SEARCH

SEARCH BY CITATION

Keywords:

  • frugivory;
  • fruit nutritional content;
  • fruit size;
  • intraspecific variation;
  • plant secondary metabolites;
  • Rhamnus alaternus (Mediterranean buckthorn)

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • • 
    We studied within-species variation in and interrelations among morphological and chemical traits of ripe Mediterranean buckthorn ( Rhamnus alaternus ) fruit, a bird-dispersed species.
  • • 
    Principal component analysis revealed that larger fruits tended to be relatively rich in nonstructural carbohydrates (NSC), water and P but poor in protein and most minerals. Small fruits tended to be relatively rich in protein, structural carbohydrates, K and Zn while intermediate size fruits tended to be rich in lipids, Mg and Ca. Variation in chemical traits (organic compounds and minerals) was typically much higher than in morphological traits (e.g. fruit size) with the exception of NSC and water content, which varied little. This discrepancy might be explained by differences in environmental conditions between plant microsites that imposed greater variability on fruit nutrient composition than on fruit-morphological traits; and by lower selective pressure by birds on fruit chemical traits than on morphological traits.
  • • 
    Secondary metabolite (emodin) concentration was positively correlated with concentrations of NSC, supporting the nutrient/toxin titration model, which predicts that high levels of secondary metabolites in fruits should be off set by high nutritional rewards for dispersers.
  • • 
    Emodin concentration in leaves was much higher than in fruit pulp, which may indicate its differential adaptive roles in seed dispersal and against herbivores.

Introduction

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

Many plants throughout the world produce fleshy pulp surrounding seeds to entice potential dispersal agents (e.g. Van der Pijl, 1982; Herrera, 1995). The fruit selection by potential seed dispersers may depend on many nonnutritive factors, such as fruit abundance, accessibility, competition, predation risks, fruit size, color, and secondary compound content (Murphy, 1994), although the nutritional value of the fruit and the digestive ability of the bird still play a major role in fruit choice (Martinez del Rio & Restrepo, 1993). Indeed, several studies have demonstrated the importance of nutritional fruit traits in fruit selection across plant species (Johnson et al., 1985; Izhaki, 1992; Fuentes, 1994). Among nonnutritive factors, fruit size seems to be a common determinant of avian fruit choice among and within plant species (Herrera, 1981, 1984, 1988; Jordano, 1987, 1989, 1995; McPherson, 1988; Avery et al., 1993; Izhaki, 2002a, but see Johnson et al., 1985).

Among other nonnutritive factors, fruit secondary metabolites may have a significant effect on frugivore preference (Tsahar et al., in press). For example, phenols (especially tannins) have been widely adopted as feeding deterrents because they can interfere with protein digestion (Robbins, 1993) and are associated with food taste (Rosenthal & Janzen, 1979). Thus, the presence of secondary metabolites in ripe fruit pulp seems to be an evolutionary paradox because such compounds can deter seed dispersers from consuming the fruit. But, in some cases, secondary metabolites in fruit may actually enhance seed dispersal and hence plant reproduction (reviewed by Cipollini & Levey, 1997a,b; Cipollini, 2000). Cipollini & Levey (1997a,b) proposed the nutrient/toxin titration model, predicting a positive correlation between toxin and nutrient content in ripe fruit pulp. Their reasoning was that high nutritional quality would compensate dispersers for ingesting high levels of toxins. Yet, this model has little support (Cipollini, 2000). Furthermore, the adaptive approach of Cipollini & Levey was challenged by Eriksson & Ehrlén (1998), who claimed that the presence of secondary metabolites in fruits is a consequence of their presence throughout the plant and therefore nonadaptive in the contest of fruit–frugivore interactions. Their assumption that secondary metabolites in ripe pulp are actually a byproduct of plant metabolism has not yet been tested. It was also suggested that secondary metabolites in plants possess multiple adaptive roles (Cipollini, 2000; Izhaki, 2002b).

