• community structure;
  • competition;
  • functional traits;
  • response groups;
  • similarity


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    The response of an annual plant community to protection from grazing as a function of variation in site productivity was studied in a semi-arid Mediterranean rangeland in Israel over 4 years (1996–99). The abundance of species was compared in grazed vs. ungrazed plots (exclosures) in four neighbouring topographic sites (south- and north-facing slopes, hilltop and Wadi shoulders), representing a gradient of resource availability and productivity.
  • 2
    Above-ground potential productivity at peak standing crop in spring (i.e. inside exclosures) varied considerably between years and topographic sites. Productivity was similar among the hilltop, south- and north-facing slopes, and was typical of semi-arid ecosystems (10–200 g−2). Productivity in the Wadi was consistently greater (up to 700 g−2) and reached the range of subhumid grassland ecosystems.
  • 3
    The effect of grazing exclusion on the composition of the annual vegetation was productivity-dependent. Lower similarity (Sorenson's quantitative similarity index) between grazed and ungrazed subplots was observed in the productive Wadi compared with the less productive sites. The small-scale variation in grazing impact on species composition, due to differences in productivity, is consistent with models predicting similar trends in perennial grasslands across larger scale gradients.
  • 4
    The relationship between plant size (above-ground dry-weight), site productivity and response to fencing was analysed for the 36 most abundant annual species. Large species were more abundant in more productive sites, and small species at lower productivity, although few species were restricted to particular productivity levels. The response of individual species to protection from grazing was productivity dependent, with plant size playing a central role. Larger species tended to increase and small ones to decrease in abundance after fencing, with a mixed response in species with intermediate size.
  • 5
    A conceptual model is presented relating the response to protection from grazing along gradients of productivity to species plant size.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Grazing by large herbivores has been shown to interact with primary productivity in determining plant community structure (Milchunas & Lauenroth 1993; Proulx & Mazumder 1998; Osem et al. 2002). In their conceptual model, Milchunas et al. (1988) predicted that modifications in plant community composition caused by grazing are more likely to occur, and will be larger, with increasing productivity, due to ‘divergent selection’ for grazing vs. competition for light. Competition for light will select for traits such as taller growth forms with larger leaves and faster growth (Keddy 1989; Gaudet & Keddy 1995; Grime 2001), but should make plants more vulnerable to grazing (Noy-Meir et al. 1989; McIntyre et al. 1995; Sammul et al. 2000; Díaz et al. 2001). On the other hand, adaptations to aridity such as shorter plants, small leaves, basal meristems and annual life cycle, should increase tolerance to, or avoidance of, grazing (‘convergent selection’) (Coughenour 1985; Milchunas et al. 1988). Therefore, moderate or small changes in the composition of plant communities in response to grazing are expected in semi-arid rangelands with lower productivity in which plant growth is usually limited by soil resources.

The conceptual model of Milchunas et al. (1988) was generally supported by an extensive meta-analysis of the effects of grazing on community structure over a global range of environments (i.e. large scale) (Milchunas & Lauenroth 1993). However, when comparing ecosystems from different geographical regions confounding effects may mask the interactive effects of productivity and grazing on community structure. This is due to other factors, such as vegetation composition and physiognomy, evolutionary history, grazing regime and dominant grazers. We propose that a small-scale approach, based on comparisons among neighbouring topographic sites differing in productivity, in which the plant community evolved under similar biotic and abiotic conditions, may allow a finer resolution of the interactive effects of productivity and grazing on plant community structure. We hypothesize that the trends observed in community structure due to grazing across ecosystems differing in primary productivity (i.e. large scale) also occur among neighbouring topographic sites differing in their productivity (i.e. small scale).

In their model Milchunas et al. (1988) proposed that plant growth rate and plant size are major traits involved in the mechanistic interactions between productivity and grazing, as they are related to both competitive ability and tolerance or avoidance of grazing and drought (Coughenour 1985; Grime 2001). This model and the supporting meta-analysis (Milchunas & Lauenroth 1993) were based mainly on grazing responses observed in perennial grasslands and shrublands, due to scarcity of information on annual grasslands, particularly those from semi-arid regions. However, the relative importance of growth rate and plant size traits along productivity gradients may differ between perennial and annual grasslands, as annual species lack temporal continuity in competitive interactions. Therefore, changes in vegetation structure of annual grasslands due to grazing may depend to a larger extent on seed-bank dynamics and seedling establishment than in perennial grasslands, particularly in semi-arid regions with lower plant density (Briske & Noy-Meir 1998).

Grassland species exhibiting similar responses to grazing have been traditionally categorized into grazing response groups, such as increasers, decreasers and invaders (Dyksterhuis 1949; Noy-Meir et al. 1989; McIntyre et al. 2003). Within these groups, species frequently share traits regarding life cycle, growth form, plant size, palatability and defence mechanisms that are presumably related to their tolerance to, or avoidance of, grazing (Noy-Meir et al. 1989; Díaz et al. 2001; McIntyre & Lavorel 2001; Vesk & Westoby 2001). The importance of these traits in understanding the dynamics of vegetation change and associated ecological processes in response to grazing is well established (Noy-Meir et al. 1989; Fernandez Alés et al. 1993; McIntyre et al. 1995; Lavorel et al. 1999). Plant size and height are traits frequently considered as robust predictors of species response to grazing (Noy-Meir et al. 1989; Lavorel et al. 1997; Hadar et al. 1999; Lavorel et al. 1999; Sternberg et al. 2000; Díaz et al. 2001; Dupre & Diekman 2001; McIntyre & Lavorel 2001). However, plant size may vary considerably along environmental gradients determining productivity, particularly in the case of annuals (Aronson et al. 1990). This variation should be taken into account when studying species response to grazing, as species ranking by size and their categorization in size groups may change along environmental gradients regardless of grazing impact. A related, cardinal question for species categorization into grazing response groups is whether their response to grazing, or to protection from grazing, is consistent (i.e. repeatable under different conditions), or dependent on grazing regime and habitat conditions (i.e. context dependent). Vesk & Westoby (2001) proposed that inconsistent responses might result from context characteristics, such as abiotic conditions, neighbour species, competitive relationships, grazing regime and the grazing herbivore. McIntyre et al. (2003) also reported that environmental variables, such as soil disturbances and water enrichment, influence species response to grazing. Lack of consistency in grazing response will complicate predictions about the effects of grazing on the behaviour of individual species, composition of response groups and assumptions about vegetation dynamics (Gitay & Noble 1997; Garnier et al. 2001).

