Experiences of Indigenous and Minority People Worldwide Discussion

Why do Gibbons Sing? — Columbus Zoo Qs –

YouTube

Violent chimpanzee attack – Planet Earth – BBC wildlife –

YouTube

Mother Chimpanzee Protects Her Cute Babies from Other Suspicious Monkeys | BBC Studios – YouTube

America’s Black Babies Are Paying For Society’s Ills. What Will We Do To Fix It? | LAist

When a 200,000-Year-Old Culture Met the Modern Economy – The Atlantic

FactCheck: are first Australians the most imprisoned people on Earth? (theconversation.com)

Behav Ecol Sociobiol (2001) 50:9-19
DOI 10.1007/s002650100338
ORIGINAL
ARTICLE
Rebecca Bliege Bird *Eric Alden Smith
Douglas W. Bird
The huntinghandicap:costly signalingin humanforagingstrategies
Received: 31 March2000 / Revised: 10 January2001 / Accepted: 5 January2001 / Publishedonline: 31 March2001
C Springer-Verlag2001
Abstract Humans sometimes forage or distributethe rates, provide widely sharedgoods without repayment,
productsof foragingin ways that do not maximizeindi- or incuran increasedrisk of injury.These decisionshave
vidualenergeticreturnrates.As an alternativeto hypoth- often been explainedby the benefitsof a sexual division
eses that rely on reciprocalaltruismto counterthe costs of labor in parental investment (e.g., Hurtado et al.
of inefficiency,we suggest that the cost itself could be 1992), the optimization of macronutrients(e.g., Hill
recoupedthroughsignal benefit. Costly signalingtheory 1988), or reductionof consumptionvariance(e.g., Smith
predicts that signals can provide fitness benefits when 1988). In particular,the practicecommonto many huntcosts are honestly linked to signalerquality,and this in- er-gatherersocieties of widespreador group-widesharformationis broadcastto potentialmates and competi- ing of large prey capturedby any subset of the groupis
tors. Here, we test some predictionsof costly signaling conventionallyexplained as a form of risk reduction
theoryagainstempiricaldataon humanfood acquisition where all ultimatelybenefit from a reciprocalsharing
and sharingpatterns.We show that at least two types of of unpredictableharvests (reviews in Hawkes 1992;
marineforaging,turtlehuntingand spearfishing,as prac- Winterhalder1997).
While some types of food seem to be distributedin
ticed amongthe Meriam(a Melanesianpeople of Torres
Strait,Australia)meet key criteriafor costly signaling: ways that conformto delayedreciprocity(Gurvenet al.
signal traits are (1) differentiallycostly or beneficial 2000), recent studies (Hawkes 1993; Bliege Bird and
in ways that are (2) honestly linked to signalerquality, Bird 1997) have challenged its ability to explain the
and (3) designedto effectively broadcastthe signal. We wide variety of sharing patternsto which it has been
conclude that relatively inefficient hunting or sharing claimed to apply. The cases that seem least congruent
choices may be maintainedin a populationif they serve with strategiesof delayedreciprocityare those associatas costly and reliable signals designed to reveal the ed with publicdistributionsof food in whichthe acquirer
does not controlaccess to the harvestor its distribution,
signaler’squalitiesto observers.
and may not even reserve a portionfor him- or herself
Keywords Costlysignaling*Humanbehavioralecology (Wiessner1996). Underthese conditions,acquirersseem
*Hunting *Handicap models
unlikely to be able to direct sharesto reciprocatorsand
withhold them from free-riders,as is requiredto solve
the collective action problem(e.g., prisoner’sdilemma)
Introduction
associated with enforcing delayed reciprocity.Similar
challengesto the primacyof reciprocalaltruism(RA) on
Human foragers often make decisions that lead them both theoreticaland empiricalfronts are also found in
to bypass alternativeactivities with higher energy gain recent literatureon non-humanbehavior (reviews in
Dugatkin1997; Pusey and Packer1997). Many theorists
have suggested the paradigm may require extensive
Communicatedby M. BorgerhoffMulder
revisions (Clements and Stephens 1995; Connor 1996;
R. Bliege Bird (1) *D.W. Bird
Dugatkin1997;Roberts1998).
Departmentof Anthropology,270 S. 1400 E., University of Utah,
An alternativeexplanationfor such “economicallyirSalt Lake City, UT 84112, USA
rational”
decisionsis thatundersome circumstancesthey
e-mail: r.bird@mindspring.com
could serve as an honest signal of one or more dimenTel.: +1-801-5814494, Fax: +1-801-5816252
sions of fitness-relatedquality (Neiman 1998; Boone
E.A. Smith
1998; Sosis 2000; Smith and Bliege Bird 2000). Costly
Departmentof Anthropology,University of Washington,
Seattle, WA 98195, USA
signaling theory (CST) provides a powerfulframework
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
10
for explainingtwo paradoxicalobservations:the persis- successfulturtlehunterssignal strength,risk-taking,and
tence of wasteful phenotypeswhen naturalselection is (in the case of hunt leaders) a variety of cognitive and
assumedto creategreaterefficiency,andthe evolutionof leadershipabilities to potentialallies, mates, and comhonest communicationdespite the pervasiveconflicts of petitors.
interestunderlyingevolutionaryprocesses(Zahavi1975,
1977; Grafen1990; Johnstone1995, 1997; Getty 1998).
CST proposes that communicationbetween individuals Methods
with conflicting interestscan be evolutionarilystable if
(a.k.a. MurrayIsland) is a small (1.6×2.2 km) island on the
the signal honestly advertisesan underlyingquality of Mer
northernend of the GreatBarrierReef, 140 km from New Guinea
interestto observers.Advertisingis kept honest andthus in Australia’sTorres Strait. The island’s currentpopulationis 430
mutuallybeneficialto both signalerand observeras long individuals of TorresStrait Islanderdescent, scatteredin approxias the cost or benefit of advertisementis so closely tied mately 85 households. The Torres Strait as a whole is adminisby the State of Queensland and the Commonwealth of
to the quality of the signaler that faking it costs more tered
Australia.
beIf
this
the
correlation
worth.
the
is
than
holds,
signal
Prior to about 1975, when Australian welfare payments were
tween the qualityof a signalerand the qualityor intensi- first made available to all indigenous Australians, the Meriam
ty of the signal will be maintainedby differencesin mar- were nearly full-time subsistence horticulturalistsand marine forginal cost or benefit, allowing recipientsto reliably dis- agers planting tropical yams, bananas, sugar cane, and introduced
new world crops such as manioc, sweet potatoes, and corn, and
criminateamong competingsignalers.When these con- harvesting marine fish, shellfish,
and sea turtles. Today, fishing
ditions are met, honest signals and reliablecommunica- and shellfish collecting remain a critical component of Meriam
tion will be evolutionarilystable,even when signalerand subsistence economy: mean per capita after-sharingconsumption
rates average 630 kcal of meat and 40 g of protein.More than 80%
recipientare antagonistsor competitors.
of these calories are supplied by turtle when in season. For addiThe hypothesisthathuntingmight serve as a form of tional
ethnographic description and previous work among the
status competition among men interestedin “showing Meriam, see Haddon(1906), Beckett (1988), Sharp(1993), Bliege
off’ to a public audience is not new (Hawkes 1990, Bird et al. (1995), Bird and Bliege Bird (1997, 2000), Bliege Bird
1991). Here, we recast the “show-off’ model using a and Bird (1997), and Smith and Bliege Bird (2000).
Observationsof Meriam foraging choices, time allocation, and
payoff structurecompatiblewith CST (SmithandBliege food-sharing
strategiesreportedhere were conducted over several
Bird 2000). We proposethat individualsof high pheno- periods totaling 27 months between January 1993 and July 1999.
typic quality might reap higher benefits or pay lower Much of this field research was specifically designed to test hycosts to acquireskill-basedresourcesor to uncondition- potheses related to the origin and maintenanceof the human sexually sharetheir harvest.These benefits (material,politi- al division of labor,and to determinethe natureof the tradeoffsafmen’s and women’s foraging decisions.
cal, andreproductive)flow fromobserverswho find it in fecting
Time allocation to intertidal activities (spearfishing, shellfish
theirinterestto behave in ways thatimprovethe relative collecting) was measuredthroughfocal individual follows occuradvantage,status,or social dominanceof signalers,ulti- ring during randomly selected days during the spring ebbing tide.
matelyenhancingsignalerfitness. Analyzedaccordingto Locations along the foreshorewere observed from the midpointof
ebbing tide, before the reef was fully exposed, for at least 2 h
CST, this is not delayed reciprocityinvolving an ex- the
or until the last foragerhad left the reef. During the intertidalsamchangeof substance(to the observer)for social status(to pling period, we recorded the moment-by-momentbehavior of at
the signaler),but rathera form of by-productmutualism. least one individualto arriveduringthe sample, and if the individThe payoff to the observercomes from the usefulnessof ual remained after the interval ended, we remainedto completely
the informationinferred from the signal: he or she record the episode. We observed the reef flat habitatfor a total of
118.5 h over 41 spring ebb-tide days, recording 338.1 foragershould be able to evaluatethe signaler’ssuitabilityas a hours of subsistence activity in 210 partial and full foraging folcompetitor,mate,or ally by attendingto the signal rather lows of 94 men, women, and children entering the reef flat. We
thanthroughmore costly or unreliablemeans of assess- then analyzed the percentage of total foraging time each forager
ing the signaler’sabilities or hidden qualities.The high devoted to all potentialintertidalactivities duringthe follow.