Considerable interspecific variation in nutritional content and morphological traits of fruit has been demonstrated in plant communities (Debussche et al., 1987; Herrera, 1987; Jordano, 1992; Corlett, 1996; Izhaki, 2002a). Such among-species variation is a consequence of genetic and environmental factors (Obeso & Herrera, 1994). Despite this remarkable variation, there is evidence of patterns of nutritional and morphological fruit traits across plant species (Johnson et al., 1985; Debussche et al., 1987; Herrera, 1987; Jordano, 1992; Corlett, 1996; Izhaki, 2002b). Currently, there is little evidence to support the hypothesis that these relationships are the result of selective pressure from frugivores; instead, they seem to be governed by phylogeny and environmental factors (Herrera, 1987; Debussche et al., 1987).

Nevertheless, the indispensable raw material for natural selection of fruit traits is not variation among fruit species but rather within them. Recently, Alonso & Herrera (2001) found considerable within-population variation in some nutrient concentrations of mature leaves of Prunus mahaleb. They emphasized the importance of this variation for habitat selection by herbivores and their fitness. Such evolutionary and ecological relevance of nutrient variation to herbivore–plant interactions is likely applicable to the interaction between fruits and dispersal agents. However, only a few studies have assessed intraspecific variability in fruit traits and most of them are limited to fruit morphology (Herrera, 1981, 1982; Jordano, 1984, 1991, 1995; Willson et al., 1990; Obeso & Herrera, 1994). Intraspecific variation of nutritional content of fleshy fruits are rare (Jordano, 1987, 1989; Denslow, 1987; Cipollini & Whigham, 1994) and no study has documented the intraspecific variation of secondary metabolites in wild fruits.

The aim of this study was to assess the intraspecific variability and the intercorrelations of nutrients, secondary metabolites, and morphological fruit traits of the Mediterranean Buckthorn (Rhamus alaternus). Specifically, we addressed the following questions: (1) Do fruit traits relate to each other within species? (2) Do Rhamnus plants that contain higher concentrations of secondary metabolites in their fruit also contain higher concentrations of fruit nutrients? Hence, does the nutrient/toxin model apply within species? (3) Do secondary metabolite concentrations in fruits and mature leaves differ? The answer is relevant to the multiple function approach.

Materials and Methods

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

Rhamnus alaternus

Mediterranean buckthorn (Rhamnus alaternus L., Rhamnaceae) is a fleshy-fruited, native dioecious evergreen shrub or tree, 2–6 m in height, inhabiting the Mediterranean area. Flowering occurs between February and April, and ripe, juicy, berry-like black fruit appear between May and September (Tsahar et al., in press). The fruit are commonly consumed by permanent resident birds, such as Yellow-vented bulbuls (Pycnonotus xanthopygos), Blackbirds (Turdus merula) and Sardinian warblers (Sylvia melanocephala), and by transient birds, such as Blackcaps (S. atricapilla) and Orphean warblers (S. hortensis) (Tsahar, 2001).

The genus Rhamnus is characterized from a phytochemical point of view by the abundance of phenolic substances, especially anthraquinones, but also tannins and flavonoids (Coskun, 1992a,b). Four anthraquinine aglycones (emodin, chrysophanol, alatenin and physcion) were isolated in R. alaternus from the above-ground parts of the plant, with emodin being the most abundant of these and the only aglycon detectable in the ripe pulp (Abou-Chaar & Shamlian, 1980). Emodin has several biological activities (reviewed by Izhaki, 2002b) including purgative effects in humans (Rauwald, 1998), feeding deterrent to insects and birds (Sherburne, 1972; Trial & Dimond, 1979), allelopathy (Nishimura & Mizutani, 1995), and antibacterial and antifungal effects (e.g. Wang & Chung, 1997).

Fruit traits

We collected many ripe fruits (usually hundreds) from each of 26 plants near the University of Haifa at Oranim campus, Israel (32°43′ N, 35°07′ E) in July 2000. All plants were naturally established and inhabited a typical east Mediterranean scrubland under natural environmental conditions. The study site was located in a relatively remote area with minimal human interference and thus represents a natural system. To reduce within plant variation among branches, we picked fruits from as many infructescences as possible at different heights and aspects. At the same time, we also collected mature leaves of 14 plants from this group. Morphological traits, based on 50 fruits of each individual, included fruit wet mass, fruit diameter and the ratio of seed fresh mass to pulp wet mass. Pulp and leaves were oven-dried at 60°C to constant mass for nutritional analyses and water content calculation. The dried pulp and leaves were ground to a powder that was then used for chemical analyses. In addition, wet pulp and leaf samples from each plant were kept frozen for emodin analysis (see below).

