Nordic Society Oikos

How Specialized Are Fruit-Bird Interactions? Overlap of Frugivore Assemblages within and between Plant Species Author(s): Marcelino Fuentes Reviewed work(s): Source: Oikos, Vol. 74, No. 2 (Nov., 1995), pp. 324-330 Published by: Blackwell Publishing on behalf of Nordic Society Oikos Stable URL: http://www.jstor.org/stable/3545663 . Accessed: 18/11/2011 12:24 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Blackwell Publishing and Nordic Society Oikos are collaborating with JSTOR to digitize, preserve and extend access to Oikos.

http://www.jstor.org

OIKOS 74: 324-330. Copenhagen1995

How specialized are fruit-bird interactions? Overlap of frugivore assemblages within and between plant species Marcelino Fuentes

Fuentes,M. 1995. How specializedare fruit-birdinteractions?Overlapof frugivore assemblageswithinand betweenplantspecies.- Oikos 74: 324-330. Both location and plant species identitysignificantlyinfluencethe compositionof frugivoreassemblagesof local plantpopulations.This is revealedby an analysisof assemblagesof pulp-eatingbirdsof Europeanplants.Whencomparingassemblagesof plantpopulationsat differentsites, thosefromthe sameplantspeciesaremoresimilar to each otherthanthose fromdifferentspecies.Thus,interspecificdifferencesin fruit or planttraitsseem to cause recurrentdifferencesin frugivoreassemblages.On the other hand, the similaritybetween the frugivoreassemblagesof plant populations belongingto differentspecies is significantlyhigherif the populationsgrow at the same site. This site effect seems to be as strongas the species effect. The similarity within sites is not statisticallydifferentfrom the similaritywithin species. This indicatesthat the interactionsof plants and pulp-eatingbirds are little specialized. Interspecificdifferencesin most fruittraits(includingcolor, pulp composition,etc.) probablydo not generallyresult in great or predictabledifferencesin avian fruit consumers. M. Fuentes, Dept of Rangeland Resources and the Ecology Center, Utah State Univ., Logan, UT 84322-5230, USA (present address: Facultade de Ciencias, Universidade da Coruna, E-15071 A Coruia, Spain).

Our view of species interactions can be greatly improved by a geographical approach (see e.g. Jordano 1993, 1994, Thompson 1994). In the case of fruit-bird interactions, new processes are detected as we shift from the local to the geographic scale. Within habitats, interspecific differences in the characteristics of fleshy fruits commonly result in plant species having different assemblages of avian fruit consumers (e.g. Crome 1975, Herrera 1984, Pratt and Stiles 1985, Wheelwright 1985, Davidar 1987, Loiselle and Blake 1990, Fuentes 1994). But site-related variables also affect frugivore assemblages independent of fruit characteristics. Across habitats, any given plant species is faced with changes in the relative abundances of bird species and their food resources (including other fruiting plants) that result in variation in the assemblages of frugivores visiting that species (e.g. McDiarmid et al. 1977, Howe and Vande Kerckhove 1979, Bronstein and Hoffmann 1987, Keeler-Wolf 1988, Wheelwright 1988,

Fuentes 1990, Guitian et al. 1992, Jordano 1993, 1994, Schupp 1993, Traveset 1994). Now, how important are both sets of influences (plant traits versus ecological setting) in shaping fruit-frugivore interactions? The purpose of the present study was to assess quantitatively the relative importance of plant traits and geographic location by comparing frugivore assemblages within and between plant species. I used information on assemblages of pulp-eating birds of European plant species. The specific objective of the comparisons was to discriminate between two potential scenarios. In the first scenario, plant species identity is more important than location in determining the frugivore assemblages of local plant populations: the assemblages associated with local populations of a given plant species are consequently more similar to each other than to assemblages of different plant species. In the second scenario, site-related variables are more important than species identity: coex-