analyzed the reef as a bounded habitat within which there
cost or low benefit of faking the signal guaranteesthat areWe
“hunttypes” sensu Smith (1991). All hunt types are simultasignalers will not engage in false advertising,and that neously available options within the ecological habitat”reef flat at
observerswill pay attention.Costly signaling can thus low tide”: reef flat collecting, rocky shore harvesting, and spearspread(by naturalselection or imitation)because of its fishing (Bird and Bliege Bird 1997, 2000). Reef flat collecting involves mobile search of shellfish, with in situ processing to inmutualbenefitsto bothsignalersandobservers.
crease the utility of the load (Bird and Bliege Bird 1997). SpearHere, we test the costly signaling hypothesisamong fishing involves search and travel across the reef looking for enthe Meriam(a Melanesianpeople of TorresStrait,Aus- counterswith mobile prey (small fish and squid 250 g+) to the extralia). We evaluate two candidate foraging activities clusion of all other prey; when prey are detected, the hunterstalks
(spearfishingon the reef at low tide, and huntingturtle the prey and launches his spear from a distance. Both hunt types
occur in the same
duringthe same period of time, while the
for public feasts), to determinewhetherthese hunttypes reef is exposed at patch
low tide. The hunt types vary in how spatially
meet key criteriafor costly signalingin being (1) differ- exclusive they are within the habitat: spearfishing and shellfish
entiallycostly or beneficialin ways thatare (2) honestly collecting are the only two major hunt types in which the forager
linkedto signalerquality,and (3) designedto effectively engaged in exclusive search for prey frequentlyencountersand ignores prey in the other hunt type. These are not the complete combroadcastthe signal to the intendedaudience.We pro- plement
of hunt types simultaneously available, but they are the
pose thataccomplishedspearfisherssignal such qualities three in which adults spend more than 90% of their foraging time
as hand-eye coordination,stealth, and patience, while while on the reef at low tide. Other minor hunt types are handline
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
I1
fishing from the reef edge and in deeper lagoons, netting sardines, turtle shares are distributedin large, uncooked portions (10-11 kg)
hunting octopus, and diving from the reef edge to take underwater of meat, fat, organs, and eggs among nearbyhouseholds, with the
prey by hand. Gross foraging returnrates for each hunt type (not size of portions kept by the butcheringhousehold determinedpriincluding energetic expenditure- assumed to be roughly equal for marily by the numberof demanders(hereaftertermed “household
all reef hunt types) were calculatedfrom the moment an individual consumption”sensu Bliege Bird and Bird 1997). While hunting a
stepped onto the exposed reef and began to engage in targeted turtle is a costly activity in which the benefit is acquiredthrough
search for particularprey types. The macronutrientcontent of fin- the social value of the hunt, collecting a turtleis an activity which
fish and shellfish was obtained through published sources and has little signaling potential and in which the benefit is primarily
through individual analysis of samples collected on Mer (Hirth nutritional.However, there may be some signal value of displays
of generosity among neighbors as turtle portions are shared ac1971; Sidwell 1981; BrandMiller et al. 1993).
For a detailed description of methods used to collect turtle cording to the Meriamethic of debe tonar (“the good way,” which
hunting and sharing data see Bliege Bird and Bird (1997); the re- involves sharingwithoutexpectation of return).
sults presentedhere on turtle hunting and sharing are a reanalysis
All statistical analyses were performedusing Statview (SAS).
of those data. The acquisition of over 120 turtles was noted Large-samplemeans tests were performedaftertesting for normalthroughdaily sampling of the entire village duringthe nesting sea- ity and equality of variance using either one- or two-tailed t-tests
son (October-April) in 1994-1995, and ad lib sampling of the vil- dependingupon the predictiontested, while tests on small samples
lage duringthe 1994 and 1998 hunting seasons (May-September). and those violating the assumptionsof parametrictests used twoNearly every successful hunt or collection of turtle was recorded, tailed Mann-WhitneyU-tests. All means are reportedwith associwith informationobtained on acquirer(s),place acquired, method ated SEs.
of acquisition, and subsequentdistributionthough informantinterview.
There are two primary types of marine turtle acquisition on
Mer: turtlehunting (nam deraimer) and turtle collection (nam ter- Results
pei). Turtle hunting occurs throughout the year, but is the only
way to acquire turtles between May and September(Kob Kerker), Spearfishingas costly signal
when green turtles (Chelonia mydas) feed and mate on shallow
reefs about 16-20 km from Mer. Field observations indicate that
turtles capturedon hunts range from 100 to 150 kg live weight, Thereare significantsex differencesin time allocationto
with an average edible yield of 50.1 kg (Bliege Bird and Bird hunt types in the intertidal:on average,men spend 63%
1997). Hunters head out to the hunting grounds in open boats of their reef foraging time spearing,while for women
powered with outboardmotors, often in cooperation with at least this value is
only 9% (t=6.00, df=66, P=0.001), and
one other hunting boat. Among Meriam turtle hunters, there are
women
spend
76% and men 31% of theirtime shellfish
three distinct roles: hunt leader (ariemer-le), jumper (arpeir-le),
and driver,or tiller-man(korizer-le).Hunt leaders organize and di- collecting (t=4.50, dfr66, P=0.001). The majority of
rect the hunt; there is always only one leader per hunt, regardless men never collect shellfish at all, nor do they combine
of crew size. Hunt leaders bear the cost of organizingthe hunt and spearingwith shellfish collecting in a single visit to
the
ensuringboats and fuel to spare. They direct the crew to particular
locations, decide whether prey encountered is worth pursuing, reef at low tide: 78.6% of 21 men in the time allocation
orchestratethe chase, and direct jumpers when to jump from the samplespentall of theirforagingtime spearfishing.
boat to secure the turtle. Hunt leaders are invested with public
Is spearfishinga signal? The decision men make to
recognition and receive full credit for the kill regardless of spear fish nearly exclusively (ratherthan collect shellwhetheror not they directlyparticipatein capture.
The hunt proceeds as follows. While one man drives, the rest fish) violates simple energy-maximizingprey choice
of the crew stands toward the bow scanning the reef for signs of models, because(1) on average,continuingto searchfor
turtle. The hunt leader directs the driver and coordinates with fish to spear(292?135 kcal/h,n=26) offers lower overall
crews in other boats if present. When a turtle is spotted, the hunt energy returns than shellfishing (1,492?173 kcal/h,
leader makes a decision whether to pursue it based on its size
n=47) while in the reef flat at low tide (t=4.672, df=71,
(large turtles have more meat) and sex (female turtles have more
fat). The boat(s) then give chase, keeping the turtle away from the P=0.0001) and (2) on-encounterreturns for the most
reef edge. When the turtle tires, the hunt leader usually directs his skilled spearfisher(2,505?778 kcal/h) are equivalentto
jumper to launch himself from the bow of the boat with a rope at- the on-encounterreturnsfor the lowest-rankedshellfish
tached to his upper arm. The jumper then attempts to secure the prey still in the optimaldiet breadth(2,214?414
kcallh)
turtle by locking his arms around the flippers and, if successful,
Table
(see
1).
the crew then pulls him and the turtleon board.
One explanationfor this patternmight be that spearTurtle hunting occurs primarilyin the context of public feasting events: hunters choose to hunt in response to a request from fishers are not maximizingcalories, but other macronufeast organizersto provide turtles for consumptionat a previously trients.But this appearsnot to be the case: proteinreturn
announced feast. The biggest and most elaborate Meriam feasts
occur in the context of coming-of-age celebrations and funeral rates from shellfish collecting at 284?31 g/h are higher
ceremonies (see Smith and Bliege Bird 2000 for furtherdetails). than from spearfishingat 6.6?2.9 g/h (t=6.78, dft68,
In contrast to turtle hunting, turtles are also collected (n=88 P=0.0001). Fat returnrates are also higher for shellfish
events), primarily in the context of household provisioning, but collecting at 22?3 g/h than for spearfishingat 1.6+
also for feasts, by men of all ages, women, and children. This oc0.7 g/h (t=4.808,dfr68, P=0.0001).
curs only when they can be harvested on beaches during the nestAnother possibility is that spearfishersprefer other
ing season (Nam Kerker: October-April), although during these
months some turtlesare also huntedon nearbyreefs. In the nesting benefits suppliedby reef fish: they may be more valuseason, turtles are collected at night or during the early morning able as trade goods than shellfish prey. But this seems
hours as they crawl onto sandy beaches above the mean high wa- unlikely:
shellfish collecting produces larger harvests
ter mark to lay their eggs. Turtles are acquired by flipping them
onto their backs, trussing their flippers with ropes, and hauling (1,962?247 g, n=44) than spearing(356?100 g, n=26)
them by boat back to the acquirer’shousehold where they are kept (t=24.857, df=68, P=0.0001). Shellfish prey are also
alive until butchered.When butcheredfor “private”consumption, more likely to be shared:following a harvestof shellfish,
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
12
Table 1 Mean?SE returnrates and mean time allocated to intertidal hunt types by Meriammen and women. On-encounterreturns
for each prey type are calculated from the moment a forager committed to pursuing a particularindividual prey item. For shellfish,
pursuitsbegin when a forager bends down to pick up the item; for
spearing,pursuitsbegin when a forager spots the signs of prey and
drops into a stalk stance to pursue it; for rocky shore harvesting,
encountersare defined as the moment a productivepatch is found.