In our research we hypothesized that: (a) the response of annual species to grazing (i.e. magnitude and direction of change in their abundance) is context dependent, and may vary with productivity level among years differing in rainfall and across neighbour topographic sites with different microclimatic and edaphic conditions; (b) species plant size and range of variation play a central role in determining the response of annual species to grazing along gradients of productivity; and (c) grazing impact on species composition of annual plant communities in semi-arid regions increases with increasing productivity and associated changes in plant size, as proposed by the Milchunas et al. (1988) model for larger scales.

We studied these hypotheses in an annual plant community, with and without protection from grazing, in neighbouring topographic sites representing a gradient of primary productivity. The study was conducted in a Mediterranean semi-arid rangeland in Israel. In this type of semi-arid rangeland primary productivity is limited by soil resources, mainly water and nitrogen (van Keulen & Seligman 1992), which are patchily distributed due to a large heterogeneity of annual rainfall and habitat conditions (Noy-Meir 1973; Ludwig 1986). This heterogeneity, together with the dynamic nature of annual plant populations regenerating every year from the seed-bank, causes a wide spatial and temporal variation in primary productivity, within the range of arid and subhumid rangelands (Osem et al. 2002). Such rangelands provides a good opportunity to study, at the small scale: (a) changes in species response to grazing due to variation of plant size as a function of productivity; and (b) the interactive effects of grazing and productivity on species composition of annual plant communities under similar climatic conditions, grazing regime and evolutionary history.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

the study area

The research was carried out at the LTER Lehavim Research Station (the Bedouin Demonstration Farm), located in the Goral Hills at an altitude of 350–500 m a.s.l., near Beer-Sheva in the Northern Negev of Israel. Climate is Mediterranean semi-arid, with an average annual precipitation of 305 mm (range over 44 years: 78–504 mm year−1), occurring mainly during December to March (Baram 1996). Annual precipitation in the research years 1996, 1997, 1998 and 1999 was 223, 247, 235, and 40 mm, respectively. Average daily temperatures range from 10 °C in the winter to 25 °C in the summer (Baram 1996). Soils are loamy textured, of a brown desert skeletal type that develops on Eocene limestone, dolomite and chalk (Ravikovitch 1981). The area is grazed by a flock of about 600 Awassi sheep, starting in late January, after rainfall onset and plant establishment, and continuing until May (green pasture). During the summer the herd migrates to stubble fields and returns in August to graze on dry pasture until December. A shepherd moved the flock continuously over the range to ensure complete utilization of the available vegetation. Thus, grazing intensity varied according to the availability of above-ground biomass, with average residual biomass ranging between 25% and 50% in high and low productive sites, respectively.

natural vegetation

Vegetation in the study area corresponds to the Irano-Turanian phytogeographical region (Zohary 1973). Physiognomically it consists of a sparse shrub-land in which Sarcopoterium spinosum (L.) Spach, Corydothymus capitatus (L.) Reichenb. and Thymelaea hirsuta (L.) Endl. are the dominant shrubs. The herbaceous vegetation appears shortly after the first rains and persists as green forage for 3–5 months, depending on the amount and distribution of rainfall. This vegetation is highly diverse, mostly composed of annual species that represent 56% of the regional flora (Danin & Orshan 1990). About 130 annual species have been found at the Lehavim Station, of which 15 species contribute 85% of the total average abundance and above-ground biomass of the annuals (Osem et al. 2002). The predominant species are grasses (46% of total abundance and 51% of above-ground biomass) and legumes (9% and 21%, respectively). This annual vegetation has been foraged by domestic grazers (mainly sheep and goats) since prehistoric times (5000–8000 years), and is mostly composed of grazing resistant species (Noy-Meir & Seligman 1979). Less resistant species probably became extinct long ago (Perevolotsky & Seligman 1998).

experiment design and sampling

This work is part of a long-term study on rangeland management at Lehavim. The experiment was conducted in a heavily grazed area near the sheep pen, in four different neighbouring topographic sites: Wadi (dry stream) shoulders, hilltop, and south- and north-facing slopes (incline 12–15°). Although the sites are adjacent, they differ in abiotic conditions due to their contrasting aspect and inclination, which determine solar radiation load, temperature, soil properties and water balance. In particular, the Wadi shoulders differ from the other sites: soil is deeper, with higher organic matter content and higher water holding capacity, and they receive additional water run-off from the slopes during rainfall. These differences in abiotic conditions cause variation in the density and composition of the herbaceous and woody vegetation in the different sites. In addition, shrubs are absent in the Wadi shoulders, largely due to past cultivation and use as fuel. The study area is characterized by a patchy distribution of rocks and perennial shrubs (except the Wadi), with the open patches between them partly covered by herbaceous vegetation.

Four permanent exclosures (10 × 10 m) that prevented sheep grazing were established in each topographic site in 1993. The most distant exclosures were separated by only 370 m. Vegetation samples were collected in each site at peak standing crop (spring, April) over 4 years (1996–9), inside and outside the exclosures, to provide data on above-ground biomass, species composition and plant density. Samples were taken from 20 × 20 cm quadrats, randomly distributed in open patches. This quadrat size is commonly used to study Mediterranean herbaceous communities (Montalvo et al. 1993), where it is appropriate to the spatial scale of plant interactions. At each topographic site, samples were taken from the four plots (each exclosure and its surrounding area are considered a plot ‘nested’ within the site). In each plot, five samples were collected inside the exclosure (ungrazed subplot) and five outside it (grazed subplot). Annual plants represented most of the harvested biomass. They were identified and counted in each sample. Above-ground dry biomass was determined separately for each species after 48 h oven drying (70 °C).

parameters and definitions

Productivity was defined as the above-ground dry biomass of the annual plants per unit area (g m−2) in open patches at peak standing crop. In this system, in which annual species are the main component of the herbaceous vegetation, spring biomass is a good indicator of annual productivity. Dry biomass was determined after 48 h drying at 70 °C in an air forced oven. The average biomass of five samples within an exclosure was taken as the actual primary productivity of the ungrazed subplot and as the potential primary productivity of the adjacent grazed subplot. The average biomass of the 20 samples from the four exclosures in each site was taken as the potential annual productivity of the site. The standing biomass of the annual vegetation (i.e. productivity) is considered as an indicator of resource availability in the whole plot (grazed and ungrazed). Abundance is the average number of plants per sample (20 × 20 cm quadrat). Relative abundance is the average proportion of plants of a given species or group from the total number of plants in the sample.

Plant size was represented by its above-ground dry-weight inside the exclosure at peak standing crop. To enable categorization into distinct plant size groups we calculated the average plant weight across the low (< 200 g m−2) and high productivity plots (> 200 g m−2). Plant height is more commonly used as a measure of plant size in grazing studies, but (a) most species in our system are short and still foraged by the sheep that are able to graze at ground surface level, (b) many of the annual species, including some of the larger and more competitive ones, have a prostrate or rosette growth form, and (c) greater plant height is frequently due to erect inflorescences that develop at the end of the growing season. In such situations plant weight is a good measure of accessibility to sheep, and also of growth rate and species competitiveness.