Accepted20 March1995 Copyright? OIKOS1995 ISSN 0030-1299 Printedin Denmark- all rightsreserved 324

OIKOS 74:2 (1995)

Table1. Localitiesincludedin the analyses.Methods:OF,observationsof fruitstaken;OV, observationsof fruit-eatingvisits;FS, analysis of fecal samples. When differentmethodsapply to differentspecies at the same site, this is so indicated(species abbreviationscorrespondto the first letterof genus and threefirst lettersof species epithet;see Appendixfor full names). Site

Authors

Location

Method

CM F1,2 Fl F3 G1 G2 G3,4,5 GI H1 H2 HJ IS IS Ji J1 J1 J2 OB SO SS TR

Courtneyand Manzur1985 Fuentes1990 Fuentes1992 Fuentes1995, unpubl. Guitian1983 Guitian1987 Guitianet al. 1992 Guitianet al. 1994 Herrera1984 Herrera1984 Herreraand Jordano1981 Izhakiand Safriel 1985 Izhakiet al. 1991 Jordano1984 Jordano1987b Jordano1989 Jordano1982 Obeso 1985 Sorensen1981 Snow and Snow 1988 Traveset1993, unpubl.

N England NW Spain NW Spain NW Spain NW Spain NW Spain NW Spain Iceland SE Spain SW Spain SE Spain Israel Israel SW Spain SW Spain SW Spain SW Spain SE Spain S England S England SE Spain

OV OF: Cmon,Csan,Letr FS: Pter,Rulm OF OV OF OF FS FS FS OF OF OF FS FS: Oeur FS: Plen OF FS OF OV OF

isting plant species have relatively similar frugivore assemblages, and there is considerable intraspecific geographic variation in frugivore assemblages.

Methods The data base was all studies I could locate before 1994 describing the avian frugivore assemblages of Palearctic plant species (Table 1). I have considered only birds that feed on fruit pulp, as many of the studies do not provide adequate information on seed predators. Throughout this paper a frugivore "assemblage" consists of the percent contributions of bird species to the consumption of fruits of a plant population. Thus, the assemblage of a plant population is a list of the bird species that feed on its fruits, together with the percent of fruits eaten (or fruiteating visits made, see below) by each of those species, the total for the list adding to 100%. I have considered only studies that describe assemblages both quantitatively and completely, i.e., that list the proportional contributions of all bird species detected feeding on the fruits. The only significantly incomplete assemblages included are those of the Cazorla study (site HI; see Table 1) of Herrera (1984), which includes all disperser species except Turdus philomelos; the included species together accounted for 87.4% of total fruit consumption. Studies differ in the methods employed for documenting frugivore assemblages (Table 1). More importantly, two ways of presenting the data have been used. Guitian (1983), Courtney and Manzur (1985) and Snow and Snow (1988) give information on fruit-eating visits of birds to fruiting plants, while the remaining studies give OIKOS 74:2 (1995)

the proportion of fruits eaten by each bird species. These differences may affect the results of the comparisons by decreasing similarity between sites, especially for the northern set of plants (see below). Also, what is considered a single site in some studies (especially the one by Snow and Snow 1988) would be broken down into multiple sites by other authors. As a consequence, the importance of location as an influence on frugivore assemblages may be underestimated.The same bias may apply to studies based on fecal samples, as the latter may contain remains of fruits eaten at different sites. However, I have focused on very simple analyses and general patterns, and the main conclusions should be insensitive to such biases. For studies based on direct observations of fruit feeding, a record was a fruit-eating visit (data from Guitian 1983, Courtney and Manzur 1985, and Snow and Snow 1988) or a fruit taken by a bird (remaining studies; see Table 1 and Appendix). For studies based on fecal samples, a record was each occurrence of a fruit species in a bird dropping (I. Izhaki pers. comm., Fuentes 1992) or each whole-fruit equivalent found in the droppings (remaining studies). As a measure of similarity (overlap) between pairs of frugivore assemblages I used the index of proportional similarity (PS, see e.g. Hurlbert 1978, Jordano 1994), which is computed as follows: n

