittell, R. C., Miliken, G. A., Stroup, W. W., Wolfinger, R. D., &Schabenberger, O. (2006). SAS for mixed models (2nd ed.). Cary,NC, USA: SAS Institute.

alo, J. E., & Suárez, F. (1995a). Herbivorous mammals as seed

dispersers in a Mediterranean dehesa. Oecologia, 104, 246–255.

alo, J. E., & Suárez, F. (1995b). Establishment of pasture specieson cattle dung: The role of endozoochorous seeds. Journal ofVegetation Science, 6, 169–174.

2 pplied E

M

M

M

M

M

O

P

R

S

S

S

S

T

V

W

W

W

26 C. Faust et al. / Basic and A

itlacher, K., Poschlod, P., Rosén, E., & Bakker, J. P. (2002).Restoration of wooded meadows – a comparative analysis alonga chronosequence on Öland (Sweden). Applied Vegetation Sci-ence, 5, 63–73.

ouissie, A. M. (2004). Seed dispersal by large herbivores – impli-cations for the restoration of plant biodiversity. PhD dissertation,University of Groningen, The Netherlands.

ouissie, A. M., Vos, P., Verhagen, H. M. C., & Bakker, J. P.(2005). Endozoochory by free-ranging, large herbivores: Eco-logical correlates and perspectives for restoration. Basic andApplied Ecology, 6, 547–558.

uller, F. M. (1978). Seedlings of the North-Western Europeanlowland. The Hague: Junk Publishers.

yers, K., & Poole, W. E. (1963). A study of the biology of thewild rabbit, Oryctolagus cuniculus (L.) in confined populations.Journal of Ecology, 51, 435–451.

zinga, W. A., Römermann, C., Bekker, R. M., Prinzing, A., Tamis,W. L. M., Schaminée, J. H. J., et al. (2009). Dispersal failurecontributes to plant losses in NW Europe. Ecology Letters, 12,66–74.

akeman, R. J., Engelen, J., & Attwood, J. P. (1999). Rabbit endo-zoochory and seedbank build-up in an acidic grassland. PlantEcology, 145, 83–90.

amos, M. E., Robles, A. B., & Castro, J. (2006). Efficiency ofendozoochorous seed dispersal in six dry-fruited species (Cis-taceae): From seed ingestion to early seedling establishment.Plant Ecology, 185, 97–106.

chaalje, G. B., McBride, J. B., & Fellingham, G. W. (2002). Ade-quacy of approximations to distributions of test statistics incomplex mixed linear models. Journal of Agricultural, Biologi-cal, and Environmental Statistics, 7, 512–524.

W

cology 12 (2011) 215–226

tammel, B., & Kiehl, K. (2004). Do hoof prints actually serve asa regeneration niche for plant species in fens? Phytocoenologia,34, 271–286.

troh, M., Storm, C., Zehm, A., & Schwabe, A. (2002). Restorativegrazing as a tool for directed succession with diaspore inoc-ulation: The model of sand ecosystems. Phytocoenologia, 32,595–625.

üss, K., Storm, C., Zehm, A., & Schwabe, A. (2004). Succession ininland sand ecosystems: Which factors determine the occurrenceof the tall grass species Calamagrostis epigejos (L.) Roth andStipa capillata (L.)? Plant Biology, 6, 465–476.

er Heerdt, G. N. J., Verweij, G. L., Bekker, R. M., & Bakker, J. P.(1996). An improved method for seed-bank analysis: Seedlingemergence after removing the soil by sieving. Functional Ecol-ogy, 10, 144–151.

allejo, G., Fernández, P., Herrero, F. J., & Conejo, N. M. (2004).Alternative procedures for testing fixed effects in repeated mea-sures designs when assumptions are violated. Psicothema, 16,498–508.

elch, D. (1985). Studies in the grazing of heather moor-land in North-east Scotland. Journal of Applied Ecology, 22,461–472.

essels, S., & Schwabe, A. (2008). Testing the potential seed avail-ability in dung samples: Comparison of two seedling emergencemethods. Flora, 203, 429–436.

essels-de Wit, S., & Schwabe, A. (2010). The fate of sheep-dispersed seeds: Plant species emergence and spatial patterns.

Flora, 205, 656–665.

isskirchen, R., & Haeupler, H. (1998). Standardliste der Farn-und Blütenpflanzen Deutschlands. Stuttgart: Ulmer Verlag.