willson_traveset 2000 - the ecology of seed dispersal.pdf

7/26/2019 Willson_Traveset 2000 - The Ecology of Seed Dispersal.pdf

1/26

Chapter 4

The Ecology of Seed Dispersal

Mary F. Willson1 and Anna Traveset215230 Terrace Place, Juneau, Alaska, USA; 2Institut Mediterrani dEstudis Avanats (CSICUIB),

Esporles, Balearic Islands, Spain

Introduction

Seed dispersal has long been a topic ofinterest to naturalists, but it has not beenuntil the last three decades that the ecologyof dispersal has received much rigorousscientific attention. Many theoretical and

empirical advances have recently beenmade, although important lacunae in ourunderstanding still need to be filled beforedispersal ecology becomes a coherent bodyof knowledge. A major goal of this chapteris to review the existing literature on seeddispersal, highlighting these essential butmissing kinds of information.

This review is divided into two sec-tions, the first dealing with the evolution ofdispersal mechanisms and the second with

the consequences of dispersal at popula-tion and community levels. Dispersal canoccur in both space and time, but only theformer will be treated here, except wheresome relationship between the two axes isknown (Venable and Brown, 1988; Leck etal., 1989; Eriksson and Ehrln, 1998a; seealso Murdoch and Ellis, Chapter 8, thisvolume).

The seed shadowThe spatial distribution of dispersed seedsaround their source is called a seed

shadow (Janzen, 1971). The term is mostcommonly (and perhaps properly) used inreference to the postdispersal distributionof seed around the maternal parent, but itcan also be used to refer to the distributionof seeds around a source composed of mul-tiple parents. Although the unit of disper-

sal may be technically a fruit or a group offruits and, thus, the generic label should bediaspore or propagule, we shall continue touse the more euphonious term of seedshadow, with the intent of conceptuallyincluding all diaspores of any morphologi-cal and genetic derivation.

Seed shadows exist in two horizontaldimensions for most seed plants (a salientexception is found in epiphytic plants, inwhich the third, vertical, dimension is pre-

sumably well developed). The shape of theseed distribution in the horizontal planemay be asymmetric with respect to thesource (if, for example, the wind carriesmost seeds in a downwind direction). Thissimple descriptor of shape is generally aug-mented by information on seed density,such that the seed shadow resembles atopographic map, with peaks of high seeddensity. Two factors, then, can be used todescribe seed shadows the relationship ofseed numbers or density to distance fromthe source and the directionality withrespect to the source.

Although directionality is clearly

CAB International2000. Seeds: The Ecology of Regeneration in Plant Communities, 2nd edition 85(ed. M. Fenner)


2/26

significant for many ecological questions, itis common to discuss seed distributionschiefly in terms of distance from thesource. Conventional wisdom describes the

seed number/distance relationship as lep-tokurtic (with a higher peak and longer tailthan a normal distribution); from the peakoutward, seed numbers are generally con-sidered to decrease monotonically, fitting anegative exponential curve (or sometimes anegative power function: Okube and Levin,1989; Fig. 4.1). Most measured seed shad-ows conform to this expectation (Fig. 4.2;Willson, 1993a). The nature of the source(single or multiple individuals) can influ-

ence the location of the peak of the curve,and the use of seed density rather thannumbers can change the overall shape ofthe curve, including the proximity of thepeak to the source (Peart, 1985; Greene andJohnson, 1996).

Deviations from the conventional seedshadow shape can result from patchinessof habitat structure (Hoppes, 1988;Debussche and Lepart, 1992; Debusscheand Isenmann, 1994; Kollmann and Pirl,

1995; Aguiar and Sala, 1997) and other

ecological factors, including behaviour pat-terns of the seed vectors that lead to nucle-ation processes (Willson and Crome, 1989;McClanahan and Wolfe, 1993; Verd and

Garca-Fayos, 1996, 1998; Julliot, 1997).For species with polymorphic seeds (i.e.with and without dispersal devices, orwith two or more different kinds of disper-sal devices), the shape of the seed shadowfor each seed type may differ, such that thecombined seed shadow for a given parentmay have a very unconventional shape.

Many factors can alter the location ofthe peak and the slope and shape of the tailof the seed shadow for particular species

and individuals (e.g. Rabinowitz and Rapp,1981; Johnson, 1988; Debussche andLepart, 1992; Verd and Garca-Fayos,1996, 1998; Julliot, 1997). Some factors are,from the plants perspective, strictly envi-ronmental and thus outside the control ofthe plant (e.g. the strength and direction ofthe wind, the social behaviour of animaldispersal agents, the patterns of rainfalland relative humidity). Other factors, suchas plant height, have a strong environmen-

tal component (Greene and Johnson, 1996)

86 M.F. Willson and A. Traveset

Fig. 4.1. Idealized curves commonly used to describe the distribution of seeds at increasing distances(arbitrary units) from the seed source. Real seed shadows often peak at some distance from the source; inthat case, the curves refer to the part of the seed distribution from the peak outward. For a given set ofcoefficients, the negative exponential curve (y= axmx) drops less steeply than the negative power function(y= axm); it is converted to a straight line on a semilogarithmic scale. (From Okube and Levin, 1989.)


3/26

but may also have a genetic component.Still others are probably controlled both byenvironment and the genetic constitutionof the parent plant, the balance depending

on the species and circumstances (e.g. fruitsize, seed size, ease of dehiscence orabscission). All such factors can contributeto variation in the size and shape of the

seed shadow among species and amongconspecific individuals.

Few data yet exist, either from thetropics or from the temperate zone, to com-

pare seed shadows generated by differentdispersal modes (but see Gorchov et al.,1993; Portnoy and Willson, 1993; Willson,1993a). Even less is known about how the

Ecology of Seed Dispersal 87

Fig. 4.2. Seed shadows of three individuals of Lithospermum caroliniense. The tails of all three seedshadows fit a negative exponential curve, but the slope of that curve (on a semilog plot) varies from 1.47to 1.72, and the location of the peak differs. (From Westelaken and Maun, 1985.)


4/26

loss or the addition of a dispersal agentalters the seed shadow of a plant.Moreover, we cannot yet make any general-ization about the relative ecological impor-

tance of different portions of the seedshadow for the evolutionary ecology ofplants.

In order to understand the ecologicaland evolutionary consequences of variationin the length and shape of the seed shad-ows, we need to experimentally modifyseed distributions and monitor the fitnessof recruits in different parts of the distribu-tion, for a variety of species and circum-stances (Portnoy and Willson, 1993). The

tail of the distribution may be at least asimportant as the modal portion of thecurve, although little theoretical effort hasbeen devoted to this question (but see, forinstance, Cain et al., 1998). Propagules inthe distributions tail potentially spread theparental genes more widely, and planttraits that affect the behaviour of such tailscan be subject to selection. The examina-tion of 68 data sets has shown a lack ofassociation between tail shape and disper-

sal mode, suggesting that, in most circum-stances, selection for tail behaviourcontributes little to the evolution of thedispersal mode itself (Portnoy and Willson,1993).

The evolution of dispersal

Why are seeds dispersed?

If the dispersal of offspring increases thefitness of a parent, we should expect thatdispersed offspring survive and reproducebetter than undispersed offspring, eitherbecause they avoid detrimental conditionsnear the parent or because they reach betterconditions farther away (which amounts tothe same thing, from a different perspec-tive). If seeds fall directly beneath thecanopy of a parent, the physical separationhardly constitutes real dispersal, but fallenseeds are normally treated as part of theseed shadow, for purposes of comparingseed fates. Van der Pijl (1982) actuallytreats simple seed fall as a separate mode

of dispersal for species that have no evi-dent special means. The principal factorsthat favour dispersal are avoidance of nat-ural enemies or sibling interactions and the

probability of finding a physically suitableestablishment site.

1. Some natural enemies of seeds andseedlings respond to density and/or dis-tance from the parent (or other conspe-cific). Pathogens, postdispersal seedpredators, parasites and herbivores oftenconcentrate their activities where theirresources are common, and more distantseeds/seedlings may survive better thanthose close to the parent (e.g. Howe and

Smallwood, 1982; Augspurger, 1983a, b;Augspurger and Kelly, 1984; Howe, 1993;Peres et al., 1997; Hulme, 1998); the magni-tude of this effect varies among species(Augspurger, 1984; Howe, 1993). Theimpact of such consumers depends, inpart, on their specificity to genotype or tospecies of resource some attackers mayspecialize in the offspring of particular par-ents or in particular taxa. The density orproximity of other genotypes or species of

seed would have little impact on the avail-ability of suitable resources for such spe-cialists, and thus the effect of density- anddistance-responsive enemies must oftenvary with their specificity.

The ability of natural enemies todepress seed and seedling density alsodepends, of course, on other factors limit-ing their abundance and activity. Thesewill vary among consumers, and evenamong populations of the same species of

consumer. Thus, although numerous casesof density- and distance-responsive attack-ers have been reported, we do not yet havea general picture of which plant species aresubject to such attacks and in what circum-stances (e.g. habitat, season, geographicalregion, adult densities).2. Because the seeds and seedlings of anyone parent are genetically related (at leasthalf-sibs), they are subject, potentially, tosibling competition. Conventional wisdomsuggests that sib competition may often bemore severe than competition with non-sibconspecifics, because their patterns ofresource use are probably more similar (see



5/26

references in Ellstrand and Antonovics,1985; see also McCall et al., 1989).However, in a number of cases, sib compe-tition does not have a detectably different

outcome from that of non-sib competition,and sibs may even profit, at some earlystages of the life history, from the proxim-ity of genetic relatives (Smith, 1977;Williams et al., 1983; Willson et al., 1987;Kelley, 1989; McCall et al., 1989).Furthermore, self-fertilized seeds disperseless well than outcrossed seeds inImpatiens capensis and Amphicarpaeabracteata (Schmitt and Ehrhardt, 1987;Trapp, 1988), which is the opposite of what

would be expected if dispersal were anadaptation to reduce sib competition. Also,if sib competition were critical, geneticallyvariable offspring should have higher fit-ness than clonal progeny where sib densi-ties are high, but this was not the case inAnthoxanthum odoratum (Kelley et al.,1988).

