The University of Chicago
Seed Dispersal by Animals: Contrasts with Pollen Dispersal, Problems of Terminology, and
Constraints on Coevolution
Author(s): Nathaniel T. Wheelwright and Gordon H. Orians
Source: The American Naturalist, Vol. 119, No. 3 (Mar., 1982), pp. 402-413
Published by: The University of Chicago Press for The American Society of Naturalists
Stable URL: http://www.jstor.org/stable/2460937 .
Accessed: 24/12/2013 11:41
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].
.
The University of Chicago Press, The American Society of Naturalists, The University of Chicago are
collaborating with JSTOR to digitize, preserve and extend access to The American Naturalist.
http://www.jstor.org
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
March 1982
The AmericanNaturalist
Vol. 119, No. 3
SEED DISPERSAL BY ANIMALS: CONTRASTS WITH POLLEN
DISPERSAL, PROBLEMS OF TERMINOLOGY, AND
CONSTRAINTS ON COEVOLUTION
NATHANIEL
T. WHEELWRIGHT
AND GORDON
H. ORIANS
Departmentof Zoology, Universityof Washington,Seattle, Washington98195
SubmittedFebruary27, 1981; Accepted September28, 1981
Recent thinkingabout seed dispersal and the coevolutionof fruiting
plants and
frugivoreshas been stronglyinfluencedby ideas developed throughthe studyof
pollen dispersal by animals. Analogies are commonlydrawn between the two
systems (Howe 1977; Van der Pijl 1972; Howe and Primack 1975; Howe and
Vande Kerckhove 1979; Howe 1979; Howe and De Steven 1979; Howe 1980).
Both systemsrepresentcoevolved relationshipsof mutualbenefitto plants and
animalsthatinvolvethe disseminationof pollen or seeds and the provisionby the
plant of some incentivesuch as nectar or fruit.The evolutionof highdegrees of
mutualdependence and morphologicaland behavioral specialization,evident in
some pollinationsystems,is assumed also to be likelyin certain seed dispersal
systems (Howe and Primack 1975; McKey 1975; Howe 1977; Howe and Estabrook 1977). A rich array of evocative metaphorscurrentlyused in studies of
interactions("loyalty," "theft," "reliability," "faithfulness,"
plant-frugivore
"reward," and so on; see McKey 1975 and Howe and Estabrook 1977) reflects
conceptual biases apparentlyderived in part fromimplicitand explicitanalogies
withpollen dispersal. Althoughsuch termsare sometimesuseful,theymay cloud
interactions.
the true natureof fruit-frugivore
Pollen and seed dispersal are fundamentallydissimilar,particularlyin most
characteristicslikelyto influencechoice of diet and foragingitinerariesby animal
vectors,and patternsof productionof pollen and fruitby plants. The purpose of
this paper is threefold:(1) to outline clearly the relevant differencesbetween
pollen and seen dispersal,(2) to indicatesome of the problemsof usingterminology borrowedfromstudiesof pollen dispersalto describe seed dispersalsystems,
and (3) to consider the factorsthat may explain the virtualabsence of obligate
mutualismsin seed dispersal systems(Howe and Vande Kerckhove 1979). Because of our greaterfamiliaritywith avian frugivoresand their importancein
theoreticaland empiricalstudies of seed dispersal, most of our referencesdeal
principallywithbirds. Futureworkon seed dispersalby mammals,ants, reptiles,
and fishmay offernew insights.
Am. Nat. 1982. Vol. 119, pp. 402-413.
All rightsreserved.
? 1982 by The Universityof Chicago. 0003-0147/82/1903-0016$02.00.
402
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
COEVOLUTION
THE
THEORY
OF FRUITING PLANTS AND FRUGIVORES
OF FRUIT-FRUGIVORE
403
COEVOLUTION
Snow (1971) firstconsidered the consequences of evolutionaryinteractions
thatdispersetheirseeds and recognized
plantsand the frugivores
betweenfruiting
theunique attributesoffruitas a food resource.Relativeto other"prey," fruithas
evolved to be generallyaccessible, conspicuous, and easily digestiblein orderto
attract seed dispersers (Snow 1971). Following Snow's contribution,McKey
coevolutionto explain the varietyof
(1975) proposed a model of plant-frugivore
strategiesplantsuse to dispersetheirseeds in whichhe introducedthe concept of
dispersalquality.At one extreme,some plantsproduce manysmall, nutritionally
poor fruitsthatattracta wide varietyof "poor quality" dispersers.Alternatively,
superiorfruitswhose seeds
othersproducea smallernumberoflarge,nutritionally
are dispersedby a limitednumberof species delivering"high dispersalquality."