Major organic compounds

Crude lipid content was determined by the gravimetric Soxhlet method, using petroleum ether for extraction (AOAC, 1984). Total nitrogen was determined by the Kjeldahl procedure (AOAC, 1984). Total crude protein was determined by multiplying nitrogen content by a conversion factor of 3.19, which is appropriate for assessing the true protein content of R. alaternus (Izhaki, 1993). Neutral detergent fiber (NDF) and acid detergent fiber (ADF) of the pulp were analyzed with an Ankom Fiber Analyzer (ANKOM220). NDF is the cell wall fraction of the total fiber content and includes cellulose, hemicellulose and lignin. The ADF fraction contains cellulose and lignin and is closely related to indigestibility of food items (Robbins, 1993). The difference between NDF and ADF is hemicellulose content. The percentage of total Non-Structural Carbohydrates (NSC hereafter) was calculated as 100 − %protein − %lipids – %NDF − %ash.

Minerals

Pulp was prepared for mineral analysis either by dry ashing (high-temperature combustion) for Fe, Zn, Mn, Cu or by wet acid digestion for K, Na, Ca, Mg, and P. Determination of K, Na, and Mg was done with flame spectrophotometery. Ca, Fe, Zn, Mn and Cu were determined by atomic absorption spectrophtometery, and P by colorimetry (Walinga et al., 1989).

Secondary metabolites

Total phenolic compounds were measured using the Folin-Ciocalteu method (Waterman & Mole, 1994). For emodin analysis fresh-frozen pulp was soaked in cold methanol at a ratio of 1 : 10 and stirred for 1 h, then filtered under vacuum. This process was repeated four times. The extraction was then rotoevaporated to dryness. Using this method only the free emodin (aglicon) was extracted. The concentration of free emodin was determined using HPLC. The HPLC analysis was done by Migal R & D Laboratories, Kiryat, Shmona, Israel. HPLC conditions were based on Coskun (1992a,b).

Statistics

Consideration of dimensionality predicts that coefficient of variation (CV, hereafter; standard deviation/mean) of fruit mass (a cubic function) will be larger than those variables of linear measures (Lande, 1977). Therefore, we calculated CV on the log-transformed scale of all variables, which is more appropriate when CV comparisons involve variables differing in dimensionality (Obeso & Herrera, 1994). However, while it is valid to compare log-transformed values of CVs of organic and mineral content of fruits, it is still expected that their CVs will be lower than the CV of fruit mass and mass-related variables. Differential measurement error of variables in different categories (e.g. metric triats vs macronutrients vs micronutrients) may have also contributed to observed differences in CV.

Fruit traits of 17 Rhamnus plants were analyzed by Principal Component Analyses (PCA) to detect patterns among morphological traits (fruit mass, fruit diameter and seed to pulp), organic compound and water content (NSC, lipids, protein, ADF, NDF, and water) and mineral content (K, Na, P, Ca, Mg, Fe, Zn, Cu, Mn). PCA produces compound variables (principal components) that are linear combinations of the original variables. Following PCA, Varimax orthogonal rotation was applied to construct a new, more easily interpretable pattern of component loadings. In the ordination diagram of the PCA analysis we plotted the distribution of each fruit-trait variable according to its PCA score coefficient. This plot was used to detect related fruit traits and relationships among different fruit traits. We did not include the secondary metabolite (phenols and emodin) with the PCA analysis since only 10 samples of secondary metabolites were taken from the same 17 trees, which were used in the PCA. Therefore, we revealed relationships between secondary metabolites and fruit traits by Pearson correlation coefficient. Because we used the same correlation matrixes, we adjusted all P-values with the use of the sequential Bonferroni test to control for type I error (P = 5%). Statistical analyses were done with SYSTAT (1998).