Another potential disadvantage of highsibling densities is the possibility ofinbreeding when (and if) the offspringreach adulthood (Ghiselin, 1974). The rela-tive effects of extreme inbreeding and out-crossing on the genetic variance ofoffspring are not predictable in any simpleway, because they depend on many aspectsof the genetic system, as well as pastepisodes of inbreeding (McCall et al.,1989). As a result, the degree of offspringsimilarity and the intensity of potential sibcompetition are likewise difficult to pre-dict. A degree of inbreeding may be advan-

tageous under certain circumstances (e.g.Shields, 1982; Jarne and Charlesworth,1993). Furthermore, even if inbreeding isdisadvantageous, there are other ways forplants to reduce inbreeding (e.g. throughchanges in the floral biology and matingsystem), so it is difficult to assess theimportance of inbreeding avoidance as afactor that selects for offspring dispersal.Moreover, dispersal has its own costs (e.g.Cohen and Motro, 1989), which may out-

weigh the costs of some inbreeding (Waseret al., 1986); the benefit/cost ratio is likelyto vary among species (e.g. Augspurger,1986).

Attacks by parasites and pathogensmay be more devastating when sibs growin close proximity to each other (seeAlexander and Holt (1998) for a recent

review on the interaction between plantcompetition and disease). Just as con-specifics growing in a monoculture areoften more heavily hit by pests than theyare when growing in a mixed stand, so alsoare the genetic monocultures of closelyrelated individuals sometimes more heav-ily hit by certain kinds of pests than standsof mixed parentage (e.g. Parker, 1985;Burdon, 1987). To the extent that a speciesis subject to such attacks on particular

genetic lineages, there may be selection fordispersal, which lowers the concentrationof any one lineage in a given area (Fig. 4.3).3. Some species have special physicalrequirements for germination and estab-lishment that are met only in scatteredlocations, such as fallen logs, tree-fall gapsor badger mounds. In the absence of well-developed dormancies and the ability towait for suitable conditions to arrive, selec-tion may favour dispersal in order toincrease the probability of finding the nec-essary kind of location (Platt and Weis,1985; Reid, 1989; Sargent, 1995). Evenwith good dormancy mechanisms, disper-sal should enhance the probability that awaiting seed will eventually encounter theright conditions for establishment.Theoretically, dispersal generally enhancesthe likelihood that at least some offspringreach appropriate sites (Hamilton and May,1977; Comins et al., 1980). These theoreti-cal expectations need to be examined for avariety of particular cases. If the seedshadow is adapted to the distribution ofsafe sites, species with widespread seedshadows should have more far-flung estab-lishment sites than species with restrictedseed shadows (Green, 1983; see also Geritzet al., 1984; Horvitz and Le Corff, 1993).

What portion of the seed shadow ismost effective in yielding successful off-spring and how does this vary with species

and conditions? The peak portion of theseed shadow, where the most seeds aredeposited, often receives the most attentionfrom ecologists. Although the peak may be



6/26

ecologically important as a source of foodfor seed predators and other consumers, we

seldom know if it is the most important interms of parental fitness. In part becauseconsumers or intense competition can levelthe peak and in part because rare eventscan be ecologically important, it seemsessential that the tail of the distributionshould receive more attention in the future.Moreover, we need to know if there are anypatterns (among species, for instance) inthe relative importance of the distribu-tional tail. A number of studies have

shown that seed and seedling survivalincreases with distance from the parent(see above), or that more distantly dispers-ing seeds are more likely to reach goodsites (Platt and Weis, 1985). However, a fewstudies have shown that seeds at the end ofthe seed shadow often do poorly (Horvitzand Schemske, 1986; Augspurger andKitajima, 1992) and some authors haveargued (without experimental documenta-tion, however) that selection may actually

oppose dispersal in some species (Zohary,in van der Pijl, 1982; but see Ellner andSchmida, 1981). Recent studies by Russelland Schupp (1998, and references therein)

show that patterns of initial seed-fall den-sity are more affected by distance from a

seed source than by the physical structureof the microhabitat, at least for wind-dis-persed plants. On the other hand, investi-gations by Donohue (1997, 1998), who hasdecoupled the fitness effects of density anddistance from the home site, have shownthat selection on dispersion patterns islikely to be through density rather than dis-tance effects. This author has also investi-gated the maternal environment effects onseed dispersal within an evolutionary con-

text. Although maternal characters (e.g.fruit and seed traits, architectural traits,plant size, fruit production) are known toinfluence seed dispersal, only a few studiesaddress the evolution of dispersal withinthe context of maternal character evolution(see also Thiede and Augspurger, 1996).

How are seeds dispersed?

If dispersal is advantageous, we wouldexpect to find that diaspores have adapta-tions that enhance dispersal (Ridley, 1930;van der Pijl, 1982). Morphological devices


Fig. 4.3. A bear garden seedlings of Ribes bracteosumgrowing from a faecal deposit of an Alaskanbrown bear. The small squares in the grid beside the garden are 3 cm on a side. Both sib and non-sibcompetition must be intense, predation by rodents can be severe and survivorship may ultimately be low.Some bear deposits contain seeds of two or three species, and germination may occur a month to a year ormore after deposition. (Photo by J. Zasada.)


7/26

that enhance dispersal are often quite read-ily evident and interpretable (Kerner, 1898;Ridley, 1930; van der Pijl, 1982), althoughsome dispersal-enhancing traits (such as

buoyancy of seeds dispersed by water) areless immediately obvious. Wind-bornediaspores often have wings or plumes thatincrease air resistance and slow the rate offall. Diaspores carried by animal con-sumers commonly have edible appendagesor coverings that are consumed by animalsthat later eject the seeds (Fig. 4.4); in somecases, the seeds themselves are harvestedand either eaten (killed) or cached andsometimes abandoned by the harvester

(Sork, 1983; Price and Jenkins, 1986).Other animal-dispersed diaspores travel bymeans of hooks or sticky coatings thatadhere to the exteriors of the animal vec-tors. Some plants disperse their offspringballistically, by the explosive opening ofthe fruits or the springing of a trip-lever.Seeds of certain plant species combine two

or even three modes of dispersal (Westobyand Rice, 1981; Clifford and Monteith,1989; Stamp and Lucas, 1990; Aronne andWilcock, 1994; Traveset and Willson,

1997), as in some Viola (ballistic plus ants),Disporum, Rhamnus, Myrtus, Smilax(birdsplus ants) and Petalostigma pubescens(birds plus ballistic plus ants); a great num-ber of species are dispersed by both birdsand mammals (e.g. Herrera, 1989b;Willson, 1993b; Traveset and Willson,1997).

The dispersal potential of the differentmodes of dispersal varies greatly. Bothwind and vertebrates can potentially carry

seeds far from the parent plant, but antsand ballistic mechanisms typically gener-ate shorter seed shadows. A preliminarysurvey (Willson, 1993a) for herbaceousspecies indicates that peak and maximumdispersal distances are greater, and theslope of the tail of the seed distribution isless steep, for wind and ballistic dispersal


Fig. 4.4. Ants (Formica podzolica) picking up seeds of Viola nuttallii. The seed bears an attractive andedible appendage. Ants carry the entire seed back to their nest, eat the appendage and discard the seed.Dispersal of seeds by ants is very common in some floras, but the advantage of ant dispersal may varygreatly among species or regions (e.g. escape from predators or other destructive agents, or deposition in anespecially favourable site for germination and growth). (From Beattie, 1985, p. 74.)


8/26

than for species with no special devices onthe diaspore. Maximum distances aregreater for wind-dispersed than for ballisti-cally dispersed seeds of herbs, although

peak distances and slopes are similar. Suchresults indicate that, on average, dispersaldevices seem to work. But variation aroundthe averages was great, and both samplingmethods and environmental conditions ofdispersal affect the outcome. A much betterdatabase is needed to make good compar-isons of the seed shadows produced in dif-ferent ways.

Of course, there is also great variationin the dispersal potential within each gen-

eral mode of dispersal. The relative size ofseed and wing or plume can have enor-mous effects on the seed shadow of wind-borne seeds (Augspurger and Franson,1987; Sacchi, 1987; Benkmann, 1995). Thesize and chemistry of the edible appendageon ant-dispersed seeds may influence therate of seed removal (Gunther and Lanza,1989; Gorb and Gorb, 1995; Mark andOlesen, 1996), the array of dispersing antspecies and the eventual fate of the seeds.

Dispersal of fleshy fruits by ground-forag-ing vertebrates generates different seedshadows from those produced by flyingfruit-eaters. Nuts favoured by scatter-hoarding squirrels may be spread morewidely than less favoured species of nuts(Stapanian and Smith, 1984), but jays carryacorns much further than squirrels carryany nuts (Johnson and Webb, 1989). Seedsdispersed by frugivorous lizards also showdifferent patterns of deposition from those

dispersed by mammals (Traveset, 1995). Bycontrasting dispersal syndromes in a fam-ily (Marantaceae) of tropical understoreyherbs, Horvitz and Le Corff (1993) foundthat bird-dispersed species went furtherthan ant-dispersed species; however, dis-persion patterns did not vary among typesof dispersal, most species having aclumped spatial patterning.

For vertebrate-dispersed species, it hasbeen hypothesized that plants can exert

some kind of control over seed shadowsproduced by frugivores, by specific laxativeand/or constipative chemicals in the fruitpulp, which affect seed retention time in

the dispersers guts (Murray et al., 1994;Cipollini and Levey, 1997; Wahaj et al.,1998). In turn, seed retention time insidethe disperser, together with other factors

(reviewed in Traveset, 1998), can affect thegerminability, the rate of germination, orboth, in certain species.