Species similarity between grazed and ungrazed subplots in each topographic site and between sites was assessed by Sorenson's quantitative similarity index CN = 2jN/(aN + bN), in which aN is the total number of individuals in site a, bN is the total number of individuals in site b and jN is the sum of the lower of the two abundances for each species in both sites (Magurran 1988). This index takes into account not only the presence and absence of species, but also the quantitative changes in their abundance.

statistical analysis

anovas were applied to study the effects of year (1996/97/98/99), site (Wadi, hilltop, south- and north-facing slopes) and fencing (grazed, ungrazed) on productivity, abundance, relative abundance and similarity. The analysis was fully factorial. Sampled plots were nested within the site. Residuals of productivity, plant density, abundance and relative abundance lacked normality (Shapiro-Wilk W-test). Hence, data were ranked (ties averaged) and the anova was applied to the ranked data. The analyses were therefore non-parametric (Conover & Iman 1981). Regression analysis was applied to study the relationship between productivity and variation in plant above-ground biomass of species. The statistical package JMP (SAS Institute 1995) was used.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

temporal and spatial variation in productivity

Productivity under protection from grazing varied greatly between years (F3,292 = 152, P < 0.0001) and among topographic sites (F3,12 = 10.3, P < 0.0012), ranging between 10 and 494 g−2 (Fig. 1). Productivity in the Wadi was consistently greater (two- to fivefold) than in the other sites (hilltop, south- and north-facing slopes), between which there were no significant differences. The two-way anova model explained 72% of the temporal and spatial variance in productivity, with year accounting for 64% and topographic site for 24% of the explained variance. The year–site interaction was low (3% of the explained variance; F9,292 = 2.68, P = 0.0053) and was mainly due to the extremely dry 1999/2000 season, that resulted in a larger biomass reduction in the most productive site (the Wadi).


Figure 1. Above-ground biomass of annual plants in ungrazed subplots, harvested at peak standing crop in April in four topographic sites: Wadi (•), hilltop (□), north- (▵) and south-facing (◊) slopes. Bars indicate SE.

Download figure to PowerPoint

variation in community composition in response to grazing exclusion and site topography

Analysis of vegetation similarity based on Sorenson's quantitative index, showed that similarity between the grazed and ungrazed subplots was significantly smaller (F3,292 = 15.8, P < 0.0001) in the Wadi (0.25 ± 0.02) compared with the less productive sites (hilltop 0.49 ± 0.04; south-facing slope 0.51 ± 0.02; north-facing slope 0.48 ± 0.04), which again did not differ from one another (Fig. 2). Similarity between grazed and ungrazed subplots in each topographic site was not, however, significantly affected by year, in spite of the large differences in productivity among years (Fig. 1). The two-way anova explained 56% of the variation in similarity between grazed and ungrazed adjacent subplots, with site accounting for 79% of the explained variation. We defined grazing intensity in each plot according to the fraction (%) of consumed biomass and compared similarity level in plots with biomass removal in the range 60–80%. In these plots similarity between grazed and ungrazed subplots was still significantly lower in the Wadi (0.25 ± 0.02, n = 13) compared with the less productive sites (0.45 ± 0.04, n = 12) (P < 0.0001), even though grazing intensity was comparable. The lower similarity found in the Wadi could not therefore be explained by more intense grazing.


Figure 2. Similarity in species composition (Sorenson's quantitative index) between grazed and ungrazed subplots, at peak standing crop, in four topographic sites: Wadi (•), hilltop (□), north- (▵) and south-facing (◊) slopes. Bars indicate SE.

Download figure to PowerPoint

We tested whether the lower similarity in the Wadi could be due to differences in composition of the vegetation before fencing by comparing similarity in species composition among sites. Under grazing, similarities between the Wadi vs. hilltop, north- and south-facing slopes were comparable with those between the less productive sites themselves (0.49 ± 0.03 and 0.52 ± 0.02, respectively). Grazing exclusion decreased similarity between the Wadi and the less productive sites (0.32 ± 0.03) (P < 0.05), but not among these sites (0.47 ± 0.02). Thus, exclusion of grazing leads to distinctive vegetation developing in the Wadi compared with the low productivity sites.

response groups and site productivity

Of the 128 annual species recorded in the study area, the frequency and abundance of most species was too low to allow statistical analysis of their response to protection from grazing. Therefore, only 36 species that reached either presence in 15% of the sampling units or 1% of the total abundance in at least one of the topographic sites were included in the response group analyses. These species, however, represented 96%, 96%, 93% and 87% of the total abundance of annuals in the south- and north-facing slopes, hilltop and Wadi, respectively.

Two-way anova showed that of these species, 28 were significantly affected by site and 27 showed a significant (P < 0.05) response to protection from grazing or to the interaction of protection × site. Several patterns of variation, defined according to significant behaviour of species, emerged from this analysis and were used to define response groups (Table 1). Species were included in a response group if they showed consistently similar trends of behaviour even if these were not significant (P > 0.05).

