ARTICLE IN PRESS Basic and Applied Ecology 7 (2006) 360—369
www.elsevier.de/baae
Seedling establishment after endozoochory in disturbed and undisturbed grasslands E. Cosynsa, B. Bossuyta,, M. Hoffmanna,b, H. Vervaeta, L. Lensa a
Department of Biology, Ghent University, Terrestrial Ecology Unit, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium Institute of Nature Conservation, Section Landscape Ecology and Nature Management, Kliniekstraat 25, B-1070 Brussels, Belgium
b
Received 20 December 2004; accepted 22 August 2005
KEYWORDS Cattle dung; Gap creation; Horse dung; Safe site; Sod-cutting; Soil disturbance; Species richness
Summary Local plant community composition and structure may be largely influenced by germination and seedling establishment from seeds dispersed in animal dung, through seed input, gap creation and nutrient enrichment. With an experimental approach we assessed (1) what the effect is of dung deposition on the number of seedlings in the plant community 3 months and 1 year after dung deposition, (2) what the effect is of this seedling establishment on the local plant community characteristics such as species richness and (3) if this effect interacts with largescale soil disturbance which removes the close canopy, such as sod-cutting. Viable seeds of monocotylous species were abundantly present in the dung, and dung deposition led to a higher number of monocotylous seedlings after 3 months. However, this effect was no longer significant after 1 year. Moreover, the proportion of viable monocotylous seeds that effectively established in the field after 3 months was less than 5%. A lower number of viable seeds of the less-dominant dicotylous species was dispersed in the dung but they had a higher cover and species richness after 1 year. This resulted in an increased total small-scale species richness and diversity after dung deposition through a decreasing dominance of monocotylous species. Sod-cutting had a pronounced effect on seedling emergence: viable seeds dispersed by dung had a higher probability of successful establishment when the dung was deposited in large gaps. This indicates that an increase of safe sites associated with disturbance strengthens the effects of seed dispersal and gap creation by dung deposition. ¨ kologie. Published by Elsevier GmbH. All rights reserved. & 2006 Gesellschaft fu ¨r O
Corresponding author. Tel.: +3292645094; fax: +3292648794.
E-mail address:
[email protected] (B. Bossuyt). ¨ kologie. Published by Elsevier GmbH. All rights reserved. 1439-1791/$ - see front matter & 2006 Gesellschaft fu ¨r O doi:10.1016/j.baae.2005.08.007
ARTICLE IN PRESS Establishment after endozoochory
361
Zusammenfassung Die lokale Zusammensetzung von Pflanzengemeinschaften und ihre Struktur ko ¨nnten vor allem durch die Keimung und Keimlingsetablierung aus Samen, die durch Tierdung verbreitet werden, durch den Sameneintrag, durch die Schaffung von Lu ¨cken und die Na ¨hrstoffanreicherung beeinflusst sein. Wir scha ¨tzten mit einem experimentellen Ansatz ab, (1) welchen Effekt hat die Deposition von Dung auf die Anzahl der Keimlinge drei Monate und ein Jahr nach der Dungdiposition, (2) welchen Effekt hat die Keimlingsetablierung auf die Charakteristiken der lokalen Pflanzengesellschaft (Z. B. Artenreichtum) und (3) ob dieser Effekt mit großra ¨umigen Sto ¨rungen des Bodens interagiert, die, wie Sodenstechen, die geschlossene Deckung beseitigen. Keimfa ¨hige Samen monokotyler Arten waren im Dung ha ¨ufig vertreten und die Deposition von Dung fu ¨hrte zu einer gro ¨ßeren Zahl von monokotylen Keimlingen nach drei Monaten. Dieser Effekt war jedoch nach einem Jahr nicht mehr signifikant. Daru ¨ber hinaus war der Anteil der keimfa ¨higen, monokotylen Samen die sich im Gela ¨nde nach drei Monaten etabliert haben, geringer als 5%. Mit dem Dung wurde eine geringere Anzahl von keimfa ¨higen Samen der weniger ha ¨ufigen Dikotylen verbreitet, aber sie hatten nach einem Jahr eine gro ¨ßere Deckung und Artenzahl. Das resultierte in einem erho ¨hten Gesamtartenreichtum und einer erho ¨hten Diversita ¨t auf kleiner Skala nach der Dungdeposition aufgrund der abnehmenden Dominanz der monokotylen Arten. Das Sodenstechen hatte einen deutlichen Effekt auf das Keimlingsaufkommen: keimfa ¨hige Samen, die mit dem Dung verbreitet wurden, hatten eine ho ¨here Wahrscheinlichkeit fu ¨r eine erfolgreiche Etablierung, wenn der Dung in großen Lu ¨cken deponiert wurde. Dies zeigt an, dass eine Zunahme von sicheren Orten, die mit Sto ¨rungen assoziiert sind, die Effekte der Samenverbreitung und der Schaffung von Lu ¨cken durch die Dungdeposition sta ¨rkt. ¨ kologie. Published by Elsevier GmbH. All rights reserved. & 2006 Gesellschaft fu ¨r O