Although many species exhibit themorphological devices that are presumedto enhance dispersal, large numbers ofspecies lack any evident dispersal device(Ridley, 1930; Willson et al., 1990a;Willson, 1993a; Cain et al., 1998). Theseeds of some of these species are so smallthat they are easily wind-borne without

special devices (e.g. orchids). Others havesmall, hard seeds that are consumed byherbivorous vertebrates along with thefoliage and dispersed after passing throughthe animals gut, but whether or not thisconstitutes an evolutionary design can bedebated (Janzen, 1984; Collins and Uno,1985; Dinerstein, 1989). Still other specieshave small, round seeds that are shakenout of the fruit when the plant is stirred bya passing breeze or animal (e.g. Papaver);

this is considered to be a special mecha-nism for dispersal by some (van der Pijl,1982). Yet many species remain that lackany apparent device for dispersal in spaceand that also appear to have little capacityfor dispersal in time (Willson, 1993a).Although some of these species may be dis-covered to be dispersed by one of the majormodes, it is reasonable to ask why so manyspecies seem to lack dispersal devices.How do such species achieve effective dis-

persal? Or is dispersal less advantageous inthese species?Variation in dispersal devices is often

evident, both within species and amongclosely related species. Examples ofwithin-species variation are provided bymany Asteraceae (seeds with and withouta device for wind dispersal) and amphi-carpic species, which have both above-ground seeds, which may be dispersed byany of the normal vectors, and below-

ground seeds, which may not disperse atall or may be harvested and cached byrodents (van der Pijl, 1982). Examples ofvariation among related species are many,



9/26

one of the most striking being found in theAcacia genus, which is widespread

through Latin America, Australia, Asia andAfrica. Most Australian acacias exhibitmorphological adaptations for dispersal byants or birds (Fig. 4.5); the Americanspecies are dispersed by birds or largemammals (ODowd and Gill, 1986); theAfrican species are dispersed by largemammals and, reportedly, by wind (Coeand Coe, 1987). Different populations ofAcacia ligulata in Australia seem to beadapted to dispersal by ants or birds

(Davidson and Morton, 1984). Even if thisenormous genus is eventually split intoseveral new genera, the basic observationof diversity of dispersal within a set ofrelated species remains valid. In the genusPinus, seed mass is associated with themode of dispersal; pine seeds weighingless than about 100 mg are wind-dispersed,while heavier seeds usually have adapta-tions for bird dispersal (Benkman, 1995). Inlineages where long-distance seed disper-

sal predominates, clonal propagation hasevolved on frequent occasions, possibly tomake genet fitness less dependent on localdispersal by seed (Eriksson, 1992).

The mode(s) of dispersal of any plantspecies must reflect many different pres-

sures and constraints. Because naturalselection must work with existing variationand many plants have very long generationtimes, there are inevitable phylogeneticconstraints. Entire families or genera some-times exhibit only slight variations on asingle mode of dispersal. However, exten-sive variation within families (e.g.Liliaceae), genera (e.g. Acacia, Pinus) andspecies (e.g. Spergularia marina,Heterotheca latifolia) demonstrates that

such constraints are neither universal nortotally confining.Constraints on the evolution of dias-

pores also emerge from the many, some-times potentially conflicting, selectionpressures that impinge on diaspore design(e.g. Ellner and Schmida, 1981; Benkmanet al., 1984; Westoby et al., 1991;Armstrong and Westoby, 1993; Leishmanand Westoby, 1994a, b; Kelly, 1995; Winnand Miller, 1995; see also Leishman et al.,

Chapter 2, this volume). For example,seeds must be endowed with adequateresources to accomplish germination andestablishment. If these processes must take


Fig. 4.5. Variations on a theme Acacia seeds apparently adapted for dispersal by ants (above) and bybirds (below). The food reward in the appendage on the seed is larger and usually more colourful (red,orange, yellow versus white) in bird-dispersed species. (Photo courtesy of D.J. ODowd.)


10/26

place in sites where the seedling can ini-tially generate few resources on its own orin sites where intra- or interspecific compe-tition is severe, there can be selection to

increase the energy reserves within theseed, often resulting in a larger seed (e.g.Foster and Janson, 1985; Winn and Miller,1995). Large seeds are generally harder todisperse than small ones they need largeranimals, stronger winds or more powerfulpropulsion (Kelly, 1995). Thus, selectionfor large seed size may bring with it selec-tion for altered dispersal devices or mayconstrain the array of efficacious dispersalagents. Very large seeds cannot go far bal-

listically or by adhering to animal exteri-ors, they cannot be carried by smallanimals (such as ants) and they need verylarge wings to be successfully wind-dis-persed. One option may be dispersal byvertebrates (Willson et al., 1990a), andlarger seeds generally require larger verte-brates to carry them (Foster and Janson,1985; Wheelwright, 1985; Hammond andBrown, 1995). Even the chemical composi-tion of seeds is associated with the disper-

sal mode; wind- and animal-dispersedspecies generally have greater proportionsof fat and less of protein and carbohydratesthan passively dispersed seeds (Lokesha etal., 1992).

Furthermore, seeds must be protectedagainst the physical environment and fromnatural enemies (see Chambers andMacMahon, 1994), and the demands ofprotection may sometimes interfere withdispersal by certain means. Also, fruits are

often photosynthetic; selection to enhancethe photosynthetic capacity of the fruitcould affect fruit size, colour, shape andother design features that influence disper-sal. In addition, a long style can increasethe intensity of competition among malegametophytes, but it also affects dispersaldistance in the ballistically dispersedGeranium (M.F. Willson and J. gren,unpublished). The timing of dispersalaffects the probability and pattern of dis-

persal and the susceptibility to enemies,and hence can affect the selection pres-sures on diaspore morphology. The physio-logical costs of the various modes of

dispersal are generally unknown, but theyconstitute potential constraints on the evo-lution of dispersal devices.

Plant size and growth form show some

correlations with dispersal mode, whichmay affect the evolution of dispersal traits(Thompson and Rabinowitz, 1989; Willsonet al., 1990a). For instance, few plants thatare dispersed by ants or externally on ani-mals are tall in stature. One reason may bethat tall plants typically have large crowns,and ants commonly carry seeds for rela-tively short distances, such that the seedsof large-crowned, tall plants would seldombe carried beyond the crown of the parent

and dispersal would be relatively ineffec-tive. Ant dispersal of small acacia trees inAustralia may be the exception that provesthe rule and thus worthy of special study.Given that a species has become tall, therange of efficacious modes of dispersal maybe limited. Most ballistically dispersedplants are small in stature, but at least afew trees use this mode. Dispersal by windis frequent among species that are rela-tively tall within their respective habitats

(e.g. trees in forests, tall forbs in fields).Although some understorey plants are alsodispersed by wind, the common inferenceis that relatively short stature often renderswind dispersal less advantageous thanother modes. Stature and growth form aresometimes correlated with seed size (Fosterand Janson, 1985; Willson et al., 1990a; butsee Kelly, 1995), which may thus influencedispersal mode indirectly (through growthform), as well as more directly.

Whatever the array of constraints ondiaspore evolution may be, it is also neces-sary to ascertain the occurrence and magni-tude of selection on dispersal traits. Atleast two fundamental approaches are use-ful.

First, studies explicitly designed tomeasure selection on dispersal traits areessential. Seemingly small differences inthe design of dispersal devices can haveprofound effects on dispersal ability: on

aerodynamic performance by wind-dis-persed species (Augspurger and Franson,1987; Matlack, 1987; Sacchi, 1987); oncapacity for attachment for diaspores car-



11/26

ried externally by animals (Bullock andPrimack, 1977; Carlquist and Pauly, 1985;Sorensen, 1986); on foraging preferences ofavian fruit consumers (Howe and vande

Kerckhove, 1980; Wheelwright, 1985;Whelan and Willson, 1994; Traveset et al.,1995; Loiselle et al., 1996; Rey et al., 1997;Traveset and Willson, 1998); and on theaverage distance of seed dispersal in a bal-listically dispersed herb (M.F. Willson andJ. gren, unpublished). Furthermore, indi-vidual plants often differ in the allocationof resources to dispersal devices on thediaspore (Willson et al., 1990b; Jordano,1995a), although the extent to which such

differences are inheritable is seldom estab-lished (but see Wheelwright, 1993).Individual variation in seed shadows hasbeen documented for a few species(Augspurger, 1983a; McCanny and Cavers,1987; Thiede and Augspurger, 1996;Donohue, 1997, 1998). Such informationneeds to be brought together, so that weknow the extent and pattern of individualvariation in dispersal devices: how thevariation affects the seed shadows of the

respective parents; how the seed shadowaffects parental fitness; and how these rela-tionships vary among species and condi-tions. Several studies have shown thatselection by avian dispersal agents may berelatively weak (e.g. Manasse and Howe,1983; Herrera, 1984c, 1987, 1988; Jordano,1987, 1993, 1994, 1995a, b; Guitin et al.,1992; Willson and Whelan, 1993; Traveset,1994; Whelan and Willson, 1994),although, in most cases, fitness is indexed

by removal rates rather than by the even-tual pattern of offspring dispersion.Secondly, a less direct but still useful

approach is to document patterns of varia-tion in the array of dispersal modes presentin plant communities (i.e. the dispersalspectra of those communities). Examinationof the pattern can help generate hypothesesabout the relative advantage of differentdispersal modes in different regions andhabitats. A few patterns have begun to

emerge, but we seldom know how generalthey are. One consistent trend is that a highproportion of species in tropical wet forestis dispersed by vertebrate consumers (see

references in Willson et al., 1989), althoughthere are biogeographical differences in thestrength of the trend (Karr, 1976; Snow,1981; Fleming et al., 1987). In temperate

zones, forests commonly have more verte-brate-dispersed species than other habitats,and the frequency of fleshy-fruited speciesis especially high in certain southern-hemi-sphere forests (Willson et al., 1990a).Vertebrate dispersal apparently increaseson moister sites, on fertile soils (inAustralia) and in floras dominated byshrubs and/or trees (Willson et al., 1990a).The causal factors for such patterns are notyet clear. In contrast, when comparing the

seed dispersal spectra of five differenttypes of communities on the IberianPeninsula (potential woodland, forestfringe, scrubland, nitrophile communitiesand montane communities) betweenMediterranean and Eurosiberian regions,no significant differences are found for anytype of community, although biotic disper-sal appears to be consistently more preva-lent at mature stages of succession (Guitinand Snchez, 1992).