Table 1.  The abundance of annual plant species in grazed and ungrazed subplots in four topographic sites. Species are classified according to their response to grazing exclusion in the different sites. * P < 0.05, ** P < 0.01, *** P < 0.001. Values are number of plants per 20 × 20 cm quadrats
SpeciesTwo way anovaSouth-facing slope Hilltop North-facing slope Wadi 
FencingSiteFenc. × SiteGrazedUngrazedGrazedUngrazedGrazedUngrazedGrazedUngrazed
Wadi increasers
 Hordeum spontaneum******** 0.00 0.00  0.00 0.00  0.00 0.00  0.01 2.59***
 Avena sterilis******** 0.00 0.00  0.00 0.01  0.01 0.09  0.03 1.53***
 Medicago truncatula******** 0.14 0.08  0.25 0.08  0.07 0.13  1.20 4.74***
General increasers
 Aegilops spp.*****  3.3610.30*** 0.20 1.91*** 0.43 2.72* 1.17 5.96***
 Onobrychis crista-galli*** ** 0.34 1.96*** 0.32 1.62*** 0.57 1.42** 0.43 3.36***
 Hymenocarpus circinnatus*** ** 0.58 1.14* 0.32 1.43** 0.93 1.28  0.47 3.04***
 Crithopsis delileana *** 5.83 6.42  5.91 7.04**16.8730.16* 4.3421.89 
 Stipa capensis ** 0.25 0.30  0.67 1.63* 0.59 0.50  1.96 6.61 
 Carrichtera annua******** 0.61 0.66  0.41 1.83*** 0.38 0.75  0.08 0.46 
 Calendula arvensis *  0.03 0.00  0.01 0.04  0.03 0.26* 0.25 0.32 
 Scabiosa porphyroneura*   0.22 0.21  0.01 0.01  0.16 1.11  0.13 0.24 
 Medicago minima *  0.05 0.03  0.21 0.33  0.28 0.46  0.39 0.68 
Wadi decreasers
 Crepis sancta********* 0.00 0.01  0.01 0.00  0.11 0.01  1.20 0.17***
 Crepis aspera********* 0.00 0.00  0.00 0.00  0.07 0.03  1.71 0.42***
 Minuartia hybrida******* 0.08 0.05  0.17 0.00  0.01 0.00  0.74 0.17**
 Polycarpon tetraphyllum******** 0.00 0.00  0.00 0.00  0.00 0.04  0.63 0.01***
General decreasers
 Trisetaria macrochaeta******  1.53 1.32 16.3816.57 12.71 4.62*** 1.38 0.17**
 Filago spp.******* 1.92 1.29*10.80 5.14* 4.18 3.34 11.75 1.17***
 Catapodium rigidum    0.13 0.08  0.70 0.33  0.14 0.20  0.86 0.07*
Site dependents
 Hippocrepis unisiliquosa *** 0.61 0.89  0.12 0.38* 0.13 0.12  0.33 0.01***
 Sedum pallidum****** 0.03 0.37* 1.29 2.71* 0.11 0.14  0.11 0.00*
 Trigonella monspeliaca ****** 0.16 0.39  0.42 0.45  0.47 1.13** 0.72 0.20**
 Linum strictum  *** 0.62 0.32  0.47 0.13  0.08 0.49*** 0.36 0.03**
 Erodium malacoides **** 0.00 0.03  0.00 0.07* 0.21 0.12  0.28 0.05**
 Anthemis hebronica ****** 0.07 0.14  0.18 0.33  0.58 0.78  2.22 1.76***
 Plantago coronopus* *** 0.57 1.21  1.99 2.66  1.41 1.97  2.49 0.61***
 Bromus fasciculatus  * 0.04 0.00  0.26 0.99  0.64 5.33* 0.58 0.17 
 Trifolium campestre *  0.00 0.01  0.00 0.00  0.12 0.89* 0.46 0.21 
 Salvia horminum****** 0.01 0.18** 0.03 0.00  0.62 1.26** 0.14 0.05 
 Plantago cretica*******15.28 9.84**15.0824.18* 8.0419.96**10.36 0.88***
 Brachypodium distachyon**** 0.86 2.32* 1.46 0.11  0.28 5.21* 0.00 0.00 
 Bromus alopecuros    0.05 0.01  0.30 0.29  0.41 0.16  0.41 0.17 
 Torilis tenella **  0.00 0.00  0.03 0.01  0.11 0.16  0.79 1.14 
 Chaetosciadium trichosperm ***  0.00 0.01  0.04 0.05  0.50 0.74  0.14 0.39 
 Anagallis arvensis *  0.24 0.25  0.03 0.01  0.00 0.00  0.17 0.05 
 Erodium cicutarium *  0.00 0.00  0.04 0.05  0.11 0.04  0.32 0.08 
Total increase     13.1  20.4  43.9   41.5
Total decrease     −6.9  −8.1  −9.6  −30.7
Number of increasers      6  10  11    6
Number of decreasers      2   1   1   15

Six response groups were identified:

  • 1
    Wadi protection increasers: three species, present in the high productive Wadi, where they increased in abundance after fencing, but absent or infrequent in the other less productive sites.
  • 2
    General protection increasers: nine species, present in all topographic sites, which increased in abundance after fencing both in the Wadi and in one or more of the low productivity sites.
  • 3
    Wadi protection decreasers: four species, present in the Wadi, but absent or with very low presence in the other sites, which decreased in abundance after fencing.
  • 4
    General protection decreasers: three species, present in all topographic sites, which decreased in abundance after fencing in one or more of the low productivity sites, as well as in the Wadi.
  • 5
    Site productivity-dependent species: 10 species, present in all sites, whose abundance after fencing increased in one or more of the low productivity sites and decreased in the high productivity Wadi. Two species with slightly different site-dependent responses are also included.
  • 6
    Uncategorized species: five species, which were present in all or most sites, but lacked a clear trend in their response to protection from grazing.

As expected, all Wadi-specific responders (i.e. groups 1 and 3) had a strong interaction between site and grazing exclusion (P < 0.001), as well as nine out of the 10 site-dependent species (due to their opposite response to protection in the Wadi vs. the less productive sites). In the general increasers group, seven out of nine species showed a significant response to fencing (P < 0.05), usually strongest in the Wadi, although two, Crithopsis delileana and Stipa capensis, were significantly affected by fencing only in the less productive sites (Table 1). These two species showed highly patchy distribution in the Wadi, resulting in high variability in abundance in this site and thus lack of significance, despite their large increase after fencing. In the general decreasers group, two out of three species decreased significantly after grazing exclusion (P < 0.001), while the other one decreased in all sites but only significantly in the Wadi (P < 0.05).

Two species, Plantago cretica and Brachypodium distachyon, showed atypical site-dependent responses. P. cretica is the most abundant species in the study area, contributing 13–39% and 2–27% of the total abundance in the grazed and ungrazed subplots, respectively. Protection caused a significant decrease in the Wadi (P < 0.0001), and significant increases in the hilltop (P < 0.05) and in the north-facing slope (P < 0.01) but, unlike other site-dependent species there was a decrease in the south-facing slope (P < 0.01) (Table 1). Brachypodium distachyon behaved as an increaser in the low productive sites, but was absent in the Wadi.

The numbers of species showing a significant (P < 0.05) increase or decrease in abundance in response to fencing varied among topographic sites. In the less productive sites the number of increasers after grazing exclusion was much larger than the number of decreasers (6–11 vs. 1–2), while an opposite trend was observed in the more productive Wadi (6 vs. 15, respectively) (Table 1). The principal or only decreaser varied between low productivity sites (Filago spp., Plantago cretica and Trisetaria macrochaeta in the hilltop, south- and north-facing slopes, respectively) and these species were also the main decreasers in the Wadi, together with Plantago coronopus, Anthemis hebronica, Crepis sancta and Crepis aspera.

