Symbiosis, Instability, and the Origins and Spread of Agriculture: A New Model [and
Comments and Reply]
Author(s): David Rindos, Homer Aschmann, Peter Bellwood, Lynn Ceci, Mark N. Cohen,
Joseph Hutchinson, Robert S. Santley, Jim G. Shaffer and Thurstan Shaw
Reviewed work(s):
Source: Current Anthropology, Vol. 21, No. 6 (Dec., 1980), pp. 751-772
Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological
Research
Stable URL: http://www.jstor.org/stable/2742515 .
Accessed: 28/09/2012 11:48
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 support@jstor.org.
.
The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaborating
with JSTOR to digitize, preserve and extend access to Current Anthropology.
http://www.jstor.org
CURRENT ANTHROPOLOGY
?
Vol. 21, No. 6, December 1980
1980 by The Wenner-GrenFoundation for Anthropological Research 0011-3204/80/2106-0002$02.55
and the Originsand
Symbiosis,Instability,
A New Model
Spread of Agriculture:
by David Rindos
INTRODUCTION
The originsofagriculture
and its subsequentspreadhave been
of concernto both naturaland social scientistsforover 100
years.The studyofagricultural
originsis farmorethanmerely
the investigationof a particularshiftin man's subsistence.
is seenby mostpeopleas thefoundation
Agriculture
ofmodern
civilization.It is frequently
heldthatonlyagriculture
permitsa
societytheluxuryofexcessproductionand thatalongwiththis
excessproductioncomes the developmentof the hallmarksof
civilization:writing,metalworking,
and the establishmentof
well-defined
economicclasses. Despite the significance
of agriculturalorigins,our understanding
of thisphenomenonis far
fromsatisfactory.
We lack any unifiedtheoryto explainboth
howagriculture
aroseand whyit shouldhave arisenat all.
Variousapproacheshave been takentowardstheexplanation
of the initial appearance of agriculturalsubsistence.Most
workersin thesocial scienceshave begunby lookingintoman's
cultureforthe factorswhichcould explainthe developmentof
pristineagriculturalsocieties. Bachofen (1967 [1861]) and
Frazer (1912) believedthatagriculturedevelopedout of traditionalsexual rolesand the divisionsoflabor that theycreated.
Allen (1897) and Hahn (1909) presentedargumentsseekingto
understandagricultureas an outgrowthof ritualisticand religiouspractices.This approachhas been recentlydevelopedin
greatdetail by Isaac (1970). Otherworkershave ascribedthe
originof agricultureto the influenceof climaticevents (Pumpelly 1908,Childe 1951,Wright1968). Finally,certainauthors
have alwaystakenwhatmaybe describedas a "commonsense"
view of the originof agriculture:that agriculture,like other
is theresultofan "invention."Carter
technological
innovation,
(1 977) is a contemporary
authoradheringto thisviewpoint.
Recently,ecologicalparadigmshave cometo dominateagriculturaloriginstheories.Ecological theoristshave developed
thepioneeringworkofthebiologistsDe Candolle(1959 [1886])
and Vavilov (1926) and thegeographerSauer (1936, 1969).The
is a graduatestudentat CornellUniversity
and
teaching
and research
assistantat theUniversity's
L. H. Bailey
Hortorium
(467MannLibrary,
CornellUniversity,
Ithaca,N.Y.
14853,U.S.A.).Bornin 1947,hereceived
hisB.S. inruralsociology
from
Cornell
in 1969andhisM.S. inplanttaxonomy,
witha minor
in anthropology,
in 1980.He has beenpalaeoethnobotanist
with
theCornellArchaeological
Projectin Cyprusand in Honduras.
His research
interests
are thesystematics
oftheCompositae
and
theinteraction
ofplantsand man.His Master'sthesisis entitled
" GenericDelimitation
in theVerbesinoid
Heliantheae(Compositae).I. The GenusZexmenia."
The present
paperwas submitted
in finalform8 I 79.
DAVID RINDOS
Vol. 21 * No. 6 * December1980
hallmarkof the ecologicaltheoristis thathe places major emphasis on the interactionbetween man and his immediate
comworkersstressdifferent
environment.
Of course,different
Workersplacinggreatemphasison
ponentsofthisrelationship.
a theoretical"equilibrium"betweenpopulationand the environment
includeBinford(1968),Meyers(1971), Cohen(1975,
1977),and Flannery(1965,1973).Most oftheseworkersexplain
the originor adoptionof agricultureas the resultof attempts
to correcta disturbedequilibriumbetweenpopulationand the
environment.
Others,such as Lewis (1972) and Harris (1969,
1972), along with Flanneryin a paper otherthan those just
cited (1969), stressthe importanceof the developmentof new
techniquesand culturalpatternsto the originof agricultural
systems.
Most of thistheorizing
on the subjectofagriculturalorigins
has been permeatedby two fundamentalconcepts:that agriin originand that its adoptionand
culturewas revolutionary
subsequentdevelopmentweredue to intentionalpractices.As
notedby Harlan, de Wet,and Stempler(1976:3), "the idea of
an agriculturalrevolutionas elaboratedby V. GordonChilde
fortheNear East was basicallyappliedto thesocialand cultural
economies."Onlyrecentlyhas
consequencesof food-producing
thisidea begunto be challenged.The agricultural
way oflifeis
now beginningto be seen as the resultof a seriesof gradual
changes(cf. MacNeish 1964,Bray 1974,Higgs 1976, Pfeiffer
1976). Even today,most modelsforthe originof agriculture
containat least some elementsof culturalor individualintention-eitherto explaintheinitialappearanceof thetechniques
and plants of agriculturalsystemsor to elucidatethe survival
and dispersalofagricultural
systems.Corollaryto,and perhaps
underlying,
theseconceptsis the beliefthat agriculturefacilitatesthedevelopment
ofa "superior"economy-onewithclear
and substantialadaptivebenefitsto thesociety.This adaptiveinvokedto explainthe maintenanceand
nessis also frequently
it is held,is maineconomies.Agriculture,
spreadofagricultural
meansto feedlargenumtainedbecauseit is themosteffective
bersofpeople;agriculture
spreadsbecauseits greateradaptiveness allows it to spread at the expenseof more "primitive"
meansofproduction.
I shallhereintroducea newmodelfortheoriginand developmentof domesticatedplantsand agriculturalsystems.Rather
thanrelyingupon the assumptionsjust mentioned,thismodel
views agricultureas the outgrowthof evolutionarypotentials
feedsupon
whichmaydevelopwheneveran animalconsistently
any set offoodplants.Intentand inventionare notdenied,but
they are regardedas unnecessaryto the model. Parsimony
would suggestthat if agriculturaloriginsmay be explained
751
withoutuse ofintentor invention,
thentheseconceptsmay,for
thepurposesofthismodel,be set aside.'
Given the evolutionary
means by whichagriculturalplants
and systemsdevelop,the originof agricultureis anythingbut
theultimateeffects
revolutionary-although
oftheprocessmay
be describedas a revolutionofthefirstorder.The adaptiveness
of agricultureis not stressedin this model,althoughit is not
denied that agriculturepermitsgreaterdensitiesof human
populationthan mostotherformsof subsistence.Instead,emphasisis placed upon theinstabilities
whichfirstdomestication
and then agriculturalproductionintroduceinto subsistence
strategiesand whichare the basis foragriculture's"success."
Finally,the relationship
betweenthe originof agricultureand
the riseof moderncivilization,whileaddinga certainurgency
to the study of the problem,is not addresseddirectlyhere.
Discussionof thisdifficult
problemwouldof necessitytake us
farbeyondthe consideration
ofagriculturalorigins.Nevertheless, if thismodelhas any heuristicvalue, implicationsmay be
drawnfromit whichgo beyondthe meredescriptionof a new
modeoflifeforman.
Modernagriculturaldevelopmentschemesare a responseto
theshortagesoffoodand resourcesthatplague so muchof the
world.Allagricultural
systemsappearproneto occasionalcrises
in production.Besides their immediateeffectsupon human
healthand survival,shortagesoffoodfrequently
have deleterious effects
uponthestabilityofpoliticaland economicsystems.
The importanceof theinteractionofpopulationgrowth,dwindlingnaturalresources,and recurrent
shortagesof foodplaces
the studyof agriculturalsystemsin the mainstreamof human
concern.The prospectof foodshortagesencouragesthe breeding of improvedcrops and the developmentof moreefficient
croppingsystems.Success in theseprogramsencouragesreliance upon fewerand fewerspeciesofplantsand upon an ever
decreasingnumberof varietiesof a givenspecies.Agricultural
productionalso tendsto becomemorelocalized,to take advantageoffavorableenvironmental
conditions,
and to allowforthe
exploitationof economiesof scale. Yet, as is well recognized,
monocultures
are extremelyvulnerableto catastrophicfailure
as a resultofdisease,pests,and climaticextremes.
Today, man reliesuponabout 20 speciesofplantsto provide
most of his food (National Academyof Sciences 1975). Yet
hundredsof species of plants have been domesticatedand
thousandsutilized.Remarkably,no majorcrophas beendomesticatedfromthewildsincetheearliestdaysofagriculture.
Even
our best attemptsat improvement
of existingcropshave been
less than totallysuccessful.Recentlywe have begunto understand that plant breedingand the successfulintroductionof
new,improvedcultivarsmay inadvertently
be acceleratingthe
loss ofmuchofthevariationin thecropsgrownby man. Attention has come to be directedtowardsthe importanceof the
conservationof gene pools forcultivatedplants. We are also
to understandthatthebreedingofimprovedvarieties
beginning
of crop plants is, paradoxically,oftenaccompanied by increased susceptibilityof the crop to previouslyunknownor
unimportant
pests and diseases. We are finallybeginningto
recognizethatagricultureis a dynamicsystem-thatthe agriculturalecologyexistingin any regionis the productof the
interactionofnumerousfactorsoverlongperiodsoftime.Only
certainparts of the agriculturalecologyare completelyunder
humancontrol.As we beginto understandthefunctioning
and
the evolutionof theagriculturalecology,however,our control
over it is likelyto increase.Part of this understanding
must
come froma considerationof the originand developmentof
agriculturalsystemsand an investigation
of the factorswhich
controland limitthedomestication
ofplants.
1 Elsewhere I have given the issues of intent and inventionextendedtreatment(Rindosn.d.). The viewofhuman behaviorpresented here should not be confusedwith "sociobiology"; instinctand
geneticsare also set aside.
752
If themodelforagriculturaloriginsand thedomestication
of
plants presentedhere proves useful,its greatestcontribution
maycomefromtherecognition
thatwe are notfacinga qualitativelydifferent
set ofproblemsfromthosefacedby our distant
forebears.The historyof agricultureis a historyof instability
inproductionand ofagriculturally
inducedcrises.Identification
of the factorsresponsibleforthis instabilitymay allow us to
takeactionto reduceit. Seen fromtheperspectiveofthemodel,
manyof ourbestattemptsat increasingagricultural
productivitycan be expectedto increasethevulnerability
of thesystem
to failureas well;thebestattemptsofourancestorsto improve
the productivity
of theirsystemshad similareffects.Whilein
the past we have survivedagriculturalcrisesand even benefitedfromthem,however,we are no longerin a strictlycomparable situation.The consequenceof agriculturalcrisisis no
but thepossibilityof starvation.By apprelongeremigration,
ciatingthe historyof agriculturalsystemswe may be better
preparedto developourpresentsystemsto satisfythedemands
of the future.
In the model, domesticationoccurs beforethe originand
developmentof agriculturalsystems.Domesticationis thereason agricultural
systemsdevelop.Althoughthe importanceof
domesticationper se to a societydecreaseswith the rise of
agricultural
systems,domestication
does not cease. It continues
withinthe agriculturalcontext,furthermodifying
the plants
undercultivation.It may also occurwithplantsoutsideof the
agriculturalecology.Domesticationis theresultoftheevolution
ofa symbiosis
between
man and plant.
The best way to understandthe changesthat have occurred
in domesticated
plantsis to lookat therelationship
betweenthe
plantand theanimalwhichfeedson it. Two ofthemostimportant effectsof domestication
are (1) intensification
of the mutualisticrelationship
betweenanimal and plant and (2) exclusion ofotheranimalswhichmightalso be competingforaccess
to theplant.These effects
are usuallymediatedby morphological change.One of the mostimportantfactorsin the evolution
ofthecultivatedplantis theisolationofthepotentialcultivated
plant, by eithergeneticor spatial mechanisms,fromthe protakenhere mayalso
genitorspecies.The viewofdomestication
be used to accountforthe evolutionof weedsand forthe nondomestication
ofpotentiallyvaluable crops.