A conspicuous and well-documentedobservation is the extraordinarily high fre-quency of ant-dispersed species inAustralia and South Africa, particularly insclerophyll vegetation on infertile soils.Several hypotheses (reviewed in Westobyet al., 1991) have been proposed to explainsuch patterns, although few have beentested thoroughly (Hughes et al., 1993).Seed size, the cost of dispersal structuresand the availability of dispersal agents are

among the potentially important factors.The availability of potassium and nitrogenin the soil may be limiting for the produc-tion of fleshy fruits and elaiosomes, respec-tively (Hughes et al., 1993). Theimportance of seed deposition in nutrient-enriched ant mounds is debated (e.g. Bondand Stock, 1989) and may, indeed, varyfrom place to place. Seed burial may alsoprotect seeds from fire or from surface-for-aging seed predators (Bennett and Krebs,

1987). See Stiles (Chapter 5, this volume)for further discussion of myrmecochory.External dispersal on vertebrates is

common in riparian zones in arid parts of



12/26

Southern Africa and in disturbed andgrazed habitats (Sorensen, 1986; Milton etal., 1990; Willson et al., 1990a; Fischer etal., 1996). This pattern may reflect, in part,

the level of activity of terrestrial mammalsin such areas (i.e. the availability of disper-sal agents).

If dispersal spectra are constructedwith some measure of abundance (numberof stems, percentage cover) instead ofspecies, quite different trends may appear.For example, ants disperse 29% of theherbaceous species in a North Americandeciduous forest, but ant-dispersed speciesconstitute 5060% of the stems (Handel et

al., 1981). However, such comparisons are,at present, even rarer than those basedon species composition (Frantzen andBouman, 1989; Willson et al., 1990a;Guitin and Snchez, 1992). The construc-tion of comparative dispersal spectra basedon both species counts and abundances formany vegetation types in diverse regionswould be heuristically productive.

When are seeds dispersed?

Less seems to be known about the evolu-tionary ecology of the when of dispersalthan of the why and how. Numerousecological factors may contribute to disper-sal phenology. Ideally, seed maturation anddispersal would be timed to match the sea-sonal availability of good dispersal agents(where required) and the availability ofgood germination conditions (for seeds

lacking dormancy). Constraints on theideal may derive from selection to avoidseed predators or to shift the floweringtime, as well as the length of time requiredfor fruit maturation. In addition, there isenvironmental variation in the time of fruitmaturation and the timing of vector activ-ity in a given area with concomitant differ-ences in rates and quality of dispersal.

A few general patterns in dispersalphenology have been described. Wind-dis-

persed neotropical trees often mature theirseeds during the dry season, when tradewinds are strong and trees are leafless(Foster, 1982; Morellato and Leitao, 1996).

This contrasts with the production offleshy or dry fruits throughout the year (DeLampe et al., 1992). Fleshy-fruited plantsin the north temperate zone commonly

produce mature fruit crops in late summerand autumn, when avian frugivores areusually abundant, but, a little furthersouth, more fruit maturation occurs in win-ter, when flocks of wintering migrant birdsare foraging (Thompson and Willson, 1979;Willson and Thompson, 1982; Herrera,1984a, 1995; Skeate, 1987; Snow andSnow, 1988). In contrast, ant-dispersedplants in central North America generallymature their seeds in early summer, at a

time when avian frugivores are relativelyfew but ants are very active; given that theybloom in early spring, if they held theirfruits until autumn, their low staturewould keep the fruits inconspicuous tobirds, beneath the foliage of other plants,and the maturing fruits would be exposedto predators all summer long (Thompson,1981).

As these patterns of phenology in rela-tion to disperser availability have emerged,

evidence has appeared that they may notbe entirely interpretable as adaptations todispersal. Fruiting patterns in westernEurope tend to match bird phenology at thecommunity level, because abundantspecies with strictly northern distributionsfruit earlier than those with strictly south-ern distributions, but wide-ranging plantspecies show no latitudinal shift in fruitingtimes, as would be expected if their fruit-ing seasons were adapted to disperser phe-

nology (Fuentes, 1991; French, 1992;Willson and Whelan, 1993). Markedannual variation in the seasonal timing offruit maturation of temperate plants indi-cates that events during earlier phases ofreproduction can have a large impact onfruit timing and serves as a reminder thatcompromises may be required between thetiming of flowering and the timing of fruitproduction (see Fenner, 1998).

Eriksson and Ehrln (1998a) have

examined structural and nutritional fea-tures of fleshy fruits of temperate plants inrelation to phenology, finding that somesecondary compounds containing nitrogen



13/26

decrease during the season, i.e. are moreabundant in early- than in late-fruitingspecies. Whether this pattern is adaptiveremains an open question. In contrast, no

phenological trends in lipid or carbohy-drate contents were found. The adaptivevalue of secondary compounds is contro-versial. Cipollini and Levey (1997, 1998)suggested that they probably have a spe-cific function in the fruits, postulating dif-ferent adaptive hypotheses, whereas Ehrlnand Eriksson (1993) and Eriksson andEhrln (1998b) argue that the distributionpatterns of secondary compounds in planttissues do not call for any adaptive expla-

nation of their presence in the fruits.Another pattern that appears when com-paring early- with late-ripening species isin seed number per fruit, which decreasesseasonally (Eriksson and Ehrln, 1998a);assuming trade-offs of numbers and size,this is attributed to developmental con-straints imposed by a demand for a longdevelopmental time in large seeds or, alter-natively, to a better dissemination ofsmall seeds early in the season.

Future investigation must be directedto unravelling both the ecological causes ofseasonal patterns of dispersal and the con-sequences of variation in dispersal phenol-ogy. We know very little about how muchvariation occurs among individuals in thetiming of dispersal and whether this is theresult of genetic or site differences in condi-tions controlling fruit maturation (but see,for instance, Heideman, 1989; see alsoreview in van Schaik et al., 1993). We know

even less about the possibility that differ-ences in timing of seed maturation result inaccompanying differences in seed dispersaland offspring success. From the perspectiveof animal dispersal agents, variation in tim-ing and abundance of fruit production canhave great effects on consumers, resultingin massive population movements (e.g. van-der Wall and Balda, 1977; Leighton andLeighton, 1983; Ostfeld et al., 1996; Selas,1997; Hansson, 1998) or catastrophic mor-

tality (see below), and such major effects onthe consumer community are likely to havereciprocal, lingering effects on seed disper-sal for several years.

For certain, usually long-lived,species, fruits are produced only onceevery few years in an unpredictable pat-tern. This phenomenon, called masting,

has received much attention from ecolo-gists and from evolutionary biologists(Kelly, 1994, and references therein). Threetypes of masting are distinguished: (i) strictmasting, when the population reproducessynchronously and the distribution of seedcrop size among years is bimodal (bam-boos, Strobilanthes); (ii) normal masting,when synchrony is poor and there aremany overlapping cohorts (e.g. imperfectlysynchronized monocarps and many poly-

carps; Fagus, Quercus, Pinus); and (iii)putative masting, where variation in seedoutput is due only to environmental varia-tion, without any evolutionary signifi-cance. A number of hypotheses (reviewedin Kelly, 1994) have been postulated toexplain this phenomenon, the most widelyaccepted being related to economy of scale(i.e. larger reproductive efforts are moreefficient in terms of successful pollinationor seed production and survival) (e.g. Sork,

1993; Tapper, 1996; Kelly and Sullivan,1997; Shibata et al., 1998). The great varia-tion in fruit production from year to yearcan have strong effects not only on therecruitment of the plant population itself(e.g. Schupp, 1990; Jones et al., 1994;Crawley and Long, 1995; Shibata andNakashizuka, 1995; Forget, 1997a), but alsoon the animal populations that consumetheir seeds (Ostfeld et al., 1996, and refer-ences therein; Selas, 1997; Hansson, 1998;

see also Crawley, Chapter 7, this volume).The usefulness and ecological signifi-cance of the masting concept have beenquestioned by Herrera et al. (1998), whoargue that a critical re-examination of pat-terns of annual variability in seed produc-tion is necessary, as most speciesapparently fall along broad continua ofinterannual variability in seed production,with no indication of multimodality (Kelly,1994). In reviewing almost 300 data sets,

Herrera et al. (1998) failed to identify dis-tinct groups of species with contrastinglevels of annual variability in seed output,although most polycarpic plants had alter-



14/26

nating supra-annual schedules, consistingof either high- or low-reproduction years.Annual variability in seed productionappears to be weakly associated with polli-

nation mode (wind versus animal pollina-tion). In contrast, animal-dispersed speciesare less variable than those dispersed byeither inanimate means or animals thatusually act as seed predators (Herrera etal., 1998). These associations certainlydeserve further investigation.

Ecological consequences of dispersal

Population structure

Dispersal of offspring away from the natalsite is one way that genes move through apopulation or into new populations.Movement of genes also occurs at pollina-tion in outcrossing species. Paternallytransmitted genes in outcrossing speciesmove twice in each seed generation, onceduring pollination and again during seeddispersal; maternally transmitted genes in

outcrossers and all genes in self-fertilizedseeds move only once, during seed disper-sal. Thus, in any one seed generation,paternally transmitted genes are likely tomove farther from their source than mater-nally transmitted genes. Lloyd (1982) usedthis observation as a factor favouring theevolution of cosexuality in seed plants. Healso noted that the difference in paternaland maternal gene movement is less inplants with very effective seed dispersal

(e.g. by birds), which might facilitate theevolution of dioecism in bird-dispersedplants.

Gene movement is often limitedwithin a population, such that many plantpopulations consist of genetic neighbour-hoods of more or less related individuals(Levin, 1981; Gibson and Wheelwright,1995). Dispersal by caching animals(Furnier et al., 1987) or by animals that usesleeping sites (Julliot, 1997) or that are

attracted to infected hosts (e.g. mistletoedispersers: Larson, 1996) can lead to clus-ters of related plants within populations,even when the seeds have been carried

long distances. Microdifferentiation oflocal populations can occur on a very smallspatial scale, in response to localized selec-tion and/or very restricted gene flow (e.g.

Schemske, 1984; Turkington and Aarssen,1984; Parker, 1985; Berg and Hamrick,1995; Linhart and Grant, 1996; Nagy, 1997;Nagy and Rice, 1997). Thus, the dispersalpattern of seeds contributes to the geneticstructure of populations and to the poten-tial for both genetic drift and responses tonatural selection. Although some correla-tions of dispersal mode with the degree oflocal differentiation have emerged, thesecorrelations are not very tight, and other

factors must also contribute to observedgenetic structuring (Hamrick and Loveless,1986; Hamrick et al., 1993; and see reviewin Linhart and Grant, 1996; Schnabel et al.,1998).