Patterns of change in abundance of each response group show that the rise in the total abundance of increasers due to protection from grazing was greater than the decline in the total abundance of decreasers in each topographic site (Table 1). Thus, total plant density was increased by protection in all sites, although the effect was more marked in the less productive sites, where the change in abundance of increasers was two- to fourfold that of decreasers, than in the Wadi, where the decline of decreasers was much greater.

plant size and response group

Ordering the 30 species with categorized response to protection from grazing according to their size (i.e. average above-ground weight) shows that their response was size dependent (Fig. 3). Except for the composite Crepis aspera, large species, of which two-thirds were grasses and legumes, responded to fencing as increasers. Medium-sized species showed site-dependent responses. Smaller species, apart from the increaser Crithopsis delileana, responded to fencing either as site-dependents or decreasers (including Crepis sancta, Minuartia hybrida and Polycarpon tetraphyllum, which had a very low frequency outside the Wadi). In spite of their small size, these Wadi decreasers probably lack physiological tolerance to the more xeric conditions in the less productive sites, where they are too rare to show a clear response.


Figure 3. Average plant above-ground biomass of annual species and classification in groups according to plant size and response to protection from grazing. W = species restricted to the Wadi site (absent or infrequent in the low productivity sites). Large species: Horspo =Hordeum spontaneum C. Koch.; Aveste =Avena sterilis L.; Onocri =Onobrychis cristagalli (L.) Lam.; Aegspp =Aegilops spp.; Scapor =Scabiosa porphyroneura Blakelock; Medtru =Medicago truncatula Gaertn; Hymcir =Hymenocarpus circinatus (L.); Sticap =Stipa capensis Thunb.; Calarv =Calendula arvensis L.; Creasp =Crepis aspera L.; Medmin =Medicago minima (L.) Bartal. Medium species: Carann =Carrichtera annua (L.) Asch; Salhor =Salvia horminum L.; Placor =Plantago coronopus L.; Hipuni =Hippocrepis unisiliquosa L.; Eromal =Erodium malacoides (L.) Willd.; Antheb =Anthemis hebronica Boiss. Small species: Cridel =Crithopsis delileana Roshey; Trimon =Trigonella monspeliaca L.; Tricam =Trifolium campestre Schreb.; Placre =Plantago cretica L.; Cresan =Crepis sancta (L.) Borum; Brofas =Bromus fasciculatus Presl.; Filspp =Filago spp.; Linstr =Linum strictum L.; Trimac =Trisetaria macrochaeta (Boiss.) Maire; Minhib =Minuartia hybrida (Vill.) Siskin; Catrig =Catapodium rigidum (L.) C.E. Hubbard; Polter =Polycarpon tetraphyllum L.; Sedpal =Sedum pallidum MB.

Download figure to PowerPoint

The mean plant above-ground biomass was 363 ± 95, 176 ± 33, 50 ± 8, 42 ± 22 and 19 ± 4 mg for Wadi increasers, general increasers, site-dependents, Wadi decreasers and general decreasers, respectively. Mean size of increasers (two groups combined) was significantly larger than that of the other response groups (P < 0.05), but no other differences were significant.

plant size, productivity and response to grazing exclusion

Plot productivity in different sites and years was categorized according to above-ground dry biomass, as low (< 150 g−2), moderate (150–250 g−2) or high (> 250 g−2). These levels of productivity are assumed to represent a range in which competition shifts from below- to above-ground resources (Osem et al. 2002). Species were categorized into size groups by their individual average above-ground dry-weight, as small (< 50 mg plant−1), medium (50–100 mg plant−1) or large (> 100 mg plant−1) (Fig. 4). Classification of species into the three size classes was robust and maintained across the productivity gradient, even though plant size was affected by both site and year. Regression analysis of variation in plant size as a function of productivity showed that, in 20 of the selected 36 species, in which a significant linear relationship (P < 0.05) was found, species affiliation to a particular size class did not change with productivity (Fig. 4). Categorization by size was maintained across the productivity gradient because the extent of size change with productivity was a linear function of average plant size (R2 = 0.92, P < 0.0001, n = 20). Of the 16 species in which the plant size vs. productivity relationship was not significant, eight were small species that did not show any clear trend of increase along the productivity gradient and were therefore always small. The remaining species had a relatively low frequency (Calendula arvensis, Trifolium campestre, Salvia horminum, Anagallis arvensis, Erodium cicutarium, Erodium malacoides), or were limited to the Wadi (Crepis sancta, Crepis aspera), precluding analysis of size changes with productivity.


Figure 4. Relationship between average plant size and productivity. Plant size is represented by above-ground biomass per plant. Lines represent species in which plant size is significantly related to productivity (P < 0.05), with broken lines delimiting the size groups.

Download figure to PowerPoint

The relative abundance of each size group was determined for the 640 vegetation samples collected in the different sites over 4 years, and averaged according to the three productivity levels (Fig. 5). When grazed, the relative abundance of the size groups changed relatively little with increasing productivity: small species decreased from 63% to 54% (P < 0.05), large species increased from 9% to 15% (P < 0.05), and the proportion of medium species remained constant at about 20%. Small species were thus much more abundant (P < 0.0001) than medium and large species, regardless of productivity level. Protection from grazing, however, caused significant changes in the relative proportion of large and small species, with the extent and direction of change being productivity dependent (fencing and productivity × fencing for both groups P < 0.0001, Fig. 5a). At low productivity fencing caused an increase from 9% to 24% in large species and a decrease from 63% to 50% in small species but had little effect on medium species. Despite these changes small species remained the most abundant group under protection from grazing at low productivity (P < 0.0001). But, increasing site productivity caused a gradual reversal in the relative abundance of the size groups in ungrazed subplots. In the high productivity sites, small species decreased from 55% to 15% (P < 0.0001), while large species increased from 15% to 45% (P < 0.0001), thus becoming the most abundant group (P < 0.0001), and medium species had no significant response.


Figure 5. Relative abundance (± SE) of small (○, < 0.05 g plant−1), medium (□, 0.05–0.1 g plant−1) and large (▵, > 0.1 g plant−1) species in (a) ungrazed and (b) grazed subplots from plots with low (< 150 g m−2), moderate (150–250 g m−2) and high (> 250 g m−2) productivity, sampled at peak standing crop in 1996–99.

Download figure to PowerPoint


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

effects of grazing exclusion on species composition are productivity dependent

Productivity-dependent effects of grazing are indicated by the lower similarity between the grazed and ungrazed subplots in the productive Wadi, compared with the less productive sites (hilltop, north- and south facing slopes). The trends imposed in this annual plant community by small-scale variation in productivity were comparable with those reported at regional scales for perennial plant communities along a precipitation gradient (Milchunas & Lauenroth 1993), where grazing tends to have smaller effects on species composition in dry grasslands than in humid grasslands and savannas. Taken together the results support the assumption that productivity is a major factor influencing the extent of changes in vegetation composition generated by grazing.