Encounterreturnrates for all prey types include time spent handling (pursuitand processing) the item while in the foraging habitat: no cooking times are included since cooking methods for each
item can vary from minutes to hours dependingupon the dish prepared. Mean time allocation per person shows the mean percentage of total “reef habitat at low tide” time devoted to each hunt
type averaged for each of 19 women and 21 men observed more
than once on the reef duringthe randomtime allocation scans
Reef collecting
Reef spearing
Rocky shore
Tridacnagigas: 13,064?4750
Hippopus:6,859?464
Tridacnamaxima:4,418?708
Trochus:3,904?467
Lambis:3,412?205
Cypraea:2,214+414
All fish: 2,505?778
Nerita: 1,106?465
Asaphis: 455?52
Hunt type returnsby macronutrient
kcal/h
Protein(g/h)
Fat (g/h)
1,492?173
284?31
22?3
292?135
6.6?2.9
1.6?.7
575?56
88?9
9?+95
Mean time allocation
19 Women
21 Men
0.76?0.07
0.31?0.07
0.09?0.05
0.63?.08
0.14?0.05
0.04?0.03
On-encounterreturns(kcal/h) by prey type
on average22% of the take is sharedto anotherhouse- throughcooperativegenderspecializationin macronutrihold, while only 7.5%of a spearingharvestis shared.
ent harvesting,nor throughreciprocalsharing.We proFinally, could choosing to spear fish instead of col- pose thatthe benefits of spearfishingare gainedthrough
lecting shellfish as women do providegreaterconsump- the honest signal value of acquiringthe prey ratherthan
tion benefits in the long-term?The short-termcosts are through consumption,and that honesty is maintained
inescapable:while women maximizetheir patch returns throughdifferentialbenefits: men of higher phenotypic
and the size of their meat harvestsby stoppingto take quality benefit more than lower-quality individuals
shellfishwhen encounteredon the reef, male spearfishers because they can signal more intensely each time they
take a large cut in protein, fat, and energy income by signal.
ignoring shellfish and continuingto search for fish to
spear.Therecould be long-termbenefitsof such specialization if by dividing labor and pooling harvests,male Prediction1: menwho signal morefrequently
spearfishers and female shellfish collectors maintain obtaingreaterbenefits
long-term shellfishing productivity on the reef. This
could happenif the cooperativepooling unit (the house- If spearfishingis a competitivedisplay,signalersshould
hold) defendeda reef territory,excluding other pooling reap social benefits associatedwith morefrequentspearunits from foragingso thatthe futurebenefits of conser- fishing, such as the benefits from gaining statusthrough
vation could be realized.But they do not: while reef ter- building a reputationas a skilled spearfisher.In interritories are owned, the group sharinguse-rightsto sec- views with 33 Meriam men and women, none would
tions of reef is not equivalentto the pooling unit. Sec- nominatea slate of “thebest shellfishcollectors,”claimtions of reef are considered extensions of residential ing that “being better than others”(i.e., getting larger
plots: use-rights to residentialplots are shared by all harvests)dependssolely on workinglong hours, not on
members of a patriline. Only bounded clam gardens qualitiesintrinsicto the forager,and most nominatedthe
within reef territoriesare excludableand defendableby most frequentlyobservedwomanon the reef as “hardestsingle households,and these gardensare approximately working shellfish collector.”Across individuals,shell2-4 m in diameter,not largeenoughto permitspearfish- fish-collectingharvestsizes are stronglycontingentupon
ing. Any gains in long-termshellfishproductivitywould foragingtime, whereasspearfishingharvestsizes are not
have to be sharedby the entire patriline,and all men (Fig. 1). All interviewees were willing to nominate a
withinthe patrilinewould have to forgo shellfishcollect- slate of “thebest spearfishers.”Because there was little
ing, which is unlikely since men do vary in the extentto variancein signal frequencyamongthe majorityof men
which they collect shellfish. We cannot definitively ex- observedon the reef, we divided the signalersinto the
clude this possibility,only note thatit seems unlikely.
most frequent(1 man with 10 observedbouts) and the
Spearfishersinterested in maximizing return rates least frequent(14 men with 19 bouts total). The most
should not ignore shellfish while foraging on the reef, frequentsignaler obtained 23 out of 61 (37.7%) total
but most of them do. The costs of choosingto spearfish nominationsby the 32 respondents(who nominatedas
do not appear to be balanced by benefits received many as three individuals).More than 75% of respon-
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
13
2
A. Spearfishing
B. ReefShellfishCollecting
10?0?16001400-
6000-x
1200 –
X
~~~1000
0
800-300U) 600-
-200 -25
/
5000
0
25
75
125
175
225
20
60
100
140
180
220
ForagingTime (minutes)
nal elementscontainedin spearfishingare subtleanddifficult to observe.Yet reputationsare still built in the absence of any overt boastingor tale-tellingon the partof
the foragerhimself, who when successful, displays the
typical behavior of a male solitary hunter:he returns
home carryinghis catch quietly,albeit very visibly. The
signal elements incorporatedin spearfishingare conveyed to observersin threedistinctways. (1) Reef activities are highly visible to a large section of the village
surroundingone’s residence, which contributesto the
ease of observerperceptionof boutfrequency.(2) Spearfishermenmakeperceptionof harvestsize easierby cardentsnamedthe most frequentsignaleras the best spearrying theirharvestsopenly in hand,ratherthancarrying
fisherman.None of the less frequent signalers in our
them in a bucket,as line fishers and shellfish collectors
sample obtainedany nominationsat all. The remainder
do. (3) Although bout frequencyis the signal element
of the nominationswent to deceasedindividualsor those
most easily observed, harvest size is usually made
currentlynot active as spearfishers.
known to a large numberof individualsthroughverbal
communication.Huntswith large harvestsare recounted
many times over, and storiesaboutsuch success become
Prediction2: signal intensityis linked
partof local legend.
to phenotypicquality
Fig. 1 Relationship between harvest size and bout length for
spearfishing(A) and shellfish collecting (B). Foragerscan predictably expect larger harvests with longer foraging time in shellfish
collecting but not spearfishing.Each point representsone randomly selected, fully observed bout, with the open circle representing
the most active individual. Episodes are distributedthroughoutthe
tidal season of 1994. A Linearregression with 95% confidence intervals for mean harvest size in kcal (y) on reef patch foraging
time in minutes (x). Harvest size=155+2,x; r2=0.074, P=O.1990.
B The same regression for mean reef shellfishing (female foragers
only) harvest size in kcal (y) on reef flat collecting time in minutes
(x). Harvest size=-6682+21.86x; r2=0.479, P=O.OO1
For the qualitiessignaledby spearfishingto be honestly
advertised,high-qualityspearfishersshould signal more Turtlehuntingas costly signal
intensely,because they gain greatermarginalbenefits or
endurelower marginalcosts per unit signal. In addition Turtle hunting could serve as a competitive display if
to signal frequency(see above), good measuresof sig- foragerscan distinguishthemselvesfromothersbasedon
naling intensity are harvest size and success rate. The their skill. The signal(s) thus producedby turtlehunting
meanharvestsize per bout (in kcal) of the most frequent would dependupon the kinds of costs incurred.If turtle
spearfisher(680?197, n=9) was largerthanthatobtained hunting is associated with lower foraging returnsthan
by other spearfishers(209?101, n=15; U=32, P=0.05). otheroptions,signals of skill sent throughhuntingcould
His hunt type return rates were also higher at honestly indicate the ability of the forager to expend
616?356 kg/h versus 137?75 kg/h (U=32, P=0.49), time and energy in a wasteful pursuit.If huntinghas a
althoughhis bouts were of equalmeanlength as those of high returnbut is associatedwith wide, unconditional,
the less frequentsignalers(106?17.2 vs 102?15.6 min). and costly distributionsof food, signals of skill could
The most frequentsignaler’sreturnrateswere still lower honestly indicate one’s quality (prosocialtendencies,or
than he could achieve shellfish collecting (U=72, abilityto gain resourcereservesthroughskill in otheracP=0.002). The “frequentsignaler”also had a muchlower tivities) by expending time and energy in providing
bout failure rate than other men (failure rate per goods for publicconsumption.
