hamilton's rule and kin competition: the kipsigis case

ior 28 (2007) 299–312
Evolution and Human Behav
Hamilton's rule and kin competition: the Kipsigis case☆

Monique Borgerhoff Mulder⁎

Department of Anthropology/Center for Population Biology/Graduate Group in Ecology, University of California at Davis, Davis, CA 95616, USA

Initial receipt 4 October 2006; final revision received 29 May 2007

Abstract

Evolutionary studies of human behavior have emphasized the importance of kin selection in explaining social institutions and fitnessoutcomes. Our relatives can nevertheless be competitors as well as sources of altruism. This is particularly likely when there is localcompetition over resources, where conflict can lead to strife among nondispersing relatives, reducing or even negating the effects ofrelatedness on promoting altruism. Here, I present demographic data on a land-limited human population, utilizing large within-populationvariation in land ownership to determine the interactions between local resource competition and the benefits of kin in enhancing childsurvival, a key component of fitness in this population. As predicted, wealth affects the extent of kin altruism, in that paternal relatives(specifically father's brothers) appear to buffer young children from mortality much more effectively in rich than in poor households. Thisinteraction effect is interpreted as evidence that the extent of nepotism among humans depends critically on resource availability. Furtherunanticipated evidence that maternal kin play a role in buffering children from mortality in situations where paternal kin control fewresources speaks to the important role that specific local circumstance plays in shaping kin contributions to child welfare.© 2007 Elsevier Inc. All rights reserved.

Keywords: Kinship; Local resource competition; Cooperative breeding; Child survival

1. Introduction

Since the first applications of evolutionary theory tohuman behavior (Wilson, 1975), kin selection has played akey explanatory role. Studies of parental investment (Daly &Wilson, 1988), food sharing (Gurven, Allen-Arave, Hill, &Hurtado, 2001), residence decisions, and violence (Chagnon,1979) show that individuals favor close relatives over distantones (or nonrelatives) as targets of altruism, consistent withinclusive fitness theory (Hamilton, 1964). However, kinaltruism can be disrupted if there is local competition overresources because this can lead to competition amongnondispersing relatives, reducing or negating the effects ofrelatedness on promoting altruism (Boyd, 1982, Frank, 1998,Hamilton, 1967). While research on nonhumans demon-strates that the extent of nepotism among kin can depend

☆ Fieldwork was funded by the National Geographic Society.Additional funding for analysis came from the University of California atDavis.

⁎ Department of Anthropology, University of California at Davis, 1Shields Avenue, Davis, CA 95616, USA.

E-mail address: [email protected].

1090-5138/$ – see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.evolhumbehav.2007.05.009

critically on resources available to parents or sibships ininsects (Griffin, West, & Buckling 2004; West, Murray,Machado, Griffin, & Herre, 2001), our understanding of thesensitivity of human kinship relations to resource competi-tion derives largely from folklore (Cinderella's scrubbing ofthe kitchen floor on the night of the prince's ball at the behestof her stepsisters) or historical anecdote (the antics of thebattling sons of Eleanor of Aquitaine).

Here, I present demographic data from Kenya, leveraginglarge variation in land ownership among Kipsigis agropas-toralists, to determine the interactions between local resourcecompetition and the role of kin in enhancing child survival.As predicted, in this patrilineal, patrilocal population, wherepolygyny generates large coresident aggregations of paternalrelatives competing for inheritances, there is an interactionbetween resource availability and the extent to whichpaternal relatives apparently buffer young children frommortality. Additional unanticipated evidence suggests thatmaternal relatives also appear to protect children againstmortality, although only in situations where paternal kincontrol inadequate resources to raise children. These resultsshed light on the ecological and social factors affecting whenkin might affect the success of their relatives' reproductive

http://dx.doi.org/10.1016/j.evolhumbehav.2007.05.009

300 M. Borgerhoff Mulder / Evolution and Human Behavior 28 (2007) 299–312

careers. While there has been discussion of some of thepossible contingencies influencing why certain categories ofkin might (or might not) help (e.g., Beise, 2005; Hames &Draper, 2004; Hrdy, 2005; Leonetti, Nath, Heman, & Neill,2005; Sear & Mace, 2006), this study presents an empiricalanalysis of how the resources available to lineages affectthe behavior of kin in such a way as to influence childsurvival outcomes.