The effects of temporal (i.e. interannual) and spatial (i.e. topographical) variation of productivity on similarity in this semi-arid annual plant community were not, however, equivalent. A large reduction in productivity in the extremely dry year (1999) did not affect similarity between grazed and ungrazed subplots in the different sites, whereas, at a particular date lower productivity sites had higher similarity. Milchunas et al. (1989) also reported a smaller effect of interannual rainfall variation than of topography on community composition. The differences between temporal and spatial effects were probably related to seed-bank dynamics (Y. Osem, unpublished results), as interannual effects can be buffered by a relatively stable seed-bank, whereas effects of site productivity on composition accumulate over the years, resulting in divergent species compositions.

The larger changes in vegetation composition observed after grazing exclusion in the more productive Wadi, could be due to differences in either the initial species composition, or grazing intensity (i.e. relative biomass removal). However, under grazing, similarity in all paired comparisons between the four topographic sites was similar and quite high (c. 0.5). Furthermore, similarity levels in plots with comparable relative biomass removal were still lower in the Wadi than in less productive sites. The assumption that productivity is the main factor behind the differential response to fencing among topographic sites is therefore further supported.

species responses to grazing exclusion: the importance of productivity and plant size

Species are characterized by unimodal patterns of abundance distribution imposed by gradients of productivity (resource availability) (Austin 1990). Availability of soil resources may act as an environmental filter, selectively determining the establishment of annual species according to their growth requirements (Litav 1965; Kutiel & Noy-Meir 1986). Better availability of soil resources generally allows establishment of larger species, increasing competition for light and leading to competitive displacement of smaller, less competitive species (Newman 1973; Grubb 1985; Tilman 1988; Grime 2001). Thus, within the range of productivity in which each species occurs, its abundance increases with increasing availability of limiting soil resources, but decreases when higher levels of resources allow the establishment of larger and more competitive species.

Response patterns of our semi-arid annual species to protection from grazing were indeed dependent on plant size and productivity level. After fencing, large species were more abundant the higher the site productivity, while small species became more abundant with decreasing productivity. Generally, large species were palatable grasses and legumes that consistently increased with protection across the productivity gradient (Noy-Meir et al. 1989; Hadar et al. 1999; Sternberg et al. 2000), although they differed in their ability to establish and complete their life cycle in low productivity sites (Litav 1965; Kutiel & Noy-Meir 1986). Chances for the establishment of the largest and more mesic species in poorer sites were extremely low, even after fencing, and their distribution was practically restricted to the Wadi exclosures, leading to classification as Wadi protection increasers. Other large species, probably more stress-tolerant, were able to establish themselves even in the low productivity sites and increased with protection across the whole productivity gradient. Medium-size species usually had mixed productivity-dependent responses (typically an increase in the low productivity sites and a decrease in the more productive Wadi). Small species responded either as consistent protection decreasers (n = 6) or site-dependents (n = 5). Overall, a trend from increaser to decreaser was observed with diminishing size, with mixed responses in intermediate and some small species. Plant size, as presented in this study, is not an intrinsic property of the species, because it was measured in the field in competition with other plants, under protection from grazing. Thus, species whose population size increased after grazing exclusion were those who could achieve larger individual size in the absence of grazing. Comparable trends were also found in the seed-bank (in autumn, just before germination) of this annual plant community (Y. Osem et al., unpublished data). A decrease in the abundance or cover of short species and an increase of tall species after fencing has been reported for other Mediterranean (Noy-Meir et al. 1989; Fernandez Alés et al. 1993; Hadar et al. 1999; Lavorel et al. 1999; Sternberg et al. 2000) and semi-arid regions (Noy-Meir 1990). Diminishing size allows both avoidance from grazing and better adaptation to low resource availability, but it reduces competitive ability under higher productivity (Coughenor 1985).

A larger number of species responded to protection as decreasers in the Wadi compared with the less productive sites (15 vs. 1 or 2), because the combined effect of additional large species, reinforced by their greater increase in size in the productive Wadi, resulted in a stronger competitive displacement of the smaller species (Laffarga & Leiva 1991) and thus extensive changes in vegetation composition. Similar trends with increasing productivity have been reported in other systems (Fernandez Alés et al. 1993; Turkington et al. 1993; Sammul et al. 2000; Dupre & Diekman 2001). The variation in response to productivity as a function of species size represents the C–S axis in Grime's CSR scheme, with larger species as Competitors able to exploit increasing resource availability, and small species as Stress-tolerators able to survive at low resource availability (Grime et al. 1988).

productivity, plant size and grazing responses: a conceptual model

In order to simplify the analysis of community structure and prediction of community change in response to disturbances, relevant environmental variables and species functional traits are usually categorized into distinct classes. However, plant size and response to grazing, as well as productivity, grazing intensity and competition intensity, are all continuous variables. Thus, a conceptual model relating the response to protection from grazing along gradients of productivity and plant size was elaborated (Fig. 6), based on our quantitative data (Table 1; Figs 3 and 4).


Figure 6. Conceptual model of species response to protection from grazing along gradients of productivity and species plant size. The magnitude and direction of response to grazing, represented by the ratio of abundance in protected vs. grazed subplots, is indicated by contour lines (values of ratios (ungrazed/grazed) on the right of each line). Domains of response groups are delimited by thick lines (names on the right). The dashed equality line indicates, for a given productivity level, a plant size above which the response to grazing exclusion is positive and below which the response is negative. The dotted threshold line indicates productivity levels below which species of a given plant size do not thrive.

Download figure to PowerPoint

Relative changes in abundance, indicated as contours, are predicted along axes representing gradients of productivity and plant size. Larger species appear gradually with increasing site productivity, as indicated by the threshold productivity, below which species of certain plant size do not thrive, because of limits imposed by resource availability. The model predicts a larger positive response to protection (increase in abundance) with increasing plant size and productivity level, and larger negative response to protection (decrease in abundance) with decreasing plant size and higher productivity. The equality line in Fig. 6 represents the plant size at which no change in abundance occurs in response to protection: species that occur above this line should increase in abundance after grazing exclusion, while those below the line should decrease. Of our response groups, Wadi-increasers includes the largest plants that appear above a certain level of productivity (c. > 200 g m−2) that occurs mostly in the Wadi, whereas general-increasers are large plants that are also able to grow at less productive sites. Both groups occur in the model only above the equality line, with a stronger relative increase at higher productivity. Site-dependents are mostly medium size species that occur along a wide range of productivity. Their response to fencing is reversed from increase to decrease at higher productivity and they can therefore occur both above and below the equality line. General decreasers are small species that occur in all productivity levels and consistently decrease in response to protection (below the equality line). Species showing no response to protection at any productivity (neutral species occurring on the equality line) were not found in the topographic sites included in this study.