Results

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

Fruit traits

Variation in fruit mass and seed : pulp ratio among individual plants was seven times higher than that of fruit diameter (Table 1). The major organic constituents in pulp were nonstructural carbohydrates (NSC), comprising > 80% of dry mass and structural carbohydrates comprising 5% and 4% dry mass (NDF and ADF, respectively, Table 1). Variation of NSC was low (CV = 2%, n = 23) and variation of NDF and ADF was much higher (CV = 24% and 26%, respectively, n = 26, Table 1). Although fruits were relatively low in lipids (< 6%) and proteins (1%), plants varied considerably in their lipid (CV = 30%, n = 25) and protein contents (CV = 24%, n = 24). Water was the main constituent of wet pulp (68%) and was also the most constant variable in this study (CV = 1%, n = 26). Potassium was the most abundant macroelement in Rhamnus pulp, followed by Ca, Mg, P and Na (Table 1). Fe was the major trace element followed by Zn, Mn and Cu (Table 1). The concentration of minerals was extremely variable between plants with the range of CV = 16% for Zn to CV = 85% for Na (Table 1).

Table 1.  Mean, standard deviation, coefficient of variation (CV) and sample size ( n ) for all variables measured for Rhamnus alaternus fruits and leaves
 MeanSDCV*n (%)
  • *

    Calculated on log-transformed data.

Fruit morphology
Fruit mass (mg)23855 2017
Fruit diameter (mm)  5.3 0.57  317
Seed/pulp  2.53 0.83 2026
Pulp organic compounds and water
NSC (% d. wt) 82.2 5.8  223
Lipids (% d. wt)  5.7 3.5 3025
Protein (% d. wt)  1.2 0.4 2424
ADF (% d. wt)  4.2 2.1 2626
NDF (% d. wt)  5.3 2.4 2426
Water (% f. wt) 68.4 2.6  126
Pulp mineral content
K (mg g−1) 12.90 3.77 2124
Na (mg g−1)  0.23 0.20 8524
P (mg g−1)  0.56 0.53 8424
Ca (mg g−1)  3.61 1.57 3724
Mg (mg g−1)  0.92 0.55 5424
Fe (mg kg−1) 49.1832.42 1826
Zn (mg kg−1)  5.80 2.84 1626
Cu (mg kg−1)  1.92 1.04 3626
Mn (mg kg−1)  3.54 1.57 2326
Total ash (% d. wt)  5.2 3.4 2026
Pulp's secondary compounds
Total phenols (% d. wt)9.61 × 10−31.59 × 10−3  317
Emodin (% d. wt)1.71 × 10−30.64 × 10−3 3712
Emodin (% f. wt)1.17 × 10−30.84 × 10−3 7426
Emodin in mature leaves (% f. wt)4.04 × 10−25.03 × 10−211914

Intercorrelation among fruit traits

The first three principal components accounted for 65% of the total variation. The first axis explained 37% of the variance and was negatively correlated with the three morphological variables (Fig. 1). Hence, the increase in the seed : pulp ratio was associated with an increase in fruit mass (seeds + pulp) and therefore the contribution to fruit mass of the seeds was higher than that of the pulp. The first component was also negatively correlated with NSC, water and P and positively correlated with proteins, lipids, ADF, NDF, K and Zn, Na, Ca, Mg. The second axis explained 15% of the variance and was positively correlated with Mn and Fe and negatively correlated with Mg, Cu, lipids and seed : pulp ratio (Fig. 1). There was a pronounced negative trend between NSC, lipids and protein, whereas NSC-rich fruits contained more water and P, protein-rich fruits contained more NDF, ADF, K and Zn, and lipid-rich fruits contained more Mg and Ca. The NSC-rich fruits were also the largest while the protein-rich fruits were the smallest and the lipid rich fruits were of intermediate size (Fig. 1). Those fruits that were relatively rich in P were relatively poor in all other minerals. The third component explained 13% of the variation and was positively correlated with lipids and Mg and negatively correlated with NSC and water.

image

Figure 1. PCA ordination diagram of morphological and chemical traits of ripe fruits of Rhamnus alaternus ( n  = 17 plants) in northern Israel.