To place the somewhat counterintuitive expectations thatpaternal Kipsigis kin are not necessarily very helpful whenresources are strained, I review some previous research onthis population, adding ethnographic observations. Armedwith the principal scholarly source on the Kipsigis(Peristiany, 1939), I first arrived in a Kipsigis neighborhood(Tabarit) expecting to find strong localized patrilineageswith marked patterns of respect and economic obligation.Almost immediately, I learned of intense fraternal strife,conflicts over land and bridewealth (livestock) distributionsthat are often taken to the neighborhood moots (or courts).When a young man from a neighboring community turneddown funds to study journalism in Canada because of fear“that my brother and cousins will take my land and cattle iffather dies in my absence” (Kab Gelegele resident,February, 1982), the truly pervasive influence of intraline-age conflict dawned on me. Empirical analyses confirm theextent to which brothers hinder each other's marriage,inheritance, and reproductive chances (whereas sisters arean asset, Borgerhoff Mulder, 1998).

Ethnographically, the story is more complicated. Over themonths, I began to see instances of extraordinary cooperationover blood feuds, over the hosting of the all-importantcircumcision ceremonies, and in crises resulting from severeillness, theft, or livestock loss. Agnates would appear fromfar away and deal with the disaster. Looking back intoPeristiany (1939), I found that he, too, saw the same things:“(T)he economic obligations between paternal relations areonly enforced during moments of exceptional need” (p. 98)and “the solidarity of the paternal family can be seenfunctioning… in the case of a murder committed by one of itsmembers” or “if the harvest is bad” (pp. 98 and 100).

With this knowledge of how paternal kin and lineagesoperate in contemporary and traditional Kipsigis commu-nities, and receiving letters from Kipsigis friends whoselives were being destroyed by competitive kinsmen, Iwatched with some amazement a burgeoning literature onthe cascades of positive kin effects on survival and growth(reviewed in Hrdy, 2005, Sear & Mace, 2006). Previousanalyses of Kipsigis data had already shown negativeeffects of siblings (Borgerhoff Mulder, 1998) and cowives(Borgerhoff Mulder, 1990, 1997) on rates of child survival,indicative of intrafamilial resource competition. Themotivation for the present study therefore lay in continuingto investigate such competition beyond the confines ofsibships and polygynous marriages, guided by ongoingdevelopments in kinship and reproductive skew theory towhich I now turn.

1.1. Child mortality and social support

Although child mortality in the developed world is rare(6 deaths per 1000 live births), the developing world exhibitsa 29-fold higher rate (175 per 1000 for sub-Saharan Africa;Black, Morris, & Bryce, 2003). This mortality level reflectsthe coincidence of an evolved set of life history traits,specifically the rapid production of altricial young requiringhigh levels of parental investment (Kaplan & Lancaster,2003), and the poor socioeconomic conditions typical ofmany regions within the developing world, including foodinsecurity, high pathogen exposure, low education, andnegative effects of global markets (Armelagos, Brown, &Turner, 2005; Cesar et al., 2003). While within-populationvariation in child welfare and survival are clearly a functionof household income, parental education, season of birth,maternal age, and child and maternal nutrition, public healthscholars have recently extended their investigations to theimportance of social support networks, often made up of kin,in promoting positive health outcomes through bufferinghouseholds against risk of food shortages (Cohen, Under-wood, & Gottlieb, 2000).

As regards such familial networks, biologists expect kinto assist their relatives in successfully raising offspring, evenif at personal cost, because of inclusive fitness benefits.Hamilton's kin selection theory (Hamilton, 1964) providesan explanation for such altruism: altruistic behavior isfavored wherever rb−cN0, where r is the genetic relatednessbetween actor and beneficiary, b is the benefit of receivingthe altruistic behavior, and c is the cost of performing thebehavior. Evolutionary anthropologists have made specifictests of Hamilton's rule regarding nepotism and investmentwithin human families (e.g., Alexander, 1979, Chagnon,1982, Chagnon & Bugos, 1979) and, in some cases, foundquite close fits between predicted patterns of altruism andempirical data (Bowles & Posel, 2005).