PS =

min (Pi,bi), i=l

where for n bird species Pai is the percent of records (number of fruits eaten, number of visits, or number of 325

For some of the analyses I sorted the plant species into two groups, one including the northernlocalities (mainly England) and the other including the southern localities (southern Spain and Israel). The plant species of each group are mostly typical of, respectively, Eurosiberian and Mediterranean vegetation. Only two localities (Fl, F2 and OB; see Table 1) shared plant species with both northern and southern localities. Both localities were included in both groups (see Appendix for a list of the assemblages included in each group).

w

60 O 40I- )0 C 00

On

20-

m

o 50

60

70

80

90

100

110

120

SAMPLE SIZE

Results

Fig. 1. Suitabilityof differentsample sizes (numberof fruits eaten,numberof visits, or numberof occurrencesin birddroppings; see Methods)for documentingfrugivoreassemblages. Shown are the median (horizontallines; N=37 assemblages) and95%of range(bars)of percentagesof randomsampleswith similaritieslower than85%to the originaldata. occurrences in bird droppings) by bird species i on plant species a. PS ranges from 0 (no overlap between assemblages) to 100 (complete overlap). I excluded assemblages based on insufficient number of records. In order to establish the minimum number of records needed for adequately describing an assemblage, I conducted a bootstrap procedure on the 37 assemblages based on more than 200 fruit-eating records. For each of the 37 assemblages, I drew samples of 50 fruit-eating records randomly and with replacement, and computed the similarities of each of the new random assemblages to the original ones. I calculated the proportion of 1000 random samples that had PS < 85 to the original assemblages. I repeated this procedure using samples of increasing number of records (see Fig. 1). I decided to include in the analyses 9 studies based on 63-100 records, 23 on 100-200 and 37 on more than 200 (see Appendix). In view of the results of the bootstrap analysis (Fig. 1), I consider these sample sizes adequate for the kind of general analyses I was interested in. Courtney and Manzur (1985) did not specify the number of observations they made, but other data in their paper suggest that it was large.

My first approach was to analyse general patterns of similarity in frugivore assemblages between plant populations by performing Mantel tests on each of two data matrices: one for the northern group of species, and another for the southern group. Mantel tests compare a data matrix (in this case, a matrix of pairwise PS between plant populations) with a hypothesis matrix that describes a certain relationship between the objects of study (plant populations). The comparison is mathematically equivalent to a product-moment correlation (Mantel r also ranging from -1 to 1), but the statistical significance of this correlation is calculated by a randomization procedure that accounts for the lack of independence of data within each matrix. For each geographical group of species I constructed two hypothesis matrices in order to answer two different questions. First, I tested each data matrix against a matrix of species relationships. Cells in this matrix were 0 for pairs of plant populations belonging to different species, 1 for pairs of the same species. Second, I tested each data matrix against a matrix of site relationships: cells were 0 for pairs of plant populations at different sites, 1 for pairs at the same site. Results are similar for the northern and the southern group. Both plant species identity and site significantly influence the similarity between frugivore assemblages (Table 2). The species effect is less significant than the site effect, perhaps because of the smaller number of possible comparisons. The following analyses used plant species that allowed simultaneously for within-species and within-site com-

Table 2. Mean values of similarity between pairs of frugivore assemblages of plant populations of different species and growing at different sites (unmatched), of the same species (within-species), and of different species growing at the same sites (within-site). Also shown are the normalized Mantel correlation coefficients (r) and associated probabilities (based on 5000 permutations) of the effects of species and site on assemblage similarities.