Introduction Large numbers of plant seeds are potentially dispersed via animal dung (e.g. Cosyns, Delporte, Lens, & Hoffmann, 2005; Cosyns & Hoffmann, 2005; Malo & Sua ´rez, 1995a; Mouissie, Vos, Verhagen, & Bakker, 2005; Pakeman, Digneffe, & Small, 2002; Welch, 1985). The species composition of the seeds dispersed seems to differ between dispersal vectors (ungulates, horses or rabbits) (Cosyns, Claerbout, Lamoot, & Hoffmann, 2005; Cosyns & Hoffmann, 2005). The effect of dung deposition has three main components: (1) a source of colonisers in the form of seeds, (2) gap creation, as a result of the possible death of the vegetation under the dung, with potential favourable conditions for germination and seedling growth and (3) local soil nutrient enrichment (Olff & Ritchie, 1998). Since safe sites for recruitment are rare in temperate semi-natural grasslands (Austrheim & Eriksson, 2003), gap creation is of key importance in maintaining and restoring their species richness. Several studies (Burke & Grime, 1996; Carson & Pickett, 1990; Collins, 1987; Jakobsson & Eriksson, 2000) already emphasised the role of disturbance in enhancing plant establishment. Gaps created by dung deposition are likely to be a regeneration site both for
seeds dispersed in the dung and for seeds in the local soil seed bank or seed rain. The importance of this effect may depend on dung type. Increased seed germination and seedling establishment after dung deposition will affect local plant community composition and structure by increasing species richness or by changing relative abundance values. It has already been demonstrated that several plant species had a larger cover on cattle dung than in the surrounding vegetation due to germination from seeds in the dung (Malo & Sua ´rez, 1995b; Welch, 1985) and that seeds in cattle dung were a main seed source for colonisation of the gaps that remained after dung decomposition (Dai, 2000; Malo & Sua ´rez, 1995b). Pakeman, Engelen, and Attwood (1998) estimated that rabbit dispersed seeds accounted for 15% of the developing vascular plant cover and 10% of the species richness in experimental disturbances in a temperate acidic grassland. Small-scale species richness is found to be higher on sites where dung was deposited, although this was only a short-term effect (Traba, Levassor, & Peco, 2003). Very little is known, however, about the proportion of the seeds still viable after gut passage that is effectively establishing in the plant community. It is likely that only a very small proportion of the
ARTICLE IN PRESS 362 viable seeds will finally contribute to the plant community and that this proportion, and hence the effects on the plant community, will depend on the initial conditions on the site at the time the dung is deposited. Seedling establishment success will be higher when competition from the vegetation on the site, or from other seedlings, is smaller. The probability of surviving may also depend on speciesspecific characteristics, such as seed size, growth rate, light stress tolerance and competitive ability (Austrheim & Eriksson, 2003; Jensen & Gutekunst, 2003). With an experimental approach, we will answer the following research questions:
What is the effect of dung deposition on the number of seedlings in the plant community 3 months and 1 year after dung deposition? Is the effect of dung deposition different between monocotylous species and dicotylous species and between sod-cut and undisturbed grassland? What are the consequences of dung deposition for local plant community characteristics (cover, species richness, diversity)?
Material and methods Study area The experiment ran from August 2001 until July 2002 in a dune slack in the southern part of the Westhoek (340 ha), a nature reserve situated at the Belgian coast (511040 5000 N–21340 1900 E). Since 1997 this area (61 ha) is grazed year-round by, on average, five Scottish Highland cattle and 19 Shetland ponies. The experiment itself was carried out in a site dominated by perennial grasses. From the start of the experiment onwards one part (10 50 m) of this site was exclosed from large herbivore grazing. Before the start of the experiment it was assured that no dung of large herbivores was present.