Qualityrefersto seed size carriedand the probabilitythata seed removedfroma
plantis deposited in viable conditionin a site suitableforgerminationand establishment(McKey 1975). A numberof recent studies, both theoretical(Morton
1973; Howe and Estabrook 1977; Howe 1979; Thompson and Willson 1979;
Fleming 1979) and empirical(Smith 1975; Howe 1977; McDiarmid et al. 1977;
Thompson and Willson 1978; Howe and De Steven 1979; Howe and Vande
Kerckhove 1979; Howe 1980; Howe 1981; Snow 1981) have refinedhypotheses,
evaluated some predictions,and providedobservationson fruitremovalrates and
trees.
behavior of frugivoresat fruiting
Centralto currenttheoryis the propositionthatfrugivoresdifferin behavioral
and ecological traitsthataffecttheirsuitabilitiesas seed dispersers.If so, under
certaincircumstances(e.g., predictableecological requirementsforseeds) natural
selectionis expected to lead to the mutualinterdependenceand tightassociation
of individual plant species with "one particularlyeffectivedisperser" (Howe
1977),or a small numberof obligatefruigivores(McKey 1975; Howe and Primack
1975; Howe and Estabrook 1977). However, empirical studies have generally
failedto discover eitherhighlycoevolved frugivoresor plantsadapted to exclude
all but a single disperser, as in some (though much fewer than generallyapcounterpartsto
preciated) pollen dispersal systems.The lack of plant-frugivore
specialized pollinationsystems,such as orchids (Dressler 1968; Dodson 1975) or
Dalechampia (Armbrusterand Webster 1979) and euglossine bees, or yucca
species and the mothTegeticula (Powell and Mackie 1966), is striking(see also
Galil 1977; Sazima and Sazima 1978; Thien 1969). The majorityof studiesof birds,
at least, have documentedinstead a tremendousdiversityof frugivoresat fruiting
trees (Eisenmann 1961; Land 1963; Willis 1966; Diamond and Terborgh 1967;
Jenkins1969; Leck 1969; Haverschmidt1971; Cruz 1974; Wheelwrightet al., in
prep.).
Not onlydo manybirdspecies feed on thefruitofa singleplantspecies, but also
they may represent up to 13 families (McDiarmid et al. 1977). Even the
"specialized," high reward fruitsof some Lauraceae (McKey 1975) may have
theirseeds dispersedby morethan 17 species of birdsfromat least eightfamilies
(Wheelwright,in prep.). Morphologicallyas well as taxonomically,birdsfeeding
et
on thefruitsofthe same treemaybe quite different
(Ricklefs1977; Wheelwright
al., in prep.).
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
THE AMERICAN NATURALIST
404
The discoveryofassemblages of manyspecies ofdiversefrugivores,
ratherthan
limitedassociations, spawned a deprecatingterminologyto referto species presumed to be poor dispersers: "opportunists,exploiters" (McKey 1975), "fruit
thieves" (Howe and Estabrook 1977), and "parasites" (Janzen 1977). Such
metaphorsreflectthe expectations of the theorymore than they do the actual
effectsof the frugivores;data do not yetexistto evaluate accuratelythe effectiveness of dispersalby different
species. Despite thefactthatmanystudiespurportto
deal with seed dispersal, most reportonly observationson who removes fruit,
how regulartheyare as visitors,and, in some cases, whetheror not theydestroy
seeds. The above terminologyobscures the fact thatfrugivoresrepresenta continuumin termsof the numberof seeds each species deposits in suitablesites and
the probabilitythat it gives an individualseed favorabletreatment.
SEED
DISPERSAL
VERSUS
POLLEN
DISPERSAL
Pollen dispersaldiffersfromseed dispersalin several fundamentalrespectsthat
preventcomparableprecisionin seed dispersal (table 1). First,pollen has a very
specific"target," the stigmaof a nonspecificflower.These flowersare usually
easily recognizedbecause of theirdistinctivecolors, scents, and morphology.In
contrast,a seed's "objective," a site suitableforgermination,
establishment,and
ultimatelyreproduction,is more difficult
to characterizein space and time. Appropriatesites for seeds have few distinguishing
featuresperceptibleby frugivores, althoughfrugivoresmaypreferentially
frequentcertainhabitattypes. On a
gross scale, some habitat types (flood plains, riverbanks)may be predictable
(Howe and Estabrook 1977), but withinthem sites for seeds are not. Moisture,
sunlight,soil nutrients,seed and seedlingpredators,herbivores,fungalassociates,
allelopaths,and othersubtle factorsdeterminewhethera site is a good one fora
seed. Moreover, conditionsmay easily change and characteristicsof a site at the
time a seed is deposited may be poor indicatorsof futurequality.
Second, pollen dispersaldiffersfromseed dispersalin the temporalavailability
of suitable sites. Antherdehiscence and pollen dispersal are coordinated with
stigmareceptivityin synchronouslyfloweringpopulations,but opportunitiesfor
successfulseedlingestablishment(e.g., lightgap formation)may occur randomly
in time and space (Poore 1967; Knight 1975; Hartshorn1978).