It is often forgotten that Hamilton also recognized thepotential for competition among kin in viscous populationswhere dispersal is limited (Hamilton, 1967). In studies ofnonhumans, competition between relatives over resourceshas been convincingly shown to reduce selection forcooperation among relatives (Griffin et al., 2004; Westet al., 2001) and to bias sex ratios away from the competingsex (Clarke, 1978, Gowaty, 1993); this is because althoughlimited dispersal raises levels of relatedness among interact-ing individuals, it can also lead to more local competitionamong relatives (Frank, 1998). In a separate theoreticalliterature on reproductive skew, some of the transactionalmodels in which dominants are assumed to control thereproduction of subordinates, specifically the concessionsmodel (Johnstone, 2000, Keller & Reeve, 1994), predict thatrelatedness within a group can exacerbate reproductivedifferentials among kin, leading to the prediction that somerelatives can suffer from living in high aggregations of kin.

In humans, there is clear evidence that within-familyinequities exist (Boone, 1986, Hrdy & Judge, 1993) and that

301M. Borgerhoff Mulder / Evolution and Human Behavior 28 (2007) 299–312

families can be suffused with tension (Alexander, 1979,Emlen, 1995). Furthermore, psychologists have exploredhealth and educational differentials among siblings (Hertwig,Davis, & Sulloway, 2002), in some cases within theframework of resource dilution (Downey, 2001). Never-theless, there has been no systematic examination of theimplications of local resource competition for constrainingthe role of relatives as cooperators or competitors, with theexception of Hadley's (2004) study of Tanzanian Pimbwe inwhich he examined statistical interactions between socio-economic status and kin effects. Ultimately, a perspectivefocusing on the intersections of kin effects and resourcescould be brought to bear on explaining cross-culturalvariation in kinship systems, but in this analysis, I focus onthe more tractable question of within-population variability.

1.2. Predictions for a Kipsigis community

The Kipsigis of Kenya are a Nilotic agropastoralpopulation, inhabiting the fringes of the White Highlands,living on now demarcated plots of land on which they keeplivestock and grow crops for subsistence and cash. Theyhave, at least, since the late 18th century, been patrilineal,polygynous, and patrilocal (Mwanza, 1977). Marriages aresealed with a bridewealth payment from the groom to thebride's lineage, whereupon a woman moves from her natal toher husband's homestead and, as a new lineage member,becomes entirely dependent on her husband's resources(Peristiany, 1939). A woman's sons inherit land exclusivelyfrom their father.

The study area on the borders of Kericho and NarokDistricts was settled in the 1930s and 1940s by Kipsigis menand their families escaping the overcrowded “NativeReserves” established by the British colonial government(Borgerhoff Mulder, 1990). The nontraditional practice ofland enclosure (through fencing, Manners, 1967) thatemerged throughout the century was exacerbated by a1980s World Bank initiative for land registration in the area.Land ownership, together with rapid population growth andinterethnic territorial tensions, have, to some extent,inhibited the traditional pattern of married sons establishingfarms adjacent to that of their father, and, increasingly, allsons settle on their father's land or on a plot of landpurchased by their father. Unsurprisingly, many demo-graphic traits are affected by the size of a man's landholding, including polygynous marriage (BorgerhoffMulder, 1987a; Borgerhoff Mulder, 1990), fertility, andoffspring survival (Borgerhoff Mulder, 1987b). Perhaps,unsurprisingly, there can be intense competition amongbrothers over inheritances and reproductive opportunities,including bridewealth (Borgerhoff Mulder, 1998). Ascooperation among kin may be restricted when localcompetition is intense, I predict that the effects of paternalkin on child survival will be stronger in wealthier than inpoorer households. This hypothesis predicts wealth inter-action effects, not necessarily a reverse in the sign of thestatistical relationship across wealth categories.