The transitionfrommutualisticdomesticationto evolved
agriculturalsystemswas mediatedby environmental
manipulation.Humanactivitiessuchas thefellingoftreesorthesetting
offirescannothelpbut have had effects
uponthelocal environment.The major effectof theseactivitieswas the concentration of domesticatedplants in localized areas. This permitted
therealizationofa newseriesofevolutionary
potentials.I shall
presenta dynamicand interactivemodel to account for the
originand subsequentspreadofagriculturalsystems.
is based on environmental
Agriculture
manipulation.Ecological limitsupon plant productivity
are thusreduced.Agricultureis a set of integrated
whichaffects
theenvironment
activities
inhabited
its lifecycle.Agribythedomesticated
plantthroughout
cultureservesto increasethe domesticity
of the plant. However,it has major effectsupon the communityof plants that
are utilizedin agriculturalsystems.It tends to increasethe
average yield of domesticatedplants. Competitionwithina
speciesof domesticatedplantstendsto selectthoseindividuals
best adapted to agriculturalpractices.Competitionbetween
species of domesticatedplants tends,in a similarmanner,to
select forthosespecies best adapted to agriculturalpractices.
Thus, over time,althoughyieldincreases,it comesfromfewer
plant species.
Localizationof agriculturalproduction,however,is accomforsurvivaland
panied by convergencein the requirements
reproductionin species of domesticatedplants. This convergenceand,indeed,thelocalizationofproductionitselfintroduce
new instabilitiesinto agriculturalproduction:what is a bad
year forany givenagricultural
plant is likelyto be a bad year
CURRENT
ANTHROPOLOGY
many
by concentrating
plants.Agriculture,
forall agricultural
plants in one area, may also encourageexcessivedamage by
disease,insects,and herbivoresand thusdecreaseyield.Instaare expressedas periodsof greatlyrebilitiesin productivity
duced yield. These crises cause the emigrationof the excess
populationfromagriculturalsocieties.These emigrantpopulaforthedispersaloftheagriculturalway of
tionsare responsible
life. Selectionhas favoredthe spread of agriculturalsystems
which maximizeinstabilityin production.Maximization of
instabilitybringswithit maximizationof dispersals.Thus the
mostsuccessfulagriculturalsystemshave been thosethat are,
at least in the broadestsense,the least "adapted" agricultural
systems.
Rindos:ORIGINS
AND SPREAD OF AGRICULTURE
Hartzell (1967) has reportedon one particularant/acacia
relationship.The ants inhabithollow,enlargedthornson the
acacia and receivea sugaryexudatefromspecialfoliarnectaries
at the base of the leaves. They also harvestmodifiedleaf tips
calledBeltrambodies,whichare richin bothproteinsand lipids.
Theseare usedto feedtheant larvae.The antspatrolthefoliage
insects.Whenants were
and removeand feeduponherbivorous
experimentallyremoved from acacias, the plants were so
and insectsthatall died within
severelyattackedby herbivores
a year.Otheracacias are protectedby theirants fromherbivore
theplantis attackedand given
predation;anyanimaldisturbing
numerouspainfulbites (Hockeng 1975).
The loss of chemicaldefensemechanismswithina speciesof
DOMESTICATION AND SYMBIOSIS
acacia involvedin a mutualisticrelationshipwithants is apparentlycommon.Ant-acaciaslack the biochemicaldefense
ANIMAL-PLANT INTERACTIONS
mechanismpresentin otheracacias (Rehr,Feeney,and Hanzen
plantshave
Domesticationis best understoodas an interaction 1973). In an analogousfashion,manyhuman-crop
Coevolution.
is exeggplant
wild
the
systems:
defense
chemical
lost
their
betweenman and theplantson whichhe feeds.2It wouldseem
wild
of
flesh
bitter
the
1976:80);
(Choudhury
bitter
tremely
as an adaptaless thanaccurateto describesuchan interaction
(1975)
and
Bemis
for
Whitaker
is
sufficient
species
Cucurbita
tion of or by man to the demandsof continuedsurvival-we
to postulatean originaldomesticationof thesefruitsfortheir
may just as easily describethe processof adaptationof or by
seeds; the domesticationof lettuce (Latuca sativa) has been
the plant, forit too benefitsfromthe association.By placing
in latex
the
emphasisupon theactionsof man we neglect fundamental accompaniedby a reductionofspininessand a decrease
can
oleraceae)
(Brassica
cabbage
wild
1976:39);
content
(Ryder
contribution
made by the domesticatedplant to the developdanof
bitter
and
quantity
the
much
as
four
times
as
contain
exist
withcould not
mentof agriculturalsystems.Agriculture
gerousglucosimatesas cultivatedstrains(Josefsson,cited in
out domesticatedplants. Domestication,as seen here, is a
1976:49); cassava (Manihot esculenta)has two
Thompson
natural evolutionaryprocessby means of whichanimalsand
one
poisonouswhen raw and the otherlacking
highly
forms,
plantsare able to increasetheirfitness.Coevolution-a typeof
of beingcooked and eaten withprelimiand
capable
poisons
evolutioninvolvingtwo geneticallyunrelatedspecies-occurs
of the organismspositively narytreatment-thepresenceof toxinsprobablybeing correwheneverthe interrelationship
lated, at least in part,withmethodsof cultivationin differing
affectstheir potential for survival. The relationshipwhich
describedas a symbiosis. agriculturalsystems(Rogers and Appan 1973, Rogers 1965).
resultsfromcoevolutionis frequently
The hoardingof seeds is widespread
Storageand harvesting.
Coevolutionis widespreadin nature.Pollination,seed discan
persal, and even predation frequentlyhave coevolutionary in theanimalkingdom,and, as it relatesto plantdispersal,
of
plants.
history
in
the
evolutionary
significance
be
of
major
withinwhich
aspects.Coevolutionprovidesa usefulframework
Various birds,includingnutcrackers,jays, and woodpeckers,
to observedomestication.By means of his interactionswith
and mor- are knownto storeseeds of such plants as pine, beech, oak,
plants,man inevitablyinfluencestheirdistribution
phology.These changes,of course,do not occur by the direct chestnut,filbert,and variousPrunus species. Small mammals
storeedibleseedsin their
and chipmunks
suchas mice,squirrels,
influenceof man's activityon the plants. Rather,theyoccur
1972,Stebbins1971).
der
Pijl
caches
(van
in
special
and
nests
over manygenerationsas certainplants-the morefit-leave
dried
made
of
grassesand weeds.
up
a
hay
and
store
guard
Pikas
greaternumbersofprogeny.
Guardingof storesis also commonin various birds (Emlen
The most strikingevidenceforthe coevolutionaryview of
1973:164). Janzen(1971) notesthatseedsofstoredspecies can
and agriculturaloriginsis the existenceof widedomestication
soils near
be
foundgrowingto adult statusin thenitrogen-rich
spread nonhumanagriculturalsystems.Domesticatedplants
were
"lost"
seeds
these
that
assumed
be
It
can
nest
entrances.
have establishedrelationshipswithmany animals otherthan
resource
a
useful
provide
these
plants
Clearly,
or
"rejected."
man. In a fewinstances,theseotheranimalshave incorporated
fortheinhabitantofthenest.Muskrattrappersin centralNew
manipulationintotheirbehavioral
techniquesofenvironmental
rice
and we may thereforecall the resultantsystems York reportthatlargestoresofarrowheadtubersand wild
repertoires,
dens.
muskrat
in
found
can
be
agricultural.Agriculturehas generallybeen characterizedby
forour purposesis the phenomenonof
Far moreinteresting
such actions as sowing,protection,and harvestingor by the
Harvesterants
seed
(myrmecochory).
dispersal
ant-mediated
existenceofmorphologically
distinctcultivatedplantswhichare
have establisheda symbiosiswithspeciesof plants frommore
None ofthese
adapted to thebehaviorsinvolvedin agriculture.
63 familiesof flowering
plants
phenomenais restrictedto the human-plantinterrelationship. than 223 genera representing
(Nesomn.d.). These ants collectand therebyoftendisseminate
Whilenumerousexamplesofprotectionof plants
Protection.
theseedsoftheseplants.The antsare "apparentlyable to conby animalsmightbe given,I wouldlike to describeherea paroftheseedstheycollectand also exertcontrolthegermination
ticularlycomplicatedformof protectionactivitywhich has
on thevegetationin thevicinityofthenest"
influence
siderable
evolvedin the behaviorof certainants. Here we finda highly
(Hockeng1975:83).
developedmutualisticrelationshipbetweenants and plants in
Harvesterants have developedquite sophisticatedformsof
whichthe plants providedomiciles,food,or both forthe ant.
behaviorto protecttheirseed stores.At least some species of
In "return,"theant protectstheplant.
ants willremoveseeds thathave becomewetfromthenestand
drythemin thesun.Many speciesoftheants "thresh"theseed
2 In this analysis I am restricting
myselfto the originof planttheygather,and moundsof chaffmay be foundin the vicinity
human relationships.To add an analysis of animal domestication
of the nest.It appears that bothant and plant benefitby this
would substantiallyextendthe lengthof thisessay. Furthermore,
all
agriculturalsystemsare ultimatelybased upon plant domestication;
association(Hartzell 1967:127):
human-animalrelationships
lackingthiselementare generallytreated
The
anddrought.
coldperiods
Theantis assureda foodsupplyduring
as a separate phenomenon(i.e., pastoralism).Finally, the treatment
against
and it is protected
plantspeciesis aidedin its distribution
hereis at least in part dictatedby my trainingas a botanist.
Vol. 21 * No. 6 * December1980
and coldby beingstoredin theunderground
galexcessive
drought
collecttheirfoodsupplyin
leriesoftheants.Sincetheantsnormally
thisis an important
inthesurvival
factor
excessoftheirrequirements,
of the plantspeciesduringclimaticchangesthatwouldnormally
themaboveground.
destroy
is characterized
Stebbins(1971:243) notesthatmyrmecochory
modifications
in the plant:
by a wholeseriesof morphological
to theirrelatives
In comparison
... [these]plantshavethefollowing
theflower
stalksarerelatively
characteristics:
low,and thepeduncle
whentheseedis ripeso thatthematurecapsuleis
becomesrecurved
closeto theground.The capsule... does notdehisceregularly
by
and overa relatively
longperiodof
valvesor pores,butirregularly
time.... Thismakespossiblerepeatedvisitsbytheantsto thesame
haveseveralmodifications.
Sometimes
plants.The seedsthemselves
theyhave a specialoilyseedcoat; theantsremovethisafterthey
have carriedthe seeds to theirnests,and thentheydiscardthe
fat-bearing
strippedseeds.Moreoften,each seed has a particular
appendagewhichthe ant clipsoffwithits mandiblesand carries
ofthe
afterithas transported
theseedsto thevicinity
underground,
nest.
ofant-dispersed
Changesin themorphology
plantsare adaptaof the symbiosis.Traits of
tionsnecessaryto the functioning
domesticatedcropssuch as indehiscencemay be viewedin the
same manner.
Transportationof the seeds to the vicinityof the nest can
of fairlylarge populationsof the
allow forthe establishment
plantin an area whereit mostbenefitstheants.Like theestablishmentoffoodplantsin thevicinityof mammaliannests,this
formof ant dispersalbears strikingsimilarities
to the "dumpheap" modelsfortheoriginofagriculture.
Furthermore,
myrmecochory
may allow forthe survivalof
otherwiseunadaptedorganisms.Handel (1978) has extensively
studied the competitiverelationshipsof threecloselyrelated
United
speciesofsedges(Carex)whichgrowin thenortheastern
and extensiveexperimental
States.Bothfieldwork
investigation
have shownthat one ant-dispersed
speciesis a relativelypoor
competitor.It has survivedbecause it does not have to compete directlywiththe otherspecies.This is because it is dispersedintoa habitat (rottinglogs) whichis unavailableto the
others.Antsnestingin rottinglogs dispersethe speciesin the
vicinityof theirnestsand providea refugeforit. The sedge
thususes theant as its dispersalagentin a race to keep ahead
of the other,morecompetitivespecies.The ant-dispersed
species mustmoveintonewhabitatsas theybecomeavailable,for
oncethehabitatis no longera newhabitat(thatis,oncethetree
is completelydecomposed)the plant will be replaced by its
relatives.