On the other hand, the at least occa-sional passage of genes out of a localneighbourhood or between conspecificpopulations is important in maintainingthe genetic diversity of the recipient popu-lation and presumably slows the rate of

population differentiation (e.g. Slatkin,1987; Hamrick et al., 1993; Linhart andGrant, 1996). To the extent that outcrossingis advantageous, the most effective out-breeding in populations with neighbour-hood structure will occur when genes passfrom one neighbourhood to another. Thus,in neighbourhood-structured populations,the best outcrossing is rare, by definition.When neighbourhoods reflect ecotypic dif-ferentiation to local conditions, however,

the best outcrossing may occur betweenindividuals that are not too close togetherand yet not too far apart. The concept ofoptimal outcrossing has been controver-sial, and the extent to which selection mayfavour a degree of inbreeding is stillunclear (Shields, 1982; Waddington, 1983;Jarne and Charlesworth, 1993, and refer-ences therein). Seed dispersal patternshave a clear potential to affect the level ofoutcrossing achieved.

The demographic and evolutionaryconsequences of seed dispersal began toreceive attention only a few years ago(Houle, 1992, 1995, 1998; Herrera et al.,



15/26

1994; Horvitz and Schemske, 1994, 1995;Jordano and Herrera, 1995; Schupp, 1995;Schupp and Fuentes, 1995; Shibata andNakashizuka, 1995; Kollman and Schill,

1996; Forget and Sabatier, 1997; Valverdeand Silvertown, 1997; Carlton and Bazzaz,1998; Dalling et al., 1998). The scarcity ofinformation available on the causes andconsequences of spatial patterns of disper-sal at a variety of scales from seeds to newadults is certainly a major gap in ourknowledge on the ecology of seed disper-sal. Most of the studies that consider themultistaged nature of recruitment find nostrong and consistent relationships

between seed and seedling spatial patternsof abundance. The causes of this uncou-pling are mainly attributed to the spatio-temporal variation in the relativeimportance of mortality factors (e.g. preda-tion, pathogens, competition) for seeds andseedlings (Houle, 1992, 1995, 1998).Seedseedling conflicts occur, for instance,in those microhabitats where the probabil-ity of seed survival is low but seedling sur-vival is high (Jordano and Herrera, 1995).

These conflicts probably play an importantrole in structuring many natural systems,as they appear to be rather common(Schupp, 1995). Plant population dynamicsin patchy environments depends not onlyon stage-specific survival and growth indifferent patches, but also on the degree ofdiscordance in patch suitabilities acrossstages (Kollmann and Pirl, 1995; Schupp,1995; Schupp and Fuentes, 1995;Kollmann and Schill, 1996; Aguiar and

Sala, 1997; Forget, 1997b; Russell andSchupp, 1998). Such discordance can havemajor impacts on both the quantity and thespatial patterning of recruits. Furthermore,site suitability may not be independent ofseed arrival (due to density-dependentmortality factors).

In the case of animal-dispersed plants,the influence of frugivorous animalsdepends on the extent of coupling of thedifferent stages in the recruitment process,

which can vary among sites and popula-tions. Factors acting at the end of therecruitment process can potentially screenoff the effects acting at the beginning,

making less predictable the demographicconsequences of seed dispersal (Herrera etal., 1994; Schupp, 1995). Intra- and inter-population variation in the composition of

disperser assemblages visiting a plantspecies has been little documented (Snowand Snow, 1988; Reid, 1989; Guitin et al.,1992; Jordano, 1994; Traveset, 1994;Loiselle and Blake, 1999), despite its poten-tial demographic importance. Differentspecies of frugivores generate characteristicseed shadows, depending on foragingbehaviour, seed retention times, patterns offruit selection and response to the vegeta-tion structure (Herrera, 1995; Rey, 1995). In

order to evaluate the effect of seed vectorson plant demography we need to know thedisperser effectiveness, i.e. the proportionof the seed crop dispersed by a particularspecies (Schupp, 1993), and to examinehow suitable the microsite is where seedsare deposited for germination and estab-lishment. As effectiveness is difficult toestimate in the field, Bustamante andCanals (1995) have proposed a model toestimate it indirectly.

Colonization and plant community structure

Dispersal mode is one factor that affects theability of a plant species to colonize a newarea, especially one at some distance fromthe seed source. Long-distance dispersalcapacity is poorly developed in ballisticand ant-dispersed species and much betterdeveloped in wind- and vertebrate-dis-

persed species. Wind and birds account forthe arrival of most species in an isolatedcloud forest in Colombia (Sugden, 1982).But wind dispersal is insufficient to resultin frequent colonization of extremely dis-tant islands, where many colonists arriveinside avian guts or stuck to the feathers(good numbers also arrive, without specialdevices, in the mud on birds feet, andsome come on ocean currents (Carlquist,1974)). Birds may have been responsible

for post-Pleistocene colonization of habitatislands on mountain-tops in western NorthAmerica by conifers (Wells, 1983). Manycolonists in the island flora of the Great



16/26

Lakes are bird-dispersed, and a similar pro-portion may travel by water (Morton andHogg, 1989). Likewise, a study byWhittaker and Jones (1994) showed that

30% of the flora of Krakatoa island hasarrived and expanded, since the volcaniceruption in 1883, by endozoochory (specif-ically by birds and bats). Thus, the compo-sition of island floras reflects, in part, thedispersal ability of potential colonists. Theinitial colonization of debris avalanchesafter a volcanic eruption on Mount StHelens, Washington, was accomplishedprimarily by wind-dispersed species,although colonization was independent of

distance to the source area (Dale, 1989).The ability of a species to establish a newpopulation at unoccupied sites is a criticalfeature in the maintenance of biologicaldiversity. Current habitat fragmentationcreates barriers to dispersal, however,impeding the natural dispersal of somespecies out of their range in response toglobal climate change (Primack and Miao,1992).

Harper (1977) modified the original

model of van der Plank (1960) and sug-gested that patterns of colonization maydiffer as a function of the shape of the seedshadow. If the slope of the regression ofseed number versus distance (on a loglogscale) is steeper than 2, Harper proposedthat colonization would frequently occurby fronts of invasion, in which phalanxesof colonizers gradually invade new areasrelatively close to the seed source. But, ifthe slope is less steep, colonization may

occur chiefly by far-flung outposts of estab-lishment. There is a weak association ofdispersal mode with steepness of theloglog slope of the seed shadow tail(Portnoy and Willson, 1993). However,many other factors also affect colonizationpatterns (e.g. postdispersal seed predation,germination requirements, conditions fordispersal).

After colonization has occurred, thespatial distribution of the colonizers may

persist for decades or centuries, with reper-cussions for the establishment of subse-quent colonists (e.g. Yarranton andMorrison, 1974). The presence of small

trees and shrubs in an old field or pastureoften increases the deposition of bird- orbat-dispersed seeds beneath them (see ref-erences in Willson, 1991; Debussche and

Isenmann, 1994; Verd and Garca-Fayos,1996, 1998) and decreases the depositionof wind-dispersed seeds (Willson andCrome, 1989). Clusters of individuals offleshy-fruited species often persist evenafter the initial perch tree has died. On theother hand, the early colonizers mayinhibit further colonization if they estab-lish themselves so densely that few otherplants can grow beneath them. Thus, someaspects of the spatial patterning of plant

succession can be related to dispersal. Inthe Mediterranean region, dispersal offleshy-fruited plants by birds appears to beunimportant for plant dynamics in openherbaceous communities and in denseforests, but it is crucial when woodypatches appear with succession in the opencommunities or when grassy patchesappear in the forest (Debussche andIsenmann, 1994; see also Kollmann, 1995).In desert playas of western North America,

what seems to be limiting the initiation ofprimary succession is not seed dispersalbut the low rates of seed entrapment inthese habitats (Fort and Richards, 1998).

Plant dispersal and animal communities

Plant propagules (i.e. the dispersing phaseof the life history) are critical foodresources for a vast number of animal

species. Legions of insect species have spe-cialized to a life of seed predation (bothpre- and postdispersal), and some of theprodigious radiation of insects is associ-ated with these specializations. Whole tax-onomic groups of birds and mammals alsouse seeds as central resources. In turn,these predators have exerted selectionpressures on plants to develop and diver-sify chemical and structural defences.

Fleshy-fruited plants engage in mutu-

alisms with their dispersal agents; theserelationships are quite generalized, veryancient, extremely widespread and extraor-dinarily frequent in certain communities



17/26

(see references in Willson et al., 1989,1990a; Willson, 1993b). Many vertebratepopulations rely on fleshy fruits as food formigration, breeding and winter mainte-

nance. Fruit resources are thought to becrucial in sustaining certain vertebratepopulations in some tropical areas (e.g.Terborgh, 1986; Gautier-Hion andMichaloud, 1989; Julliot, 1997). Heavy useof fruit resources may account for part ofthe great diversity of tropical vertebrates(Karr, 1980) and may have been related tothe radiation of certain tropical bird fami-lies (Snow, 1981). In turn, the biotic disper-sal of seeds seems to have contributed to

some extent to angiosperm diversification(Tiffney and Mazer, 1995, and referencestherein; but see also Ricklefs and Renner,1994).

Non-mutualistic animals also exploitmutualistic interactions and effectivelybecome parasites on the mutualistic system.Both vertebrate and invertebrate consumers(plus fungi and microbes) capitalize onfleshy fruits, without dispersing the seeds.Some insects have become fruit-pulp spe-

cialists to the extent that the radiation ofcertain families (e.g. Tephritidae: Bush,1966) is associated with this kind of para-sitism. The effect of invertebrate parasitismof fruit pulp on potential vertebrate disper-sal agents varies. Although microbial andfungal infestations generally depress effec-tive dispersal (Knoch et al., 1993), infesta-tion by insect larvae can either increaseor decrease fruit consumption by birds,depending on the bird species (Willson

and Whelan, 1990; Traveset et al., 1995,and references therein). In addition, theforaging of frugivores may decrease theabundance and change the distribution ofinsect frugivores, with reciprocal, oftenbeneficial, effects on plant reproduction(Herrera, 1984b, 1989b; but see Traveset,1992, 1993). In sum, dispersal mutualismsbetween plants and animals have hadprodigious and ramifying effects on theanimal community.