Third,floweringplantsbenefitfromhabitatspecificityand sequentialvisitation
of nonspecificplants by theirpollen vectors because visits to flowersof other
species may resultin pollen transferto inappropriate"targets" and consequent
waste of pollen. A similaradvantage has been suggestedforfruiting
plants when
host-specificfrugivorescarryseeds to the habitatsoccupied by nonspecificplants
(Howe and Primack1975; Howe and Estabrook 1977). Yet, ifspecies-specificseed
and seedling predationis greaternear nonspecifictrees (Janzen 1970; Connell
1971), selection should not favorspecialized diets and movementsby dispersers
betweentreesof the same species. Moreover,the presence of an adultnonspecific
plantis not necessarilyevidence of an appropriatesite fora seed, particularlyin
thecase of shade-intolerant
plants,whichcomprisethemajorityof canopy species
in some tropicalforests(Hartshorn 1978) where up to 90% of plant species are
animal dispersed (Frankie et al. 1974). Once a tree reaches the canopy, its shade
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
COEVOLUTION
OF FRUITING PLANTS AND FRUGIVORES
405
TABLE 1
GENERALDIFFERENCESBETWEENPOLLEN DISPERSALAND SEED DISPERSAL
PollenDispersal
Suitable site for dispersal ....
Characteristics
of siteas predictorsof suitability.......
stigmaof conspecificflower
Seed Dispersal
site appropriateforgermination
and establishment
distinctive:
color,shape,etc.; unpredictable:
manysubtle
oftenapparentat a distance
factorsinvolved;present
characteristics
oftenpoorindicatorsof future
Temporalpattern
of suitable
quality
sites .....................
synchronous
withpollendisunpredictable:
oftenindepersal
pendentof habitattypeor
phenology
of nonspecific
plants
Advantage
forplantofdietand
habitatspecificity
by animal
vector ...................
high:most pollen lost ifvisitsto low: presence of adult conother species of plants interspecificplantoftenan unsuitvene between visits to conable site because of densityspecifics
dependentseed predationor
different
habitatrequirements
Abilityof plantto directanimal
of seed and tree
vectors to suitable site ....
high:incentives(nectar,pollen, low: no incentiveforfrugivore
etc.) providedat suitable site
to deposit seed in favorable
site; seed representsheavy
and space-consumingballast
thatis profitably
discarded as
quickly as possible.
has changed the favorablemicrohabitatthat allowed its own establishmentas a
seed and seedling.
Perhaps the most fundamentaldistinctionbetween pollen dispersal and seed
dispersal systems is the abilityof the plant to manipulatethe behavior of the
animal vector. Floweringplants can controlpollinators'movementsto a greater
extentby providingnutritional
(Baker and Baker 1975) and reproductive(Dodson
1975) incentivesat the appropriatesite (flower)forpollen transfer.There is, in
effect,"paymentupon delivery." There is no similarincentiveforseed dispersers
to drop seeds in appropriateplaces. Fruitingplants sufferthe costs of "advance
payment":Once frugivores
have strippedseeds oftheirnutritiousfruit,thereis no
furtheradvantage in carryingthe "ballast" (Snow 1971) which, unlike pollen,
or defecationof
may be heavy and space consuming.Hence, rapid regurgitation
the seed benefitsthe disperser,irrespectiveof its advantageto the plant.For both
pollinatorsand frugivores,the delayed returnbenefitsof "altruistic" pollen or
seed dispersal(such as futureincreases in local floweror fruitavailability)maybe
discounted (see also Howe and Estabrook 1977).
Given these limitations,to whatdegreemightplantsbe able to manipulatetheir
potentialseed dispersers?It has been suggestedthatmorphological,chemical,and
phonologicalcharacteristicsof plants and theirfruitsmay induce frugivoresto
leave fruiting
trees afterconsumingfew fruits.Amongthe conceivable incentives
are limited levels of fruitproduction (Howe 1977), mild toxicity,nutritional
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
406
THE AMERICAN NATURALIST
inadequacy of fruits,bulkyor slowlydigestiblefruits,and structuralmodifications
of the plantto increase the riskof predation.Whetherthese are "adaptations" or
consequences of otherlifehistorydemands (defense againstfungal,bacterial,or
insect fruitand seed parasites; limitationsof energyor nutrients;photosynthetic
considerations,etc.) remainsuncertainat this point. In any event, plants cannot
directthe dispersal of seeds to a particularlocation with a degree of exactness
comparable to pollen dispersal,thoughpossibly theycould favoranimal vectors
withparticularbehaviors(Docters van Leeuwen in McKey 1975), habitatpreferences, or probabilisticpatternsof seed dispersal (Smith 1975; Howe 1977; Howe
and Primack 1975).
of directingseed vecand the difficulty
Thus, environmentalunpredictability
constrain
seed
if
suitable
sites
were
dispersal systems
"knowable,"
tors, even
fromachievingthe precisionand specializationof some pollen dispersalsystems.