Typical of the syndromeof a weedycolonizingspecies,the
ant-dispersed
sedgehas evolvedseveralothertraitswhichcontributeto its successin thisrole.It has, on the average,twice
the seed set of its relatives.It is also typifiedby immediate
germinationof the seeds; the othertwo species requireoverwintering
beforetheywill germinate.This is also probablyan
adaptationto greaterseed productionand thusthe maximizationof the total seed crop duringthe relativelyshortlifespan
ofthedisturbedhabitat.The weedycolonizingspeciesmustput
a great deal of energyinto the productionof seeds. Only by
maximizingthe potentialfordispersalcan this type of plant
survive.
are not limstorage,and controlovergermination
harvesting,
ited to the relationshipbetweenplants and man. Similarly,
of the plant throughcultivationis
modification
morphological
withman.Numerousspecies
to plantsinteracting
notrestricted
of insectshave establishedsymbioticcultivationrelationships
withfungi.More than40 speciesofbeetles,knownas "ambrosia
beetles,"as wellas somewoodwaspscultivatefungusas sources
of food.These insectsboreinto,but do not directlyeat, wood
and otherplant tissue.Instead,theyinoculatetheplant tissue
withsmallamountsoffungus.This fungusis usuallycarriedby
the insectsto the hostplant in specializedorgans,mycetangia.
organsin the various
Mycetangiaare derivedfromdifferent
groupsof ambrosiabeetles,and we may thusassume that the
in originevenforthis
habitoffunguscultivationis polyphyletic
groupofinsects.In all cases,fungusservesas thechiefsourceof
stagesoftheinsect.
foodforall thedevelopmental
The fungithat are cultivatedcome frommany different
generabut can be broadlyclassifiedintotwo groups.Some are
knownoutsideof the symbiosis,but the majorityare highly
specializedand are not knownto existindependently(Baker
1963). These fungiare generallyhard to classifybecause they
do nothave a truesexual cycle,insteadbeingadapted to vegetativedispersalsby theirrespectiveinsectvectors.3It is important to note that the funguswill not growin the host plant
withoutthepresenceoftheinsect.The insectin factmaintains
pure culturesof its associatedfungiby removingany competremoved,the
fungus.If theinsectis experimentally
ing,foreign
site is usuallytakenoverby alien growth.The insectsalso frequentlypreparespecialbedsofwoodchipsand fecesto enhance
the growthof the fungi(Francke-Grosman1967). Growing
thesefungiunderlaboratoryconditionshas provento be exComplexmedia that includevitamins,lipids,
tremelydifficult.
and amino-acidsare required.It appearsthatmanyofthefungi
are dependentuponinsectsecretionsforgrowth.Underlaboratoryculture,thefungusloses thegrowthpatternand morphologypresentin theambrosialstate.
Antsand termiteshave also developedsophisticatedcultivationsymbiosiswithfungi,someofwhichare highlycoevolved.
For example,manyspeciesoftropicaltermitesmaintainfungus
"gardens" in theirnests. Unlike most termites,these species
lack theintestinalmicroorganisms
necessaryforthedigestionof
cellulose.Instead,theircarefully
nurtured
fungusprovidestheir
foodsupply (Trager 1970). It is again among ants, however,
that we findthe mosthighlydevelopedcultivationanalogues
(Weber1966,Hartzell1967,Hockeng1975,Trager1970). Cultivatorantspreparespecialbeds,generallyofplantdebris,cut-up
in special chambersin the ant
and excrement,
leaves, flowers,
nest.The ants are meticulousabout growthconditionswithin
thechamber;numerousventilationpassagesare dug,and these
and humidare openedor closedto regulateboth temperature
ity.To constructthebeds,theants chewthesubstratematerial
to make a pulpymass and depositit in layersin the chamber.
The bed is thenplantedwithpropagulesfrompreviouslymaintainedbeds. Constantcare is giventhe beds. The ants remove
alien fungiand add anal and salivary secretionswhich apon thegrowthofthefungi.These
parentlyhave a positiveeffect
cultivationactivitiesencouragetheproduction,by the fungus,
ofsmallwhitishroundbodies,theso-calledkohlrabistructures.
These structures
are theprincipalfoodoftheant colony.While
underthecare oftheants,thefungusproducesno sexualstage.
mushroom
If the ants are removed,however,the reproductive
ofremovalof
however,theeffect
mayappear. More frequently,
of the beds by alien fungiand other
the ants is contamination
organismsto thequickexclusionoftheant-fungus.
It is clear thatagriculture
way,be
cannot,in any meaningful
If we tryto findany
interaction.
restricted
to thehuman-plant
activityor collectionofactivitiesto definethe man-plantrela-
The similarities
betweentheant-dispersed
colonizingspecies
and the cultivatedplant are ratherstriking,
but theyare not
unexpectedifwe recognizethat theyare the resultsof a commonprocess.The problemsconfronted
by the cultivatedplant
identical
duringtheearlyphasesofdomestication
are essentially
to thoseofany otherplantmaximizing
itspotentialas a weedy
colonizingspecies.Total crop of seeds must be maximizedto
allow forthetransportofthe plant to a newlyopened habitat.
However,the creationof a new habitat is diagnosticforadvanced agriculturalsystems.It is withinthis habitat that
I It would be hard to hold that this is necessarilya resultof the
actionssuchas weedingbecomeimportant.
association;thereare numerousfunginot associatedwithinsectsthat
Morphological
changethrough
cultivation.
Clearly,
protection, possessno sexual stage (the so-calledfungiimperfecti).
754
CURRENT
ANTHROPOLOGY
tionship,we are at the same timedefiningotheranimal-plant
To say that man has a different
relationships.
attitudeor approachto the cultivatedplant seemsinsufficient
to distinguish
the human-plantrelationship
fromtheseothers.Further,even
if we grant a qualitative difference
in agriculturalbehavior
betweenman and otheranimals,it seemsunreasonableto assumethatman was necessarilyas consciousofthe evolutionof
agriclultural
systemsin the earlieststages of plant domesticationas he is today.
THE EVOLUTION OF DOMESTICATED
MORPHOLOGIES
Thedomesticated
plant.As we have just seen,domestication
is a
particulartype of plant-animalinteractionwhichfavorsthe
evolutionof certainmorphologicalchangesin the plant. The
is also characterized
relationship
by specificsetsofbehaviorby
the animal. It is necessaryto stress,however,that domesticationdoes notevolveorthogenetically
or inevitablyintoagriculture.Indeed, the degreeof developmentof the domestication
relationshipmay vary amongstthe total set of plants with
whichan animalmay be interacting.
Differentdegreesof domesticationmay be seen in manyof
thecropswhichare cultivatedby man. Sincedomestication
is a
naturalevolutionary
process,thisis notsurprising;
it is a functionofthefactthattheplantis notexperiencing
selectivepressure onlyfromits relationship
withman (foran extendeddiscussion,see Rindos n.d.). For example,certain varietiesof
wheat currentlyundercultivationscattersome of theirseeds
withoutrequiringthreshing.
The rachisto whichthe seeds are
attachedis said to be partiallydehiscent.Thus the seeds are
capable of limiteddispersalwithoutthe intervention
of man.
This is especiallysignificant
whenone considersthewidelyheld
beliefthat the indehiscentrachis is the hallmarkof domesticated grains.Likewise,in manyvarietiesofoats thereis a tendencyforthe ripe grain to shatter.This is a major problem,
sinceit reducesthetotalyieldwhichcan be recoveredat harvest
(see Schwanitz1966:34-35). Despite millenniaof interaction
betweenman and thesecrops,"wild" characteristics
have not
been completelyeliminated.In the same way, "natural" dispersal mechanismscharacterizemost of our cultivatedornamentalplants and many of our vegetable,forage,and fruit
crops (Schwanitz1966:96).
One way to understandthe great variationwhichmay be
foundin cultivatedplantsis to lookto theevolutionary
tendencies whichare favoredby varioustypesofinteractions
between
man and plant.Perhapstheeasiestway to approachthisproblem is to considerthe interactions
fromthe perspectiveof the
organsof theplant thatare utilizedby man. (These responses
are not restricted
to human-plantinteraction,
but are possibilitiesin all herbivore-plant
interactions.)
Threemajorclassesof
plant organsmay be recognized:vegetativestructures,
reproductivepropagules,4
and vegetativestructurespossessingsecondaryadaptationsas dispersalpropagules.
In one type of herbivore-plant
interaction,the animal consumes vegetative
structures
of the plant: roots,stems,petioles,
leaves,and flowers.(Flowers,whilestrictlysexual organs,may
be consideredvegetativeorgans until the fruitripens.) The
generalevolutionarytendencywhenan animal feedson these
parts of a plant is towardsnonedibility;
the potentialforthe
developmentof a cultivationsymbiosisis ratherlimited.Typical evolutionary
sequencesincludethedevelopment
ofchemical
defensessuch as poisons and tanninsand the elaborationof
physical deterrentsagainst predation such as trichomes,
sclerids,and thorns.Amongannual plants, a pressureexists
towardsthe evolutionofprecociousflowering,
whichdecreases
the timeduringwhichthe plant existsand thusis susceptible
to consumption.Anotherresponseof the plant is colonization
I A propaguleis any plant
organadapted to the sexual or vegetative reproductionof the motherplant.
Vol. 21
N 6 * December
No.
1980
Rindos:ORIGINS
AND SPREAD OF AGRICULTURE
in regionsof highlydiversified
florasin whichthe individual
plant becomesless conspicuous.It wouldthusseemreasonable
to assume that plants whosevegetativeorgansare utilizedby
man were domesticatedafterprotectivebehaviorshad begun
to be practiced.At such a time,plants not possessingdefense
mechanisms
mightbe sparedbecauseoftheirdesirability,
and it
wouldbe possibleforthemto pass thenonprotected
stateon to
succeedinggenerations.
Whenan animal feedsupon the reproductive
propagulesof a
plant,thepotentialforthe developmentofdomestication
symbiosesis great.The plantmaybe able to utilizethepredatoras a
dispersalagent.The developmentand elaborationof dispersal
symbioseshas been extensivethroughout
nature.Confirmation
of the importantrole of the reproductivepropagule in the
developmentof domesticationsymbiosesmay be seen in the
typesofcropswhichman generallygrows.Thereare numerous
plants whose vegetativeparts are highly defended,either
mechanicallyor chemically,but whosereproductive
propagules
are edible.This edibilityis frequently
advertisedby the developmentof conspicuouscolorationwhen the propaguleripens.
Of course,as would be expected,this advertisingand defense
syndromeis not restrictedto cultivatedplants. It is also extremelywidespreadin trulywildplants(thatis, thosethathave
establishedrelationships
withanimalsotherthanman).
Thereare twobasic modesin whichplantsadapt to theproblemsand opportunities
ofan animal-mediated
dispersalsystem.
The firstand mostcommoninvolvesthe elaborationof an attractiveaccessorystructurefor consumptionby the animal.
The accessorystructureis "expendable"tissuewhichservesto
attract the dispersalagent. The propaguleitselfis variously
protectedand carriedpassivelywith the accessorystructure.
Many of the propaguleswhichwe call "fruits"representthis
modeofadaptation.The fleshofthefruitservesto attractand
feedthe animal dispersalagent,whilethe actual reproductive
unitsare inedibleand are discarded,and thusdispersed,after
thefruitis consumed.The seedsin stonefruitssuchas cherries
and plumsand the corewhichcontainsthe seed in apples and
pears are dispersedin thisway. The abundanceof feralapple
trees growingin hedgerowsthroughoutthe northernUnited
States and the feralpeaches of the southare evidenceof the
effectiveness
ofthismodeofdispersal.
The otherbasic mode of plant responseto potentialanimal
distribution
agentsmaybe describedas aggregation.
It involves
the collectionof propagulesinto a largerstructure.Here the
propagulesthemselvesare consumed,and dispersaloccursdespitepartiallossesfromthetotalpropagulepopulation.The indehiscentrachisof the small grains,the ear of maize,and the
heads producedby sorghumsmay all be seen as examplesof
this type of attractiveaggregation.The developmentof large
seed massesin the New Worldcucurbitsis a further
example.
Accordingto Whitakerand Bemis (1975), theseplants were
originally
valuedfortheirseeds.Selectionforhighseed number,
and thusforlargerfruits,allows forsecondaryadaptationsby
theplant.Anaccessorystructure-thefleshofthegourditselfbecomesthemeansby whichtheplant attractsman. Even the
indehiscentlegumeof the cultivatedbean may be seen as an
example of this tendencytowardsaggregationin cultivated
plants. Aggregationand the developmentof accessorystructuresmay be describedas responsesto thepredatorybehavior
ofanimals.Eithercan occurwithoutthepresenceofdeveloped
agriculturalbehaviors.