Northernspecies

Southernspecies

(N= 40)

Unmatched Within-species Within-site 326

(N= 36)

Mean

r

P

Mean

r

P

25.8 41.1 41.8

0.061 0.304

0.061 0.001

32.7 43.5 49.2

0.086 0.233

0.027 0.001 OIKOS 74:2 (1995)

Table 3. Mean values of similarityof frugivoreassemblagesfor selected plant species, and Wilcoxon signed ranks tests (T, conductedon arcsine-transformed values;two-tailedprobabilityvalues are shown in brackets)betweenthe sets of similarities.

1) Between-speciesbetween-site 2) Within-species 3) Within-site Wilcoxontests 1-2 1-3 2-3

Northernspecies

Southernspecies

All species

(N= 7)

(N= 7)

(N= 14)

31.4 36.5 38.6

37.9 52.5 48.5

34.6 44.5 43.6

10 (ns) 6 (ns) 11 (ns)

0 (0.02) 0 (0.02) 9 (ns)

18 (0.05) 6 (0.001) 48 (ns)

parisons (see Appendix). For each of these species I computed the mean similarity between the frugivore assemblage at each site X and the frugivore assemblages of (1) other plant species at site X, (2) conspecific plant populations at other sites Y, and (3) other plant species at those other sites Y. For example, I computed the mean similarity between the frugivore assemblage of Crataegus monogyna at site SS and the assemblages of (1) remaining species at site SS, (2) C. monogyna at other sites (CM, F1 and SO), and (3) other species at sites CM, Fl and SO (see Table 1 for abbreviations and references). I made the same calculations for the three other assemblages of C. monogyna (CM, Fl and SO) and then computed the overall means for within-site similarity, within-species similarity, and between-species betweensite similarity. When comparing frugivore assemblages of southern plants at different sites, those from the same plant species are significantly more similar to each other than those from different species (Table 3). Thus, in this case there is a significant effect of species identity on similarity. There is also a significant effect of site: frugivore assemblages of different species are more similar to each other if they are at the same site (Table 3). I did not detect species or site effects among the northernplants, but both effects are significant when northern and southern plants are grouped (Table 3). I found no evidence that either of these effects is greater than the other: within-species overlap is not significantly different from within-site overlap. This is true for both southern and northern species (Table 3). I made additional analyses using five plant species that are shared by sites H2 and J1, and four species that are shared by sites SO and SS (see Appendix). In this case, I used Mantel tests to compare similarity matrices with one another: H2 with J1, and SO with SS. In the case of H2 and J1, I found a positive and slightly significant correlation between the two matrices (Mantel r=0.76, onetailed P=0.06 based on 1000 permutations). This means that the two matrices have a like structure: if frugivore assemblages of species a and b in H2 are very similar to one another (cell cabin matrix H2 has a relatively high PS value), then assemblages of a and b in Ji will also tend be similar to one another (cell cab in matrix J1 tends to have OIKOS 74:2 (1995)

also a high PS value). The matrices of SO and SS are not significantly correlated (Mantel r=-0.05, one-tailed P =0.46).

Discussion A high degree of specialization in the interactions of fruits and birds (i.e., interspecific differences in fruit and plant traits, such as those discussed by Herrera (1987), resulting in strong and predictable differences in fruit consumers) would result in two interrelated patterns. First, it would produce high heterogeneity of frugivore assemblages among plant species. The frugivore assemblage of a plant population would be different from those of populations of other plant species, regardless of whether these populations grow at the same site or not. Second, different populations of a plant species would have similar frugivore assemblages. My results show a substantial departurefrom this hypothetical situation. Both location and plant species identity significantly influence the frugivore assemblages of local plant populations. When comparing frugivore assemblages of plant populations at different sites, those from the same plant species are significantly more similar to each other than those from different species. This species effect is probably due to interspecific differences in fruit or plant traits causing recurrent differences in frugivore assemblages. Different species of European pulp-eating birds show different fruit preference patterns based on traits such as nutritional profitability (Herrera 1984), seed packaging (Sorensen 1984), fruit size (Jordano 1987a, Snow and Snow 1988), mineral composition of pulp (Jordano 1988), plant lifeform (Snow and Snow 1988), pulp texture (Snow and Snow 1988) and fat content of pulp (Fuentes 1994). Besides this, the strong seasonality of both bird communities and bird foraging behavior in Europe, and the differences between plant species in fruit ripening phenology together play a major role in generating within-species similarity and between-species dissimilarity (Jordano 1988, Snow and Snow 1988, Fuentes 1990, 1994, Boddy 1991). The role of plant traits in influencing interactions is 327