Download figure to PowerPoint

Mean total phenolic concentration in pulp was 6 times higher than free emodin concentration (d. wt, Table 1). No significant correlation was found between emodin concentration and total phenols (r = −0.32, n = 12, P > 0.1). Free emodin concentration on d. wt basis was 45% higher than that on f. wt basis (Table 1). Because there was a strong correlation between emodin concentrations as calculated on d. wt and f. wt bases (r = 0.98, n = 12, P < 0.0001) and because the sample size of emodin based on f. wt was larger than on d. wt, we used only the values of emodin on f. wt basis for further calculations.

Emodin concentration in the leaves was 35 times higher than in the pulp (f. wt) and had the highest variation (CV = 119%, n = 14) in this study (f. wt, Table 1). Emodin in the pulp was negatively correlated with lipids (r = −0.44, n= 25, P < 0.05), ADF (r = −0.46, n = 26, P < 0.05), and NDF (r = −0.55, n = 26, P < 0.05), and positively correlated with NSC (r = 0.42, n = 23, P < 0.05). No significant correlations were found between emodin and protein and minerals. Total phenol was not significantly correlated with any organic compound and mineral. No significant correlation was detected between emodin concentrations in pulp and in leaves (r = −0.06, n = 14, P > 0.05).

Discussion

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

Intercorrelation of fruit traits within a species

Our study is the first to document correlations among a full profile of chemical traits including secondary metabolites of fruits within a species. There were trade-offs between the main organic compounds (NSC, lipids and proteins). Analyses among species have similarly revealed negative correlations between NSC and proteins (Izhaki, 2002a) and between NSC and lipids (Debussche et al., 1987; Herrera, 1987; Jordano, 1992; Corlett, 1996; Izhaki, 2002a; but see Eriksson & Ehrlén, 1991). Debussche et al. (1987) argued that the content of the major organic components in fruit actually depend on water content. They attributed it to differing solubility of these organic compounds, as NSC are the most soluble in water, lipids are hydrophobic, and proteins may be soluble to some extent. Indeed, the strong positive correlation between water and NSC and the negative correlation between water and lipids that we found at the within-species level and others found at the between-species level (Debussche et al., 1987; Herrera, 1987; Izhaki, 2002a) support this notion. Furthermore, the low variation of water content contrasts with the broad variation of fats and proteins perhaps indicating that a slight change in water content imposes a great change in the composition of the organic compounds. Thus, water content of fruit may, in fact, be the key factor in explaining the observed trade-offs among the main organic compounds in R. alaternus.

The negative correlation that was found between fruit size and protein content within the interspecific level of several east-Mediterranean species (Izhaki, 2002a) was also found within the intraspecific level of R. alaternus. Furthermore, relatively large Rhamnus fruits contained more NSC, water, and P, intermediate size fruits contained more lipids, Mg and Ca and the smaller fruits contained more proteins, ADF, NDF, K and Zn. One possible explanation for the three fruit-types detected here is that fruit composition reflects the soil properties of the microsite of each plant. For example, a microsite with a relatively high water availability in the soil may support the production of larger fruits with high water content and thus with higher NSC content. However, further study is needed to verify this theory.

The interrelationships among fruit nutrient may indicate synergistic and antagonistic interactions between nutrients (Cecil et al., 1995), some of which might also be explained by plant-environment relations. For example, the positive correlation between Mg and Ca detected in Rhamnus fruit was also found in other plant parts (Garten, 1976; Hocking, 1986; Alonso & Herrera, 2001). This correlation is probably a consequence of the close association of Ca and Mg in metabolism and photosynthesis (Garten, 1976).

Secondary metabolites in fruit

Our data show that fruit pulp always contained much less emodin than leaves. High concentration in leaves might be needed to deter herbivores while low concentration in the pulp may still deter seed predators and may also promote seed dispersal by other means (Tsahar et al., in press) and hence increase plant fitness. It was already suggested that emodin exhibits multiple adaptive functions (Izhaki, 2002b) and different functions in different tissues likely require different concentrations of secondary metabolites. The positive association we found between emodin and NSC supports the nutrient/toxin titration model, which predicts that high levels of secondary metabolites should be off set by high nutritional rewards for dispersers (Cipollini & Levey, 1997a,b; Cipollini, 2000).