McKey (1975) has recentlyreviewedthe coevolutionof
tropicalfrugivorous
birdsand the plants on whichtheyfeed.
These plantshave adapted to specializeddispersalby a limited
numberof species of birds.Their fruitsare generallyhighin
proteinsand lipidsand thusprovidea verynutritiousdiet for
the birdswhichfeedupon them.It is assumedthat the birds
provide the plant with a "higher-quality,
more 'predictable'
755
dispersal."This seems reasonable,because only one or a few
species may feed on any plant. Thus dispersalbehaviorwill
show less variationthan the total set of dispersalbehaviors
exhibitedby all speciesofbirds.
The various morphologicalmodifications
exhibitedby the
birdsare interesting
in and of
fruitsdispersedby frugivorous
themselves.They also show a type of adaptation which is
of the evolutionof
especiallyimportantforour understanding
cultivatedplants. The fruitsdispersedby frugivorous
birds
generallypossess morphologicalor biochemicalmodifications
whichpreventdispersalsby birdswhichhave not coevolvedto
are most comfeed upon them.Morphologicalmodifications
mon. Generallythe specializedfruitsare too large formost
birds of the tropicshave
birds to ingest.Many fruit-eating
specializedguts whichare adapted to the handlingof these
fruits
large fruits.Other types of specialized bird-dispersed
seedswhichare groundup in thegizzards
produceintoxicating
of most otherbirds,producinga negativeexperienceto disof the fruits.Thus not onlydoes
couragefurther
consumption
whichencouragethe relationcoevolutioncause modifications
ship betweentheplant and its dispersalagent,but othermorphologicaland chemical traits may develop which serve to
excludedispersalsby otherthanthecoevolvedagent.
Certaintraitsof cultivatedplants havinga symbioticrelationshipwithman mayhave originally
arisen,and are certainly
in part maintained,as means forthe exclusionof nonhuman
dispersalagents.The largefruitsofthecultivatedapple,mango,
and peach are effective
meansforexcluding
pear, watermelon,
most dispersalagents otherthan man. Of course,birds may
peck at apples or stonefruits,but theyare generallyincapable
of carryingoffthe fruitand thus actually dispersingit. The
indehiscentrachis of the small grainsand the cob of maize
serveanalogousfunctions
in discouraging
nonhumandispersal
agents. As was implicitin my remarksabout the chemically
defendedfruitsof certainbird-dispersed
speciesof the tropics,
a certainamountof outsideinterference
withthe systemmay
occur,but overevolutionary
timeits effects
are minimal.Thus
it maybe seenthatcertaintraitsofthecultivatedplantare not
solelyforman'sbenefit,
butalso servetheplantin itsdispersals.
The major evolutionaryeffectof thistypeof specializationin
the relationship
betweenplant and animalis an intensification
of thesymbioticrelationship.
An animal's consumptionof asexual propagulesalso allows
foran adaptive responseon the part of theplant. Analysisof
this typeof symbiosis,however,presentsmoreproblemsthan
the analysisof the evolutionof domesticatedplants whosereproductivepropagulesare consumed.The firstissue which
must be clarifiedis the distinctionbetweenreproductiveand
storageorgans.The domesticationof crops such as manioc,
turnips,and carrotsis best consideredwithinthe schemepresentedearlierforthedomestication
ofplantsin whichthevegetativestructureis consumed.These plantsare not propagated
by the structuresthat are generallyconsumed.However,the
potentialof the top of the carrotor turnipto continuegrowth
and eventuallyproducea seed stalkcannotbe discounted.The
growingpoint of these roots is generallydiscarded (at least
today) in preparation.The discardingof thesevegetabletops
on a rubbishheap mightallow forseed productionand thus
colonizationofthearea by theplant.
Plantssuchas thepotato,in whichtheorganconsumedhas a
natural dispersalfunction,are capable of developinga symbioticdispersalrelationship.
The gatheringof the tuberswith
subsequentinadvertentlosses wouldpermitthoseplants producingthemostacceptablefoodto humansto spreadpreferentially.Since thepropagulesare asexual,and thusreproducethe
parentalphenotype,desiredcharacteristics
oftheplantswould
be establishedin thecropmuchmorerapidlythanwhensexual
propagulesare beingdispersed.Plantswhoseasexualpropagules
are utilizedhave an advantage that is only approachedby
plants whose sexual propagulesare apomictic:any particular
756
fixedifit is successfulin the
formis immediately
morphological
is elimsymbiosis,since variationfromsexual recombination
inated.
terms,the mostlikelyresponseof a
In generalevolutionary
plantlike thepotato to thepredatoryactivitiesofan animalis
ofsmalltuberswidelyspaced fromtheparenthedevelopment
tal plant.Individualspossessinglargetuberslocatedcloseto the
motherplants would be undera negativeselectionpressure,
predatorcouldeasilyretrievethem.If we
sincea nondispersing
during
assumethatno plantingbehaviorby man was occurring
the early stages of potato utilization,it would seem that the
"desirable"typesofpotato wouldexperiencea loweredchance
of the naturaldispersalsof wild
ofsurvival.An understanding
the evolutionof the
potatoes could help us in understanding
possiblethatpotato tubers
cultivatedplant.It is theoretically
werealreadycoevolvedwithanotherdispersalagentand man
was able to "steal" the cropfromthatanimal.
The domesticationof the potato is furthercomplicatedby
of potato plants also occurs
the fact that sexual reproduction
withgreat frequencyin the regionto whichtheyare native.
Most of the potatoes growingin the Andes set fruitsabundantly.This may be more than a climatic response.Brush,
Carney,and Huaman (1980) reportthatthepresenceofvolunteerand feralpotatoescontributesgreatlyto the diversityof
thepotato genepool in its nativehabitat.Fruitingabilitymay
selectionforthe trait,since fruitbe a resultof unintentional
ing potatoesmay be morelikelyto colonizefields.This is because a seed willalmostcertainlybe overlooked,whilea tuber
willlikelybe harvestedand thusbe unavailableforgrowththe
nextseason. Fruitingabilitymay also be a relicofthe original
domesticationof the crop. It is possiblethatpotatoesbearing
desirabletuberswere originallycarefullydug and that reproduction of the desirable types occurredby the inadvertent
plantingof the fruitsin disturbedgroundthat occurredas a
activity.Sinceconsiderabledisturbance
resultoftheharvesting
ofthesoiloccursduringthisprocess,thefruitwouldbe situated
in an ideal positionforfuturegrowth.Competitionfromother
plants wouldalso be reducedbecause of thisdiggingactivity.
theevolutionof
in understanding
An additionalcomplication
cropslike the potato is the factthat theymay be reproduced
froma portionof the tuber.Thus, as in the case ofthe carrot,
thediscardingofa sectionofthetubercouldallowforcolonizaall ofthecomplicationofthecropin dumpheaps. Considering
investigation
tionspresentedby a croplike thepotato,further
of the behaviorsof the plant in the wild and in its regionof
of
could do muchto advance our understanding
domestication
domesticationas an evolutionaryprocess.Such investigation
ofthetypes
couldbe especiallyimportantforan understanding
of selectivepressurewhichmay have existedbeforeplanting
behaviorsby man werewellestablished.
Independentsupportfora coevoluWeedsand domestication.
may be foundin theexistenceof
tionaryviewofdomestication
weeds specificallyadapted to the conditionsof cultivation.
Weeds showmanyofthe traitsofcultivatedplantsdespitethe
factthattheyare clearlynotbeingselectedconsciouslyby man
forthesetraits.Weeds may be analyzedin thesame manneras
thattheyare "paracultivatedplants,but withtherecognition
sites" upon thehuman-plantsymbiosis.
Wickler(1968) gives several examplesof the evolutionof
weeds under cultivation.He notes that the weed Camelina
sativahas divergedfromits nonweedancestorCamelinaglabra
by virtueof the selectiveprocessesoperatingin flaxfields.C.
The
sativais largerthanC. glabraand generallynonbranching.
to flaxleaves
resemblance
leaves ofC. sativaalso bear a striking
in both shape and color. Clearly,this resemblanceserves to
protecttheplant frombeingweededout of theflaxfields.The
to
fruitof the weed is also indehiscentand requiresthreshing
separate the seed fromthe husk. While the seeds of C. sativa
fromthoseofflaxin bothsize and weight,thecombination
differ
ofthesetwofactorsis such thattheycannotbe separatedfrom
CURRENT
ANTHROPOLOGY
flaxseedsby winnowing.
C. sativais, in essence,a mimicof flax
in bothvegetativeand dispersalphases ofitslifecycle.
In a similarmanner,theagriculturaltechniquesofman have
selectedstrainsof dodder (Cuscutaepilinum),whose mimicry
offlaxoccursonlyduringthedispersalstage.Weedystrainsof
dodderproduce a double-seededfruitwhichapproximatesa
flaxseed in size and weight."Wild" strainsof doddercan be
but the weedy
removedfromthe linseedcrop by winnowing,
strainsare dispersedwiththe crop.
Teosinte so mimicsthe vegetativestructureof maize that
cannot distinguishbetweenthe plants
native agriculturalists
untiltheflowering
stage.At thispoint,ofcourse,thelargesize
oftheplantand thelate stagein thegrowingcyclepresentgood
reasonsnot to botherremovingit; removalwouldbe pointless
or perhapseven damagingto adjacent maizeplantsbecause of
rootdisturbance.Thus teosintehas adapted in sucha way that
man unintentionally
protectsit. Many believethat maize and
teosintehave a commonevolutionaryhistory(Galinat 1974,
1975,Doggett1965,Mangelsdorf1974). If thisis thecase, then
we mayviewthehistoryofthesetwoplantsas one ofdivergent
evolution.Two paths developedin a partiallysharedgenepool
by adaptationto different
aspectsofthehuman-plant
relationship.Both teosinteand maize have takenadvantageofman as
a protectiveagent.As longas thecropwas an incidentaldomesticate,man may have dispersedboth of them,but once maize
became obligatelydispersedby man it was dispersedwithout
theteosinteportionofthegenepool. Thus themaizecob served
to allow maize to divergefromboth its originalhome and its
closest relative,teosinte.In the area in which teosintewas
incidentallydispersed,however,man continuedto protectit,
eventhoughtheintentoftheprotection,
at leastat a laterdate,
was directedtowardsquitea different
plant.
Isolationand domestication.
The divergenceof a cultivated
plant fromits wild ancestorrequireseffective
isolation.Introgressionbetweenthe wild and the domesticatedpopulation
would continuallyoppose the developmentof the cultivated
plant. Isolation throughthe activitiesof differing
dispersal
agentshas alreadybeenmentioned.
I wouldnowliketo consider
spatialisolation.In populationgeneticsthisprocessis generally
referred
to as allopatricspeciation.
Strongevidenceexists that certaincultivatedplants have
becomedomesticatedoutsideof the regionto whichtheyare
indigenous.Two examplesare the sunflower
and the tomato.
The sunflower
(Helianthusannuus) is almostcertainlya native
ofthewesternUnitedStatesand providedan importantsource
ofwildfoodto theearlyinhabitantsofthatregion."It has been
postulatedthat,in time,thesunflower
becamea camp-following
weedand was introduced
fromthewesternto thecentralUnited
States. Somewherein the latterarea the sunflower
appears to
have been domesticatedand, as a domesticatedplant,was carried both eastwardsand to the southwest"(Heiser 1976:37).
The tomatois apparentlydescendedfromwildspeciesofLycopersiconindigenousto theAndes.Yet withintheAndesthereis
no evidencethatthetomatowas knownas a domesticated
fruit.
Instead,all of the evidence-linguistic,botanical,biochemical,
and historical--pointsto the area of domesticationfor the
tomato as having been Central America and Mexico (Rick
1958, 1978).
Since in naturebothof thesecropsare stronglyoutcrossing,
spatial isolationofthecultivatedpopulationwas a prerequisite
formorphological
divergence
ofthecropfromthewildancestor.
While the sunfloweris still stronglyoutcrossing,cultivated
strainsof the tomatohave becomelargelyself-fertilizing.
This
enhancesthepotentialforthe evolutionof the cropas a cultivated plant,since variationis greatlylimitedin self-fertilizing
plants.Self-fertilization
as a meansofisolationfordomesticated
plants is extremelycommon;it may be seen, forexample,in
peas, beans,peppers,and certainofthesmallgrains.