Most of the species observed fit this conceptual model, but three exhibited patterns of response that can be attributed to factors other than plant size. Sedum pallidum, an extremely small and fragile species, behaved as a protection increaser in the low productivity sites, probably due to its susceptibility to trampling. Crepis aspera, a relatively large composite restricted to the Wadi, behaved as a protection decreaser: its flat rosette shows low competitive potential (Lavorel et al. 1999) and large size is achieved later, during flowering. The grass Crithopsis delileana behaved as a protection increaser in both low and high productivity sites despite being only of small size. Its expansion in the Wadi after fencing was probably associated with production of relatively large seeds, allowing successful patch occupation (Y. Osem, unpublished observations).

missing strategies

This semi-arid community of Mediterranean annuals has been exposed to grazing by wild and domestic herbivores since prehistoric times, but several common strategies involved in grazing avoidance are missing among the more abundant species. Thus these species lacked morphological defence mechanisms, such as spines or tough leaves, and, as sheep consumed all the larger species, these are unlikely to have chemical defences. Spines do, however, occur in a few less abundant annual composites (Carthamus tenuis, Crupina crupinastrum, Atractylis cancellata). The cost of these defence mechanisms may be prohibitive for annual species with a short growing season both in sites with low resource availability and under high competitive pressure in the more productive sites. A negative relationship between plant growth rate and presence of chemical defences has frequently been reported (Hartley & Jones 1997). Diminishing plant size rather than morphological and chemical defences is apparently the favoured response when grazing pressure is high (Rosenthal & Kotanen 1994) or when water and nutrients are scarce (Herms & Mattson 1992). Grazing tolerance due to an ability to re-grow has not been studied in these annual species.