Intraspecific variation of R. alaternus fruit traits

The R. alaternus plants we studied exhibited substantial variation in their fruit traits (CV range = 1–119%). This variation likely represents inherent genetic differences and responses to microsite environmental conditions. But our results indicate that although the variation of mass-related traits is expected to be larger than that of chemical traits (see Materials and Methods), the variation of fruit morphological traits (CV = 3–21%) was actually lower than variation in fruit organic compound content (usually, CV ≥ 24%), with the exception of NSC and total phenols. Two nonexclusive hypotheses may explain this discrepancy. (1) Differences in environmental conditions between plant microsites imposed greater variability on fruit nutrient composition than on fruit morphological traits. (2) Lower selective pressure by birds on fruit chemical traits than on morphological traits. It is currently believed that fruit traits had only marginally been shaped by dispersal agents (Herrera, 1987). Nonetheless, some fruit attributes such as fruit size may still be under heavier selective pressure than other attributes (Jordano, 1992). Hence, we suggest that the relative low variation we observed in fruit size of R. alaternus reflects selective pressures of its potential avian dispersers. By the same token, we suggest that the relatively widely observed variation in the chemical composition of R. alaternus fruit indicates that nutritional preferences of birds are much less focused than their preferences for morphological traits. The exceptionally low variation in NSC content of R. alaternus fruit may therefore reflect a strong selection pressure generated by high carbohydrate requirements of migratory frugivores.

To conclude, many of the fruit attributes within the species level are intercorrelate. Although physiological and biochemical plant constraints, which depend upon environment-related conditions may explain some of these associations, the selective pressures generated by frugivores may further shape some of these attributes and the correlation among them. Furthermore, the different concentration of secondary metabolites in different Rhamnus parts (fruits and leaves) may indicate that other selective forces such as herbivory are also involved in shaping chemical concentrations throughout the plant.