Nondomestication.
Yet anotherindependent
sourceofsupport
forthecoevolutionary
modelforagricultural
originsis thenonVol. 21 * No. 6 * December1980
Rindos:ORIGINS
AND SPREAD OF AGRICULTURE
domestication
of certainfoodplants. Clearly,manymorefood
plantsare available to man thanhave beendomesticated.Why
should certain crops become domesticatedwhile other, apparentlyequally valuable crops remainwild? As a paradigm
forthisproblemI wouldliketo briefly
considertheacorn.
Acornshave been used as botha stapleand a majoralternative source of food by many culturesthroughoutthe world.
While the acorn-basedculturesof the west coast of North
Americaare perhaps the best-known,
acorns have also been
extensively
utilizedas humanfoodand as fodderin bothEurope
and theNear East. Indeed,it is likelythat theacornhas been
utilizedwhereverit is present.It is a large,conspicuousfood
sourceand one thatneedsonlyminorprocessing(suchas leaching) beforecooking.The acornsof certainspeciesof oaks are
palatable withoutany processing.In manywaystheoak seems
ideallysuitedto beingone of thebasic cropsofan agricultural
civilization.
Since theoak has a longlifecycle,and thusa longtimemust
pass betweengenerations,
it wouldseem thatit wouldbe difficultto domesticate:theresultsofselectionwouldbe too longin
appearingto encouragefurtherconsciousselection.Othertree
crops,however,suchas olives,thevarioustreefruits,
dates,and
numerouspalms,have developedinto domesticatedplants despite theirlong life cycles. Anotherreason which mightbe
advanced for the oak's nondomestication
is that, since it is
capable of providinga large and reliableharvest,thereis no
reasonto select"improved"cultivatedforms.Yet, wildstands
ofwheatsin theNear East are quiteas productiveas theolder
varietiesof wheat (Harlan 1967). Thus productivity
does not
seemto be necessarilya controlling
variable.
Within the contextof a coevolutionarymodel for plant
domestication,
however,we may considerthe possibilitythat
thenondomestication
oftheoak had littleto do withits potential as a food source forman. Looking at the relationships
already existingbetween oaks and otherpotentialdispersal
agents in the environment,
we may hypothesizethat divergenceof the oak intowildand cultivatedpopulationswas preventedbypreexisting
coevolvedrelationships
betweenoaks and
variousotheranimalssuch as squirrels.Squirrelsnot onlyharvest acorns,but also plant them.We may call the planting
behaviorof squirrelshoardingand considerany plantingaccomplishedmerelyincidental,but as faras thepropagationof
the oak is concernedthe distinctionis meaningless.It would
have been verydifficult
forman to upsetthesquirrel'sagriculturalrelationship
withoaks. Furthermore,
it is hardto imagine
a morphologicalchange in oaks that would exclude squirrels
fromfeedingupon the acornsand thus dispersingthe plants.
Withoutexclusionof otherdispersalagents,domesticationdid
not occur.
AGRICULTURAL ORIGINS AND DISPERSALS
INTRODUCTION TO THE MODEL
Domestication,as viewedhere,is an evolutionary
processthat
is the result of predator/preyinteractions.Domestication
oforganand distribution
changesthemorphology,
physiology,
isms.It resultsin a mutualism-a relationshipwhichbenefits
geneticallyunrelatedorganisms.It is, however,neitherineviThe treatment
thatfollowsis proneto a
tablenororthogenetic.
fatalerrorin interpretation.
I shall be providinga modelfora
certainchangein subsistencepattern.The modelis a dynamic
one and places great emphasison feedbackprocesses,but it
shouldnot be read to condonethe view that theprocess,once
begun, must proceed inexorablyto certain ends. There are
numerouswaysin whichtheprocessesleadingto domestication
and agriculture
may be subverted.
757
In a
Likewise,thismodeldoes notpretendto be all-inclusive.
sense, I am tryingto followa particularthread of cultural
developmentsbackwardin time; I am not attemptingto describethefabric.Thus, muchofthevariationwhichmayoccur
Little
in humanculturesmustbe excludedfromconsideration.
attentioncan be givento societieswhichwereagriculturalor
domesticatory
at certaintimesbut laterabandonedthe behavior. Clearly,it would be extremelyusefuland interestingto
deal with such phenomena,but it would go far beyond the
boundsofthisessay.
Finally,we mustbeginto refinetheconceptofdomestication
advancedearlier.I have attemptedtoprovidean understanding
so
of whydomestication
occurs,but the view of donmestication
far presentedis so broad as to be, in the finalanalysis,uninteresting.
Coevolutionary
domestication
is a processthathas
occurredin almostall culturesat mostperiodsofhumanexiswith
tence.It gains its significancefromits interrelationship
agriculturalorigins.
In orderto appreciatetheroleofdomestication
inagricultural
origins,it is necessaryto anticipatesomeofthelaterarguments
and firstpresenta definitionfor agriculture:agricultureis
characterizedby an integratedset ofactivitieswhichaffectthe
environment
inhabitedby the domesticatedplant throughout
itslifecycle.As we shallsee,themajoreffect
is to
ofagriculture
forthe doincreasethe carryingcapacity of the environment
mesticatedplant, and this increasein plants and thus plant
has majoreffects
productivity
upon humanpopulations.
Agriculture
is a level or type of behavior.Like manyother
phenomena,it is frequently
easier to recognizethan to define.
To attemptto defineit solelyon thebasis ofa certaintechnique,
such as plowingor weeding,tendsto createa falseimpression
of the importanceof any particulartechniquewithinthe integrated schedule of activitiescomprisingagriculturalsubsistence.Frequentlythisleads us intobelievingthatculturessharinga particulartechnique,suchas theuse ofa diggingstick,or
evenparticularplants,suchas wheat,musttherefore
have had
a commonorigin.
of domestication.
is an outgrowth
Thus it is imAgriculture
to definea momentoftransition.
At an early
possiblerigorously
stagein thedevelopmentofagriculture,
it would be impossible
to identify
theprotoagricultural
confidently
society.The interactions with domesticatedplants of the early agricultural
fromthoseofthedomesticasocietywouldbe indistinguishable
matorysociety.The significant
techniquesof environmental
nipulationmightnotbe beingperformed
forconsciouslyagriculturalends.Firesencouraging
thegrowthofdomesticated
plants
mightbe beingset to drivegame;forestsmightbe beingcleared
to furnishbuildingmaterials.We could only be sure that we
wereindeeddealingwitha protoagricultural
societyifwe knew
howtherelationship
woulddevelopin thefuture.If we need to
define,for any particularsociety,the momentof transition
froma domesticatory
to an agriculturalway of life,I would
suggestthat it is the point at whichagriculturalinteractions
became more importantto the society than domesticatory
interactionswith plants. "Importance" may be read in two
ways.We may considerit to representtherelativesignificance
of agriculturalbehavior to the futuredevelopmentof the
thateach form
society.We mayalso viewit as thecontribution
of behaviormakes to the overallsubsistenceof the societyat
is of criticalimany givenmoment.This latterinterpretation
theevolutionand subsequentspread
portancein understanding
ofagriculturalbehavior.
0~~~~~~
Population
Solidline,totalfood;
relationship.
1. The typicalpredator/prey
broken
line,yieldper predator;a, pointat whichcompensatory
or all availableregions
behaviorof predatorbecomesineffective
ofpreyarefilled.
togrowth
favorable
FIG.
betweenpredapredatordecrease.In a mutualisticrelationship
torand prey (fig.2), as the numberofpredatorsincreasesthe
totalamountofavailablefoodalso increases.The energeticcost
to thepreyinherentin providingsubsistenceto thepredatoris
compensatedforby actions of the predatorthat provide inforthe survivalor dispersalof the prey.
creasedopportunities
Seen in otherterms,mutualismsserveto increasethe carrying
forboththepredatorand theprey.
capacityoftheenvironment
The actionsof thepredatorwhichtendto increasethenumber
of the prey indirectlypermita greaternumberof predators.
Domesticationtendsto increasethe yieldofpreywithinthe
area inhabitedby thepredator.Ofcourse,thistendencywillnot
Numerousfactorsmay act to limitthe
continueindefinitely.
numberofpreyand thusofpredator.Eventuallythecompensaor all
tory behaviorof the predatorwill become ineffective
available regionsfavorableto the growthof the prey will becomefilled;thusthe yieldwilllevel offor even assumea negative slope. Yet thisstabilizationin yield,and thusin predator
numbers,is also subjectto change.Most ofthefactorslimiting
variablesand
thesuccessofthepreyspeciesare environmental
are thus external to the mutualisticrelationshipbetween
predatorand prey. Since few,if any, species are naturally
limitedin numbersby theirintrinsicrates of increase,any
will permitincreasesin the
"relaxation"of the environment
numbersof the prey species. Changes in the environment
for
directlyincreasethe carryingcapacity of the environment
permitgreaternumbersof predathepreyand thusindirectly
tors.
ofthepreyspeciesmay
The changethatleads to proliferation
/
__
~~~~~//
/
DOMESTICATORYRELATIONSHIPS
Predator/prey
i-nteraction.
Mutualisticrelationships
such as doPopulation
mesticationhave thepotentialto changetherelationship
which
existsbetweenpredatorand prey.In thetypicalpredator/prey FIG. 2. A mutualistic
Solid line,total
relationship.
predator/prey
relationship(fig.1), as the total numberofpredatorsincreases food;broken
change
ofeffective
b, moment
line,yieldperpredator;
ofthepreyspecies.
leadingto proliferation
both the total and relativeamountsof food available to the
758
CURRENT
ANTHROPOLOGY
be a changein the morphology
ofthepreyor a modification
of
thebehaviorofitscoevolvedagent,includingchangein theway
theagentrelatesto the environment.
At a timebeforeagriculturalbehaviorswerewell established,changesin the morphologyofplantswereprobablyofmajorimportanceforthefurther
developmentof the human-plantrelationship.Traits such as
indehiscence,gigantism,and aggregationcould exclude previouslyeffective
dispersalagentsand thus allow fora greater
available yieldto humans.This changein theplants' morphology would enhance the developmentof the human-plant
mutualism.
As we have just seen,the
Domesticates,
yield,and instability.
majoreffect
ofthemutualismbetweenmanand plantis greater
of the environment
forman. This increaseis not,
productivity
however,based on equal increasesin productivityof all componentsof the environment
(all potentialor actual preyspecies). Rather, a very small subset of potentialfood sources
providesall of the increasein productivity.As domestication
proceeds,so does man's relianceupon an eversmallersubsetof
potentialfoodsources.Increasein theavailabilityofa resource
is accompaniedby a corresponding
increasein its utilization.
The effectof thisdietaryshiftis a reductionin theimportance
of nondomesticates.
This simplification
of the subsistencepatternhas majoreffects
upon its stability.
To understandtherelationship
betweenyieldand instability
in yield,we mayfirstconsidertherelationship
betweenmanand
any one domesticatedplant. The increasein yieldthat results
is onlythe averageincreaseovertime,that
fromdomestication
is, as the mutualisticrelationshipdevelops.Any given plant
whichis subjectto periodicand norgrowsin an environment
and its yield will vary about the mean from
mal fluctuation,
Thus the contribution
year to yearbecause of thisfluctuation.
of any particularspeciesofplant to the totalplant foodavailable to man will vary in responseto environmental
factors.
Whileaverageyieldis increasingovertimebecause oftheelaboration of coevolutionarydomesticationrelationships,the
absoluteyield at any given momentis a functionof specific
environmental
conditions.
beforethedevelopDuringtheearlystagesofdomestication,
mentof agriculturalsystems,this fluctuationin yield should
have littleeffecton the total yieldavailable to man. The contributionof any givendomesticatedplant to the total subsistence patternwill be relativelysmall. Membersof any given
domesticatedspeciesare likelyto be growingin any and all of
the locationswhichare conduciveto theirgrowth.Also, the
variousspeciesof plants withwhichman has establishedmutualisticrelationships
are likelyto be fairlyuniformly
distributed throughout
the environment,
no particularnichehaving
establisheditselfas the principal one for all domesticated
plants.Thus,whileenvironmental
fluctuation
willhave adverse
effectson the yieldof any particularspecies,the overallyield
will tend to remainreasonablyconstant.A bad year forone
plant is likelyto be a good yearforanotherplantgrowingin a
different
habitat;a bad yearforonespeciesis likelyto be a good
yearforanotherspecieswithdiffering
edaphicand physiological
This compensationin yield will permitproducrequirements.
tivityto remainrelativelystablefromyearto year,muchas it
does in any nondomesticatedecology.As we shall see, this
standsin starkcontrastto theinteraction
ofyieldand environmentwhichoccursin theagriculturalsetting.