also revealed when the patterns of similarity among a particularset of plant species at one site are compared to the patterns at other sites where the same species also occur together. The patterns of overlap of frugivore assemblages between plant species are similar between two sites in southwestern Spain on which the test could be performed. For example, at both sites the assemblage of Olea europaea is more similar to that of Pistacia lentiscus than to that of Rhamnus lycioides. I did not find such a constancy of similarity relationships for two sites in south central England. On the other hand, the similarity between the frugivore assemblages of plant populations belonging to different species is significantly higher if the populations grow at the same site. This site effect seems to be as strong as the species effect. The overlap within sites is not statistically different from the overlap within species. This indicates that the interactions are little specialized. At each site, a few abundantand/or highly frugivorous bird species consume a high fraction of fruits of most plant species, regardless of differences in fruit traits (Fuentes 1994). As a consequence, the composition of the frugivore assemblage of any plant largely reflects the relative abundance and degrees of frugivory of bird species at each location. Furthermore,were it not for the marked seasonal fluctuations in the abundance and degree of frugivory of many bird species in Europe, the site effect (within-site overlap) would probably be even higher, and the species effect less significant. Bronstein and Hoffmann (1987) and Restrepo (1987) have shown just how large seasonal variation in the frugivore assemblages of particularplant species can be. In this respect, it would be interesting to quantify the relative influence of fruiting phenology (which acts in concert with bird seasonality) and of the remaining fruit and plant traits on the differences and similarities of frugivore assemblages within species and within sites. Interactions of plants and pulp-eating birds elsewhere seem to be as unspecialized as in Europe (e.g. Wheelwright et al. 1984, Gauthier-Hion et al. 1985, Willson 1986, Jordano 1987c, Dowsett-Lemaire 1988, Palmeirim et al. 1989, Loiselle and Blake 1990). Only a few groups of species (most notably mistletoes, see Reid 1991) are exceptional in having restricted and relatively constant assemblages of avian frugivores. These species typically bear fruits that have viscous pulp or protective husks, such that can only be dealt with by a few frugivorous species possessing certain digestive/behavioral traits (Janson 1983, Gauthier-Hion et al. 1985, Pratt and Stiles 1985, Reid 1991; see also Place and Stiles 1992 for the case of wax-covered Myrica fruits). Interspecific differences in most other fruit traits (including color, pulp composition, etc.) probably do not generally result in great or predictable differences in avian fruit consumers. Apparently, much of the observed disparity of fruit traits among plant species is not the result of adaptive responses to (or adaptation for) disparity in frugivore assemblages. 328

- I thank Ido Izhaki, Ram6n Obeso and Acknowledgements AnnaTravesetfor kindlysupplyingunpublisheddata;Jose Guifor helpingwith tian,PedroJordano,andEmilioRolan-Alvarez suggestionsanddiscussions;andPilarAmezquita,JaimeJimenez and Eugene Schupp for making commentsthat greatly helped to improvethe initial manuscript.Duringpart of this studyI enjoyedpredoctoral(FPI)and postdoctoral(MEC/Fulbright)fellowshipsfrom the SpanishGovernment.