Acknowledgements

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

The authors thank Miloda Laboratories and Soil, Water and Plant Laboratory, Ministry of Agriculture, Israel for the nutritional analyses, José Ramón Obeso, Doug Levey and two anonymous reviewers for critical and helpful comments, Emma Kvitnitsky, Moshik Inbar, Gidi Ne’eman and Zvia Shapira for help in different stages of the study and the Department of Zoology, University of Florida, where the manuscript was written while I. Izhaki was on sabbatical leave.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Abou-Chaar CI, Shamlian SN. 1980. A chromatographic study of the anthraquinones of Rhamnus alaternus L. extraction, isolation and identification of the aglycones. Quarterly Journal of Crude Drug Research 18: 4955.
  • Alonso C, Herrera CM. 2001. Patterns made of patterns: variation and covariation of leaf nutrient concentrations within and between populations of Prunus mahaleb. New Phytologist 150: 629640.
  • AOAC. 1984. Official methods of analysis of the association of official analytical chemists. Washington, DC, USA: AOAC.
  • Avery ML, Goocher KJ, Cone MA. 1993. Handling efficiency and berry size preferences of Cedar Waxwings. Wilson Bulletin 105: 604611.
  • Cecil JS, Barth GE, Maier NA, Chvyl WL, Bartetzko MN. 1995. Leaf chemical composition and nutrient removal by stem of Leucadendron cvv. Silvan Red and Safari Sunset. Australian Journal of Experimental Agriculture 35: 275283.
  • Cipollini ML. 2000. Secondary metabolites of vertebrate-dispersed fruits: evidence for adaptive functions. Revista Chilena de Historia Natural 73: 421440.
  • Cipollini ML, Levey DJ. 1997a. Why are some fruit toxic? Glycoalkaloids in Solanum and fruit choice by vertebrates. Ecology 78: 782798.
  • Cipollini ML, Levey DJ. 1997b. Secondary metabolites of fleshy vertebrate-dispersed fruits: Adaptive hypothesis and implications for seed dispersal. American Naturalist 150: 346372.
  • Cipollini ML, Whigham DF. 1994. Sexual dimorphism and cost of reproduction in the dioecious shrub Lindera benzoin (Lauraceae). American Journal of Botany 81: 6575.
  • Corlett RT. 1996. Characteristics of vertebrate-dispersed fruits in Hong-Kong. Journal of Tropical Ecology 12: 819833.
  • Coskun M. 1992a. HPLC Analysis of the anthroquinones from Rhamnus species growing in Turkey. International Journal of Pharmacognosy 30: 151156.
  • Coskun M. 1992b. HPLC Separation and quantitative determination of antheraquinones in the bark of some Rhamnus species. Pharmacia – JTPA 32: 1321.
  • Debussche M, Cortez J, Rimbault I. 1987. Variation in fleshy fruit composition in the Mediterranean region: the importance of ripening season, life form, fruit type and geographical distribution. Oikos 49: 244252.
  • Denslow JS. 1987. Fruit removal rates from aggregated and isolated bushes of the red elderberry, Sambucus pubens. Canadian Journal of Botany 65: 12291235.
  • Eriksson O, Ehrlén J. 1991. Phenological variation in fruit characteristics in vertebrate-dispersed plants. Oecologia 86: 463470.
  • Eriksson O, Ehrlén J. 1998. Secondary metabolites in fleshy fruits: are adaptive explanations needed? American Naturalist 152: 905907.
  • Fuentes M. 1994. Diet of fruit-eating birds: what are the causes of interspecific differences? Oecologia 97: 134142.
  • Garten CT. 1976. Correlations between concentrations of elements in plants. Nature 261: 686688.
  • Herrera CM. 1981. Fruit variation and competition for dispersers in natural populations of Smilax aspera. Oikos 36: 5158.
  • Herrera CM. 1982. Interspecific variation in fruit shape: allometry, phylogeny, and adaptation to dispersal agents. Ecology 73: 18321841.
  • Herrera CM. 1984. Adaptation to frugivory of Mediterranean avian seed dispersers. Ecology 65: 609617.
  • Herrera CM. 1987. Vertebrate-dispersed plants of the Iberian Peninsula: a study of fruit characteristics. Ecological Monographs 57: 305331.
  • Herrera CM. 1988. Habitat-shaping host plant use by a hemiparasitic shrub, and the importance of gut fellows. Oikos 51: 383386.
  • Herrera CM. 1995. Plant-vertebrate seed dispersal systems in the Mediterranean: Ecological, evolutionary and historical determinants. Annual Review of Ecology and Systematics 26: 705727.
  • Hocking PJ. 1986. Mineral nutrient composition of leaves and fruits of selected species of Grevillea from south western? Australia, with special reference to Grevillea leucopteris Meissn. Australian Journal of Botany 34: 155164.