Huntingturtlesis indeeda competitivepursuit,with a
bout=22.2% vs 66.7%; X2=4.444, P=0.03).
very differentcomplementof participantsthan collecting. As the Meriamputit, anyonecan collect turtlein the
nesting season, but only certainmen have the ability to
Prediction3: the signal is designedto be perceived
succeed at turtlehunting.Older and youngermen, chilby the intendedaudience
dren,and women of all ages participatein turtlecollectSince spearfishingis a solitaryhuntingactivitythatpro- ing duringthe nesting season: 21% of turtlecollections
duces no large,widely consumedcommongood, the sig- involve adultfemales. In contrast,for turtlehunting,the
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
14
only participantsare males between the ages of 16 and per capita net returnof 4,653 kcal/h if they divided the
47 (meanage=31.6, n=38). Thougha substantialpropor- turtleamong themselves;however,all hunts duringthis
tion of Meriammen participatein turtlehuntingat some seasonprovisionfeasts, and hunterskeep no shareof the
point in their lives, relativelyfew do so with regularity. turtle they provide. Huntersdeliver the turtle whole to
In our sampleof 37 turtlehuntsextendingover two peri- the feasting location. Huntersthus obtain negative per
ods (1994-1995, and 1998), there were a total of 102 capita consumptionreturnsof -1,086 kcal/h, including
man-daysexpended,or 2.76 huntersper hunt (all hunts the costs of travelingto the patch (see Bliege Bird and
last 1 day or less). For the 87 man-dayswhere hunters Bird 1997 for details of the method of calculatingnet
are individually identified, there were a total of 40 returnsfromturtlehunting).
unique individuals,or 2.18 hunts per hunter.But the 3
most active men participateda total of 16 times, or
18.4% of the hunter-daysof known individuals.Thus, Nesting season hunts
44.4% of the 90 Meriammales aged 16-47 hunted at
least once in the study period, but the 3 most active During the remainderof the year, hunts are undertaken
participants(3.3% of males) were over five times more for two reasons:to provisionpreviouslyarrangedfeasts
likely to engage in a turtlehuntthanthe averageMeriam (n=13), and to provisionhouseholds(n=9). In this seamalein this age range.
son, huntingis much easier and hunterstake on fewer
Turtle hunting entails a variety of costs that could costs: turtlesare found on nearbyreefs waitingto crawl
ultimately be linked to signaler quality in a way that onto the beaches to lay eggs at night, the tradewinds
ensures signals sent by hunting will remain honest have largelyceased, andin betweenmonsoonstorms,the
(Smith and Bliege Bird 2000). Huntingmay involve (1) wateris clear, calm, and visibility is excellent, allowing
high opportunitycosts in the formof passingover oppor- huntersto dog turtlesmore closely, to lose fewer, and to
tunities to acquireother resourceswith a higher rate of more finely discernsize and sex. Consideringonly sucreturn,(2) low consumptionreturnrates(Bliege Birdand cessful hunts,huntersin the nesting season could potenBird 1997) as huntersdistributemeat to non-huntersfol- tially obtain per capita returnsof 8,061 kcallh hunting
lowing the hunt,or (3) high energetic,monetary,or time for householdconsumptionif they did not sharethe meat
investment costs of preparingfor and conducting the with others.But hunterssharemore widely than collechuntthatreduceenergeticreturnratesbelow thatof oth- tors duringthe same season,leavingtheirpersonalreturn
er more easily acquiredresources.There may be other rate from huntingfor a feast at -1,633, and those from
costs or risks of a social naturethat provide a link be- hunting for household consumptionat 814 (Table 2).
tween turtlehuntingand hunterquality,particularlythe Huntsduringthe nestingseason shouldnot happenat all
loss of social statusshoulda hunterfail to deliverturtles if hunterswere interestedsolely in maximizingenergetic
to a feast (see below).
returnrates: duringthe nesting season, collected turtle
offers higher after-sharingenergeticreturnsthanhunted
turtle. In addition, other available resources are also
Hunting season hunts
more energeticallyproductive:sardinenetting provides
11,008?1,705 (n=28) kcal/h before sharing – sardine
During the huntingseason, between April and October, harvestsare sharedonly 5%of the time, leaving on averthe goal of all hunts is to captureturtlefor a previously age 10,864 kcal/h.If huntingturtlewere solely aboutenarrangedfeast. Duringmajorfeasts, all those who come suring feast-goerscan eat plenty of meat and fat, men
must be fed by the feast-giving family and their allies, should net 50 kg of sardineswith no chance of failure
and the family can never predict precisely how many ratherthan spendingthe entireday chasingturtleto risk
people will come to the feast itself. Theoretically,the comingbackto the feast empty-handed.
entire populationof the island (over 400 men, women,
Turtlehuntingin both seasons thus satisfies two of
and children)plus many off-islandvisitors could attend, our criteriafor considerationas an honestandcostly sigthoughattendanceby 200 or fewer fromboth on and off nal. Unlike turtle collecting, observerscan distinguish
the island is typical(mean=174.9,range=49-343,n=54). skilled fromunskilledindividualsandparticipationis reFailure to provide turtle is a substantialcost to the strictedto a subset of the community;second, hunting
hunter’sstatussince all feast-goerswill note the absence has the potential to generate high returnrates among
of turtlemeat at the feast. Since huntsare only undertak- simultaneouslyavailable foraging alternatives(second
en a day or two priorto a feast, there are only as many only to sardinenetting),but does not resultin muchconhuntsas thereare feasts duringthis season:in 1994 there sumptionbenefit to the hunterssince they sharewidely
were ten huntsfor public feasts. Duringthe huntingsea- andtakeno portionfor themselves.As with spearfishing,
son, the turtlepatchis moredistant,encounterswith tur- we hypothesizethat the benefits of huntingare gained
tle are less common,and turtlesare more difficultto de- throughthe honest-signalvalue of acquiringthe prey,
tect and follow underwaterthatis often murkyand bro- rather than through consumption,and that honesty is
ken by swells and whitecapscreatedby the nearlycon- maintainedthroughdifferentialcosts and benefits:men
stant 20-knot southeasttradewindsduring this season. of higher phenotypicquality benefit more or pay less
After a successfulhunt,hunterscould expect to obtaina thanlower-qualityindividualsper unitsignal.
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
15
Table 2 Turtle hunting per capita mean?SE net retums (kcal/h),
numberof hunts/collections (n), and total numberof participating
acquirersby consumptiontype (house or feast), season (huntingor
nesting), and acquisition method (hunting or collecting). Net return rates are calculated as gross edible energy capturedminus energy expended in travel divided by time spent in travel, search,
and handling. Since travel costs are primarilythe cost of fuel expended in outboardmotors, the cost of fuel used (1)is convertedto
calories of meat that could have been purchasedin the local shop
with the money used to buy fuel for the hunting trip. Furtherdetails in Bliege Bird and Bird (1997)
After Sharing
Before sharing
Nesting
season
Hunting
season
House
Feast
Total numberof hunts
Total numberof collections
Numberof turtles
Collecting (108 acquirers)
Hunting(27 acquirers)
Collecting (56 acquirers)
4,922?345 (n=10)
Hunting(40 acquirers)
10
0
13
21,875+2279 (n=44)
8,061?970 (n=9)
16913?1842 (n=44)
10,607?3566 (n=13)
22
88
116
Hunting
season
Nesting
season
-468?660
5,068?922
814?526
1329?1127
-1,633?325
agility,leadershipandorganizationalabilities,andprosocial tendencies(e.g., expendingtime and energy in providing goods for public consumptionat feasts). Such inThe relative costs and benefits of hunting turtle vary formationis presumablyof interestto potentialfemale
with the season:hunts are less costly in the nesting sea- mates,theirfamilies,potentialpoliticalalliancepartners,
son because the prey are found closer to home, but they or competitorsfor social status.
As detailedabove, the most active huntersare a subare moreambiguousas a signal, since turtlesmay also be
collected at this time. Since thereis the potentialfor dis- set of the adult male population.The best turtle hunt
honest signaling (acquiringa turtle throughcollecting leaders are well known to communitymembers:in a
and passing it off as a huntedturtle),we predictthatthe series of 32 interviews we conducted with men and
benefits of huntingduringthe nesting season are lower women in which informantswere askedto nominateany
as well. We would thus expect that more skilled hunters threeindividualsof theirchoice for “bestturtlehunters,”
would be over-representedamonghuntingseason hunts, 3 men (all ones we had observed as the most frequent
which have the highest cost and the lowest signal ambi- huntleaders)garnered38 (39.6%)of the 96 nominations
guity, while the least skilled hunterswould be over-rep- (a total of 30 men were nominatedfor “besthunter”out
resented among nesting season hunts, when hunts are of a pool of 90 or moremales aged 16 andolder).While
less costly as well as harderto discriminatefromcollect- several deceased individuals or older men no longer
ing. Among turtle hunt participants,age is usually a active as huntleaderswere namedamongthe “besthuntgood proxy for skill, since hunt participantsmove up in ers,” no men currentlyactive solely as jumperswere so
the hierarchyof roles as they gain in skill over time, and named. Jumpers are rarely individually credited by
since the skills necessaryto be a good huntleaderarenot othersfor acquiringa turtle,and driversare rarelypubdependentupon size and strength.We thus expectjump- licly creditedwith participationin the hunt. Feast-goers
ers to makethe move to becomehuntleadersonly during can readilyname hunt leadersas providersof the turtle,
the nesting season. Huntleadersduringthe huntingsea- even several years after the feast, but when pressed for
son (n=8) are in fact an averageof 7.9 years older than the names of otherhuntparticipants,lump all otherstohunt leaders (n=22) duringthe nesting season, a differ- gether as “theboys.” Jumpersmay begin in this role as
ence thatis statisticallysignificant(U=41.5, P=O.05).In young as 15-17 years old; anecdotalevidence indicates
contrast,there is no differencein age of jumperaccord- that those who excel in this role and gain increasing
ing to season (df=37, t=-0.253, P=0.80). Of the 17 knowledge and peer respect eventually become hunt
knowncurrenthuntleaders,the 4 youngest(aged 18-22) leaders,while othersremainjumpersor cease participatwere observed as hunt leaders only during the nesting ing in turtlehunting.
season, and served only as jumpers duringthe hunting
season.