AGRICULTURAL ORIGINS
tendencies.
The primaryeffectof agricultureon a
Evolutionary
ofthedependenceofthatsocietyon
societyis an intensification
domesticatedplants.Highlydevelopedagricultural
systemsare
based upona limitednumberofcultivatedplantswhichprovide
thebulk ofthesociety'sfood.Fromtheprecedingdiscussionit
should be clear that this tendencytowardsa decreasein the
number of niants nrovidinz subsistence is an ouutgrowth of the
Vol. 21 * No. 6 * December1980
Rindos:ORIGINS
AND SPREAD OF AGRICULTURE
dynamicsof the mutualisticrelationship.There is a tendency
towards increase in yield and concomitantdecrease in the
numberofspeciesofdomesticatedplants providingthat yield.
places the major
It has also been notedthat the environment
limitationupon thedevelopmentofthisrelationship.
is composedof various niches forplant
Any environment
colonization.These nichesare welldefinedand relativelystable.
Thus the potentialnumberof places in whichany givenplant
oftheenvironby thecharacteristics
mightgrowis determined
ment. Clearly, any behavior by man that transcendsthese
on plantgrowthwillbe accompanied
limitations
environmental
by increasesin yield proportionalto the amount of "new"
nichespace. At the same time,any behaviorincreasingniche
size available forplant colonizationwillcause a corresponding
forman.
increasein the carryingcapacityof the environment
changeswhich
Any humanbehaviorcausingenvironmental
successof the domesticated
increasetheprobablereproductive
plant, and thus its yield,will have importanteffectson the
furtherevolutionof the domesticatedplant. Withinany locality,severaltendencieswill guidethe evolutionof the cultivated crop, among them (1) a reductionin diversity,both
geneticand phenotypic,(2) a tendencytowardsincreasedproAt thesame time,
and (3) autecologicalconvergence.
ductivity,
a seriesof tendencieswillbecomemanifestwithinthe total set
thedomesticatedecology,amongthem(1)
of cropscomprising
a tendencytowardsreductionin the numberof species on
whichman reliesforhis subsistence,(2) increasein total crop
As we shallsee, these
yields,and (3) autecologicalconvergence.
tendencieswill formthe basis forthe elaboration
interrelated
systems.
and dispersalofagricultural
antdyield.Changesin thedirectionofplant
diversity,
Systems,
modifiedecologyare
evolutionwithinthe early agriculturally
based on one rathersimplefactor:human agriculturaltechforplant growthand reproducnologycreatesan environment
is structurally
tion which, unlike the overall environment,
and temporallystable.
homogeneous
As numerousauthorshave observed,agriculturalsystems
ecologies:typicallythe agriculmay be describedas simplified
turalplant is a "weedyheliophyte,"that is, a colonizerof disturbed habitats, and the agriculturalfield or garden is an
in whichthe earlieststages of ecologicalsuccesenvironment
ofthiscolonizersystem
sionsare maintained.The predominance
foragricultureis based upon the factthat thereare numerous
ways in whicha disturbedhabitat can be createdby human
behavior.Fires,disturbanceofthesoil,theclearingofforestsor
the ringingof trees,and the creationof dump heaps are all
routesby whicha disturbed,and thus open, habitat may be
created. Early domesticatedplants preadaptedto this niche
evolutionintoagricultural
willthusbe favoredin theirfurther
plants.
Variousagriculturalbehaviorsserveto reducethe intensity
ofplantsthatare necesof naturalselectionforcharacteristics
saryforsurvivalin thewild.Irrigationreducesthenecessityfor
plantsto maintainmechanismsforsurvivalduringperiodically
droughts.Clearingof the land and weedingdecrease
recurring
the importanceof competitivemechanisms.Plantingencourages specificand uniformgerminationand seedlingphysioloultimatelyreducethevariagies.The techniquesofagriculture
tion to be foundin a speciesof domesticatedplant. Both the
wild and the early domesticatedplant had to maintainpleiotropicresponsesin the face of an unpredictableenvironment.
With a reductionin overall environmentalunpredictability
withinagriculturalsystems,thisvariabilityis no longerbeing
maintainedby naturalselection.Physiologicaland autecologiplantswithinany
also occursin all agricultural
cal convergence
givenlocality.As we shall see, thisdecreasein variabilityultimatelyyieldsan increasein vulnerability.
At thesame time,theselectionforhigheryieldsjust discussed
759
will continueand even increase.Plants yieldingthe greatest
numberof propagules-thoseplants best adapted to the agricultural environment-willbe most likely to survive and
spread. The relaxationof selectionpressuresfortraitssuch as
willalso bringabout a new way forplants to
competitiveness
increasetheiryield. Energythat was previouslydivertedto
such tasks as protection,the manufactureof long internodes,
and the like may now be utilizedfor
perenniating
structures,
furtherincreasesin yield. Released fromthe requirementto
the plant whichputs energyinto the
possesssuch structures,
productionofpropaguleswillbe themostfitwithinan agriculturalsystem.
Processesanalogous to those occurringat the intraspecific
levelwillalso occurat theinterspecific
level.Those plantsbest
withinthe agricultural
adapted to survivaland reproduction
ecologywill come to dominatethe system.Given the homogeneityand stabilityintroducedintothe environment
by agriculturalbehavior,it is not unreasonableto expecta reduction
in species diversity-an increasingreliance upon fewerand
fewerspeciesto providethebasis forhumansubsistence.
thistendencytowardsuniAnotherfactorwillalso reinforce
of developingagriculturalsystems.
and simplification
formity
As a food source becomes more common,feedingon it will
increase.At firstthiswill be merelya functionof availability,
but as timepasses techniquesof productionand consumption
will tend to improve.Availabilityand efficiency
interactin a
positive-feedback
manner,and furtherspecializationis likely.
Specializationwillalso be encouragedby the localizationof
agriculturalproduction.Reductionof usage will decreasethe
probabilityoftheestablishment
ofcoevolutionary
relationships
betweenhumansand new species of plants. As more time is
man willcomeintoconspentin theagricultural
environment,
tact with fewerspecies of alternativefood sources. These
changes in time allocation, generallyknown as scheduling
changes,decreasetheimportanceofenterprises
competingwith
agriculture.Thus the establishmentof agriculturewithina
society will intensifyand direct tendenciesalready existing
underthedomesticatory
way oflife.
The development
ecologies.Agriculture
creates
of agricultural
a new type of climaxformation.The agriculturalfloratends
towardsstabilityas longas humanbehavioris interfering
with
othersuccessionalprocesses.Many ofthespecieswhichinhabit
the agriculturalclimaxare derivedfromplantsadapted to the
earliest,colonizing,stage of ecologicalsuccession.Agricultural
derivedfrom"weedy" plants that tend
staplesare frequently
to existonlyin newlydisturbedground.Anothercharacteristic
of manyof the staple cropsof agricultureis that theyare descendedfromplantsthatdid not alreadyhave mutualisticdispersal systemswithany animal otherthan man. The agriculturalflorais a climaxformation
composedofhighlyspecialized
colonizerspecies.Many oftheplantsare descendedfromancestorshavinggeophysicalmodesofdispersal;secondarymodificationsare adaptationsto human-mediated
dispersal.This striking combinationof geneticheritageand evolutionarysetting
and
goes farin helpingus to understandboththeproductivity
thelimitedspeciesdiversitytypicalofagricultural
systems.
Whiletheagricultural
ecologyis bothstableand simplecomparedwiththe overallecology,it nevertheless
willshowchange
overtime.Probablythemostimportantlong-term
changethat
occurs in agriculturalecologiesis the creationof new niches
withinthe ecology.As timepasses and plants respondto the
optionspresentedby the existenceof agriculturalareas, subdivisionof the existingspace will occur.This will permitthe
entryof new, not necessarilydomesticated,plants into the
ecology.For example,weedsmay evolveto utilizeagricultural
fieldsduringfallowperiodsand then begin to be subject to
manyof the same selectiveforcesexperiencedby the domesticates.
The changesthat occur in agriculturalecologiesover time
also allow us to understaind
the evolutionof weedinessand the
760
entryof "secondary"domesticatesintothe system.Weeds are
colonizersof thisnewagriculturalhabitatjust like agricultural
betweenthemis in theattitudeman
plants.The onlydifference
has towardsthem.Vavilov (1926), amongothers,has pointed
out that the distinctionbetweenweed and domesticateis at
besta tenuousone. Secondarydomesticatesare plantsthatare
capable of establishingthemselves,like weeds,in the agriculturalecologybut thatprovideman withusefulproducts.Prime
amongtheseare plantswhoseutilityis foundin theirvegetative
parts. Edible plantswhichcan establishthemselvesin the disturbedagriculturalecologyneed not developspecial coevolved
in orderto survive.Theyneedonlyscatter
meansofdistribution
theirseed. However,since theyare growingin the same environmentwith otherearly agriculturalplants, they will be
subject to the same selectionforhighpropaguleyield. They
will also exhibitthe same tendencytowardsedaphicand autcrop.The secas any otheragricultural
ecologicalconvergence
ondarydomesticatewill evolve subject to the same selective
pressuresas theprimarydomesticate:it is evolvingin thesame
environment.
The developmentof the agriculturalecologywill,because of
the interactionbetweenit and thegeneralecology,have major
effectson the divergenceof the cultivatedplant fromits progenitorspecies.While diversitywithinthe ecologicalniche is
likelyto be, at least initially,less than that outsideof it, the
of
resultwill neverthelessbe an increasein the heterogeneity
fordisthe entireregion.This will permitmoreopportunities
thedivergenceof
ruptiveselection(Thoday 1958) and intensify
the agriculturalplant fromthe early domesticatedplant. Human interactionwith the originallydomesticatedtaxon will
withtheagricultural
decreaseas theinteraction
plant develops.
Relaxationof the amountof interactionbetweenman and his
earliest domesticatedcrops will, in the presenceof highly
developed agriculturalsystems,leave the early domesticate
"stranded."The humanagent withwhichit had developeda
coevolved dispersal or protectionsystemwill no longer be
is
as dispersalagentor protector,
and its extinction
functioning
to
almost inevitable.It is not surprisingthat it is so difficult
species" forso manyofour important
identifythe "progenitor
forextinction
have
agricultural
plants.Two majoropportunities
occurredin the historyof all primarydomesticates:(1) The
wild,uncoevolvedportionof theancestralgenepool mayhave
becomeextinctduringthe long timeperiodduringwhichcoevolutionarydomesticationoccurred.(2) The portionof the
genepool whichevolvedundertheconditionsofearlydomesticationmayhave becomeextinctduringtheperiodofintensified
evolutionwhich led to the developmentof the agricultural
plant.
DISPERSAL
OF AGRICULTURAL SYSTEMS
We have alreadynoted
andenvironmental
instability.
Agriculture
thatincreasinghumanrelianceon cultivatedplantsbringswith
it a relativedecreaseover timein the absolutenumberof taxa
producingthe major portionof the society'ssubsistence;that
fromyearto year;
taxa varyinproductivity
manydomesticated
thatconvergence
occursamongagricultural
crops;and that,by
limitationsupon the carrying
removingcertainenvironmental
fordomesticatedplants,agriculcapacity of the environment
increasein potentialyield
turalbehaviorpermitsa tremendous
and thus in potentialhumanpopulation.I would now like to
explorethe interactionof a contractingsubsistencebase, inan
crease in carryingcapacity,and variationin productivity,
have led to thespreadof
effects
interaction
whosedemographic
as a modeofhumansubsistence.
agriculture
We have discussedsomeof the consequencesof the developstableniche:thedivergence
mentofagriculture
as a temporally
the evoluof agriculturaldomesticatesfromtheirprogenitors,
tionofsecondarydomesticates,
and theevolutionofagricultural
weeds.Agricultural
ecologiesalso have locationalstability.The
CURRENT
ANTHROPOLOGY
ofagricultural
concentration
plantsin limitedregionsincreases
yieldsby allowinga morecompleteharvest.At the same time,
it intensifies
thepotentialinstability
ofthesystem,and thishas
major effectsupon the demographyof the humanpopulation
dependentupon agriculturalproduction.Locational stability
intensifies
instabilityin severalways. Prime among themare
ofmicroenvironmental
and microclimatic
theeffects
variation.