References Boddy,M. 1991.Someaspectsof frugivoryby birdpopulations using coastaldune scrubin Lincolnshire.- Bird Study38: 188-199. Bronstein,J. L. and Hoffmann,K. 1987. Spatialand temporal variationin frugivoryat a Neotropicalfig, Ficus pertusa.Oikos 49: 261-268. Courtney,S. P. andManzur,M. I. 1985. Fruitingandfitnessin Crataegusmonogyna:the effects of frugivoresand seed predators.- Oikos 44: 398-406. Crome,F H. J. 1975. The ecology of fruit-pigeonsin tropical northernQueensland.- Aust. Wildl. Res. 2: 155-185. Davidar,P. 1987. Fruitstructurein two neotropicalmistletoes and its consequencesfor seed dispersal.- Biotropica19: 137-139. F 1988. Fruitchoice and seed dissemination Dowsett-Lemaire, by birds and mammalsin the evergreenforests of upland Malawi.- Rev. Ecol. (TerreVie) 43: 251-286. Fuentes, M. 1990. Relaciones entre pajarosy frutos en un espinal del norte de Espafia:variaci6nestacionaly diferenciascon otrasareasgeogrfficas.- Ardeola37: 53-66. - 1992. Relacionesentre pajarosy frutos en espinales del noroestede Espafia.- PhD Dissertation,Univ. Santiagode Compostela. - 1994. Diets of fruit-eatingbirds: what are the causes of interspecificdifferences?- Oecologia97: 134-142. - 1995. The effect of unripefruitson ripe fruit removalby birdsin Pistaciaterebinthus: flag or handicap?- Oecologia 101: 55-58. Gautier-Hion, A., Duplantier,J. M., Quris,R., Feer,F, Sourd, C., Decoux,J. P.,Dubost,G., Emmons,L., Erard,C., Hecketsweiler,P., Moungazi,A., Roussilhon,C. andThiollay,J. M. 1985.Fruitcharactersas a basis of fruitchoice andseed dispersalin a tropicalforestvertebratecommunity.- Oecologia 65: 324-337. del acebo(Ilexaquifolium Guitian,J. 1983.Sobrela importancia L.) en la ecologiainvernalde passeriformesen la Cordillera Cantabricaoccidental.- Ardeola30: 65-76. - 1987. Hederahelix y los pajarosdispersantesde sus semillas: tiempode estanciaen la plantay eficienciade movilizaci6n.- Ardeola34: 25-35. , Fuentes,M., Bermejo,T. and L6pez, B. 1992. Spatial variationin the interactionsbetweenPrunusmahaleband frugivorousbirds.- Oikos 63: 125-130. -, Munilla,I., Guitifn,P. andL6pez,B. 1994. Frugivoryand seed dispersalby redwingsTurdusiliacusin southwestIceland.- Ecography17: 314-320. Herrera,C. M. 1984.A studyof avianfrugivores,bird-dispersed scrublands.plants,and theirinteractionin Mediterranean Ecol. Monogr.54: 1-23. - 1987.Bird-dispersed plantsof the IberianPeninsula:a study - Ecol. Monogr.57: 305-331. of fruitcharacteristics. - andJordano,P. 1981. Prunusmahaleband birds:the highefficiencyseed dispersalsystemof a temperatefruitingtree. - Ecol. Monogr.51: 203-218. Howe, H. F andVandeKerckhove,G. A. 1979. Fecundityand seed dispersalof a tropicaltree. - Ecology 60: 180-189. Hurlbert,S. H. 1978. The measurementof niche overlapand some relatives.- Ecology 59: 67-77. Izhaki,I. andSafriel,U. N. 1985.Why do fleshy-fruitplantsof OIKOS 74:2 (1995)