Prediction1: turtlehuntingis differentiallycostly
or beneficial
Prediction 3: the signal is designed to be perceived
by the intended audience
Prediction2: signal intensityis linked
to phenotypicquality
Signalsdo little good if they arenot perceivedandthe informationcontainedis not deciphered.We proposethat
Turtle hunting provides evidence of skills that could when signalsare designedto acquiregeneralsocial status
honestly signal several relevantqualities:environmental or political dominance,broadcastefficiency will be inand ethological knowledge, risk-taking, strength and creasedby directingthem at largeraudiences,ratherthan
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
16
small subsetsof interestedparties(as mightbe predicted sult, the most active spearfishergains the benefits of an
if the signal elements of huntingwere directedsolely at enhancedspearingreputationthat more infrequentsigpotentialmatingpartners).Thus, we predictthat hunted nalersdo not seem to enjoy.Spearingthus has the potenturtles should be shared more widely than collected tial to serve as an honest signal of a spearfisherman’s
turtles;in otherwords, that individualsshouldchoose to phenotypicquality.
Althoughwe have demonstratedthe potentialfor sighuntfor largeraudiences,and to collect when audiences
(the numberof householdsattractedto the turtleandcon- naling, and the bias in men’s foragingtowardresources
suming portions) are predictedto be smaller.Because which have high competitivesignalingpotential,preciseturtles cannot be collected during the hunting season, ly how both signaler and observer ultimately benefit
we controlfor the effects of season.We thus predictthat fromthe signalis not clear.We do know thatgood spearaudiencesize shouldhave a significanteffect on acquisi- ing men arewidely recognized,andthatthis addsto their
tion methodwithin the nesting season only: when larger social statuswithinthe community.Potentialcompetitors
audiencesare available,men shouldpreferto huntrather mightalso gain fromknowingwho is a bettermanwith a
thancollect. Duringthe 1994-1995 nestingseason,hunts spear, although the knowledge certainly would have
for which the numberof consumerswere known (n=22) been moreuseful in the past when spearswere the major
averaged26.7?23.7 households,while collections(n=80) means of ambushingcompetitors.Spearingsignals may
averaged18.7?19.0 householdsconsuming,a significant not be designed to appeal to potentialmates interested
difference(t=-1.649, df=100, P=0.05).
only in marryingprovisioners.A man who demonstrates
Individualswho acquireturtlesfor feasts have a large his intention to provision a household by collecting
built-in audience;however, turtles acquiredfor house- shellfish may be more attractiveto such women than a
hold provisioningwill be sharedamong a smallernum- man who demonstrateshis intentionto engage in a comber of households.If huntingserves as an honest signal petitive,status-enhancing
pursuit.
and hunterswish to broadcastthe signal widely, turtles
Turtlehuntersseem to have very differentgoals than
acquired for household provisioning should also be turtlecollectors.Comparedto collecting,huntingis more
shared more widely when hunted than when collected costly (in time, energy,andrisk),providesmeatless effi(controllingfor season). During the 1994-1995 nesting ciently, and is associated with wider distributionsof
season, there were 9 hunts and 44 collections to supply meat and largeraudiencesto witness the hunters’prowturtlesfor householdconsumption(Table2). As predict- ess. Hunterskeep no meat for themselvesunless (quite
ed, hunted consumption turtles were shared to more rarely) hunting for household consumption,in which
households(7.5?3.0) than collected consumptionturtles case they still keep less and sharemore than turtlecol(4.6?2.5), andthe statisticaldifferenceis stronglysignif- lectors.Hunterstake on a varietyof costs for which they
icant (U=64, P=0.015). Because the numberof house- are not materiallycompensated:they expend more time
holds receivingportionsof a huntedturtleis largerthan andenergyin huntingthanthey do collecting,they spend
the numberreceiving portionsof a collected turtle,por- more money for fuel, they spend time organizingand
tions shared to each household should be smaller,but preparingthe huntingteamandits equipmentpriorto the
they do not seem to be so, becausehunterskept less for hunt,and they deliverthe meatto be consumedby large
themselvesand theirown householdsthandid collectors audiences at feasts. The ability to bear such costs ap(meanportionkept (kcal) for hunters=7,780?2,194kcal, pearsto be linked to hunterquality.Because a hunteris
n=38; for collectors=14,387?2,631 kcal, n=123) al- an organizerand decision-maker,his abilitiespeak as he
though the difference is not significant (t-test on log- gains skill and experience: Those named as the best
transformedkcal kept, t=0.682, df=159, P=0.50).
huntersare older than other hunt participants,such as
jumpers.As we predicted,when huntingwas less costly
and more difficult to discriminatefrom collecting (i.e.,
Discussion
duringthe nesting season),hunterswere younger,in fact
composed almost entirely of ambitiousjumperspreparAlthoughour sample of spearfishersis too small to de- ing to be hunt leadersthemselves.Furthermore,the sigfine a continuumin level of skill, spearfishingdoes seem nals sent by huntingare efficientlybroadcast:huntswere
to exhibit strongpotentialfor competitivesignaling.The associated with larger numbers of consumers overall
most frequent signaler obtains much higher gains per than collections during the nesting season and during
unit patch residence time than other spearfishers.Har- householdconsumptionevents. Most feast-goers(audivest size (the most easily observed indicatorof return ence members), when quizzed, know the identity of
rates) is likely to be linked to phenotypicqualityin the hunters,but not the identityof jumpers.
formof hand-eyecoordination,stealth,andpatience,and
Since the low take-homereturnsof huntersproviding
providesan immediatesignal of foragerquality.Success turtlefor householdconsumptionand the negativetakerateis a signal elementcontributingto the long-termrep- home returnsof huntersprovidingturtlefor a feast are
utationof the spearfisher.High success ratesin conjunc- due to widespreadsharing,the argumentcould be made
tion with largeharvestsize ensurethatlargeharvestsare thatthe signal cost is eventuallyrecoveredin the formof
more likely due to skill (forager-dependent
variability) meat or other goods or services returnedas paybackfor
than to luck (forager-independent
variability).As a re- the hunter’sgift. In other words, perhapsRA by the re-
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
17
The results of additionalanalysis in progress show
cipientsof a hunter’slargesseerasesthe cost of his signal
(Sosis and Hill 1997). The stablemaintenanceof RA re- thatthe displayof such qualitiesis correlatedwith higher
quiresthat the provisioningof a turtlebe contingentup- social status and higher age-specific reproductivesucon the eventual receipt of counterbalancingbenefits cess of huntersand theirmatesthan otherMeriam(E.A.
which compensatethe hunter for the marginalcost of Smith, R. Bliege Bird, D. Bird, unpublisheddata). As
giving up turtle that could have been used in other fit- explainedto us by Meriamhuntersand their mates, the
ness-enhancingways. Previoustests of predictionsbased signalssent by huntersarelikely to be partof bothpolition RA in the form of risk reductionreciprocityfor shar- cal strategies,in which huntersdemonstrateto othermen
ing of collected and hunted turtles combined received their honest intentionto work for the public good, and
little support(Bliege Bird and Bird 1997). Furthertests reproductivestrategies, in which hunters demonstrate
of alternativeforms of reciprocityare being conducted their “willingnessto work hard”in orderto gain access
to evaluate this line of explanation(R. Bliege Bird, G. to the “best girl” (Kaddy, personal communication;
Kushnik,E.A. Smith, D. Bird, C. Hadley, unpublished Passi, personal communication).Hunters know that
data),as will be detailedin futurepublications.
spreadingone’s influencewidely via the provisioningof
While we have measuredthe cost of huntingusing di- collective goods at feasts increasesocial statusover the
rect materialcurrenciessuch as time, energy,and money, long-termamong the communityas a whole, while less
there are additionalsocial costs that, thoughdifficult to public-mindedstatus-enhancingactivities (like stealing
measure,may be of equal or even greaterimportance. land or selfishly hoardingresources)provides narrow,
Turtlesprovidedfor feasts are neededby communityel- short-termstatusin only a smallpool of competitors.
ders (feasthosts) to enhancetheirown statusby ensuring
thereis plentyof meatfor guests.If successfulhuntingis
a reliablesignal of the hunter’sunderlyingquality(as de- Implicationsfor the sexualdivisionof labor
fined above), lower-qualityindividualswould be expected to fail on turtlehuntsmoreoften andthuspay a higher One intriguing result of our analysis are the gender
marginalcost per turtledeliveredto a feast (i.e., per unit differencesin participationin those activities with high
signal). Failing to deliver a turtleto a feast also entails signaling potential. With spearfishingparticularly,the
significantsocial costs, since the feast-holdersareexpect- differenceswere not consistentwith the commonnotion
ed to serve turtleto theirguests, and everyonewill know that a sexual division of labor in humansfunctions to
when a hunterhas failed. High-qualityindividualsshould maximizethe productivityof a cooperatingmale-female
have a lowerprobabilityof failureandthusa high ratioof pair (see Bliege Bird 1999). We proposethat the signalsocial benefits(fromsuccessfulhunts)to costs (fromfail- ing benefits attachedto certainforaging strategiesmay
ures). Low-qualityindividualsmay face a large enough often changethe valuationof certainpreyitems for some
risk of failing to producea net social deficit from their individuals. While foraging on the reef, males make
signalingattempts,and hence may avoid signalingat all, choices that fail to maximizetheir macronutrient
return
makingit difficultto test this hypothesis.We do have an- rates. These choices appearto be due to the tradeoffs
ecdotal evidence that some men have attemptedturtle men face over foragingfor highly productiveresources
huntingin the past, and given up when they found they which have little signalingvalue (shellfish) againstforwere “no good at it.” In addition, the facts (detailed aging for less productiveresourceswith high signaling
above)thatonly certainmen engagein turtlehunting,that value (spearedfish). Likewise, choosing to hunt turtle
only an older and more experiencedsubset of these be- for feasts provides much lower consumptionbenefits
come hunt leaders,and that hunt leaderstatusis widely than collecting turtlefor the household,but greatersigrecognizedin the community,all indicatethat success in nalingbenefits.Adultwomen only participatein cooperthis endeavoris not equallyavailable.