Experimental design The experiment had a randomised complete block design (sensu Neter, Kutner, Nachtsheim, & Wasserman, 1996) with two fixed factors: disturbance (undisturbed versus sod-cut) and dung deposition (no dung, cattle dung and horse dung). Twelve blocks were laid out along a possible soil moisture gradient (mean highest ground water levels ranged from 120–100 cm to 80–60 cm below
E. Cosyns et al. surface), each containing six plots of 0.5 0.5 m. Sod-cutting implied the removal of the upper 5 cm of soil before the start of the experiment, meaning that the organic soil horizon was removed. As a side effect a large part of the seed bank was also removed, since the upper 5 cm of the soil generally contains the highest density of viable seeds (Bekker et al., 1998). To prevent the plots being contaminated by clonal growth from the surrounding vegetation, sod-cutting was carried out over 0.7 0.7 m. Dung deposition consisted of the input of 2.5 l of fresh dung of Shetland pony or Scottish Highland cattle, grazing in the surrounding area, allowing discriminating for kind of dung. To avoid a possible confounding impact of short-distance seed dispersal, vegetation within the blocks was clipped monthly at less than 10 cm vegetation height from May to October and mowed twice between blocks during summer 2001 and 2002. Freshly deposited dung from the free ranging Shetland ponies and Scottish Highland cattle (randomly chosen individual animals) was collected on several days in August 2001. At that time dung contains the highest seed densities and a broad spectrum of plant species (Claerbout, 2001). The dung samples were pooled and mixed for each herbivore species. A 2.5 l subsample was then spread out in each of the four experimental ‘dung’ plots of one single block. For cattle dung, the layer was approximately 2 cm thick, and for the pony dung approximately 4 cm. The pony dung was spread in a thicker layer because it was not possible to cover the whole area of the plot with dung due to its more solid consistence. Twelve 2.5 l subsamples of the same homogenised cattle and pony dung each were taken to the greenhouse (see below). The average dry weight of the 2.5 l subsamples was 354 and 277 g for cattle and pony dung, respectively.
Seed content in the dung To determine the potential number of seeds in the dung, the 12 cattle and 12 pony dung samples were taken to the greenhouse with optimal conditions for seed germination. For a detailed description of the seedling germination method, see Cosyns and Hoffmann (2005) and Cosyns, Claerbout et al. (2005). Seedlings in the greenhouse were identified, counted and removed at regular intervals during 1 year. Nomenclature follows Lambinon, De Langhe, Delvosalle, and Duvigneaud (1998). A total of 18,974 seedlings of 50 plant species emerged from the 12 cattle dung samples, while
ARTICLE IN PRESS Establishment after endozoochory
363
Table 1. Average total number of seeds, total number of species, ratio of the number of monocotylous versus dicotylous seeds and number of seeds of the 10 selected species in 2.5 l cattle or horse dung (N ¼ 12) Cattle
Horse
Number of seeds Number of species Ratio number of mono/dic seeds
1581.2 28.3 8.5
(529.7) (3.4) (2.4)
900.7 24.2 6.3
(389.2) (3.2) (2.2)
Number of seeds Arenaria serpyllifolia Cerastium fontanum Epilobium spp. Galium spp. Juncus articulatus Juncus bufonius Sagina proc.+apetala Trifolium repens Urtica dioica Veronica arven.+cham.
3.8 44.3 30.2 1.3 28.6 1237.2 12.0 24.3 1.6 17.1
(2.7) (18.2) (13.8) (0.9) (25.4) (486.9) (6.4) (10.6) (1.2) (7.4)
2.8 12.0 12.1 5.5 24.7 625.9 46.6 5.6 29.0 3.8
(2.6) (3.0) (12.7) (4.3) (14.9) (292.2) (25.7) (4.0) (36.2) (3.0)
Standard deviations are given in parenthesis.
10,808 seedlings of 49 plant species germinated from the 12 horse dung samples under greenhouse conditions (Table 1). The most abundant species was by far the monocotylouse species Juncus bufonius which made up 78.2% and 69.5% of all seedlings emerging from cattle and horse dung, respectively (Table 1). A relatively small number of other species each made up more than 1% of all seedlings: Sagina spp., Poa annua, J. articulatus, Trifolium repens, Cerastium fontanum and Urtica dioica. There was on average a significantly higher number of monocotylous than dicotylous seedlings in 2.5 l dung (1088 versus 153, t ¼ 9:44, po0:0001), but the average number of dicotylous species was higher (9.2 versus 17.1, t ¼ 12:7, po0:0001).
Seedlings in the field 3 months after dung deposition Seedling emergence was recorded in October 2001, 3 months after the start of the experiment, by counting the number of monocotylous and dicotylous seedlings that emerged in each plot. Most seedlings were too small to allow exact determination. Differences in total number of seedlings, number of monocotylous seedlings and number of dicotylous seedlings between the treatments 3 months after dung deposition were analysed using a mixed model ANOVA with fixed factors DUNG (no dung, horse or cattle dung), DISTURBANCE (vegetated or sod-cut) and their interaction and with the random factor BLOCK, applying Bonferroni corrections.