2. Methods

2.1. Sample

Analyses are based on 785 births to 129 women and107 men between 1945 and 1990, collected from retro-spective interviews with members of all households in fourneighborhoods (kokwetinik; singular, kokwet) intensivelystudied in 1982–1983 and 1990–1991. Of these live births,15% failed to reach their fifth birthday, a percentage thatvaried over time: 19% of those born prior to 1980 (n=392)and 10% of those born in 1980 or later (n=393). As shownelsewhere, the principal cause of mortality for childrenbelow 5 years old is disease—malaria, diarrhea, measles, andinfluenza (Borgerhoff Mulder, 1987b)—all of which aretreated with traditional and modern medicine. Consistentwith analyses showing no gender-biased survival (Borgerh-off Mulder, 1989), child sex dropped out of all the models.

Kipsigis households, as defined here, consist of a woman,her children, and her husband. They are typically locatedwithin larger extended residential units of patrilineallyrelated kin, units that include the households of a woman'scowives, of sons and their wives, and of the husband's fatherand his wives and that retain ultimate rights of ownership toland and livestock. Each woman owns her own hut; after ayear or so of marriage, she obtains rights of use to land andlivestock and builds her own granary. Reproductive historieswere taken for all women in the four neighborhoods who hadproduced one or more children. Births to individualsdeceased at the time of the interview were recorded[10 men (9.3% of the sample) and 12 women (9.3% of thesample)] by means of questions with a living spouse who fellin the sample.

2.2. Variables

Survival status or age at death was determined for eachbirth. Of special interest were the survival status andresidence of the father's mother (FM) and father (FF) andthe number of the father's brothers (FB), as well as those formaternal kin—mother's mother (MM), mother's father(MF), and mother's brothers (MB). The number and/orstatus of kin in these categories was determined from the fulldemographic sample (Borgerhoff Mulder, 1987a,b). Thiswas straightforward for paternal kin, given the strongpatrilineal residence patterns among the Kipsigis, with menand their brothers very commonly living on the same oradjacent plots of land. Fortuitously, systematic data onwives' relatives had been collected as part of a bridewealthstudy (Borgerhoff Mulder, 1995). Additional kin effectsconsidered in the model are the survival status of the focalchild's mother and father.

Each parent and each grandparent of a child were codedas dead if this individual had died before or within thefocal child's first 5 years of life. Because the date of thespecified relative's death could not always be narroweddown to month and year, this approximation seemed more


appropriate than restricting analysis to better educatedindividuals with more precise recall of exact dates. Thenumber of a child's FB or MB was determined on thebasis of questions regarding the number of currently living(1982–1983) brothers of the father and mother and theirresidential histories; it is therefore fixed for a sib set andcan be considered only an approximation for any givenchild; these variables are accordingly treated as categoricalvariables (FBCAT, MBCAT). For FM, FF, MM, and MF, it

Table 1Nonkin variables and child survival—multivariate analyses

Variable

Model A a

nProportional hazard ratio andZ-score significance level

Wasig

1. Birth year c 582 0.974 3.2. Sex 0.Male d 300Female 282 1.012

3. Twins 7.Single d 568Twin 14 3.793 ⁎⁎

4. Mother's age (in years) at birth e 0.12–20 d 10521–27 249 0.79528–47 228 0.975

5. Previous IBI (in months) f 11.10–22 d 14723–25 147 1.37326–36 164 0.64037–150 124 0.389 ⁎

6. Birth order g 6.1–2 d 993 94 0.6964–6 228 0.363 ⁎⁎

7–14 161 0.6017. Polygyny (no. of cowives) e 3.0 d 2541 264 1.555 y

N1 64 1.8048. Education (in years) c, h 582 0.930 3.9. Number of full and half brothers 7.0–3 1124–8 338 0.557 ⁎

9–15 132 1.13010. Acres c 568 0.978 6.Full model χ2 likelihood statistics 60.371, df=17, pb.001

“Previous IBI” and “education” are dropped from Model B, leaving an n of 582 (a Reduced sample: individuals with missing data were dropped.b Full sample: variables with missing data were dropped.c Covariate.d Reference category.e Calculated at the time of birth of the child.f Data on previous IBI are missing for 146 observations.g Birth order was categorized this way to maximize capture of inverted U shh Data on education are missing for 87 observations.⁎ pb.05.⁎⁎ pb.01.⁎⁎⁎ pb.001.y pb.10.