"SPECIALIZATION"
IN PLANT-FRUGIVORE
INTERACTIONS
Justas comparisonswithpollen dispersal have led to unrealisticexpectations
about coevolved mutualismsin seed dispersalsystems,impreciseand inconsistent
has affectedtheoreticaldevelopmentsand interpretation
of empirical
terminology
observations.For example, the term"specialist" is used frequentlywithrespect
to fruit-frugivore
interactions.Basically thetermrefersto theuse ofonlya portion
ofthetotalarrayof available resources. No organismis a puregeneralist,usingall
potentialresources in proportionto theiravailability,but, because species differ
in the extentto whichthey are selective in resource use, the term"specialist,"
even thoughit is inevitablyimprecise,is useful forcomparingspecies.
A firstsource of confusionarose because the degreeto whicha frugivoreshould
be considereda "specialist" has been assessed by different
criteria:beingtotally
or mostly frugivorous(Snow 1971); feeding only fruitto its nestlings("total
frugivory";Morton1973); relyingsolely on fruitas its source of proteinand lipids
(McKey 1975); and beingan animalthatis "totallydependenton fruitforfood for
at least part of its life and which invariablyvoids or regurgitatesseeds in viable
condition" (Howe and Estabrook 1977,p. 818). Snow (1981) definesa specialistin
termsof the qualityand size of the fruitsit eats. The bearded bellbird(Procnias
averano) is oftencited as the archetypal"specialist" (McKey 1975), despite the
fact that it actually feeds on at least 40-50 species of fruit(Snow 1970) and
probablyinsects and lizards as well (based on personal observationsof its congenorP. tricarunculataand othertropicalfrugivores).All of the above notions
have value, buttheuse of a singletermto referto such disparatecharacteristicsis
certainto fostercontinuedconfusion.It would be preferableto use more precise
termsto referto particularpatternsof resourceuse, such as "principallyfrugivorous." Moreover, the term "specialist" is totallyinappropriatewhen applied to
characteristicsunrelatedto resource use, such as dispersal quality.
A second problem arises with respect to the concept of a "specialized dispersal system." Specialization in plant-animalmutualismshas two components
that must be distinguished.A fruitingplant, for example, may specialize with
regardto its seed dispersers:Its seeds maybe principallydispersedby onlyone or
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
COEVOLUTION OF FRUITING PLANTS AND FRUGIVORES
407
a few of the arrayof potentialfrugivores.A fruit-eating
animal,in turn,may be a
specialist by restricting
the major part of its diet to a small subset of the fruits
available. In principle,therefore,a plant could specialize on a few generalist
dispersers,just as a frugivorecould specialize on a plant dispersed by many
species. Moreover, there is no a prior reason to expect a strongcorrelation
betweenthe degree of specializationand the importanceof a plantto a frugivore
and vice versa. A frugivoremay be veryimportantto a plantin a situationwhere
neithertheplantnorthefrugivoreis a specialist(e.g., thetreeCasearia corymbosa
[=C. nitida] and the masked tityra,Tityrasemifasciata; Howe and Primack1975;
Howe 1977; also Noble 1975).
Pollen and seed dispersal systemsdifferin the extentto which plants benefit
fromspecialization, eitherin termsof being dispersed by a limitednumberof
potential dispersers or attractingdiet-specificanimal vectors. In the case of
floweringplants,entrainedpollinatorsreduce the incidence of hybridizationand
wasted investmentin lost pollen and inviable seeds, and increase the amountof
pollendeliveredto suitablesites. Pollinationby species-specificvectorsallows the
persistenceand reproductionof plants when populationdensitiesare low (Baker
and Hurd 1968; Janzen1971). Because of the relativeprecisionof pollen transfer,
natural selection can favor the evolution of complex flowersthat exclude or
discourageorganismsunlikelyto transferpollen to nonspecificstigmaswithlittle
adverse effecton pollinationsuccess. Not surprisingly,there are examples of
flowersthat attractonly one or a few pollinatorswith narrowdiets ("specialize
on specialists"; see referencesabove).
The potentialvalues to plants of having only a few kinds of seed dispersers
depend on both treatmentof seeds and the way they are deposited into the
environment.If onlya singlefrugivorespecies regularlyvisitsa plant,mostof the
plant's seeds receive similartreatment.Eitherseeds willbe regurgitated
or, ifthey
pass throughthe digestivetract,theywillencountersimilarchemicaland physical
to withconditions(McKey 1975). Therefore,an optimalseed coat just sufficient
butnotso tough
standtheabrasionreceivedby the seed in thegutofthefrugivore,
as to inhibitsubsequentgerminationor cause waste of resources in the construction of unnecessarilystrongcoats, can evolve. Birds thatrelyon fruitsforall or
nearly all of their nutritionalneeds are believed to have soft-linedgizzards
(McKey 1975), and many of them regurgitateseeds. However, it is likely that
soft-linedgizzards evolved to avoid digestionof toxic, thin-coatedseeds rather
is advantageous
than to provide gentletreatmentto the seeds, and regurgitation
because it quickly reduces the indigestibleballast carried by the frugivore.