rismo en altitudesmediasde la Sierrade Cazorla.- PhD Dissertation,Univ. Oviedo. Palmeirim,J. M., Gorchov, D. L. and Stolesson, S. 1989. Trophicstructureof a neotropicalfrugivorecommunity:is therecompetitionbetweenbirdsandbats?- Oecologia79: 403-411. Place, A. R. and Stiles, E. W. 1992. Living off the wax of the land:bayberriesandYellow-Rumped Warblers.- Auk 109: 334-345. Pratt,T. K. and Stiles, E. W. 1985. The influenceof fruitsize and structureon compositionof frugivoreassemblagesin New Guinea.- Biotropica17: 314-321. Reid,N. 1991.Coevolutionof mistletoesandfrugivorousbirds. - Aust. J. Ecol. 16: 457-469. Restrepo,C. 1987. Aspectosecol6gicos de la diseminaci6nde cinco especies de mu6rdagospor aves. - Humboldtia1: 65-116. Schupp,E. W. 1993. Quantity,qualityand the effectivenessof seed dispersalby animals.- Vegetatio107/108: 15-29. Snow, B. and Snow, D. 1988. Birds and berries.- T & AD Poyser,Calton,UK. Sorensen,A. E. 1981.Interactionsbetweenbirdsandfruitsin a Britishwoodland.- Oecologia50: 242-249. - 1984.Nutrition,energy,andpassagetime:experimentswith fruitpreferencein EuropeanBlackbirds(Turdusmerula).- 1989. Pre-dispersal biology of Pistacia lentiscus (AnacarJ. Anim. Ecol. 53: 545-557. diaceae):cumulativeeffects on seed removalby birds. Thompson,J. N. 1994. The coevolutionaryprocess.- Univ. of Oikos 55: 375-386. ChicagoPress,Chicago. - 1993.Geographicalecology andvariationof plant-seeddis- Traveset,A. 1993. Weakinteractionsbetweenavianand insect perserinteractions:southernSpanishjunipersand frugivofrugivores:the case of Pistacia terebinthusL. (Anacardiarous thrushes.- Vegetatio107/108:85-104. ceae). - Vegetatio107/108:191-203 - 1994. Spatialand temporalvariationin the avian-frugivore - 1994. Influenceof type of avianfrugivoryon the fitnessof Pistacia terebinthus L. - Evol. Ecol. 8: 618-627. assemblageof Prunusmahaleb:patternsandconsequences. -Oikos 71: 479-491. Wheelwright,N. T. 1985.Fruitsize, gape width,andthe diet of Keeler-Wolf,T. 1988. Fruitand consumerdifferencesin three fruit-eatingbirds.- Ecology 66: 808-818. speciesof treessharedby TrinidadandTobago.- Biotropica - 1988. Fourconstraintsin coevolutionbetweenfruit-eating 20: 38-48. birdsandfruitingplants:a tropicalcase history.- Proc.XIX Int. Orn.Congr.,pp. 827-845. Loiselle, B. A. and Blake, J. G. 1990. Diets of understory fruit-eatingbirds in Costa Rica: seasonalityand resource -, Haber,W. A., Murray,K. G. and Guindon,C. 1984. abundance.- Stud.AvianBiol. 13: 91-103. Tropicalfruit-eatingbirdsandtheirfood plants:a surveyof a CostaRicanlowermontaneforest.- Biotropica16: 173McDiarmid,R. W., Ricklefs, R. E. and Foster,M. S. 1977. 192. Dispersal of Stemmadenia donnell-smithii (Apocynaceae) by birds. - Biotropica 9: 9-25. Willson, M. F. 1986. Avian frugivoryand seed dispersalin Obeso, J. R. 1985. Comunidadesde Passeriformesy frugivoeasternNorthAmerica.- Curr.Orithol. 3: 223-279. the mediterranean scrubinterceptfall- but not spring-passage of seed-dispersingmigratorybirds?- Oecologia 67: 40-43. -, Walton, P. B. and Safriel, U. N. 1991. Seed shadows generatedby frugivorousbirdsin an easternMediterranean scrub.- J. Ecol. 79: 575-590. Janson,C. H. 1983.Adaptationof fruitmorphologyto dispersal agentsin a Neotropicalforest.- Science 219: 187-189. of Jordano,P. 1982. Migrantbirdsarethe mainseed-dispersers blackberries Rubusulmifoliusin southernSpain.- Oikos38: 183-193. - 1984.Relacionesentreplantasy aves frugivorasen el matorral mediterrnneodel area de Dofiana.- PhD Dissertation, Univ. Sevilla. - 1987a. Frugivory,externalmorphologyand digestive systemin mediterranean sylviidwarblersSylviaspp.- Ibis 129: 175-189. - 1987b.Avian fruitremoval:effects of fruitvariation,crop size, and insect damage.- Ecology 68: 1711-1723. - 1987c.Patternsof mutualisticinteractionsin pollinationand seed dispersal:connectance,dependenceasymmetries,and coevolution.- Am. Nat. 129: 657-677. - 1988. Diet, fruitchoice and variationin body conditionof scrubland.- Ardea frugivorouswarblersin Mediterranean 76: 193-209.