was possible to determine whether this relative lived in thesame set of adjacent kokwetinik as the child (generallywithin a 15-km range, an easy day's walk, termed “local”)or at a further distance (generally more than an easy day'swalk, termed “distant”); even where relatives lived beyondthe adjacent kokwetinik that constituted the sample of thisstudy, the distance rarely exceeded 25 km. In this sample,the mode, median, and range of the ego's FM's residenceand MM's residence are 0, 0, and 0–11 km and 0, 3, and

Model B b

ld χ2 andnificance level n

Proportional hazard ratio andZ-score significance level

Wald χ2 andsignificance level

117 y 785 0.957 15.975 ⁎⁎⁎

002 0.492404381 0.882

419 ⁎⁎ 31.987 ⁎⁎⁎

75629 6.304 ⁎⁎⁎

571 1.341253286 0.696246 0.715

741 ⁎⁎

050 y 3.382255111 0.989248 0.649 y

171 1.152211 10.120 ⁎⁎

391317 1.841 ⁎⁎

77 2.376 ⁎⁎

463 y

213 ⁎ 11.028 ⁎⁎

140411 0.600 y

234 1.330183 ⁎⁎ 785 0.974 11.606 ⁎⁎⁎

78.715, df=13, pb.001

there is an overlap in cases with missing values).

ape survival by parity.

Table 2Kin variables and child survival—single-variate analyses

n

Model C

Proportional hazard ratio andZ-score significance level

Wald χ2 andsignificance level

1. Parental survivalFatherDead a 19 0.09 (ns)Alive 766 1.235MotherDead a 23 19.85 ⁎⁎⁎

Alive 762 0.228 ⁎⁎⁎

2. Paternal kinFMDeceaseda 117 29.73 ⁎⁎⁎

Distant 214 0.294 ⁎⁎⁎

Local 454 0.354 ⁎⁎⁎

FFDeceased a 193 6.98 ⁎

Distant 219 0.581 ⁎

Local 373 0.600 ⁎

FBCAT0–1 a 215 3.952 177 0.8923 202 0.604 y

4–6 191 0.7453. Maternal kinMMDeceased a 84 3.93Distant 381 0.582 y

Local 320 0.740MFDeceased a 147 4.77 y

Distant 342 0.672Local 296 1.054MBCAT0–1 a 112 6.23 y

2 320 0.7323 292 0.550 ⁎

4–7 61 0.448 y

a Reference category.⁎ pb.05.⁎⁎⁎ pb.001.y pb.10.


0–128 km, respectively. Unless other information wasavailable, it was assumed that residence patterns observedin 1982–1983 and confirmed in 1991 had been constant.Very infrequent land sales and minimal labor migrationjustify such an assumption.

To study the independent effects of kin factors on focalchild survival and their interactions with wealth, thesemeasures were linked to the full demographic and socio-economic records. Wealth is measured as land available tothe child's mother, because of the significance of this factorin previous analyses of this population (Borgerhoff Mulder,1987b). Used as control variables were data on year of birth(for secular changes in survival), gender, twin status,mother's age at birth of child, previous birth interval, birthorder, polygynous status of mother, mother's education,wealth (as defined above), and the number of the focalchild's brothers (both full and half). These variables allowfor both a finer determination of the effects of kin and a morenuanced interpretation of the correlational results.

2.3. Analysis

Cox's regression, a form of event history analysis(Allison, 1984), was used to determine probability ofsurvival within the first 60 months of life, using STATA(v 8.1). This method is appropriate in that it allows for theinclusion of censored cases, in this case, children whohave survived to various ages but not yet reached theirfifth birthday. Four analytical steps were taken. First, toidentify unobserved variability between siblings born tothe same mother, all independent variables were investi-gated for shared frailty using the STATA “share” commandand entering the mother's ID as a covariate. Theta values(a single variance parameter that measures heterogeneity insurvival times across the children of different mothers)were consistently very small (0.00–0.086) and neversignificant, indicating that there is no intermother variancein frailty; in other words, all full sib sets have the sameunobserved frailty (Gutierrez, 2002). This results in partbecause many of the covariates of interest, such as wealth,status of grandparents, number of siblings, and so forth,vary much less within families than between families. Thesame analysis with similar outcomes was conducted forshared fathers and shared lineages. Each of these termswas investigated in the full model and, again, wasobserved to be very small, reflecting the fact that multipleindependent variables capture much of the variation andleave little residue to be explained as unobservedheterogeneity. Since shared frailty parameters nevercontributed significantly to any of the models, they weredropped from the presentation of the final analyses.