Dependence on animals for seed trans-port means that the plants are susceptibleto dispersal failure when their seed vectorsbecome rare or extinct. Disruption of this

mutualism can have serious consequencesfor the maintenance of the plant popula-tions. Loss of native seed-dispersing antsfrom certain habitats in South Africa

means poor seed dispersal and low seedsurvival and may lead to the extinction ofmany rare and endemic plants (Bond andSlingsby, 1984; Bond, 1994). Extinction ofthe dodo on Mauritius has probablyaffected the population structure of the treewhose seeds it dispersed (Temple, 1977;but see also Owadally, 1979; Temple, 1979;Witmer and Cheke, 1991). When a planthas many dispersal agents (as is true formany small-fruited, vertebrate-dispersed

species in North America, for instance), theloss of one species of vector may haveminor consequences for plant populationbiology. However, as both temperate andequatorial forests continue to be deci-mated, the remnant stands are losing manyof their dispersers, with potentially severeconsequences for their continued survival.Evidence is growing from some of theSouth Pacific Islands (Cox et al., 1991) andfrom Chatham Island (Given, 1995) that

disappearance of the main dispersers ofsome plant species deeply alters theirreproductive success. Likewise, the extinc-tion of the Pleistocene megafauna mayhave left many tropical trees with only afew substitute dispersers (e.g. Hallwachs,1986; but see Howe, 1985), although thepopulation consequences of the historicalchange cannot be examined.

Phylogenetic patterns in dispersal

Dispersal modes often differ greatly withintaxonomic units, and a single mode mayarise independently many times (e.g. winddispersal in the legumes: Augspurger,1989). It seems likely that some morpho-logical transformations are more easilymade than others. For instance, a plume forwind dispersal may be converted to a hookfor dispersal on vertebrate exteriors (e.g.

Anemone, sensu lato) or vice versa. Theloss of a wing contributed to a change fromwind to bird dispersal in Pinus but wasaccompanied by changes in cone structure



18/26

and seed size as well; nevertheless, theshift between these two modes of dispersalhas occurred several times (see Strauss andDoerksen, 1990; Tomback and Linhart,

1990; Benkman, 1995). A change from birdto ant dispersal in Acacia necessitatedchiefly a shift in the size and colour of thefood body attached to the seed (ODowdand Gill, 1986; see Fig. 4.5), but a similarshift in Trillium occasioned a seeminglymore complex change, from a fleshy fruitenclosing several unappendaged seeds to adry fruit with elaiosome-bearing seeds(Berg, 1958). We need phylogenetic analy-ses of whole families or genera with

respect to dispersal modes to determine: (i)how often dispersal mode has changedwithin a taxon; (ii) what the directions ofthe change are; (iii) what kinds of changesare most common; and (iv) for wide-rang-ing taxa, how the biogeographical historyof different regions influences the evolu-tion of diaspore traits. The answers to suchquestions, in conjunction with ecologicaldata, will contribute to our understandingof community dispersal spectra, patterns of

selection on dispersal devices and otheraspects of population and community biol-ogy related to dispersal. A synthesis ofmodern systematics and evolutionary ecol-ogy is a powerful tool in elucidating ques-tions about diaspore adaptation andphylogenetic radiation (e.g. Wanntorp etal., 1990; Bremer and Eriksson, 1992;

Ricklefs and Renner, 1994; Tiffney andMazer, 1995).

Conclusion

The study of the dispersal of plants hasadvanced relatively fast in the last decade,as essential elements of the evolutionaryand ecological causes and consequences ofdispersal have been examined. The linkbetween seed dispersal and its demo-graphic and genetic consequences is onemajor gap that still needs to be filled,although some recent studies are already

paying attention to it. Much remains to bediscovered yet in terms of geographical andhabitat patterns, as well as the dynamics ofcolonization, population differentiationand plant/animal interactions. Dispersalecology is a rapidly developing field thatstill offers a wealth of investigative oppor-tunity at levels ranging from good naturalhistory to sophisticated modelling and con-ceptual synthesis.

Acknowledgements

C.K. Augspurger and J.N. Thompson gra-ciously provided constructive commentson the manuscript written for the first edi-tion of this book.


ReferencesAguiar, M.R. and Sala, O.E. (1997) Seed distribution constrains the dynamics of the Patagonian

steppe. Ecology78, 93100.Alexander, H.M. and Holt, R.D. (1998) The interaction between plant competition and disease.

Perspectives in Plant Ecology, Evolution and Systematics 1, 206220.Armstrong, D.P. and Westoby, M. (1993) Seedlings from large seeds tolerate defoliation better: a test

using phylogenetically independent contrasts. Ecology74, 10921100.Aronne, C. and Wilcock, C.C. (1994) First evidence of myrmecochory in fleshyfruited shrubs of the

Mediterranean region. New Phytologist127, 781788.Augspurger, C.K. (1983a) Seed dispersal of the tropical tree, Platypodium elegans, and the escape of

its seedlings from fungal pathogens.Journal of Ecology71, 759771.Augspurger, C.K. (1983b) Offspring recruitment around tropical trees: changes in cohort distance

with time. Oikos 40, 189196.Augspurger, C.K. (1984) Seedling survival of tropical tree species: interactions of dispersal distance,

light-gaps, and pathogens. Ecology65, 17051712.Augspurger, C.K. (1986) Morphology and dispersal potential of wind-dispersed diaspores of neotrop-

ical trees. American Journal of Botany73, 353363.


19/26

Augspurger, C.K. (1989) Morphology and aerodynamics of wind-dispersed legumes. Advances inLegume Biology29, 451466.

Augspurger, C.K. and Franson, S.E. (1987) Wind dispersal of artificial fruits varying in mass, area,and morphology. Ecology68, 2742.

Augspurger, C.K. and Kelly, C.K. (1984) Pathogen mortality of tropical tree seedlings: experimental

studies of the effects of dispersal distance, seedling density, and light conditions. Oecologia 61,211217.Augspurger, C.K. and Kitajima, K. (1992) Experimental studies of seedling recruitment from contrast-

ing seed distributions. Ecology73, 12701284.Beattie (1985) The Evolutionary Ecology of AntPlant Mutualisms. Cambridge University Press,

Cambridge.Benkman, C.W. (1995) Wind dispersal capacity of pine seeds and the evolution of different seed dis-

persal modes in pines. Oikos 73, 221224.Benkman, C.W., Balda, R.P. and Smith, C.C. (1984) Adaptations for seed dispersal and the compro-

mises due to seed predation in limber pine. Ecology65, 632642.Bennett, A. and Krebs, J. (1987) Seed dispersal by ants. Trends in Ecology and Evolution 2, 291292.Berg, E.E. and Hamrick, J.L. (1995) Fine-scale genetic structure of a turkey oak forest. Evolution 49,

110120.

Berg, R.Y. (1958) Seed dispersal, morphology, and phylogeny of Trillium. Skrifter utgitt av det NorskeVidenskaps-Akademi i Oslo 1. Matematisk-Naturuidenskapelig Klasse 1, 136.Bond, W.J. (1994) Do mutualisms matter? Assessing the impact of pollinator and disperser disruption

on plant extinction. Philosophical Transactions of the Royal Society of London 344, 8390.Bond, W.J. and Slingsby, P. (1984) Collapse of an ant-plant mutualism: the Argentine ant

(Iridomyrmex humilis) and myrmecochorous Proteaceae. Ecology65, 10311037.Bond, W.J. and Stock, W.D. (1989) The costs of leaving home: ants disperse myrmecochorous seeds to

low nutrient sites. Oecologia 81, 412417.Bremer, B. and Eriksson, O. (1992) Evolution of fruit characteristics and dispersal modes in the tropi-

cal family Rubiaceae. Biological Journal of the Linnean Society47, 7995.Bullock, S.H. and Primack, R.B. (1977) Comparative experimental study of seed dispersal on ani-

mals. Ecology58, 681686.Burdon, J.J. (1987) Diseases and Plant Population Biology. Cambridge University Press, Cambridge.Bush, G.L. (1966) The taxonomy, cytology, and evolution of the genus Rhagoletis in North America

north of Mexico. Bulletin of the Museum of Comparative Zoology134, 431562.Bustamante, R.O. and Canals, L.M. (1995) Dispersal quality in plants: how to measure efficiency and

effectiveness of a seed disperser. Oikos 73, 133136.Cain, M.L., Damman, H. and Muir, A. (1998) Seed dispersal and the Holocene migration of woodland

herbs. Ecological Monographs 68, 325347.Carlquist, S. (1974) Island Biology. Columbia University Press, New York.Carlquist, S. and Pauly, Q. (1985) Experimental studies on epizoochorous dispersal in California

plants. Aliso 11, 167177.Carlton, G.C. and Bazzaz, F.A. (1998) Regeneration of three sympatric birch species on experimental

hurricane blowdown microsites. Ecological Monographs 68, 99120.Chambers, J.C. and MacMahon, J.A. (1994) A day in the life of a seed: movements and fates of seeds

and their implications for natural and managed systems. Annual Review of Ecology andSystematics 25, 263292.

Cipollini, M.L. and Levey, D.J. (1997) Secondary metabolites of fleshy vertebrate-dispersed fruits:adaptive hypotheses and implications for seed dispersal. American Naturalist150, 346372.Cipollini, M.L. and Levey, D.J. (1998) Secondary metabolites as traits of ripe fleshy fruits: a response

to Eriksson and Ehrln. American Naturalist152, 908911.Clifford, H.T. and Monteith, G.B. (1989) A three phase seed dispersal mechanism in Australian qui-

nine bush (Petalostigma pubescens Domin). Biotropica 21, 284286.Coe, M. and Coe, C. (1987) Large herbivores, acacia trees and bruchid beetles. South African Journal

of Science 83, 624635.Cohen, D. and Motro, U. (1989) More on optimal rates of dispersal: taking into account the cost of the

dispersal mechanism. American Naturalist134, 659663.Collins, S.L. and Uno, C.E. (1985) Seed predation, seed dispersal, and disturbances in grassland: a

comment. American Naturalist125, 866872.Comins, H.N., Hamilton, W.D. and May, R.M. (1980) Evolutionarily stable dispersal strategies.