A gardenarea thatoptimizesdrainageduringaverageyearswill
be too wet duringperiods of highprecipitationand too dry
duringperiodsof drought.Otherlocationaleffectsto be consideredlater includeconcentrationof resourcesand predation
and edaphiceffects.
Domesticatedand agricultural
plantsbringabout an increase
in thecarryingcapacityoftheenvironment
contributed
by domesticatedplants, hereafterreferredto as Kdom.5 While Kdom
tendsto increasegreatlyovertime,forreasonsoutlinedabove,
the changein Kdom is not alwayspositivefromone momentto
another.The components
oftheenvironment
showseasonaland
longer-term
variation. Climate, for example,is not uniform
fromyear to year, and over relativelyshortperiodsof time
differing
average climaticconditionswill prevail. Thus, yield
willvaryin responseto theseconditions.
In the later phases of nonagricultural
there
domestication,
willbe somevariationin productivity
overtime,but variability
is restrainedby severalfactors.As we have alreadynoted,the
early domesticatedplant is growingin a varietyof microenvironments.
Thus any climaticchangeneednotaffectall ofthe
microenvironments
in thesame way. For example,domesticates
growingin sitesthatare wetterthanoptimalwillprosperduring
a period of reducedprecipitation.Nonagricultural
domesticatorysocietiesalso tendto have a largernumberofspeciesconto Kdomthando agriculturalsocieties.Thus any given
tributing
changein the environment
is not likelyto affectall speciesof
domesticatesequally.Both oftheseintrinsic
factorswilltendto
reducetheeffects
ofchangesin theenvironment
on theproductivityofthedomesticatedecology.Perhapsthemostimportant
factorlimitingtheeffectsofchangesin Kdom, however,is external to thedomestication
symbiosis.The nonagricultural
society
is relyingupon domesticatedplants foronlypart of its total
foodsupply.Thus it is possible to compensatefordeclinesin
of the total
Kdom by increasedrelianceupon othercomponents
carryingcapacity (K). The failureof one domesticatedplant,
or of one domesticatedspecies of plants,will be compensated
for by increasedreliance upon other plants, both wild and
domesticated.
In a periodof earlyagriculture,
the varianceof Kdom willbe
greaterbecause of the evolutionof agriculturalecologies.The
evolutionof the cultivatedplant in the agriculturalenvironmentincreasesnotonlyyield,but also thesusceptibility
ofthat
yieldto environmentally
inducedcropfailure.The autecological
convergence
broughtabout by agricultural
selectionbringswith
it an increaseduniformity
in theresponseofagricultural
plants
to environmental
parameters.Thus a bad year forany given
memberof a cultivatedspeciesis likelyto be bad forall membersofthatspecies.And sincetheconvergence
is also occurring
between,as well as amongst,plant species,bad years forany
givenagriculturalstaple are likelyto be bad forotherstaples
also. Increase in productivity
has been boughtat the price of
in responseto the environment.
uniformity
to and intensifying
Contributing
the effectsof autecological
convergenceis the greaterlocational stabilityof agricultural
ecologies.Localization intensifies
the effectsof microclimatic
effectsupon total yield.To take an extremeexample,a hailstormjust beforethe harvestseason willhave vastlydifferent
effects
upona societyifit fallsuponthecultivatedfieldsthanif
it fallsin the woods. Since hail is frequently
a highlylocalized
5K is widelyused to stand for"carryingcapacity." Kdommay thus
be read as " that componentof K contributedby the yield of domesticatedplants."
Vol. 21 * No. 6 * December1980
Rindos:ORIGINS
AND SPREAD OF AGRICULTURE
phenomenon,the effectof hailstormsupon the total available
than it will be for
yieldwill be far greaterforagriculturalists
or a totallynondomesticatory
society.
eithera domesticatory
ecologiesincreases
Thus localizationofresourcesin agricultural
the possibilitythat all of the resourcemay be lost to a catastrophe.
Agriculturalsubsistenceis also accompaniedby a growing
specializationin diet.Thus a decreasein theyieldfroma staple
foodsupply,
cropwillhave majoreffectson the total perceived
and techas wellas theabsolutefoodsupply.Food preferences
niques of preparationwill have placed certainfoodsourcesin
positionsof prominence.Informationconcerningthe edibility
and processingof alternativefoodsuppliesmay be lost. Thus
decreasesin yield fromcultivatedplants may create the appearanceoffoodshortageeven thoughthe total available food
supplyin theregionmaynothave fallento thepointwhereit is
actuallylimitingsurvivalof the population.Animalsrespond
to theperceivedfoodsupply,not to some "objective" measure
of total available calories.
The increasedsusceptibilityof the
The basisfor instability.
agriculturalecologyto the extremesof normalclimaticvariamanipulation
tionmay be viewedas an effectofenvironmental
manipulation
itself.Many ofthesimplerformsofenvironmental
controlrelativelyconstant,predictableparametersof the enthe effects
of the
vironment.
They increaseKdom by mitigating
on the carryingcapacity of the enfundamentalrestrictions
vironmentforthe plant. Increasingcontrolover any limiting
bringswithit, however,an increase
aspect of the environment
in the vulnerabilityof the newlyheightenedKdom to those
leftunaffected
by controlbehaviors.
aspectsoftheenvironment
As we have seen, forexample,localizationof productionincreases Kdom and thus allows fora largernumberof human
of the systemto
beings,but it also increasesthe susceptibility
effects.The
negative microclimaticand microenvironmental
removalof any givenlimiton yieldallows other,uncontrolled
limitsto becomeevident.
Agriculturepermitspreviouslynonlimitingfactorsin the
of plants to expressthemselves.For
growthand productivity
fromdroughtwill not have major
example,a plant suffering
by minorinsectinfestalimitationplaced upon itsproductivity
tion. Correctionof the droughtconditionby irrigationwill
conallow forthe expressionof the limitationon productivity
tributedby the insectpredation.Thus techniquesof environto affectyield admental manipulationallow theenvironment
in newways.At any giventime,techniquesof environversely
mentalcontrolincreasethe negativeeffectsof conditionsthat
cannotbe controlledby the system.
new opportunities
also createsentirely
for limitaAgriculture
ofa crop,
The increasing
geneticuniformity
tionsonproductivity.
reductionin species diversity,edaphic and ecologicalchanges
of recreatedby agriculturalpractices,and the concentration
sourcesin a limitedarea all contributeto new potentialinstabilitiesin productivity.
of
increasesthe susceptibility
Increasinggeneticuniformity
in the prothe crop to attacks by pathogens.Polymorphisms
ductionofsecondarymetabolitesactingas biochemicaldefenses
discouragethe evolutionof specializedpathogens.A reduction
in thisdefensestrategyencouragesthe evolutionofpathogens
whichmayseriouslydamagea plantspecies(Feeny 1973).This
potentialsourceofdamageis especiallyclearwhenwe consider
crops in whichsecondarymetabolitesdistastefulto man also
serveto protecttheplantfromattackby otheranimals.
ecology
Reductionin speciesdiversitywithintheagricultural
upon the susceptimay in and of itselfhave importanteffects
bilityofthecropplant to pathogenattack.Escape frompredationmay be aided in manyplant communities
by the associaplants;the plants"hide" frompotential
tionofmanydifferent
predatorsby beinghard to findin the mosaicof diverseplant
761
No
to changes in the carryingcapacity of an environment.
populationof animals is capable of instantaneouschange in
numbersresponseto changein the carryingcapacity.Instead,
nerability of crops to insect herbivores . . . may result from
plantingin monoculture
species whichhave evolved chemical
thepotentialpopulationmust"track"thechangesin thecarrydefensesappropriateto communitiesin which the optimum ing capacity. It will be somewhatout of phase with those
strategyis beinghard to find."Competitionbetweenagricul- changesbecause oflags in reproduction
or behavioralresponse
turalcropswhichresultsin tendenciestowardsboth increased to perceivedchangesin theenvironment.
Changesovertimein
yield and reductionin species diversitythus may work to
Kdom willbe respondedto, in a delayedmanner,by changesin
counteractdefensesacquiredduringthe evolutionof theplant
the potentialpopulation.The average populationover long
in the wild.
to the effective
periodsoftime,however,willtendto correspond
The negativeeffectsof agriculturalpracticeson the land,
minimalcarryingcapacity(Kdom) (fig.3). Part of the potential
increasesinKdomcannot
amongthemerosionand changesin soilstructureand drainage humanpopulationproducedbyrecurring
patterns,are well documentedand will not be given further be maintainedduringperiodswhenKdomdropstoitslowestlevels.
here.Less well studiedare the effects
consideration
of agricul- We may view this componentof the populationas agricultural systemson the feedingpatternsof animals other than
turallyinduced "excess production"of people, hereafterreman. Agriculture,
especiallyin its later phases, may signifi- ferredto as Pd. The amplitudeofPd is clearlydetermined
by the
cantly alter the local ecologysimplyby the replacementof
interactionofa largenumberoffactors,includingthelengthof
areas ofwildvegetationwithcultivatedfields.This destruction a generation,
theamplitudeofKdom, therateof changeof Kdom,
ofwildhabitatscreatesfoodshortagesforanimalsthatrequire the availabilityof alternativefood sources,and the way in
plants growingin thesehabitats.Thus theymay turn,out of
whichchangesin Kdom are tracked.Nevertheless,forour purnecessity,to feedinguponagricultural
crops,eventhoughthese
by the interactionof Kdom and
poses treatingPd as determined
plantsmaynotprovidefavoredsourcesoffood.
heuristicvalue.
Kdom has considerable
Finally,theincreasingconcentration
ofresourcesencourages
of the demographiceffectsof
Figure 3 is an interpretation
predation.The sameconcentration
ofresourceswhichfacilitates agriculturalinstabilitytakinginto account the trackingbehaharvestofthecropby humanswillfacilitateitsharvestby nonand
viorof humanpopulations.The hypothetical
productivity
humanpredators.And whileman may delay consumptionof
resultantdemographic
by thisgraphreprechangesrepresented
theplant to optimizeharvest,mostof thesepredatorswillnot
sentconditionssuch as mighthave existedin an earlypristine
be underthe same restraints.Thus theymay attack the field agriculturalsociety.Here we may note an averageincreasein
beforethe cropripens.They may also be capable ofutilizinga
minimalKdom
and theeffective
Kdom, thepotentialpopulations,
crop at a period duringits life cycle whenit cannot be con(Kdom). The graphalso showsthreepoints(a, b, and c) at which
sumedby man,forexample,duringearlyseedlingor vegetative the actual populationis greaterthan can be maintained.We
stages.Thus loss of the total crop may occurbeforeany yield
excessproductionof
mustconsiderthefateof thistemporarily
has been givento man.
thefactor
humanbeings.In essence,we are seekingto identify
The fundamentalcause of agriculturalinstabilityis agricul- which acts to reduce the populationin a particularlocality
tureitself.All ofthenewadverseeffects
whichtheenvironment whenan environmental
crisiscauses a suddendeclinein Kdommay have on agriculturalproductivity
are inducedby agriculThe most frequentresponseof any animal populationto a
turalpractices.Yet at thesame timeagricultureis responsible dropin thecarryingcapacityis emigration.
Emigration(rather
forgreatlyincreasedaverage yield and thus permitsgreatly than starvationor declinein per capita consumption)is espeelevatedhumanpopulationlevels.Overthelongterm,it would
ciallylikelyif onlyone componentof the carryingcapacity is
seemjustifiedto say that,despitethegreaterinstability
ofagrisuddenlyreduced.Part of the animalpopulationwillleave the
culturalproduction,this increasein populationlevels is eviarea in search of a place wherethe limitingresourceis more
dence of "progress."
abundant.
However,the increasesin populationwhichaccompanyincreasesin Kdomoverlongperiodsoftimeare "successes"forthe
systemonlyatthemoment
ofchange.The agriculturally
enhanced
populationlevels now requirecontinuallyelevated levels of
productionfortheirmaintenance.Further,overrelativelyshort
periodsoftime,theenhancedpopulationlevelbecomesnothing
morethan the "normal"populationlevel. Thus the increases
in productivity
broughtabout by agricultureare absorbedby
a growingpopulation.Increasinglysophisticatedtechniquesof
environmental
manipulationare requiredforthe maintenance
of the same rate of growthand, because of increasinginstability, oftenfor maintenanceof the same level of population.