ative turtle collection, comprising 21% of all particiTo qualify as costly signaling,huntingmust not only pants,anddo not hunt.
be honest, but must also reveal signaler quality.Given
These data suggest that foragingsex differencesmay
the three distinct roles played by hunt participants,we not simply be a result of women preferringplants and
propose(but cannotcurrentlytest) that at least four dis- men meat,or womenpreferringsmallharvestswhile men
tinct dimensionsof underlyingqualitycould be signaled preferlarge ones, or even women preferringprey which
through turtle hunting: (1) physical quality (such as can be more easily harvestedwhile childrenare present
strength,stamina,agility, and risk-taking);(2) cognitive (shellfisharejust as difficultas fish to collect with small
skills (involving the ecological and ethological knowl- children).Where there is a choice in foragingmethod,
edge needed to successfully locate and captureturtles); men seem to preferto acquiremeat throughmore risky
(3) leadershipskills (charismaand organizationalabili- methodsthat more easily differentiatethe skill of indities), and (4) generosity(ability and willingness to bear vidual foragers,while women seem to preferless risky
the high immediatecost in time, money, and energy of methodscarryinglittle potentialfor discriminationof forprovidingcollective goods withoutdirectcompensation). ager quality.We proposethatsex-biasedforagingpreferWe expect the first dimensionto apply primarilyto the ences may arisedue to the differingbenefitseach sex reyoungermen who serve as “jumpers,”whereasthe other ceives from investing in competitionfor status through
threedimensionsreferprimarilyto huntleaders.
signalingcertaingenotypicor phenotypicqualities.
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
18
Foragingas communication
Acknowledgements We thank the Meriam community foremost,
especially ChairmanRon Day for his foresight and tolerance,our
Meriamfamilies, particularlythe Passis, and the turtlehunterswho
volunteeredinformationor invited us on hunts.We also wish to acknowledge the invaluable field assistance of Andrew Passi, Ron
“Sonny”Passi, EdnaKabere,Del Passi, andCraigHadley.For many
helpful discussions and comments on the manuscript,we thank
Monique Borgerhoff Mulder, Thomas Getty, Craig Hadley, Polly
Wiessner,Kristen Hawkes, Kim Hill, FrankMarlowe, Rich Sosis,
Steven Siller, and Bruce Winterhalder.Researchin 1998-1999 was
supportedby NSF grant SBR-9616096 to R.B.B. and E.A.S., NSF
grant SBR-9616887 to D.W.B. and E.A.S., and researchin 19941995 was supportedby grantsto R.B.B. and D.W.B. from AIATSIS,
the L.S.B. Leakey Foundation,the WennerGrenFoundationfor Anthropological Research, and an NSF predoctoral fellowship to
R.B.B. andan NSF DissertationImprovementGrantto D.W.B.
As Hawkes (1990, 1991, 1992, 1993) first hypothesized, some kinds of hunting may persist in human
populationsbecause some foragers gain benefits from
widely disseminating knowledge about their prowess
relative to other individuals through the hunting of
large and risky prey items. While Hawkes (1993)
stressedthe benefits that huntersobtainfrom providing
collective goods (as with turtle hunting and unconditional sharing),our results show that benefits may also
come in the absence of any materialgood providedto
observers(as with spearfishing)as a result of honestly
revealing hidden informationto interestedparties.Differential costs and benefits for signalers of higher and
lower quality serve to keep the informationhonest and References
valuableto observers.
Beckett J (1988) The TorresStraitIslanders:custom and colonialOur results suggest that CST has the potential to
ism. CambridgeUniversity Press, Cambridge,UK
account for many puzzling aspects of human foraging Bird DW, Bliege Bird R (1997) Contemporaryshellfish gathering
strategies among the Meriam of the TorresStraitIslands, Ausstrategies, and to substantiallymodify received views
tralia: testing predictions of a central place foraging model.
about optimal foraging, food sharing, and other pheArchaeol Sci 24:39-63
nomenacentralto studies of humanbehavioralecology BirdJ DW,
Bliege Bird R (2000) The ethnoarchaeologyof juvenile
in small-scale societies (Winterhalderand Smith 2000).
foragers: shellfishing strategies among Meriam children.
CST may illuminateotherwisepuzzling behaviorssuch
J. AnthropolArchaeol 19:461-476
as preferencefor prey and patch types which provide Bliege Bird R (1999) Cooperationand conflict: the behavioralecology of the sexual division of labor.Evol Anthropol8:65-75
fewer energeticbenefits thanotherpossible choices and
Bliege Bird R, Bird DW (1997) Delayed reciprocity and tolerated
are more costly on otherlevels, if the costs (or benefits)
theft. CurrAnthropol38:49-78
incurredare linked to signaler quality, so that poorer- Bliege Bird R, Bird DW, Beaton J (1995) Childrenand traditional
subsistence on Mer (MurrayIsland), Torres Strait. Aust Abquality signalers cannot consistently maintainor reach
original Stud 1:2-17
the same level of signaling as higher-qualitysignalers.
JL (1998) The evolution of magnanimity:when is it better
Signaling hypotheseshave only just begun to be tested Boone
to give than to receive? Hum Nat 9:1-21
as a cause for variabilityin humanforaging strategies. BrandMillerJ, JamesK, MaggioreP (1993) Tablesof compositionof
AustralianAboriginalfoods. AboriginalStudiesPress,Canberra
Sosis (2000) has suggested that the benefits of signaling may explain why Ifaluk men engage in energetical- Clements R, StephensDC (1995) Testing models of non-kin cooperation – mutualismand the prisoner’s dilemma. Anim Behav
ly inefficient fishing methods associated with costly
50:527-535
ritualpreparations.Such benefits may also explain why Connor RC (1996) Partnerpreferences in by-product mutualism
Ache men prefer to hunt widely shared game even
and the case of predator inspection in fish. Anim Behav
51:451-454
though palm starch collection would offer higher
on-encounter energetic returns (Hill et al. 1987; Dugatkin L (1997) Cooperationamong animals: an evolutionary
perspective.Oxford University Press, New York
Hawkes 1991).
Getty T (1998) Handicap signaling: when fecundity and viability
We are not proposingthat the signalingbenefits prodo not add up. Anim Behav 56:127-130
vide the only motivationfor huntersand meat recipients, Grafen A (1990) Biological signals as handicaps. J Theor Biol
144:517-546
but simply thatthey providean additionalfactorof evoM, Hill K, Hurtado A, Lyles R (2000) Food transfers
lutionaryand motivationalsignificanceshapinghunting Gurven
among Hiwi foragers of Venezuela: tests of reciprocity.Hum
and sharingdecisions. Thus the social and political sigEcol 28:171-218
naling benefits of at least certainforms of huntingmay Haddon AC (1906) Sociology, magic and religion of the Eastern
Islanders. Reports of the CambridgeAnthropologicalExpedicompensate for any economic “wastefulness,”in the
tion to TorresStraits,vol 6. CambridgeUniversityPress, Camsame way that Veblen(1899) explainedboth “conspicubridge, UK
ous consumption”of materialgoods and “conspicuous Hawkes K (1990) Why do men hunt? Some benefits for risky
leisure” as instances of “conspicuousexpenditure”dechoices. In: Cashdan E (ed) Uncertaintyin tribal and peasant
economies. Westview, Boulder,pp 145-166
signed to gain or maintainsocial benefitsonly indirectly
related to economic advantage.However, CST has the Hawkes K (1991) Showing off: tests of anotherhypothesis about
men’s foraging goals. Ethol Sociobiol
potentialto explain not only “wasteful”foragingactivi- Hawkes K (1992) Sharing and collective12:29-54
action. In: Smith EA,
ties, but any behaviorin which higher-qualityindividuals
WinterhalderB (eds) Evolutionary ecology and human behavior.de Gruyter,Hawthorne,NY, pp 269-300
obtain higher marginalbenefits because they can better
absorbthe cost of the behavior,andwhere such behavior Hawkes K (1993) Why hunter-gathererswork. Curr Anthropol
is efficientlybroadcastto the properobservers(potential Hill34:341-362
K (1988) Macronutrientmodifications of optimal foraging
mates,allies, or competitors).