Plant community structure 1 year after dung deposition In 2002, 1 year after the start of the experiment, most dung was decomposed and 1-year-old seedlings could not be distinguished anymore from juvenile plants that were present in the plots before the experiment started. For that reason, we decided to count all individuals of 10 plant species present in the experimental plots in March and July 2002. The species were selected according to the following criteria: (1) they germinated in 475% of the samples and with at least 20 seedlings from dung samples kept under greenhouse conditions and (2) they were recognisable and countable as individual plants in the field. The following species or species aggregates were counted: Arenaria serpyllifolia, C. fontanum, Epilobium spp., Galium spp., J. articulatus, J. bufonius, Sagina apetala+S. procumbens, T. repens, U. dioica and Veronica arvensis+V. chamaedrys. Differences in total number of individuals of these species between the treatments were also tested with the same mixed model ANOVA, including Bonferroni corrections. Furthermore, in the summer of 2002, the percentage cover of all species was estimated in all 120 plots, using a decimal scale. These data were used to calculate the cover by monocotylous and dicotylous species, the number of monocotylous and dicotylous species, the total species richness and the Shannon–Wiener diversity (Kent & Coker, 1995) of the plant community in each plot. Cover and number of monocotylous and dicotylous species and species richness and diversity of the plant community in the plots 1 year after dung
ARTICLE IN PRESS 364 deposition were tested for differences between the treatments using a mixed model ANOVA as described above, including Bonferroni corrections. For all analyses, the number of seedlings and species were square root transformed, while the relative number of seedlings and the cover values were arcsine square-root transformed to approach normality and homogeneity of variance. All statistical analyses were carried out using SPSS 11.1 for Windows (SPSS, 2001).
Results Effect of treatments on seedling emergence on dung after 3 months The results of the mixed model ANOVA indicated that a significantly larger number of monocotylous seedlings emerged 3 months after the start of the experiment in the plots where dung was deposited (po0:0001) (Table 2). There were no differences between cattle and horse dung, and the effect of disturbance was not significant (p ¼ 0:12). In contrast, disturbance had a positive effect on the number of dicotylous seedlings in all plots (po0:0001) (Table 2) but plots with deposited dung had no significantly larger number of dicotylous seedlings than plots without dung (p ¼ 0:81). Also the interaction between dung type and disturbance was not significant (p ¼ 0:49).
Effect of treatments on establishment of plant species and vegetation structure after 1 year One year after dung deposition and disturbance, the cover by monocotylous species was lower in disturbed than in undisturbed plots (po0:0001) while the opposite was true for dicotylous species (po0:0001) (Table 2). However, the number of both monocotylous and dicotylous species was higher in disturbed than in undisturbed plots (po0:0001 for both cases) (Table 2). Dung deposition had no effect on the cover or the number of monocotylous species (p ¼ 0:81 and 0.08), and on the cover of dicotylous species (p ¼ 0:11), but the number of dicotylous species was significantly higher in the plots with dung. The total number of species and the Shannon–Wiener diversity were higher in disturbed plots (po0:0001 for both variables) and in plots where dung was deposited (po0:0001 and p ¼ 0:059, respectively), although difference in Shannon–Wiener diversity is only marginally significant (Table 2).
E. Cosyns et al. Six out of 10 species had a significantly larger number of individuals in disturbed than in undisturbed plots, while there were only four species (C. fontanum, J. bufonius and V. arvensis/chamaedrys) with a higher number of seedlings in the plots with dung (Table 3).