Human populationsgrowin proportionto the effectiveness
of
in raisingthe carryingcapacityof theenvironment
agriculture
forman. Yet, the moreeffective
agriculturalproceduresare in
reducingenvironmentallimitationsupon the productivityof
y
metiatdplnt Kdm) igte old in,miimmefeciea/
theagriculturalecology,themorelikelytheyare to createnew
lat
dmetiatd
nirnmntcotrbuedb
in
cpait
o
te
opportunitiesfor failuresof the system.And as agriculture
idiat "exes" oplaio
(Kom; roenlie,poultin.Arow
createshigherand higherpopulationlevels,the effecton the
(Pd)
.~
~
~~~~
societyof the failuresof the systemwill becomeincreasingly
tragic. Successfulagriculturalsystems require increasingly
Timeabc
successfultechniquesofenvironmental
manipulationmerelyto
maintainthe statusquo.
andpopulation.
between
FIG. 3. The relationship
Heavy
productivitv
Instabilityand dispersal.The interactionof increasein pro4iicrg
ptfevn
t
tbd/
ductivityand increasein instabilityof productivity
has been
responsiblefor the tremendousspread of agriculturaltechniques.To understandthissomewhatparadoxicalsituation,it is
species (Tahvanainen and Root 1972). As noted by Feeny
to speculateas to whetherthevul(1973:14), "It is interesting
necessarv to u,nderstand] how
762
anv
animal
nonaio,T1cn
resnonAs
CURRENT
ANTHROPOLOGY
a declinein foodsupplymaybe expectHumansexperiencing
ed to respondin the same manner.A dropin Kd,,, willencourage part ofthepopulationto leave and establishitselfin a new
area. A slowdeclinein relativeproductivity
suchas thatoccurringat Pointa maybe due to ecologicalor edaphicdegradation
becauseofagricultural
occurring
practices.Whilelargechanges
in populationwill not necessarilyoccur,groupsmay leave in
in whichto farm.Suddenand
searchof "better"environments
drasticdeclinesin relativeproductivity
such as are shownat
Pointsb and c willalso cause emigration.
In thesecases we may
assumethatemigration
willbe encouragedby anotherfactorin
additionto thesearchfora "better"agriculturalenvironment.
If a drasticdrop in production(a famineor a sudden great
scarcityofcultivatedfoods)occurs,it is likelythatthepopulation will returnto the exploitationof nondomesticated
food
sources.We have alreadynotedthatthefamineneed notbe an
"objective"lack ofpotentialcaloriesin orderto be perceivedas
one by thepopulation.However,sincean agriculturaleconomy
can supporthigherpopulationlevelsthana gathering
economy,
and since at least certainnonagricultural
domesticateshave
beenlost,partofthepopulationwillbe forcedintonewareas in
searchofwildresources.Thus a declinein Kdomwillbe especially
noticeable to an agriculturalpopulation because of several
interrelated
factors:higherpopulationlevelsin any givenarea,
losses of wild food sourcesand nonagricultural
domesticates,
and changesin theperceptionofwhatconstitutes
foodscarcity.
The onlyotherlikelyand effective
responseofa societyto a
dropin K(iomis increaseddemandforenvironmental
control.If
this occurs,however,it may simplybe incorporatedinto the
graph by increasingthe amplitudeof the Kdom curve.In this
modelincreasesin productivity
alreadyincorporatetechniques
of environmental
control;thus theycannotbe reenteredas a
newvariable.Also,as has been stressed,environmental
control
is intimatelytiedup withinstabilityin yield;decreasesin Kdom
are inevitableeffectsof an agriculturalsystemexistingin a
variable and evolvingenvironment.At best, environmental
manipulation
willdelaytheneedofthesocietyto finda solution
to theproblemof theexcesspopulationinducedby theperiods
of "successful"agriculturalproduction.
It is importantto recognizethat theseemigrantgroupswill
probablytake withtheman agriculturaltradition;theinstability created by agriculturalsubsistencewill bringabout the
spreadofthesystem.Besides theconservatism
inherentin any
society,anotherfactorfacilitatestheemigration
ofagricultural
populationswithoutloss of agriculturaltechnology.The environmentalcontrolsinherentin agriculturalbehaviorpermit
easy colonizationof new regions.Emigrantgroupsfrommost
societiesrequirea definableset of preexisting
nonagricultural
conditionsiftheyare to maintaintheirmodeofsubsistence.Of
course,agriculturalsocietiesrequiredefiniteecologicalconditionsalso, but the relativelyslow spreadof agriculturalsocieties, such as posited here,will allow time foradaptation by
crops to gradientsin the environment.
Secondarydomesticationwillpermitcolonizationof the fieldsand gardensby new,
better-adaptedplants,and eventuallya wholenew ecological
zone may open up to theagriculturalist.
We mightalso briefly
note that whilethe activitiesof most nonagricultural
peoples
do not interferewith subsequentutilizationof the land by
agriculturalists,
theconverseis farfromtrue.
DISPERSAL
AND INNOVATION
Domesticatory
and agricultural
systems.
We may summarizethe
last sectionby sayingthatthedynamicinteraction
betweenan
increasingKdomand thePd places theagriculturalsocietyat an
advantagerelativeto societieswithothermodesofsubsistence
in termsofpotentialforgrowthand dispersal.Agriculture,
by
inducingenvironmental
instability,creates the conditionsfavorableforits ownspread: theenvironmental
controlinherent
Vol 21 * No. 6 * December1980
Rindos:ORIGINS
AND SPREAD OF AGRICULTURE
notonlyspawnsnewpopulationsand sendsthem
in agriculture
but also allows themto continuein
offintonewenvironments,
the same subsistencepattern.
manipulation
The close connectionbetweenenvironmental
and instabilitiesin productionwas of majorimportancein the
evolutionof domesticatoryinteractionsinto genuineagricultural systems.Clearly,agriculturedid not springfullydevelopedfromthethighofculture.The earliesttypesofagricultural
parts of existingsubsistence
techniqueswere well-integrated
strategies.Activitiesthat were performedfor other reasons
The
on the local environment.
could not help but have effects
the
importanceoftheclearingofbrushforshelterconstruction,
destructionof trees for defensiveor economicreasons, the
burningofgrasslandsas an aid to hunting,theselectivepreserusefulplant or tree,and the creation
vationofan immediately
ofdumpheaps nearhumanhabitationsfortheevolutionof the
techniquesof earlyagriculturalsystemshas been discussedat
lengthin the literature.Preciselyhow theseactivitiesbrought
about the transitionto a totallyagriculturalway of life has
been less than clear. Especially confusingis the recognition
thatmanyofthetechniquesmaybe knownto man withouthis
necessarilyutilizingthemin a subsistencesystem.The model
thus far developedfor the originand spread of agricultural
systemsmay be extended,with certain qualification,to an
interactionsevolvedinto
of how domesticatory
understanding
agriculturalsystems.
We mayfirstconsidera nullcase: A giventechniqueincreases
fortheearlydomestithecarryingcapacityoftheenvironment
an increasein
cated plant (increasesKdwn)). There is therefore
humanpopulationin the area. However,if onlythe carrying
capacityof the regionforhumanswereto rise-that is, ifnow
the same area could supporta greaternumberof humanstheinitial
populationnumberswouldremainconstantfollowing
resultantfromthis
increase.However,the"culturalfecundity"
change in techniquewould be extremelylow. If emigration
capacity,thenemigrawerea fixedproportionof reproductive
tionwouldonlyincreasein directproportionto the increasein
population.If emigrationwere the resultof extrinsicfactors,
rateswouldremainunchanged.If,however,another
emigration
formof behaviorwere to increaseboth carryingcapacityand
ofthecarryingcapacity,we couldpositthesame
theinstability
sequence of eventsas in the dispersalof agriculturalpopulations.We have spokenof the "success" attendantupon techWe maynowmodifythe
manipulation.
niquesofenvironmental
conceptof successto includethe probabilitythat certaintypes
willprovidethecause of theirown dispersal.We may
ofactivity
potential"of a
even speak, allegorically,of the "reproductive
societypracticinga givenformofbehavior.The greatestreproductivepotentialwould arise when the relationshipbetween
productivity(Kdom)and instability(expressedas Pd) is such
that the emigrantpopulation,over time,wouldbe maximized.
Theoretically,an increasein the reproductivepotentialof a
societycould come about withoutincreasein absoluteproductivityifan increasein instabilityalone wereto occur.It seems
unlikelythat thiswouldhave happenedin the developmentof
because ofthepressures,
alreadydiscussed,towards
agriculture
domestithe natural selectionof continuallyhigher-yielding
cated plants,but it shouldbe recognizedas a possibility.
We may restatethis insightby notingthat those systems
Pd is thedirect
whichmaximizePd willspreadmostsuccessfully.
expressionof the interactionof Kdom and Kdom. Thus, fromthe
behaviorswereappearing,
veryfirststagesin whichagricultural
of unstablesystemswas
a tendencytowardsthe proliferation
evident.Of course,thisis not to claim that Kdom, the average
It is probably
ofa system,is totallyunimportant.
productivity
at least today,whereperiodicshortagesin
highlysignificant,
one area are at least potentiallyamenableto solutionby the
importof food fromotheragriculturalregions.However,the
763
ofKdomintoa modelforagricultural
incorporation
originsseems
inappropriate.
Theslowevolution
ofdomesticatory
systems.
Perhapsthegreatest anomalyforthoseacceptingthe revolutionary
viewofagriculturaloriginsis the observationin the archaeologicalrecord
thatwell-domesticated
plantsare foundin eventhemostprimitiveagriculturalcontexts.This is not in keepingwiththe view
that earlyagriculturaltechniquesprovidedthe selectiveforces
whichtransformed
thewildplant intothedomesticate.If agriculturaltechniquesarose beforethe domesticatedplant, we
wouldexpectto findnumerousintermediary
formsshowingthe
selectivepowerof agriculturaltechniqueson the evolutionof
the domesticatedplant. Anotherhighlyconfusingissue for
thosebelievingin a few"centers"of agriculturaloriginis the
existenceofnumerousspeciesofplantsthatseemto have independentlyenteredagriculturalsystemsovervast areas.
Both of theseproblemsare solvedif we accept the coevolutionaryinterpretation
of the evolutionof the domesticated
plant. As we have noted,domesticatory
systems,especiallyto
the extentthat theyare developinginto agriculturalsystems,
will disperseinto new areas by means of the same instability
maximizationprocessesjust discussed.This dispersalwill be
accompaniedby opportunities
fortheevolutionand acquisition
ofnewdomesticates.
New domesticatesmayevolve domesticationsymbiosiswithmanas he movesintonewareas. Also,since
human-coevolved
plants are not culturallybound,the movement of already coevolved domesticatesbetween coexisting
domesticatory
societieswillbe easy. This willpermita society
to acquire new domesticatesand perhapsto move theminto
new regions.Of course,not all domesticatory
societiesneed
developintoagriculturalsocieties,but thosethat do willbring
withthemtheplantswithwhichtheyalreadyhave coevolved
relationships.Thus "early agricultural societies" will be
characterizedby well-developeddomesticatedplants coming
fromseveraldifferent
sources.
Implicitin muchofthelogicthusfarhas beentheassumption
that domesticatory
societiesevolve moreslowlythan agricultural ones. This assumptiondeservessome comment."Pure"
as we have alreadynoted,lack the
domesticatory
relationships,
importantinstability
presentin agriculturalsystems.Thus the
spread of domesticatory
societieswill be dependentupon factorsextrinsicto thedomesticatory
It willbe much
relationship.
morelike that of nondomesticatory
societies.Agricultural
systems,in contrast,are literallydriveninto new localitiesby
environment
crises.
periodicallyrecurrent
Anotherimportantdistinctionbetweendomesticatory
and
agriculturalsystemsis related to anotheraspect of environmentalcontrol.Whileagriculture
acts directlyupon the carryforthedomesticatedplant,the
ingcapacityoftheenvironment
purely domesticatorysociety must rely upon the inherent
forits early coevolved
carryingcapacity of the environment
plants. Thus the populationlevel attainable by the purely
fromthoseof
domesticatory
societyis, initially,littledifferent
societies.
nondomesticatory
Finally, this hypothetical,purelydomesticatorysocietyis
totallydependentupon natural,albeithuman-aided,dispersals
of the domesticatedplant into new regions.This must of
necessitybe a slowprocess.The numberofplaces available for
colonizationby anyplantin a climaxecologyis bothstableand
limited.As we ha...
Purchase answer to see full
attachment