Journal of Theoretical Biology 82, 205230.Cox, P.A., Elmquist, T., Pierson, E.D. and Rainey, W.E. (1991) Flying foxes as strong interactors in

South Pacific island ecosystems: a conservation hypothesis. Conservation Biology5, 448454.Crawley, M.J. and Long, C.R. (1995) Alternate bearing, predator satiation and seedling recruitment in

Quercus roburL.Journal of Ecology83, 683696.



20/26

Dale, V.H. (1989) Wind dispersed seeds and plant recovery on the Mount St. Helens debrisavalanche. Canadian Journal of Botany67, 14341441.

Dalling, J.W., Hubbell, S.P. and Silvera, K. (1998) Seed dispersal, seedling establishment and gap par-titioning among tropical pioneer trees.Journal of Ecology86, 674689.

Davidson, D.W. and Morton, S.R. (1984) Dispersal adaptations of some Acacia species in the

Australian arid zone. Ecology65, 10381051.Debussche, M. and Isenmann, P. (1994) Bird-dispersed seed rain and seedling establishment inpatchy Mediterranean vegetation. Oikos 69, 414426.

Debussche, M. and Lepart, J. (1992) Establishment of woody plants in Mediterranean old fields:opportunity in space and time. Landscape Ecology6, 133145.

De Lampe, M.G., Bergeron, Y., McNeil, R. and Leduc, A. (1992) Seasonal flowering and fruiting pat-terns in tropical semiarid vegetation of north-eastern Venezuela. Biotropica 24, 6476.

Dinerstein, E. (1989) The foliage-as-fruit hypothesis and the feeding behavior of south Asian ungu-lates. Biotropica 21, 214218.

Donohue, K. (1997) Seed dispersal in Cakile edentula var. lacustris: decoupling the fitness effects ofdensity and distance from the home site. Oecologia 110, 520527.

Donohue, K. (1998) Maternal determinants of seed dispersal in Cakile edentula: fruit, plant, and sitetraits. Ecology79, 27712788.

Ehrln, J. and Eriksson, O. (1993) Toxicity in fleshy fruits a non-adaptive trait? Oikos 66, 107113.

Ellner, S. and Schmida, A. (1981) Why are adaptations for long-range seed dispersal rare in desertplants? Oecologia 51, 133144.

Ellstrand, N.C. and Antonovics, J. (1985) Experimental studies of the evolutionary significance ofsexual reproduction. II. A test of the density-dependent selection hypothesis. Evolution 39,657666.

Eriksson, O. (1992) Evolution of seed dispersal and recruitment in clonal plants. Oikos 63, 439448.Eriksson, O. and Ehrln, J. (1998a) Phenological adaptations in fleshy vertebrate-dispersed fruits of

temperate plants. Oikos 82, 617621.Eriksson, O. and Ehrln, J. (1998b) Secondary metabolites in fleshy fruits: are adaptive explanations

needed? American Naturalist152, 905907.Fenner, M. (1998) The phenology of growth and reproduction in plants. Perspectives in Plant

Ecology, Evolution and Systematics 1, 7891.Fischer, S.F., Poschlod, P. and Beinlich, B. (1996) Experimental studies on the dispersal of plants and

animals on sheep in calcareous grasslands.Journal of Applied Ecology33, 12061222.Fleming, T.H., Breitwisch, R. and Whitesides, G.H. (1987) Patterns of vertebrate frugivore diversity.Annual Review of Ecology and Systematics 18, 91109.

Forget, P.M. (1997a) Ten-year seedling dynamics in Vouacapoua americana in French Guiana: ahypothesis. Biotropica 29, 124126.

Forget, P.M. (1997b) Effect of microhabitat on seed fate and seedling performance in two rodent-dispersed tree species in rainforest in French Guiana.Journal of Ecology85, 693703.

Forget, P.M. and Sabatier, D. (1997) Dynamics of the seedling shadow of a frugivore-dispersed treespecies in French Guiana.Journal of Tropical Ecology13, 767773.

Fort, K.P. and Richards, J.H. (1998) Does seed dispersal limit initiation of primary succession indesert playas? American Journal of Botany85, 17221731.

Foster, R.B. (1982) The seasonal rhythm of fruitfall on Barro Colorado Island. In: Leigh, E.G., Rand,A.S. and Windsor, D.M. (eds) The Ecology of a Tropical Forest. Smithsonian, Washington, DC,pp. 151172.

Foster, S.A. and Janson, C.H. (1985) The relationship between seed size and establishment condi-tions in tropical woody plants. Ecology66, 773780.Frantzen, N.M.L.H.F. and Bouman, F. (1989) Dispersal and growth form patterns of some zonal

paramo vegetation types. Acta Botanica Neerlandica 38, 449465.French, K. (1992) Phenology of fleshy fruits in a wet sclerophyll forest in southeastern Australia: are

birds an important influence? Oecologia 90, 366373.Fuentes, M. (1991) Latitudinal and elevational variation in fruiting phenology among western

European bird-dispersed plants. Ecography15, 177183.Furnier, G.R., Knowles, P., Clyde, M.A. and Dancik, B.P. (1987) Effects of avian seed dispersal on the

genetic structure of whitebark pine populations. Evolution 41, 607612.Gautier-Hion, A. and Michaloud, G. (1989) Are figs always keystone resources for tropical frugivo-

rous vertebrates? A test in Gabon. Ecology70, 18261833.Geritz, S.A.H., de Jong, T.J. and Klinkhamer, P.G.L. (1984) The efficacy of dispersal in relation to safe

site area and seed production. Oecologia 62, 219221.Ghiselin, M. (1974) The Economy of Nature and the Evolution of Sex. University of California Press,

Berkeley, California.Gibson, J.P. and Wheelwright, N.T. (1995) Genetic structure in a population of a tropical tree Ocotea

tenera (Lauraceae): influence of avian seed dispersal. Oecologia 103, 4954.



21/26

Given, D.R. (1995) Biological diversity and the maintenance of mutualisms. In: Vitousek, P.M.,Loope, L.L. and Adsersen, H. (eds) Islands: Biological Diversity and Ecosystem Function.Springer Verlag, Berlin, pp. 149162.

Gorb, S.N. and Gorb, E.V. (1995) Removal rates of seeds of five myrmecochorous plants by the antFormica polyctena (Hymenoptera: Formicidae). Oikos 73, 367374.

Gorchov, D.L., Cornejo, F., Ascorra, C. and Jaramillo, M. (1993) The role of seed dispersal in the nat-ural regeneration of rainforest after strip-cutting in the Peruvian Amazon. Vegetatio 107/108,339349.

Green, D.S. (1983) The efficacy of dispersal in relation to safe site density. Oecologia 56, 356358.Greene, D.F. and Johnson, E.A. (1996) Wind dispersal of seeds from a forest into a clearing. Ecology

77, 595609.Guitin, J. and Snchez, J.M. (1992) Seed dispersal spectra of plant communities in the Iberian

Peninsula. Vegetatio 98, 157164.Guitin, J., Fuentes, M., Bermejo, T. and Lpez, B. (1992) Spatial variation in the interactions

between Prunus mahaleb and frugivorous birds. Oikos 63, 125130.Gunther, R.W. and Lanza, J. (1989) Variation in attractiveness of Trillium diaspores to a seed-dispers-

ing ant. American Midland Naturalist122, 321328.Hallwachs, W. (1986) Agoutis (Dasyprocta punctata): the inheritors of guapinol (Hymenaea

courbaril: Leguminosae). In: Estrada, A. and Fleming, T.H. (eds) Frugivores and Seed Dispersal.

Junk, Dordrecht, pp. 285304.Hamilton, W.D. and May, R.M. (1977) Dispersal in stable habitats. Nature 269, 578581.Hammond, D.S. and Brown, V.K. (1995) Seed size of woody plants in relation to disturbance, disper-

sal, soil type in wet neotropical forests. Ecology76, 25442561.Hamrick, J.L. and Loveless, M.D. (1986) The influence of seed dispersal mechanisms on the genetic

structure of plant populations. In: Estrada, A. and Fleming, T.H. (eds) Frugivores and SeedDispersal. Junk, Dordrecht, pp. 211223.

Hamrick, J.L., Murawski, D.A. and Nason, J.D. (1993) The influence of seed dispersal mechanisms onthe genetic structure of tropical tree populations. Vegetatio 107/108, 281297.

Handel, S.N., Fisch, S.B. and Schatz, G.E. (1981) Ants disperse a majority of herbs in a mesic forestcommunity in New York State. Bulletin of the Torrey Botanical Club 108, 430437.

Hansson, L. (1998) Mast seeding and population dynamics of rodents: one factor is not enough.Oikos 82, 591594.

Harper, J.L. (1977) Population Biology of Plants. Academic Press, London.Heideman, P.D. (1989) Temporal and spatial variation in the phenology of flowering and fruiting in atropical rainforest.Journal of Ecology77, 10591079.

Herrera, C.M. (1984a) A study of avian frugivores, bird-dispersed plants, and their interaction inMediterranean scrublands. Ecological Monographs 54, 123.

Herrera, C.M. (1984b) Avian interference with insect frugivory: an exploration into theplantbirdfruit pest evolutionary triad. Oikos 42, 203210.

Herrera, C.M. (1984c) Seed dispersal and fitness determinants in wild rose: combined effects ofhawthorn, birds, mice and browsing ungulates. Oecologia 63, 386393.

Herrera, C.M. (1987) Vertebrate-dispersed plants of the Iberian peninsula: a study of fruit characteris-tics. Ecological Monographs 57, 305331.

Herrera, C.M. (1988) The fruiting ecology of Osyris quadripartita: individual variation and evolution-ary potential. Ecology69, 233249.

Herrera, C.M. (1989a) Vertebrate frugivores and their interaction with invertebrate fruit predators:

supporting evidence from a Costa Rican dry forest. Oikos 54, 185188.Herrera, C.M. (1989b) Frugivory and seed dispersal by carnivorous mammals and associated fruitcharacteristics in undisturbed Mediterranean habitats. Oikos 55, 250262.