Discussion Seed germination success after dung deposition Dung deposition resulted in a significant increase in monocotylous seedlings after 3 months both in disturbed and undisturbed plots. Monocotylous species were dominant in the vegetation and were abundantly present as viable seeds in the dung. However, only a low proportion (less than 5%) germinated in the field within 3 months after dung deposition. Seeds that do not germinate immediately may replenish the soil seed bank, leaving opportunities for establishment later. Seeds of dicotylous species, in contrast, were at lower densities dispersed by the herbivores, and the increase of dicotylous seedlings after 3 months on the plots with dung deposition was hence not significant. Dicotylous species are much less abundant in the vegetation, which may explain their lower contribution to seed density in the dung. The higher number of monocotylous seedlings after 3 months in plots with dung deposition can be attributed to a combined effect of seed dispersal on the one hand and gap creation and nutrient enrichment by the decomposing dung on the other hand. These gaps and the associated fertilisation effect will however also offer opportunities for germination and establishment of seeds present in other sources than the dung (Bonis, Grubb, & Coomes, 1997; Dai, 2000). The proportion of seeds germinating from the dung may hence be an overestimate. However, the number of viable seeds in the dung was very high, especially for monocotylous species. Moreover, the number of monocotylous seedlings was found to be higher on sodcut soils with dung deposition than on sod-cut soil where no dung was deposited, which is most likely to be a direct result of a higher seed availability, rather than a consequence of gap creation. It can indeed be assumed that the effect of extra gap creation by the dung on sod-cut plots is negligible in comparison to the large gaps created by sodcutting. Moreover, dung deposition may also have a negative physical effect on seed germination, by lowering the level of light penetrating to the soil or
106.1 (12.3) 11.4 (6.7) 3.3 (0.9) 5.5 (1.9) 8.8 (2.4) 1.14 (0.3)
5.3 (6.8) 52.8 (43.8) 58.1 (47.5) 98.3 (17.6) 16.2 (17.2) 3.6 (1.4) 7.6 (2.3) 112 (2.8) 1.22 (0.3)
50.9 (35.6) 61.3 (34.5) 112.2 (40.8) 101.2 (11.0) 13.7 (5.7) 3.6 (1.4) 7.4 (2.0) 11.0 (2.6) 1.27 (0.4)
42.5 (19.1) 49.8 (24.1) 92.3 (35.1) 58.0 (24.9) 19.0 (7.6) 4.0 (1.2) 8.2 (1.7) 12.2 (2.0) 1.71 (0.3)
24.8 (19.1) 80.3 (46.6) 105.2 (60.3) 66.0 (24.5) 24.9 (4.7) 4.9 (0.9) 10.3 (1.3) 15.2 (2.4) 2.01 (0.3)
49.4 (20.9) 83.5 (24.1) 132.9 (26.4)
Cattle dung
67.8 (24.5) 29.2 (6.3) 4.4 (0.9) 10.3 (2.0) 14.8 (1.5) 1.98 (0.3)
45.8 (18.7) 94.7 (46.6) 140.4 (54.9)
Horse dung
126.4***
50.9***
40.1***
20.4***
26.3***
118.2***
16.8***
15.0***
2.4
Dist. (F1,55)
4.5
10.6***
10.0***
2.6
3.3
0.2
7.3**
0.2
22.7***
Dung (F2,55)
Anova (F-statistic)
2.4
1.7
1.4
5.2***
1.4
5.9***
2.8*
2.2
2.6*
Block (F11,55)
0.9
0.1
0.04
0.9
1.4
2.2
0.9
0.7
1.9
Dist. Dung (F2,55)
The F-statistics of the mixed model ANOVA are shown on the right (Dist. ¼ disturbance, i.e. sod-cut or not, Dung ¼ no dung, cattle or horse dung) with indication of the Bonferroni corrected p-value. Standard deviations are given in parenthesis. *: 0.01o corrected p-value o0.5; **: 0.001o corrected p-value o0.01; ***: Corrected p-value o0.001.
Shannon–Wiener diversity
Species richness
Number of dicotylous species
Number of monocotylous species
Cover of dicotylous species
After 1 year Cover of monocotylous species
Total number of seedlings
Number of dicotylous seedlings
After 3 months Number of monocotylous seedlings
No dung
Horse dung
No dung
Cattle dung
Sod-cut
Intact vegetation
Table 2. Average number of monocotylous, dicotylous and total number of seedlings and proportion of monocotylous and dicotylous seeds that germinated 3 months after dung deposition, and cover of monocotylous and dicotylous species, number of monocotylous and dicotylous species, species richness and Shannon–Wiener diversity of the plant community 1 year after dung deposition for the different treatments (N ¼ 12)
ARTICLE IN PRESS
Establishment after endozoochory 365
ARTICLE IN PRESS
With or without cattle or horse dung addition (N ¼ 12). The F-statistics of the mixed model ANOVA are shown on the right (Dist. ¼ disturbance, i.e. sod-cut or not, Dung ¼ no dung, cattle or horse dung) with indication of the Bonferroni corrected p-value. Standard deviations are given in parenthesis. *: 0.01o corrected p-value o0.5; **: 0.001o corrected p-value o0.01; ***: corrected p-value o0.001.
6.0*** 1.4 0.7 1.0 2.6 2.1 1.7 1.1 2.8 1.8 0.6 17.9*** 4.6 7.6 0.002 33.1*** 1.3 5.9 1.2 8.3* 5.50 7.08 1.83 1.50 0.25 8.42 4.92 0.42 6.92 2.50 (9.1) (3.8) (1.5) (1.2) (0.3) (8.8) (6.4) (0.8) (10.7) (1.0) 4.25 7.50 1.75 0.67 0.08 12.75 4.00 0.67 9.17 1.17 (4.4) (2.7) (1.5) (0) (0.6) (1.5) (3.6) (1.2) (3.5) (1.4) 2.17 3.42 1.25 0.00 0.17 1.42 2.33 0.42 3.75 0.67 (0.8) (1.7) (0.6) (2.0) (0) (0.9) (3.2) (1.5) (0.3) (1.0) 0.25 2.58 0.75 0.75 0.00 0.33 1.58 1.00 0.08 1.00 (0.6) (3.0) (0.5) (1.0) (0.3) (1.5) (1.2) (1.1) (0) (1.4)
Cattle dung
0.25 3.75 0.25 0.50 0.08 0.92 0.58 0.92 0.00 1.33 (1.4) (1.7) (0.3) (0) (0.3) (0.9) (4.3) (0.5) (0) (0.5) 0.58 1.58 0.08 0.00 0.08 0.25 1.83 0.17 0.00 0.33 Arenaria serpyllifolia Cerastium fontanum Epilobium spp. Galium spp. Juncus articulatus Juncus bufonius Sagina proc.+apetala Trifolium repens Urtica dioica Veronica arven.+cham.