OIKOS 74:2 (1995)

329

Appendix. Plantpopulationsincludedin the analyses.N and S denote assemblagesassignedto northernand southerngroups, respectively.The asteriskmarksthe assemblagesused in the analysesof Table3. See abbreviationsfor sites in Table 1. Species

Site

Sample

Group

Arbutus unedo Arctostaphylos uva-ursi Cornus sanguinea Cornus sanguinea Cornus sanguinea Crataegus monogyna Crataegus monogyna Crataegus monogyna Crataegus monogyna Daphne gnidium Empetrum nigrum Euonymus europaeus Hedera helix Hedera helix Hedera helix Ilex aquifolium Ilex aquifolium Ligustrum vulgare Lonicera caprifolium Lonicera etrusca Malus sylvestris Myrtus communis Myrtus communis Olea europaea Olea europaea Osyris quadripartita Phillyrea angustifolia Phillyrea angustifolia Phillyrea latifolia Pistacia lentiscus Pistacia lentiscus Pistacia lentiscus Pistacia lentiscus Pistacia palaestina Pistacia terebinthus

H1 GI F1 F2 SS CM F1 SO SS J1 GI SS G2 SO SS G1 SS SS SS F1 SS H2 J1 H2 J1 H2 H2 J1 H1 H1 H2 IS J1 IS Fl

208 107 129 601 2321

S N NS N N N NS N N S N N N N N N N N N NS N S S S S S S S S S S S S S NS

330

287 5538 3704 102 2256 347 896 181 2136 425 1797 175 156 92 194 93 84 143 305 159 91 158 236 195 486 165 1183 118 76

*N * * * *N * *

* * * * *

* * * * * * * * * *

Species

Site

Sample

Pistacia terebinthus Pistacia terebinthus Prunus avium Prunus mahaleb Prunus mahaleb Prunus mahaleb Prunus mahaleb Prunus padus Prunus spinosa Prunus spinosa Rhamnus alaternus Rhamnus catharticus Rhamnus lycioides Rhamnus lycioides Rhamnus palaestinus Rosa canina Rosa canina Rubus fruticosus Rubus ulmifolius Rubus ulmifolius Rubus ulmifolius Rubus ulmifolius Sambucus nigra Sambucus nigra Smilax aspera Solanum dulcamara Sorbus aria Sorbus aucuparia Taxus baccata Vaccinium uliginosum Viburnumlantana Viburnumopulus Viburnumtinus Viscum album

F3 TR SS G3 G4 G5 HJ SS SO SS F3 SS H2 J1 IS OB SS SS F1 J1 J2 OB SO SS H2 SS SS SS SS GI SS SS H1 SS

107 720 581 178 261 296 987 165 542 781 127 551 63 182 322 78 1082 551 227 100 897 91 309 1423 109 110 641 649 2148 732 109 118 95 262

Group * * N S S S S N N N S N

* * * *

S NS N N NS S S NS N N S N N N N N N N S N

*N * *S *S

*S

OIKOS 74:2 (1995)