In general, the smallerthe numberof species of frugivoresthat visit a plant
species, the fewerthe habitattypes in whichits seeds are likelyto be deposited.
However, withinthose habitats,suitable sites forseed germinationand seedling
establishmentmay arise unpredictably.Because of seed dormancyor seedling
longevity,sites may be occupied long beforetheyarise (Howe 1980). Therefore,
unless there are strongtemporal correlationsbetween the sites visited by a
particulardisperser and the present or futureavailabilityof suitable sites, the
potentialadvantage of havingonly a restrictedset of seed dispersersmay not be
realized (Howe 1977).
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
408
THE AMERICAN NATURALIST
Set against these possible advantages of restrictingthe array of frugivores
regularlyvisitinga plant,thereare a numberof advantageslikelyto be associated
withattractingmanykindsof dispersers.First,thelargerthe numberof species of
frugivores,the greaterthe numberof dispersingindividualsand the largerthe
numberof fruitsthatcan be removed fromthe tree per day (Snow 1971). Since
fruitsmay oftenrot on trees or fallto the groundwithouthavingbeen consumed
(personal observation; Howe 1981), there are often situationsin nature when
having more visitorswould probablybe beneficial.
Second, a fruitcapable of being dispersed by a varietyof frugivoresis not
limitedto the habitat or general geographicalrange of any single disperser (a
restrictionin some pollinationsystems;Baker and Hurd 1968) and can expand its
distributionto colonize new sites (Hamilton and May 1977).
Third,plantsthatrelyon manyspecies to dispersetheirseeds may have a lower
probabilityof extinctionif a single disperserbecomes extinctor becomes very
rare (Howe 1977; Howe and Vande Kerckhove 1979; Temple 1977; Gilbert1980).
However, the evolutionof such "bet-hedging"strategieshas yetto be supported
by geneticmodels of naturalselection at the level of the individual(Slatkin and
MaynardSmith1979). High ratesof populationturnover,probablyeven greaterin
the tropicswhere manyspecies have low populationdensitiesand poor dispersal
abilities(Willis 1974), may mean thatfruiting
plantsare rarelyin association long
enough with specificfrugivoresto evolve tightobligate mutualismseven if they
were favored over the shortterm.
Fourth,even thoughthe mean probabilityof seed survivalmightbe lower in
seed distributionsproduced by dispersersof many species than in distributions
produced by a singlehighqualitydisperser,naturalselectioncould favorattractingmanydispersersratherthana singleone (fig.1). Whetheror notthe right-hand
tail of the seed distribution
curve generatedby manydisperserslies further
to the
rightthanthatgeneratedby a singledisperserdepends, of course, on the relative
availabilitiesof dispersersand on theirbehavior. But since a tree may produce
hundredsto millionsof seeds duringits lifetime,each with an extremelylow
probabilityof survival (given the fact that each plant leaves, on average, one
in an equilibrialforest),countingon a singledispersermaybe analogous
offspring
to Williams' (1975) assessmentof asexual reproduction:the multiplepurchase of
the same lotteryticket.
Unfortunately,testingthis hypothesisdirectlyis extremelydifficult.Of the
many seeds produced, findingthe survivingseedlings is difficult.Knowing their
parents and their dispersers seems almost hopeless. Althoughthe probability
of survival of any one seed is extremelylow, whetheror not the possibilityof
improvingthatprobabilityis favoredby naturalselectiondepends on the value of
alternativeuses of energyby the plant. For adult plants that have reached the
canopy or theirnormalfull size, alternativeuses of energyor nutrientsshould
normallyhave verylow marginalvalues, favoringheavy investmentsin reproduction even when the marginalvalue of incrementsto reproductiveinvestmentsare
as low as theynormallymustbe formostplants. The real meaningof "waste" in
seed dispersal (Howe 1980) is thereforeunclear.
Viewing the problemfromthe point of view of frugivoresalso suggeststhat
under most circumstanceseating a varietyof fruits,plus animal foods as well,
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
OF FRUITING PLANTS AND FRUGIVORES
COEVOLUTION
(
w
409
Dispersal by many species In addition
to 'sigh quality disperser
(I)
0~
(n
0
w
w
|
\ispersa/
by "high quality"
a/one
>(!