theory: an approachusing indifference curves applied to some
modernforagers.Hum Ecol 16:157-197
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
19
Hill K, Kaplan H, Hawkes K, HurtadoAM (1987) Foraging decisions among Ache hunter-gatherers:new data and implications
for optimal foraging models. Ethol Sociobiol 8:1-36
Hirth H (1971) Synopsis of biological data on the green turtle
Chelonia mydas (Linnaeus) 1758. FAO Fisheries Synopsis
No 85. Food and Agriculture Organization of the United
Nations, Rome
HurtadoAM, Hill K, Kaplan H, HurtadoI (1992) Trade-offs between female food acquisition and child care among Hiwi and
Ache foragers.Hum Nat 3:185-216
JohnstoneRA (1995) Sexual selection, honest advertisementand the
handicapprinciple:reviewingthe evidence. Biol Rev 70:1-65
Johnstone RA (1997) The evolution of animal signals. In: Krebs
JR, Davies NB (eds) Behaviouralecology: an evolutionaryapproach.Blackwell, Oxford, pp 155-178
Neiman FD (1998) Conspicuous consumption as wasteful advertising: a Darwinian perspective on spatial patterns in classic
Maya terminal monument dates. In: Barton CM, Clark GA
(eds) Rediscovering Darwin: evolutionary theory and archeological explanation. Archeological papers of the American
AnthropologicalAssociation No 7, pp 267-290
Pusey AE, PackerC (1997) The ecology of relationships.In Krebs
JR, Davies NB (eds) Behaviouralecology: an evolutionaryapproach.Blackwell, Oxford, pp 254-283
Roberts G (1998) Competitive altruism: from reciprocity to the
handicapprinciple.Proc R Soc Lond B 265:427-431
Sharp N (1993) Stars of Tagai: The Torres Strait Islanders. Aboriginal Studies Press, Canberra
Sidwell V (1981) Chemical and nutritionalcomposition of finfishes, whales, crustaceans,mollusks, and their products. NOAA
Technical Memorandum.National Marine Fisheries Services,
Washington,DC
Smith EA (1988) Risk and uncertaintyin the “originalaffluent society”: evolutionaryecology of resource sharing and land tenure. In: Ingold T, Riches D, WoodburnJ (eds) Hunters and
gatherers:history, evolution, and social change. Berg, Oxford,
pp 222-252
Smith EA (1991) Inujjamiut foraging strategies: evolutionary
ecology of an arctic hunting economy. de Gruyter,Hawthorne,
NY
Smith EA, Bliege Bird R (2000) Turtle hunting and tombstone
opening: public generosity as costly signaling. Evol Hum
Behav 21:245-261
Sosis R (2000) Costly signaling and torch fishing on Ifaluk Atoll.
Evol Hum Behav 21:223-244
Sosis R, Hill K (1997) Comment on Bliege Bird and Bird. Curr
Anthropol38:73-74
VeblenT (1899) Theoryof the leisure class. Macmillan,New York
Wiessner P (1996) Leveling the hunter:constraints on the status
quest in foraging societies. In: Wiessner P, Schiefenhovel W
(eds) Food and the status quest. Berghahn, Providence, RI,
pp 171-192
WinterhalderB (1997) Gifts given, gifts taken: the behavioral
ecology of nonmarket,intragroupexchange. J Archaeol Res
5:121-168
WinterhalderB, Smith EA (2000) Analyzing adaptive strategies:
human behavioral ecology at twenty-five. Evol Anthropol
9:51-72
Zahavi A (1975) Mate selection – a selection for handicap.
J Theor Biol 53:205-214
Zahavi A (1977) Reliability in communication systems and
the evolution of altruism. In: Stonehouse B, Perrins CM
(eds) Evolutionary ecology. Macmillan, London, pp 253259
This content downloaded from 137.151.38.53 on Tue, 16 Sep 2014 16:30:37 PM
All use subject to JSTOR Terms and Conditions
Sexual Division of Labor
In foraging, horticultural, and pastoral economies
In Shuar culture (Amazonian hunter-horticulturalists), men weave the baskets that women use
Definition
✤ Sexual division of labor
✤ The delegation of different tasks to males and females, the products of
which are combined
✤ The sexual division of labor is just one specific type of task
specialization. In modern industrial societies, labor is entirely
fragmented, with divisions based largely on differential knowledge and
training (e.g. a doctor, a car mechanic, a plumber, and a farmer all have
largely non-overlapping, task-specific knowledge and skills)
✤ All human societies have some sort of sexual division of labor, i.e. tasks
that are delegated to women/girls and tasks that are delegated to
men/boys
Why
specialization?
✤ Non-human primates do not exhibit
any cooperative sexual divisions of
labor (though they do exhibit sexspecific activities)
✤ In most non-human primate species,
each individual is fully selfsufficient, and provides all the
nutrients/resources required by
itself
✤ There are several features of the
human life history, diet, and and
foraging activities that make some
sort of division of labor necessary
Features of forager socioecology that
favor sexual division of labor
✤ Highly dependent offspring, and multiple simultaneously dependent
offspring (i.e. a typical woman not using contraception may have an infant,
a three-year old, a five-year old, an eight-year old, etc.)
✤ Reliance on sources of macronutrients that are obtained through mutually
exclusive means (i.e. one cannot simultaneously dig tubers, hunt, and
process foods with hard-to-extract nutrients or toxins)
✤ Skill learning for complex foraging techniques results in delays in
maximum efficiency (e.g. most forager children do not become fully
competent in foraging tasks until well into their teen years)
✤ Some tasks exhibit sex-specific comparative advantages (e.g. men, on
average, have greater upper body strength due to sex-specific musculature)
Maasai herders, Tanzania
Ju/‘hoansi women at camp, Botswana
✤
Those features of the common
human ecological context would
have favored task specialization
✤
Age-based division of labor is also
common in subsistence economies,
with children and adolescents
responsible for different tasks
from adults or elderly
Universal but
variable
✤ The existence of some sort of
sexual
division of labor is universal (it
exists in all human societies)
✤ Cultures vary in…
‣
The pervasiveness of sex-specific tasks
(are all tasks delegated to one sex or
the other, or only some?)
‣
The rigidity of sex roles (are sex-roles
formalized and violations punishable,
or are they loose expectations?)
‣
The value attributed to sex-specific
tasks (is women’s work considered less
valuable? Easier?)
✤ Sex roles can change over time
Why a division by sex?
There are actually at least two central hypotheses the sexual division of
labor in terms of foraging:
1)
A sexual division of labor is a cooperative strategy that allows for the most
efficient way to maximize nutritional returns from foraging:
‣ women are predicted to target foods that conflict least with childcare
‣ men are predicted to target whatever is left in the environment
2)
A sexual division of labor is the result of men and women are pursuing
independent foraging strategies based on their own needs:
‣ women target more reliable foods to offset their higher levels of obligate
parental investment
‣ men target less reliable, difficult to obtain foods that may serve as a
signal to potential mates of their underlying quality or competency
Those hypotheses are not mutually exclusive.
Humans are predicted to be sensitive to environmental cues to
food availability or scarcity and adjust strategies accordingly
Non-foraging tasks
✤ A sexual division of labor is
typically broader than the specific
food resources targeted by women
and men
✤ In our textbook, Low discussed 14
tasks that have been found crossculturally to be predominantly or
exclusively the work of men
✤ Across 1,215 cases, only 7 times (less
than 1% of the time) was it reported
that one of these tasks was
predominantly the work of women
• Hunting large fauna
• Ore smelting
• Mineworking/quarrying
• Metalworking
• Lumbering
• Woodworking
• Fowling
• Boat building
• Stoneworking
• Manufacture of musical
instruments
• Bone setting/surgery
Women’s tasks
✤ Tasks that tend to be
predominantly or exclusively the
domain of women are typically
those that are not technological
(rely on the use of technology)
✤ Men can and often will
participate in many of these tasks
✤ When the technology used for
the task, or the prestige
attributed to the task, increases,
men tend to take over
• Fuel gathering
• Drink preparation
• Gathering and preparing
plant foods
• Dairy production
• Spinning
• Laundering
• Cooking
• Fetching water
Reproductive variance can partially
explain this pattern
✤
Men who are most successful in
obtaining resources may also be most
successful in obtaining mating
opportunities
✤
The most successful men may have
much greater reproductive success
than the least successful men
✤
The most successful women have only
moderately greater reproductive
success than the least successful
women
✤
We predict, therefore, that men will
invest more energy and time in
obtaining resources than will women
Data from Hill & Hurtado, 1996
Ability to secure resources affects
male reproductive success
✤
Forager populations do not typically
have great variation in wealth
✤
What resource variation that they
have does not appear to have a
significant effect on offspring
survivorship (though father presence
does)
✤
But, better hunters have higher
status and tend to:
➡Marry earlier
➡Have more wives
➡Marry younger wives
➡Ultimately, have more children
Shuar man with wives and children
If the acquisition of status or resources typically gives men a
mating advantage (over other men), we should expect that
natural selection would favor traits (e.g. competitiveness,
preference for high variance rewards, showing off) that
would help men obtain resource and/or achieve status
The Hunting Handicap
Rebecca Bliege Bird, Eric Alden Smith, Douglas W. Bird
You will read this paper for Writing Assignment #2
Some hunting choices do not provide
optimal returns
✤
✤
What are some of the explanations that people have suggested for
apparently sub-optimal hunting choices?
‣
Benefits of sexual division of labor
‣
Optimization of macronutrients
‣
Risk reduction via reciprocal sharing
What is the explanation that the authors are testing?
‣
Costly signaling
‣
Honestly signals a quality about the signaler because it is too difficult
for lower quality individuals to fake
‣
Information is useful to observers
Spear fishing
✤
Why do the authors think the usual explanations are not sufficient for
the practice of spear fishing among the Meriam?
‣
Energy maximization?