Horse dung Cattle dung No dung No dung
Horse dung
Sod-cut Intact vegetation
Mean number of individuals
(9.7) (4.2) (1.6) (2.9) (0.5) (5.2) (2.6) (0.5) (10.0) (3.6)
9.7* 19.2*** 21.2*** 0.3 1.2 71.6*** 10.1* 0.08 51.0*** 0.4
Dung (F2,55) Dist. (F1,55)
Anova (F-statistic)
Block (F11,55)
0.1 1.7 0.1 0.3 0.4 21.0*** 2.2 0.7 0.9 0.1
Dist. Dung (F2,55)
E. Cosyns et al.
Plant species
Table 3.
Mean number of individuals of a selection of ten plant species that occurred one year after dung deposition in sod-cut and intact vegetation plots
366
by physically hampering the growing seedlings. Considering the nutrient enrichment, Ampe, Ngugi, and Langohr (2002) found in the same study area only significantly more exchangeable K+ and K+saturation and no differences in other soil chemical characteristics in the upper 5 cm soil layer under medium aged and old horse and cattle dung in comparison with the surrounding soil. This suggests that the fertilisation effect may be not that important. It is hence very likely that the increase of monocotylous seedlings is at least partly a result of an increased seed input, although we cannot exclude the possible relation between the observed higher seedling establishment success after dung deposition and some gap creation or fertilisation effect. Large-scale disturbance had an important effect on seedling emergence, especially for the less-dominant dicotylous seedlings. The results clearly indicate that the removal of the upper 5 cm of the soil creates safe sites for germination and seedling establishment. However, disturbance also resulted in a significant increase of seedlings in plots where no dung was deposited. This indicates that after 3 months also seeds from other sources than dung, such as the soil seed bank or long distance seed rain, will have taken advantage of the gap creation effect. Several studies found also a higher germination and seedling establishment rate in plots where the vegetation was removed or the soil was disturbed (Austrheim & Eriksson, 2003; Carson & Pickett, 1990; Jutila & Grace, 2002). This can be explained by the combined effect of a higher light penetration down to the soil, a competitive release and litter removal (Jutila & Grace, 2002). A higher light level will trigger germination and promote seedling growth. As a result of a decreased competition for light or nutrients, the probability of successful establishment is increased. Spackova, Kotorova, and Leps (1998) found that seedling recruitment is even more sensitive to competition than the established vegetation. Litter is also found to be an important factor negatively influencing seed germination and seedling establishment success (Jensen & Gutekunst, 2003). It seems that mainly dicotylous species take advantage of these safe sites created by disturbance. This indicates that these species, which are much less dominant in the vegetation, may be hampered by a shortage of safe sites for germination within the dense grass vegetation. The increase of monocotylous seedlings in disturbed plots is less pronounced, suggesting that they have more opportunities for germination and seedling establishment within the established vegetation.