/disperser
LAL
0/
W~~~~~~~~~~~~~~~~~~~~
D/
zTALL
LOW
DL(PERLER
/PECIE5)
S
SUITABILITY OF SITES IN WHICH SEEDS
X//
GH OUAL/TY
A
HIGH
ARE DEPOSITED
FIG. 1. Suitabilityof seed dispersalby a wide assemblage of dispersers(solid line), versus
by a single"high quality" disperser(dashed line). If only those seeds deposited in the best
sites (to the rightof pointA) are likelyto survive,naturalselectionwillfavortheattractionof
a diversedisperserassemblage in spiteof itslowermean dispersalquality(cf. Williams1975).
should normallybe favored. Very few plants fruitthroughoutthe year, and few
are commonin morethana limitedpartof theirrange. Moreover,mostfrugivores
(with the exception of ants) are moderate-sized,relativelylong-livedorganisms
thatcannot package theirannual cycle withina singlefruiting
season theway that
a numberof large,principallyfrugivorous
manypollinatorsdo. Not surprisingly,
birds undergo markedseasonal movementsduringthe year in response to local
variations in fruitabundance (personal observation). Few dispersers are constantlyreliable in the sense of McKey (1975) and Howe and Estabrook (1977).
Also, frugivoresthatmay providehighqualitydispersal at one timeof the year
may not do so at other times. The obligatelyfrugivorousoilbird (Steatornis
caripensis) roosts and nests in caves where most of the seeds it ingests are
deposited (Snow 1962); bellbirds and manakins typicallydrop seeds beneath
displayperches duringthebreedingseason (personal observation).Althoughthey
probablygive all seeds gentletreatment,
theyare usuallylow qualitydispersersat
such times in termsof the sites where they deposit seeds.
CONSTRAINTS ON COEVOLUTION
We have argued that finelytuned mutualisticrelationshipsin seed dispersal
systemsshould be rare forthe followingreasons: (1) the inabilityof the plant to
provide incentivesfor precision in seed dispersal; (2) the relativelysmall dif-
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
410
THE AMERICAN NATURALIST
ferencesamong frugivoresin seed dispersal quality; (3) the unpredictability
and
of recognitionof suitabletargetsforseeds; (4) the potentialadvantages
difficulty
of having a broad assemblage of dispersers; and (5) the nature of life cycles of
frugivores.There may also be evolutionaryconstraintson developinga restricted
assemblage of seed dispersers(cf. Janzen 1977). For example, in Costa Rica's
lower montane wet forest,both mountainrobins (Turdus plebejus) and whitethroatedrobins (T. assimilis), as well as manyotherspecies, feed on lauraceous
fruits.The habitat preferencesand foragingpatternsof the two species differ
markedly(personal observation). They may also differin some measure of disto imaginewhat adaptationsa plant could use to
persal quality,yet it is difficult
exclude the poorerdisperserwithoutbarringits morphologicallysimilarcongener
as well. Classes of disperserscould be favoredby alteringfruitsize (McKey 1975;
Howe and Estabrook 1977), heightof presentation(L. Best, in prep.), firmnessof
attachment,color, nutritionand manyotherfeatures.These traits,however,are
likely simplyto reduce the diversityof visitors(possibly excluding some good
dispersers), not to limitit to a single species.
Equally importantmay be the influenceof spatial heterogeneityand gene flow
(Howe and Vande Kerckhove 1979). Geographicalregionsdifferin theproportion
of habitattypes and the pool of available dispersers.Even over relativelyshort
distances totally differentfrugivoresmay feed on the fruitsof the same tree
species. Thus, ifthereis gene flowbetweenpopulationsof a tree species, it may
not evolve adaptationsto exclude poor dispersersin one area because the same
tree in anotherplace is visitedby a different
groupof brids. Gene flowprobably
constrainscoevolutionof one-to-onemutualismsin pollinationsystemsas well.
Under what conditions would one expect to find highlycoevolved mutualisms between fruitingplants and frugivores?Two accounts document obligate
mutualisms(fromthe plant's perspective;Rick and Bowman 1961; Temple 1977).
Interestingly,
bothofthese associations occur on islands,whereenvironments
are
relativelyuniform,faunasare impoverished,and populationsare isolatedfromthe
homogenizinginfluenceof gene flow.
SUMMARY
Theoretical and empiricalresearch on frugivoryand seed dispersal has been
influencedby concepts derived from the study of pollination. In particular,
explicitand implicitanalogies between seed dispersal and pollen dispersal have
led to the expectation, under certain conditions, of the evolution of obligate,
species-specificrelationshipsbetweenfruiting
plantsand theanimalsthatdisperse
theirseeds. The two systemsdifferin importantrespects,however. Plantsbenefit
by directingpollen dispersersto a definite,recognizable "target," a nonspecific
flower,and theycan provideincentivesat flowerswhichserve to attractpotential
pollinators.In effect,thereis "paymentupon delivery" of the pollen. In contrast,
for seeds the target(an appropriatesite for germinationand establishment)is
seldom readilydiscernible,and dispersalbeneatha nonspecificplantmay actually
be undesirable. Another importantdifferenceis that frugivoresare "paid in
advance." Because of these differencesand others,the outcomes of coevolution
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
COEVOLUTION OF FRUITING PLANTS AND FRUGIVORES
411
thanthose of flowerplantsand frugivoresare expected to be different
of fruiting
ing plants and flower visitors. There are thereforeproblems with drawing
derivedfromstudiesof
analogies betweenthe two systemsand usingterminology
pollinationto design and interpretstudies of seed dispersal.