concluding comments

Although our model assumes that competition among species plays a major role in the observed community responses to protection from grazing across the productivity gradient, a quantitative analysis of the intensity of below- and above-ground competition in this rangeland is lacking. This analysis should be performed at the small scale, taking into account the patchy distribution of soil resources within the topographic sites, as this is the scale at which competitive interactions occur. In the less productive sites, there are patches of shallow and deeper soil differing in productivity, probably accounting for the coexistence of species with different plant size (Naeem & Colwell 1991). Plant size appears to be a major plant trait explaining the response to grazing along productivity gradients in this community. Nevertheless, analysis of additional traits related to adaptation to site conditions and to other effects associated with grazing (e.g. trampling, soil compaction, litter removal) (Belsky 1992), may provide better predictability of species response to grazing within size groups.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This research was supported by the International Arid Land Consortium (IALC), grants 96R-19 and 98R-27. We thank Sandra Díaz, Dan Milchunas and Imanuel Noy-Meir for critical reading of the original manuscript, helpful comments and suggestions, and to Hillary Voet for statistical advice. Many thanks to Marcelo Sternberg, Rana German, Zohar Shaham, Ram Lisai, Nitsan Shamir and Rafi Yonathan for their help in the fieldwork.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Aronson, J.A., Kigel, J. & Shmida, A. (1990) Comparative plant size and reproductive strategies in desert and Mediterranean populations of ephemeral plants. Israel Journal of Botany, 39, 413430.
  • Austin, M.P. (1990) Community theory and competition in vegetation. Pespectives in Plant Competition (eds J.B.Grace & D.Tilman), pp. 215238. Academic Press, New York.
  • Baram, H. (1996) Meteorological Data. Lehavim Hills, Israel (1987–95). Volcani Institute, Bet-Dagan, Israel. (In Hebrew.)
  • Belsky, J. (1992) Effects of grazing, competition, disturbance and fire on species composition and diversity in grassland communities. Journal of Vegetation Science, 3, 187200.
  • Briske, D.D. & Noy-Meir, I. (1998) Plant responses to grazing: a comparative evaluation of annual and perennial grasses. Ecological Basis of Livestock Grazing in Mediterranean Ecosystems (eds V.P.Papamastasis & D.Peter), pp. 1326. European Commission, Science, Research and Development, Luxembourg.
  • Conover, W.J. & Iman, R.L. (1981) Rank transformation as a bridge between parametric and nonparametric statistics. American Statistician, 35, 124129.
  • Coughenour, M.B. (1985) Graminoid responses to grazing by large herbivores: adaptations, exaptations and interacting processes. Annals of the Missouri Botanical Garden, 72, 852853.
  • Danin, A. & Orshan, G. (1990) The distribution of Raunkier life-forms in Israel in relation to the environment. Journal of Vegetation Science, 1, 4148.
  • Díaz, S., Noy-Meir, I. & Cabido, M. (2001) Can grazing response of herbaceous plants be predicted from simple vegetation traits? Journal of Applied Ecology, 38, 497508.
  • Dupre, C. & Diekman, M. (2001) Differences in species richness and life-history traits between grazed and abandoned grasslands in southern Sweden. Ecography, 24, 275286.
  • Dyksterhuis, E.J. (1949) Condition and management of rangelands based on quantitative ecology. Journal of Range Management, 2, 104115.
  • Fernandez Alés, R., Laffarga, J.M. & Ortega, F. (1993) Strategies in Mediterranean grassland annuals in relation to stress and disturbance. Journal of Vegetation Science, 4, 313322.
  • Garnier, E., Laurent, G., Bellmann, A., Debian, S., Berthelier, P., Ducout, B. et al. (2001) Consistency of species ranking based on functional leaf traits. New Phytologist, 152, 6983.
  • Gaudet, C.L. & Keddy, P.A. (1995) Competitive performance and species distribution in shoreline plant communities: a comparative approach. Ecology, 76, 280291.
  • Gitay, H. & Noble, I.R. (1997) What are functional types and how should we seek them? Plant Functional Types, Their Relevance to Ecosystem Properties and Global Change (eds T.M.Smith, H.H.Shugart & F.I.Woodward), pp. 319. Cambridge University Press, Cambridge.
  • Grime, J.P. (2001) Plant Strategies, Vegetation Processes and Ecosystem Properties, 2nd edn. John Wiley & Sons, Chichester.
  • Grime, J.P., Hodgson, J.G. & Hunt, R. (1988) Comparative Plant Ecology. Unwin-Hyman, London.
  • Grubb, P.J. (1985) Plant populations and vegetation in relation to habitat, disturbance and competition: problems and generalizations. The Population Structure of Vegetation (ed. J.White), pp. 595621. Junk, Dordrecht.
  • Hadar, L., Noy-Meir, I. & Perevolotsky, A. (1999) The effect of shrub clearing and grazing on the composition of a Mediterranean plant community: functional groups versus species. Journal of Vegetation Science, 10, 673682.
  • Hartley, S.E. & Jones, C.G. (1997) Plant chemistry and herbivory, or why the world is green? Plant Ecology (ed. M.J.Crawley), pp. 284342. Blackwell Science, Oxford.
  • Herms, D.A. & Mattson, W.J. (1992) The dilemma of plants: to grow or defend. Quarterly Review of Biology, 67, 283335.
  • Keddy, P.A. (1989) Competition. Chapman & Hall, London.
  • Van Keulen, H. & Seligman, N. (1992) Moisture, nutrient availability and plant production in the semi-arid regions. Food from Dry Lands (eds ThAlberda, H.Van Keulen, N.G.Seligman & C.T.De Wit), pp. 2581. Kluwer Academic, Dordrecht, The Netherlands.
  • Kutiel, P. & Noy-Meir, Y. (1986) The effects of soil depth on annual grasses in the Judean Hills. I. The effect of soil depth on individual plant species. Israel Journal of Botany, 35, 233239.
  • Laffarga, J.M. & Leiva, M.J. (1991) The Effect of Water and Nutrients on the Growth of Different Mediterranean Rangeland Herbaceous Species. IV International Rangeland Congress. CNRS, Montpellier, France.
  • Lavorel, S., McIntyre, S. & Grigulis, K. (1999) Plant response to disturbance in a Mediterranean grassland: how many functional groups? Journal of Vegetation Science, 10, 661672.
  • Lavorel, S., McIntyre, S., Landsberg, J. & Forbes, T.D.A. (1997) Plant functional classifications: from general groups to specific groups based on response to disturbance. Trends in Ecology and Evolution, 12, 474478.
  • Litav, M. (1965) Effects of soil type and competition on the occurrence of Avena sterilis L. in the Judean Hills (Israel). Israel Journal of Botany, 14, 7489.
  • Ludwig, J.A. (1986) Primary production variability in desert ecosystems. Pattern and Process in Desert Ecosystems (ed. W.G.Whitford), pp. 517. University of New Mexico Press, Albuquerque, New Mexico.
  • Magurran, A.E. (1988) Ecological Diversity and its Measurement. Cambridge University Press, Cambridge.
  • McIntyre, S., Heard, K.M. & Martin, T.G. (2003) The relative importance of cattle grazing in subtropical grasslands: does it reduce or enhance plant diversity? Journal of Applied Ecology, 40, 445457.
  • McIntyre, S. & Lavorel, S. (2001) Livestock grazing in subtropical pastures: steps in the analysis of attribute response and functional types. Journal of Ecology, 89, 209226.
  • McIntyre, S., Lavorel, S. & Tremont, R.M. (1995) Plant life-history attributes: their relationship to disturbance response in herbaceous vegetation. Journal of Ecology, 83, 3144.
  • Milchunas, D.G. & Lauenroth, W. (1993) Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecological Monographs, 6, 327366.
  • Milchunas, D., Lauenroth, W.K., Chapman, P.L. & Kazempour, M. (1989) Effects of grazing, topography and precipitation on the structure of a semiarid grassland. Vegetatio, 80, 1123.
  • Milchunas, D.G., Sala, O.E. & Lauenroth, W. (1988) A generalized model of the effects of grazing by large herbivores on grassland community structure. American Naturalist, 132, 87106.
  • Montalvo, J., Casado, M.A., Levassor, C. & Pineda, F.D. (1993) Species diversity patterns in Mediterranean grasslands. Journal of Vegetation Science, 4, 213222.
  • Naeem, S. & Colwell, R.K. (1991) Ecological consequences of heterogeneity of consumable resources. Ecological Heterogeneity (eds J.Kolasa & S.T.A.Pickett), pp. 224255. Springer-Verlag, New York.
  • Newman, E.I. (1973) Competition and diversity in herbaceous vegetation. Nature, 244, 310.
  • Noy-Meir, I. (1973) Desert ecosystems: environment and producers. Annual Review of Ecology and Systematics, 4, 2552.
  • Noy-Meir, I. (1990) Responses of two semiarid rangeland communities to protection from grazing. Israel Journal of Botany, 39, 431442.
  • Noy-Meir, I., Gutman, M. & Kaplan, Y. (1989) Response of Mediterranean grassland plants to grazing and protection. Journal of Ecology, 77, 290310.
  • Noy-Meir, I. & Seligman, N. (1979) Management of semi-arid ecosystems in Israel. Management of Semi-Arid Ecosystems (ed. B.H.Walker), pp. 113160. Elsevier, Amsterdam.
  • Osem, Y., Perevolotsky, A. & Kigel, J. (2002) Grazing effect on diversity of annual plant communities in a semi–arid rangeland: interactions with small-scale spatial and temporal variation in primary productivity. Journal of Ecology, 90, 936946.
  • Perevolotsky, A. & Seligman, N.G. (1998) Role of grazing in Mediterranean rangeland ecosystems. Bioscience, 48, 10071017.
  • Proulx, M. & Mazumder, A. (1998) Reversal of grazing impact on plant species richness in nutrient-poor vs. nutrient-rich ecosystems. Ecology, 79, 25812592.
  • Ravikovitch, S. (1981) The Soils of Israel. Hakibbutz Hameuchad, Tel Aviv, Israel.
  • Rosenthal, J.P. & Kotanen, P.M. (1994) Terrestrial plant tolerance to herbivory. Trends in Plant Ecology and Evolution, 9, 145148.
  • Sammul, M., Kull, K., Oksanen, L. & Veromann, P. (2000) Competition intensity and its importance: results of field experiments with Anthoxacum odoratum. Oecologia, 125, 1825.
  • SAS Institute (1995) JMP Statistics and Graphics Guide. SAS Institute, Cary, North Carolina.
  • Sternberg, M., Gutman, M., Perevolotsky, A., Ungar, E.D. & Kigel, J. (2000) Vegetation response to grazing management in a Mediterranean herbaceous community: a functional group approach. Journal of Applied Ecology, 37, 224237.
  • Tilman, D. (1988) Plant Strategies and the Dynamics and Structure of Plant Communities. Princeton University Press, Princeton, New Jersey.
  • Turkington, R., Klein, E. & Chanway, C.P. (1993) Interactive effects of nutrients and disturbance: an experimental test on plant strategy theory. Ecology, 74, 863878.
  • Vesk, P.A. & Westoby, M. (2001) Predicting plant species’ responses to grazing. Journal of Applied Ecology, 38, 897909.
  • Zohary, M. (1973) Geobotanical Foundations of the Middle East. Gustav Fisher Verlag, Stuttgart.