ARTICLE IN PRESS Establishment after endozoochory
Effects on plant community composition and structure after 1 year One year after dung deposition, three species had more individuals in the dung deposited plots (C. fontanum, J. bufonius and V. arvensis/chamaaedrys), but several species had more individuals in the disturbed plots than in the undisturbed plots (Table 3). For J. bufonius, the most abundant species present in the dung, there is a significant interaction term between dung deposition and disturbance. This means that for this species, viable seeds dispersed in the dung have a higher probability of successful establishment after 1 year when the dung is deposited in large gaps. This may indicate that the gaps associated with decomposing dung do not contribute largely to a higher seedling establishment success, and that their effect is not at all comparable with the effect of the large gaps created by sod-cutting. It is however likely that this is partly a result from the artificial spreading of the dung samples. Dung deposited directly by the animals themselves on the vegetation may have a much stronger gap creating effect. For cattle dung, the experimentally spread layer was thinner than the dung pats found in the field (2 versus 4–5 cm). For pony dung, the thickness of the layer did not differ, but the density (the number of pats/area unit) was higher for naturally deposited dung. This means that the experimentally spread dung probably has a smaller gap creating effect. We indeed noticed that the gaps that had been created in the vegetation by the decomposition of the dung in the experimental plots disappeared after a few months. On the other hand, a thinner layer, or a smaller density, may also result in a higher germination success of the seeds in the dung, because a higher proportion of the deposited seeds receive sufficient light to start germination. Dung deposition had no effect on the cover or the number of monocotylous species after 1 year, neither in disturbed nor in undisturbed plots, although more viable seeds of monocotylous species were dispersed and an increased number of monocotylous seedlings was found 3 months after dung deposition. The increase of monocotylous seedlings after 3 months seems hence to be mainly a short-term effect that had no longer-lasting influence on community composition after 1 year. Although dicotylous species were less abundantly dispersed by the horses and cattle, they had a higher establishment success after 1 year, resulting in a significant increase in the number of dicotylous species in the local plant community. This indicates that stochastic events such as endozoochorous dispersal will mainly have a positive effect on the less-dominant dicotylous
367 species by increasing opportunities for successful seed germination and seedling establishment. Disturbance also caused a pronounced increase of the number of dicotylous species, both in the plots with and without dung deposition, confirming the results of other studies (Austrheim & Eriksson, 2003; Carson & Pickett, 1990; Jutila & Grace, 2002). Although total cover was still lower in all disturbed plots because 1 year is insufficient to restore total cover on the disturbed sites (Carson & Pickett, 1990), dicotylous species already had a higher cover in disturbed plots. Both soil disturbance and dung deposition hence resulted in an increase of small-scale species richness, due to the establishment of more dicotylous species. The safe sites for germination and establishment in combination with seed input in the case of dung deposition allow more species to co-exist locally due to an increase of the number of safe sites. This is here mainly expressed by an increasing importance of dicotylous species, which are able to co-exist with the more dominant monocotylous species in the plant community. This resulted in an increased species richness and diversity of the plant community on sod-cut soil, or on soil where dung was deposited. The increase of diversity was much more pronounced after sodcutting. This is however due to the locally young successional age of the plant community on sod-cut soil. It can be expected that the effects of sodcutting will become smaller with time, since competition by monocotylous species will increase when they are gradually colonising the large gap. Dung deposition implies the combined effect of dicotylous seed input and safe site creation. Since the number of dicotylous species was higher in sodcut plots where dung was deposited than in sod-cut plots without dung deposition, here also, the increase of dicotylous species richness in these plots after 1 year is likely to be a result of higher seed availability. Moreover, this means that seedling establishment after dung deposition is still enhanced by the creation of safe sites by sodcutting, and again confirms that the gaps created by the spread-out dung are not very important for seedling establishment. However, this effect may be much larger for naturally deposited dung pats.
Conclusions The most important effect of dung deposition after 3 months was an increase of monocotylous seedlings that were abundantly dispersed as viable seeds in the dung. One year later, however, it seems
ARTICLE IN PRESS 368 that mainly dicotylous species had been able to profit from the opportunities of seed dispersal and safe site creation. Dung deposition did not result in an increase of species richness at the site level, since the establishing species were in most cases already present at the site. Viable seeds of rare species were present in the dung, but they did not seem to establish. Dung deposition rather led to an increase of small-scale species richness as a result of a decrease in dominance of monocotylous species and the related change in species co-existing patterns. This was expressed in a higher small-scale diversity of the plant community 1 year after dung deposition. The increase of diversity after dung deposition was much more pronounced on sod-cut soil, which indicates that the probability of successful establishment after seed dispersal and gap creation by dung deposition is higher on disturbed soil with an increased number of safe sites.
Acknowledgements We thank the Ministry of the Flemish Community, AMINAL, Department of Nature for permitting this research project in the ‘Westhoek’. The Flemish Institute for the Sea (VLIZ) provided the greenhouse accommodation. Indispensable practical help came from Ward Vercruysse (Institute for Nature Conservation) and Frank Broucke (AWZ-WWK). We are also grateful to two anonymous reviewers for their constructive comments that helped us to improve the manuscript.
References Ampe, C., Ngugi, N. M., & Langohr, R. (2002). Impact of recently introduced large herbivores on soil properties of coastal dune soils of the ‘Westhoek’ nature reserve Belgium. In: EUROCAST (Ed.), Littoral 2002, the changing coast (pp. 433–438). Eurocast, EUCC, Porto. Austrheim, G., & Eriksson, O. (2003). Recruitment and life-history traits of sparse plant species in subalpine grasslands. Canadian Journal of Botany, 81, 171–182. Bekker, R. M., Bakker, J. P., Grandin, U., Kalamees, R., Milberg, P., Poschlod, P., et al. (1998). Seed size, shape and vertical distribution in the soil: Indicators of seed longevity. Functional Ecology, 12, 831–842. Bonis, A., Grubb, P. J., & Coomes, D. A. (1997). Requirements of gap-demanding species in chalk grassland: Reduction of root competition versus nutrient enrichment by animals. Journal of Ecology, 85, 625–633. Burke, M. J. W., & Grime, J. P. (1996). An experimental study of plant community invasibility. Ecology, 77, 776–790.