Herrera, C.M. (1995) Plantvertebrate seed dispersal systems in the Mediterranean: ecological, evolu-tionary, and historical determinants. Annual Review of Ecology and Systematics 26, 705727.

Herrera, C.M., Jordano, P., LpezSoria, L. and Amat, J.A. (1994) Dispersal ecology of a mast-fruiting,bird-dispersed tree: bridging frugivory activity and seedling establishment. EcologicalMonographs 64, 315344.

Herrera, C.M., Jordano, P., Guitin, J. and Traveset, A. (1998) Annual variability in seed productionby woody plants and the masting concept: reassessment of principles and relationship to pollin-ation and seed dispersal. American Naturalist152, 576594.

Hoppes, W.G. (1988) Seedfall pattern of several species of bird-dispersed plants in an Illinois wood-land. Ecology69, 320329.

Horvitz, C.C. and Le Corff, J. (1993) Spatial scale and dispersion pattern of ant- and bird-dispersedherbs in two tropical lowland rainforests. Vegetatio 107/108, 351362.

Horvitz, C.C. and Schemske, D.W. (1986). Seed dispersal and environmental heterogeneity in aneotropical herb: a model of population and patch dynamics. In: Estrada, A. and Fleming, T.H.(eds) Frugivores and Seed Dispersal. Junk, Dordrecht, pp. 169186.



22/26

Horvitz, C.C. and Schemske, D.W. (1994) Effects of dispersers, gaps, and predators on dormancy andseedling emergence in a tropical herb. Ecology75, 19491958.

Horvitz, C.C. and Schemske, D.W. (1995) Spatiotemporal variation in demographic transitions of atropical understory herb: projection matrix analysis. Ecological Monographs 65, 155192.

Houle, G. (1992) Spatial relationship between seed and seedling abundance and mortality in a decid-

uous forest of north-eastern North America.Journal of Ecology80, 99108.Houle, G. (1995) Seed dispersal and seedling recruitment: the missing link(s). Ecoscience 2,238244.

Houle, G. (1998) Seed dispersal and seedling recruiment of Betula alleghaniensis: spatial inconsis-tency in time. Ecology79, 807818.

Howe, H.F. (1985) Gomphothere fruits: a critique. American Naturalist125, 853865.Howe, H.F. (1993) Aspects of variation in a neotropical seed dispersal system. Vegetatio 107/108,

149162.Howe, H.F. and Smallwood, J. (1982) Ecology of seed dispersal. Annual Review of Ecology and

Systematics 13, 201228.Howe, H.F. and vande Kerckhove, G.A. (1980) Nutmeg dispersal by tropical birds. Science 210,

925927.Hughes, L., Westoby, M. and Johnson, A.D. (1993) Nutrient costs of vertebrate- and ant-dispersed

fruits. Functional Ecology7, 5462.Hulme, P.E. (1998) Post-dispersal seed predation: consequences for plant demography and evolution.Perspectives in Plant Ecology, Evolution and Systematics 1, 3246.

Janzen, D.H. (1971) Seed predation by animals. Annual Review of Ecology and Systematics 2,465492.

Janzen, D.H. (1984) Dispersal of small seeds by big herbivores: foliage is the fruit. AmericanNaturalist123, 338353.

Jarne, P. and Charlesworth, D. (1993) The evolution of the selfing rate in functionally hermaphroditeplants and animals. Annual Review of Ecology and Systematics 24, 441466.

Johnson, W.C. (1988) Estimating dispersibility of Acer, Fraxinus and Tilia in fragmented landscapesfrom patterns of seedling establishment. Landscape Ecology1, 175187.

Johnson, W.C. and Webb, T. (1989) The role of blue jays (Cyanocitta cristata L.) in the postglacial dis-persal of fagaceous trees in eastern North America.Journal of Biogeography16, 561571.

Jones, R.H., Sharitz, R.R., Dixon, P.M. and Segal, D.S. (1994) Woody plant regeneration in four flood-

plain forests. Ecological Monographs 64, 345367.Jordano, P. (1987) Avian fruit removal: effects of fruit variation, crop size, and insect damage. Ecology

68, 17111723.Jordano, P. (1993) Geographical ecology and variation of plantseed disperser interactions: southern

Spanish junipers and frugivorous thrushes. Vegetatio 107/108, 85104.Jordano, P. (1994) Spatial and temporal variation in the avianfrugivore assemblage of Prunus

mahaleb: patterns and consequences. Oikos 71, 479491.Jordano, P. (1995a) Frugivore-mediated selection on fruit and seed size: birds and St. Lucies cherries,

Prunus mahaleb. Ecology76, 26272639.Jordano, P. (1995b) Angiosperm fleshy fruits and seed dispersers: a comparative analysis of adapta-

tion and constraints in plantanimal interactions. American Naturalist145, 163191.Jordano, P. and Herrera, C.M. (1995) Shuffling the offspring: uncoupling and spatial discordance of

multiple stages in vertebrate seed dispersal. Ecoscience 2, 230237.

Julliot, C. (1997) Impact of seed dispersal by red howler monkeys Alouatta seniculus on the seedlingpopulation in the understorey of tropical rainforest.Journal of Ecology85, 431440.Karr, J.R. (1976) Within- and between-habitat avian diversity in African and neotropical lowland

habitats. Ecological Monographs 46, 457481.Karr, J.R. (1980) Geographical variation in the avifaunas of tropical forest undergrowth. Auk 97,

283298.Kelley, S.E. (1989) Experimental studies of the evolutionary significance of sexual reproduction. VI.

A greenhouse test of the sib competition hypothesis. Evolution 43, 10661074.Kelley, S.E., Antonovics, J. and Schmitt, J. (1988) A test of the short-term advantage of sexual repro-

duction. Nature 331, 714716.Kelly, C.K. (1995) Seed size in tropical trees: a comparative study of factors affecting seed size in

Peruvian angiosperms. Oecologia 102, 377388.Kelly, D. (1994) The evolutionary ecology of mast seeding. Trends in Ecology and Evolution 9,

465470.Kelly, D. and Sullivan, J.J. (1997) Quantifying the benefits of mast seeding on predator satiation and

wind pollination in Chionochloa pallens (Poaceae). Oikos 78, 143150.Kerner, A. (1898) The Natural History of Plants: Their Form, Growth, Reproduction and Distribution

(English translation by F.W. Oliver), 2 vols. Holt, New York.



23/26

Knoch, T.R., Faeth, S.H. and Arnott, D.L. (1993) Endophytic fungi alter foraging and dispersal bydesert seed-harvesting ants. Oecologia 95, 470473.

Kollmann, J. (1995) Regeneration window for fleshy-fruited plants during scrub development onabandoned grassland. Ecoscience 2, 213222.

Kollmann, J. and Pirl, M. (1995) Spatial pattern of seed rain of fleshy-fruited plants in a scrubland

grassland transition. Acta Oecologica 16, 313329.Kollmann, J. and Schill, H.P. (1996) Spatial patterns of dispersal, seed predation and germinationduring colonization of abandoned grassland by Quercus petraea and Corylus avellana. Vegetatio125, 193205.

Larson, D.L. (1996) Seed dispersal by specialist versus generalist foragers: the plants perspective.Oikos 76, 113120.

Leck, M.A., Parker, V.T. and Simpson, R.L. (eds) (1989) Ecology of Soil Seed Banks. Academic Press,London.

Leighton, M. and Leighton, D.R. (1983) Vertebrate responses to fruiting seasonality within a Borneanrainforest. In: Sutton, S.L., Whitmore, T.C. and Chadwick, A.C. (eds) Tropical Rain ForestEcology and Management. Blackwell Scientific Publications, Oxford, pp. 181196.

Leishman, M.R. and Westoby, M. (1994a) Hypotheses on seed size: tests using the semiarid flora ofwestern New South Wales, Australia. American Naturalist143, 890906.

Leishman, M.R. and Westoby, M. (1994b) The role of large seed size in shaded conditions: experi-mental evidence. Functional Ecology8, 205214.Levin, D.A. (1981) Dispersal versus gene flow in plants. Annals of the Missouri Botanical Garden 68,

233253.Linhart, Y.B. and Grant, M.C. (1996) Evolutionary significance of local genetic differentiation in

plants. Annual Review of Ecology and Systematics 27, 237277.Lloyd, D.G. (1982) Selection of combined versus separate sexes in seed plants. American Naturalist

120, 571585.Loiselle, B.A. and Blake, J.G. (1999) Dispersal of melastome seeds by fruit-eating birds of tropical for-

est understory. Ecology80, 330336.Loiselle, B.A., Ribbens, E. and Vargas, O. (1996) Spatial and temporal variation of seed rain in a trop-

ical lowland wet forest. Biotropica 28, 8295.Lokesha, R., Hegde, S.G., Shaanker, R.U. and Ganeshaiah, K.N. (1992) Dispersal mode as a selective

force in shaping the chemical composition of seeds. American Naturalist140, 520525.

McCall, C., Mitchell-Olds, T. and Waller, D.M. (1989) Fitness consequences of outcrossing inImpatiens capensis: tests of the frequency-dependent and sib-competititon models. Evolution43, 10751084.

McCanny, S.J. and Cavers, P.B. (1987) The escape hypothesis: a test involving a temperate annualgrass. Oikos 49, 6776.

McClanahan, T.R. and Wolfe, R.W. (1993) Accelerating forest succession in a fragmented landscape.Conservation Biology7, 279288.

Manasse, R.S. and Howe, H.F. (1983) Competition for dispersal agents among tropical trees: influ-ences of neighbors. Oecologia 59, 185190.

Mark, S. and Olesen, J.M. (1996) Importance of elaiosome size to removal of ant-dispersed seeds.Oecologia 107, 95101.

Matlack, G.R. (1987) Diaspore size, shape, and fall behavior in wind-dispersed plant species.American Journal of Botany74, 11501160.

Milton, S.J., Siegfried, W.R. and Dean, W.R.J. (1990) The distribution of epizoochoric plant spectres: aclue to the prehistoric use of arid Karoo rangelands by large herbivores.Journal of Biogeography17, 2534.

Morellato, P.C. and Leitao, H.F. (1996) Reproductive phenology of climbers in a so

willson_traveset 2000 - the ecology of seed dispersal.pdf

Documents