ACKNOWLEDGMENTS
We appreciatethe interestand insightscontributedby the Ecology and EvolutionaryBiology groupat the Universityof Washingtonduringvarious discussions
of these ideas. L. Best, M. Gardner,W. Haber, H. Howe, S. Koptur, E. Leigh,
offeredhelpfulcommentson an earlierdraft.A suggestionby
and A. Worthington
M. Westoby promptedthe use of fig. 1. The firstauthor received generous
financialsupportfromthe Organizationfor Tropical Studies and Sigma Xi for
research in Costa Rica.
LITERATURE
CITED
Armbruster,W. S., and G. L. Webster. 1979. Pollinationof two species of Dalechampia (Euphorbiaceae) in Mexico by euglossine bees. Biotropica 11:278-283.
coevolution.
Baker, H. G., and I. Baker. 1975. Studies of nectar constitutionand pollinator-plant
Pages 100-140 in L. E. Gilbertand P. H. Raven, eds. Coevolution of animals and plants.
Universityof Texas Press, Austin.
Baker, H. G., and P. D. Hurd. 1968. Intrafloralecology. Annu. Rev. Entomol. 13:385-414.
Connell,J. H. 1971. On the role ofnaturalenemies in preventingcompetitiveexclusion in some marine
animals and in rain foresttrees. Pages 298-312 in P. J. den Boer and G. R. Gradwell, eds.
Proceedingsof the Advanced Study Instituteon dynamicsof numbersin populations,Oosterbeek, 1970. Centre for AgriculturalPublishingand Documentation,Wageningen.
Cruz, A. 1974. Feeding assemblages of Jamaicanbirds. Condor 76:103-107.
and feedingassemblages
Diamond, J. M., and J. W. Terborgh.1967. Observationson birddistribution
along the Rio Callaria, Departmentof Loreto, Peru. Wilson Bull. 79:273-282.
Dodson, C. H. 1975. Coevolutionof orchidsand bees. Pages 91-99 in L. E. Gilbertand P. H. Raven,
eds. Coevolution of animals and plants. Universityof Texas Press, Austin.
Dresser, R. L. 1968. Observationson orchids and Euglossine bees in Panama and Costa Rica. Rev.
Biol. Trop. 15:143-183.
Eisenmann, E. 1961. Favorite foods of Neotropical birds: flyingtermitesand Cecropia catkins. Auk
78:636-638.
Fleming,T. H. 1979. Do tropicalfrugivorescompete for food? Am. Zool. 19:1157-1172.
Frankie, G. W., H. G. Baker, and P. A. Opler. 1974. Comparative phenologicalstudies of trees in
tropical wet and dry forestsin the lowlands of Costa Rica. J. Ecol. 62:881-919.
Galil, J. 1977. Fig biology. Endeavour 1:52-56.
Gilbert,L. E. 1980. Food web organizationand theconservationof neotropicaldiversity.Pages 11-34
in M. E. Soul6 and B. A. Wilcox, eds. Conservationbiology. Sinauer, Sunderland,Mass.
Hamilton,W. D., and R. M. May. 1977. Dispersal in stable habitats. Nature 269:578-581.
Hartshorn,G. S. 1978. Tree fallsand tropicalforestdynamics.Pages 617-638 in P. H. Tomlinsonand
M. N. Zimmerman,eds. Tropical trees as living systems. Cambridge UniversityPress,
Cambridge.
Haverschmidt,F. 1971. Large numbers of birds exploitinga fruittree in Surinam. Wilson Bull.
83:104-105.
Howe, H. F. 1977. Bird activityand seed dispersalof a tropicalwet foresttree. Ecology 58:539-550.
1979. Fear and frugivory.Am. Nat. 114:925-931.
1980. Monkey dispersal and waste of a neotropicalfruit.Ecology 61:944-959.
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
412
THE AMERICAN NATURALIST
. 1981. Fruitproductionand animal activityin two tropicaltrees.In E. Leigh, Jr.,A. S. Rand,
and D. M. Windsor,eds. The ecology of a tropicalforest:seasonal rhythmsand long-term
changes. SmithsonianInstitution,Washington,D.C.
Howe, H. F., and D. DeSteven. 1979. Fruitproduction,migrantbirdvisitation,and seed dispersalof
Guarea glabra in Panama. Oecologia 39:185-196.
Howe, H. F., and G. F. Estabrook. 1977. On intraspecific
competitionforavian dispersersin tropical
trees. Am. Nat. 111:817-832.