✤
‣
Macronutrient maximization?
✤
‣
On-encounter rates for most experienced spear fisher is
equivalent to the on-encounter returns for LOWEST ranked
shellfish
Shellfish provide more protein
Non-nutrient benefits, e.g. trade utility?
✤
Shellfish are more likely to be shared (25% of take is shared vs.
7.5% with spear fishing)
Turtle hunting
✤
✤
What is the main reason that turtle hunting seems difficult to explain
from an optimal foraging perspective?
‣
Hunts are organized only to provide turtles for planned feasts
‣
Hunters keep none of the return and have no control over the
distribution ⇒ cannot direct returns to those who will share with them
in the future
What might hunters get out of hunting?
‣
Reputational benefits
‣
Ability to signal their quality
What are the features of a costly
signal?
✤
✤
✤
Not everyone can do it
‣
Both spearfishing and turtle hunting are practiced by only some men
‣
Why do men who are not so good at it not try to fake it?
Signal is designed to be perceived by appropriate audience
‣
Spearfishers fish publicly, carry their returns; Turtle hunters provide
turtle meat to feasts that include hundreds of guests
‣
Who are the intended audiences for fishers vs. hunters?
Signalers obtain benefits relative to their signaling frequency
‣
People know who the good spearfishers and turtle hunters are
Coalitionary behavior
✤
The “Prisoner’s Dilemma”
✤
Cooperation in non-human
species
✤
Male-female coalitions
✤
Single-sex coalitions
What is a coalition?
✤
When two or more individuals cooperate in order to
compete with another or others
✤
One’s success depends on the cooperation of his/her
coalitionary partner(s)
✤
Coalitionary behavior (Cooperation) is RARE in
nature… Why?
Using games to model behavior
✤
Simple evolutionary models can help us understand…
๏ How different strategies affect the outcomes of individuals
๏ The conditions under which certain strategies can evolve
๏
✤
The stability of different strategies
Experimental games can help us understand…
๏ How people actually behave in different situations
๏ Where our assumptions are incorrect
Formal models vs. Simulations
✤
Formal mathematical models
๏ Help determine the abstract conditions necessary for a strategy to
evolve
๏ Are generalizable
Must be extremely simple (few independently varying parameters)
✤ Computer-based simulations
๏ Help model the process of evolution
๏ Can incorporate many different varying parameters, come closer to
realistic conditions
๏ Parameters must be pre-set; results are not necessarily
generalizable
๏
Classic Game Theory
✤
A method developed in economics
to understand strategic decision
making
✤
Games are simplified models of
social interactions that assume
complete knowledge, perfect
rationality, a finite set of
“strategies,” and a defined “payoff
structure”
✤
“Nash Equilibrium:” A state in
which no individual player can do
better by unilaterally changing his
strategy
John Nash,
Nobel Prize in
Economics, 1994
Russell Crowe played John
Nash in “A Beautiful Mind”
The Prisoner’s Dilemma game can help
explain why cooperation so rare
✤
Greatest combined payout occurs
when both players cooperate
✤
Each individual also does better
if both players cooperate than if
both players defect
✤
But, the worst outcome for an
individual is if he cooperates and
his partner defects
✤
The best strategy, on average,
then, is to assume your partner
will defect, and Always Defect
✤
This assumes no communication, or unreliable
communication (e.g. lying)
Evolutionary Game Theory
✤
The solution to the Prisoner’s Dilemma — defect — is a
Nash Equilibrium
✤
Evolutionary Game Theory is a dynamical version of classic
game theory
✤
It considers which strategies will become common (will be
favored by natural selection) in a population
✤
It is used to model behavior and can be adjusted to account
for complicated phenomenon, such as multiple interacting
agents, noise (mistakes), imperfect information, etc.
Evolutionarily Stable Strategy (ESS)
✤
The Nash Equilibrium represents the best strategy for any player in a
given scenario, but the ESS represents the strategies that will be
favored by natural selection over time
✤
Consider a two player game with two strategies, D and C, with payoff
matrix given by E
✤
W represents the fitness benefit to an individual
✤
p is the percent of the population pursuing the rare strategy C
W(D) = W0 + (1-p)E(D,D) + pE(D,C)
W(C) = W0 + (1-p)E(C,D) + pE(C,C)
✤
These equations will
NOT be on the test!
We use the scenario of one strategy being rare because we don’t need
to worry about frequency in the calculations
For those of us who don’t get the picture
from equations…
✤
ESS: when an ESS strategy is
common, a competing strategy is
unable to take over. For example:
✤
✤
If we live in a world of
defectors, a rare cooperator
will always encounter a
defector, and will always
“lose” (have less reproductive
success)
The strategy “Defect” is an ESS
Rare Cooperator will always
encounter Defector, payoff =
0,5
For those of us who don’t get the picture
from equations…
✤
On the other hand, if we live in a
world of cooperators, a rare
defector will always meet a
cooperator, and will always
“win” (have greater
reproductive success)
Rare Defector will always
encounter Cooperator,
payoff = 5,0
Generation 2
✤
Over time, defectors will become
more and more common
(because they have greater
reproductive success)
✤
“Cooperate” is NOT an ESS, it is
unstable and vulnerable to other
strategies
Generation 10
Limitations of Game Theory
Simplistic model of social interactions
✤ Typically involve limited number of players and limited
number of strategic options
✤
‣ Prisoner’s Dilemma is two-player
‣ There are two possible actions
✤
Help us to understand the payoff structure of different
strategies, but are not very good at predicting real-life
behavior
Many parameters affect the optimal
strategies
✤
Length of interaction: “one-shot” vs. repeated play
✤
Communication: cheap talk and commitment
✤
Consequences beyond the interaction: reputation and
punishment
✤
These parameters are often not considered in simple
games
If defection is always a risk, how do
coalitions ever form?
✤
The payoffs, especially in achieving reproductive access,
can be significant
✤
Three mechanisms frequently produce cooperation:
‣ Kin selection (inclusive fitness)
‣ Reciprocity
‣ Mutualism
✤
Whenever there are coalitions, there is also a great deal of
social behavior directed at signaling and assessing intent
In other species, cooperation is
common among close kin
Hamilton’s Rule:
rb>c
relatedness*benefit to kin > cost to self
✤
Belding’s ground squirrels scan
for predators and give alarms
✤
Florida scrub jays engage in
cooperative breeding: provide
food for nestlings
Limited evidence of cooperation among nonkin via reciprocal altruism
✤
Vampire bats
‣ Bats forage each night for blood
meal
‣ Sometimes return empty
‣ Full bats may share
‣ In the future, first bat will
reciprocate
✤ Impala
‣ Impala groom each other to remove
parasites
‣ Trade roles frequently
‣ Grooming is balanced within bouts
Evidence of reciprocal altruism in
primates is also limited
✤
Experimental evidence
‣ A grooms B
‣ B hears A’s call for help
‣ B responds to A’s call
‣ Responsiveness is contingent
✤
Reciprocity mainly involves shortterm exchanges of low-cost
commodities
In other species, the range of cooperative
activities is fairly limited, and usually includes
small numbers of actors
Grooming
✤ Anti-predator behavior
✤ Food sharing
✤ Territorial defense
✤ Coalitionary support
✤ Nestbuilding
✤ Babysitting & alloparental care
✤ Reciprocal altruism takes place within
dyads (pairs)
✤ Other coalitions are usually
inbreeding avoidance
✤
When males leave the group = “female bonded,”
“female philopatric,” or “matrilocal”
✤
When females leave the group = “male-bonded,”
“male philopatric,” or “patrilocal”
Chimpanzees and humans are male
bonded
✤
Most (but not all) primates are female
philopatric
✤
Chimpanzees (both common and pygmy) are
male philopatric
✤
Human societies vary, but male philopatry is
more common
✤
Philopatry affects the power and stability of
same-sex coalitions: the sex that stays home
tends to have stronger coalitions
Please view the videos posted on TITANium
Male coalitions
✤
In female-bonded groups,
males are typically unrelated
✤
Coalitions are short-lived and
in the context of immediate
mating opportunities
✤
In male-bonded groups,
coalitions are also used for
intergroup aggression
✤
Males expend energy
maintaining relationships;
reconciling
High stakes, high risk
Please view the videos posted on TITANium
Female coalitions
✤
Tend to center around parental effort
‣ Obtaining resources, protecting
offspring
✤
Most female coalitionary behavior
centers around preventing the
success of others
✤
“Female harassment” groups:
alliances of related dominant females
harass subordinate females
✤
Females tend to engage in lower-risk
coalitionary behavior
Bonobos
(pygmy chimpanzees)
✤
Equally related to humans as
common chimpanzees
✤
Very different social
arrangements
✤
Male philopatric, but females
form strong non-kin coalitions
✤
Coalitions are maintained in
part through sexual behavior:
G-G rubbing
✤
Sheds light on possible
alternative in human
evolutionary history
Summary
✤
Game theory is a tool to help us understand strategic behavior,
even when the strategies are not conscious
✤
Coalitions are beneficial because individuals can obtain greater
gains, but they are risky because each individual has an
incentive to “cheat”
✤
Culture helps humans overcome some of the risks by signaling
group membership and shared ideas of fairness
✤
Male coalitions are characterized by high potential benefits, high
risk of defection, and high maintenance expenditures
✤
Female coalitions are more often kin-based, are characterized by
lower potential benefits and lower costs of defection