E. Cosyns et al. Carson, W. P., & Pickett, S. T. A. (1990). Role of resources and disturbance in the organization of an old-field plant community. Ecology, 71, 226–238. Claerbout, S. (2001). Potentie ¨le endozoo ¨chore zaadverbreiding door enkele herbivore zoogdieren. M.sc. thesis, University Ghent, Ghent. Collins, S. L. (1987). Interaction of disturbances in tallgrass prairie: A field experiment. Ecology, 68, 1243–1250. Cosyns, E., Claerbout, S., Lamoot, I., & Hoffmann, M. (2005). Endozoochorous seed dispersal by cattle and horse in a spatially heterogeneous landscape. Plant Ecology, 178, 149–162. Cosyns, E., Delporte, A., Lens, L., & Hoffmann, M. (2005). Germination success of temperate grassland species after passage through ungulate and rabbit guts. Journal of Ecology, 93, 353–361. Cosyns, E., & Hoffmann, M. (2005). Horse dung germinable seed content in relation to plant species abundance and seed characteristics. Basic and Applied Ecology, 6, 11–24. Dai, X. (2000). Impact of cattle dung deposition on the distribution pattern of plant species in an alvar limestone grassland. Journal of Vegetation Science, 11, 715–724. Jakobsson, A., & Eriksson, O. (2000). A comparative study of seed number, seed size, seedling size and recruitment in grassland plants. Oikos, 88, 494–502. Jensen, K., & Gutekunst, K. (2003). Effects of litter on establishment of grassland plant species: The role of seed size and successional status. Basic and Applied Ecology, 4, 579–587. Jutila, H. M., & Grace, J. B. (2002). Effects of disturbance on germination and seedling establishment in a coastal prairie grassland: A test of the competitive release hypothesis. Journal of Ecology, 90, 291–302. Kent, M., & Coker, P. (1995). Vegetation description and analysis. West Sussex, England: Wiley. Lambinon, J., De Langhe, J. E., Delvosalle, L., & Duvigneaud, J. (1998). Flora van Belgie ¨ het Groothertogdom Luxemburg Noord-Frankrijk en de aangrenzende gebieden (Pteridofyten en Spermatofyten). Meise: National Botanical Garden Belgium. Malo, J. E., & Sua ´rez, F. (1995a). Herbivorous mammals as seed dispersers in a Mediterranean dehesa. Oecologia, 104, 246–255. Malo, J. E., & Sua ´rez, F. (1995b). Establishment of pasture species on cattle dung: The role of endozoochorous seeds. Journal of Vegetation Science, 6, 169–174. Mouissie, A. M., Vos, P., Verhagen, H. C. M., & Bakker, J. P. (2005). Endozoochory by free-ranging, large herbivores: Ecological correlates and perspectives for restoration. Basic and Applied Ecology, 6, 547–558. Neter, J., Kutner, M. H., Nachtsheim, C. J., & Wasserman, W. (1996). Applied linear statistical models (4th ed.). Boston: WCB McGraw-Hill.
ARTICLE IN PRESS Establishment after endozoochory Olff, H., & Ritchie, M. (1998). Effects of herbivores on grassland plant diversity. Trends in Ecology and Evolution, 13, 261–265. Pakeman, R. J., Digneffe, G., & Small, J. L. (2002). Ecological correlates of endozoochory by herbivores. Functional Ecology, 16, 296–304. Pakeman, R. J., Engelen, J., & Attwood, J. P. (1998). Sources of plants colonizing experimentally disturbed patches in an acidic grassland in eastern England. Journal of Ecology, 86, 1032–1041. Spackova, I., Kotorova, I., & Leps, J. (1998). Sensitivity of seedling recruitment to moss, litter and dominant
369 removal in an oligotrophic wet meadow. Folia Geobotanica, 33, 17–30. SPSS. (2001). SPSS version 11.01. Chicago, IL: SPSS Inc. Traba, J., Levassor, C., & Peco, B. (2003). Restoration of species richness in abandoned Mediterranean grasslands: Seeds in cattle dung. Restoration Ecology, 11, 378–384. Welch, D. (1985). Studies in the grazing of heather moorland in North-East Scotland. IV Seed dispersal and plant establishment in dung. Journal of Applied Ecology, 22, 461–464.