Howe, H. F., and R. B. Primack. 1975. Differential
seed dispersalby birdsof the treeCasearia nitida
(Flacourtiaceae). Biotropica 7:278-283.
Howe, H. F., and G. A. Vande Kerckhove. 1979. Fecundityand seed dispersal of a tropical tree.
Ecology 60:180-189.
Janzen,D. H. 1970. Herbivoresand the numberof tree species in tropicalforests.Am. Nat. 104:501528.
1971. Euglossine bees as long-distancepollinatorsof tropical plants. Science 171:203-205.
1977. Promisingdirectionsof studyin tropicalanimal-plantinteractions.Ann. Mo. Bot. Gard.
64:706-736.
Jenkins,R. 1969. Ecology of threespecies of Saltatorswithspecial referenceto theirfrugivorousdiet.
Unpubl. Ph.D. diss. Harvard University.
Knight,D. H. 1975. An analysis of late secondary succession in species-richtropicalforest.Pages
53-59 in F. B. Golley and E. Medina, eds. Tropical ecological systems.Springer-Verlag,
New
York.
Land, H. C. 1963. A tropicalfeedingtree. Wilson Bull. 75:199-200.
Leck, C. F. 1969. Observations of birds exploitinga Central American fruittree. Wilson Bull.
81:264-269.
McDiarmid,R. W., R. E. Ricklefs,and M. S. Foster. 1977. Dispersal ofStemmadeniadonnell-smithii
(Apocynaceae) by birds. Biotropica 9:9-25.
McKey, D. 1975.The ecologyof coevolved seed dispersalsystems.Pages 155-191in L. E. Gilbertand
P. H. Raven, eds. Coevolution of animals and plants. Universityof Texas Press, Austin.
in tropicalbirds.
Morton,E. S. 1973. On theevolutionaryadvantagesand disadvantagesof fruit-eating
Am. Nat. 107:8-22.
of the
Noble, J. C. 1975. The effectsof emus (Dromaius novaehollandiae Latham) on the distribution
nitrebush (Nitraria billardieriDC.). J. Ecol. 63:979-984.
Poore, M. E. D. 1967. The concept of the association in tropicalrain forest.J. Ecol. 55:460-470.
of moths and Yucca whipplei
Powell, J. A., and R. A. Mackie. 1966. Biological interrelationships
(Lepidoptera: Gelechiidae, Blastobasidae, Prodoxidae). Univ. Calif. Publ. Entomol.42:1-46.
Rick, C. M., and R. I. Bowman. 1961. Galapagos tomatoes and tortoises. Evolution 15:407-447.
functionanalysis of assemblagesof fruit-eating
Ricklefs,R. E. 1977. A discriminant
birds in Central
America. Condor 79:228-231.
Sazima, M., and I. Sazima. 1978. Bat pollinationof the passion flower,Passiflora mucronata, in
SoutheasternBrazil. Biotropica 10:100-109.
Slatkin,M., and J. Maynard Smith. 1979. Models of coevolution. Q. Rev. Biol. 54:233-263.
Smith,A. J. 1975. Invasion and ecesis of bird-disseminated
woody plantsin a temperateforestsere.
Ecology 56:19-34.
Snow, B. K. 1970. A fieldstudy of the bearded bellbirdin Trinidad. Ibis 112:299-329.
Snow, D. W. 1962. The natural historyof the oilbird (Steatornis caripensis) in Trinidad, W.I. II.
Population, breedingecology and food. Zoologica 47:199-221.
1971. Evolutionaryaspects of fruit-eating
by birds. Ibis 113:194-202.
1981. Tropical frugivorousbirds and theirfood plants: a world survey. Biotropica 13:1-14.
Temple, S. 1977. Plant-animalmutualism:coevolution with dodo leads to near extinctionof plant.
Science 197:885-886.
Thien, L. B. 1969. Mosquito pollinationofHabenaria obtusa (Orchidaceae). Am. J. Bot. 56:232-237.
Thompson, J. N., and M. F. Willson. 1978. Disturbance and the dispersal of fleshyfruits.Science
200:1161-1163.
1979. Evolution of temperatefruit/bird
interactions.Evolution 33:973-982.
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions
COEVOLUTION
OF FRUITING PLANTS AND FRUGIVORES
413
Van der Pijl, L. 1972. Principlesof dispersal in higherplants. Springer-Verlag,New York.
Williams,G. C. 1975. Sex and evolution. PrincetonUniversityPress, Princeton,N.J.
Willis, E. 0. 1966. Competitive exclusion in birds of fruitingtrees in Western Colombia. Auk
83:479-480.
. 1974. Populations and local extinctionsof birds on Barro Colorado Island, Panama. Ecol.
Monogr. 44:153-169.
This content downloaded from 139.140.232.150 on Tue, 24 Dec 2013 11:41:24 AM
All use subject to JSTOR Terms and Conditions