  • Izhaki I. 1992. A comparative analysis of the nutritional quality of mixed and exclusive fruit diets for Yellow-vented Bulbuls. Condor 94: 912923.
  • Izhaki I. 1993. Influence of nonprotein nitrogen on estimation of protein from total nitrogen in fleshy fruits. Journal of Chemical Ecology 19: 26052615.
  • Izhaki I. 2002a. The role of fruit traits in determining fruit removal in east Mediterranean ecosystems. In: LeveyDJ, SilvaWR, GalettiM, eds. Seed dispersal and frugivory: ecology, evolution and conservation. Wallingford, Oxon, UK: CAB International, 161175.
  • Izhaki I. 2002b. Emodin – a secondary metabolite with multiple ecological functions in higher plants. New Phytologist 155: 205217.
  • Johnson RA, Willson MF, Thompson JN, Bertin RI. 1985. Nutritional values of wild fruits and consumption by migrant frugivorous birds. Ecology 66: 819827.
  • Jordano P. 1984. Seed weight variation and differential avian dispersal in blackberries Rubus ulmifolius. Oikos 43: 149153.
  • Jordano P. 1987. Avian fruit removal: effects of fruit variation crop size and insect damage. Ecology 68: 17111723.
  • Jordano P. 1989. Pre-dispersal biology of Pistacia lentiscus (Anacardiaceae): cumulative effects on seed removal by birds. Oikos 55: 375386.
  • Jordano P. 1991. Gender variation and expression of monoecy in Juniperus phoenicea (L.) (Cupressaceae). Botanical Gazette 152: 476485.
  • Jordano P. 1992. Fruits and frugivory. In: FennerM, ed. Seeds: the ecology of regeneration in plant communities. Wallingford, Oxon, UK: CAB International, 105156.
  • Jordano P. 1995. Frugivore-mediated selection on fruit and seed size: birds and St. Lucie's cherry, Prunus mahaleb. Ecology 76: 26272639.
  • Lande R. 1977. On comparing coefficients of variation. Systematic Zoology 15: 214217.
  • Martinez del Rio C, Restrepo C. 1993. Ecological and behavioral consequences of digestion in frugivorous animals. Vegetatio 107/108: 205216.
  • McPherson JM. 1988. Preference of cedar waxwings in the laboratory for fruit species, colour and size: a comparison with field observations. Animal Behaviour 36: 961969.
  • Murphy ME. 1994. Dietary complementation by wild birds: Considerations for field studies. Journal of Biosciences 19: 355368.
  • Nishimura H, Mizutani J. 1995. Identification of allelochemicals in Eucalyptus citriodora and Polygonum sachalinense. In: DashkinFA, ed. Allelopathy: organisms, processes and applications. Washington, DC, USA: American Chemical Society, 7485.
  • Obeso JR, Herrera CM. 1994. Inter- and intraspecific variation in fruit traits in co-occurring vertebrate-dispersed plants. International Journal of Plant Sciences 155: 382387.
  • Rauwald HW. 1998. Herbal laxatives: influence of anthrones- anthraquinones on energy metabolism and ion transport in a model system. In: LawsonLD, BaurR, eds. Phytomedicines of Europe, chemistry and biological activity. Washington, D.C., USA: American Chemical Society, 97116.
  • Robbins CT. 1993. Wildlife feeding and nutrition, 2nd edn. San Diego, CA, USA: Academic Press.
  • Rosenthal GA, Janzen DH. 1979. Herbivores: their interaction with secondary plant metabolites. London, UK: Academic Press.
  • Sherburne JA. 1972. Effect of seasonal changes in the abundance and chemistry of the fleshy fruits of northeastern woody shrubs on patterns of exploitation by frugivorous birds. PhD thesis. Cornell University, USA.
  • SYSTAT. 1998. Systat 8.0 Statistics. Chicago, IL, USA: SPSS Inc.
  • Trial H, Dimond JB. 1979. Emodin in Buckthorn: a feeding deterrent to phytophagous insects. Canadian Entomologist 111: 207212.
  • Tsahar E. 2001. The impact of the secondary metabolite emodin in Rhamnus alaternus. Fruits on Removal and Seed Fruit Predation. MSc thesis: Tel Aviv University, Israel.
  • Tsahar E, Friedman J, Izhaki I. 2002. Impact on fruit removal and seed predation of a secondary metabolite, emodin in Rhamnus Alaternus fruit pulp. Oikos (In press.)
  • Van der Pijl. 1982. Principals of dispersal in higher plants, 2nd edn. Berlin, Germany: Springer-Verlag.
  • Walinga I, Van Vark W, Houba VJG, Van Der Lee JJ. 1989. Soil and plant analysis. Wageningen, The Netherlands: Wageningen Agriculture University.
  • Wang HH, Chung JG. 1997. Emodin-induced inhibition of growth and DNA damage in the Helicobacter pylori. Current Microbiology 35: 262266.
  • Waterman PG, Mole S. 1994. Methods in ecology, analysis of phenolic plant metabolites. Oxford, UK: Blackwell Scientific Publications.
  • Willson MF, Michaels HJ, Bertin RI, Benner B, Rice S, Lee TD, Hartgerink AP. 1990. Intraspecific variation in seed packaging. American Midland Naturalist 123: 179185.