The second step entailed identifying the effects of a rangeof control variables, defined as variables expected to affectsurvival times but not of primary interest in the currentanalysis. This was done by determining the increasedprobability of dying in the first 5 years of life as a function

of a change in the specified independent variable. Modelswere fit using the Cox's proportional hazards model inSTATA (see Table 1). Model A drops individuals for whomthere are missing data, whereas Model B uses the fullsample, dropping variables for which there is incompleteinformation. The results are largely congruent. Interpretationof the proportional hazard ratio is as follows. The estimatedhazard ratio of 0.974 for birth year (in the reduced sample,Model A) indicates that an individual born in any given yearhas a 2.6% reduced likelihood of dying than does anindividual born in the preceding year (i.e., the ratio of theirrespective hazards is 0.974). This ratio does not differsignificantly from 1, on the basis of a Wald χ2 statistic(3.117), although it shows a statistically significant trend.Now, consider birth intervals, coded as a categorical


variable. Compared to children born after the shortestinterval, the 23- to 25-month group has a nonsignificant37% greater chance of dying, the 26- to 36-month group a36% reduced chance of dying, and the longest category aneven more (61%) reduced chance of dying. While only theratio of the first to the last category is significantly differentfrom 1, there is an overall significant effect as indicated bythe Wald χ2 statistic (11.741). Hazard ratios show thatchildren with a later birth year are less likely to die (Table 1,

Fig. 1. The effects of (top) FM, (center) FF, and (bottom) the number of FBon child survival (b60 months). For statistical tests, see Tables 2 (single-variate analyses) and 3 (multivariate model).

Fig. 2. The effects of the number of FB on child survival (b60 months)splitting the population into a poorer and a richer half. Compared to thebaseline (FBCAT 0 to 1), the proportional hazard ratio for 2, 3, and 4–6 FBare 0.996, 1.161, and 1.973 (Wald χ2=6.173) for the poorer households and0.861, 0.232, and 0.124 (Wald χ2=18.603, pb.001) for the richer households,yielding a significant interaction (Wald χ2=20.823, pb.001). For themultivariate model statistics, see Table 3.

Panel 1) and that twins have higher mortality than singlebirths (Panel 3). Mortality declines with the length of thepreceding birth interval (Panel 5), among children of middlebirth order (Panel 6, Model A only), among children born towomen with no cowives (Panel 7, B only), with each year ofmaternal education (Panel 8), with an intermediate numberof full and half brothers (Panel 9), and as the size of the landholding increases (Panel 10). These are well-established andcommonly observed effects found previously both in this andmany other developing nation populations (Rutstein, 1984,2000) and receive no further discussion here except withrespect to their interactions with kin variables.

The third step entailed single-variate analyses of kineffects for the full sample of births (Table 2, Model C). Theraw survival curves are presented in Figs. 1–4. The fourthstep examined a series of multivariate models with controland kin variables considered together (Table 3, Model D),split by wealth (Model E), and with wealth interactions(Model F). The third and fourth steps are presented inSection 3.


3. Results

3.1. Parental effects

As shown in the first panel of Table 2, a child is 77%less likely to die if his or her mother is alive during the first5 years of life than if the mother dies during this period;living fathers have no significant effect on survival. Thispattern is retained in the full model (Table 3, Model D) and

Fig. 3. The effects of (top) MM, (center) MF, and (bottom) the number ofMB on child survival (b60 months). For statistical tests, see Tables 2 (single-variate analyses) and 3 (multivariate model).

Fig. 4. The effects of the number of MB on child survival (b60 months)splitting the population into a poorer and a richer half. Compared to thebaseline (MBCAT 0 to 1), the proportional hazard ratio for 2, 3, and 4–7 MBare 0.730, 0.409, and 0.197 (Wald χ2=9.539, pb.05) for the poorerhouseholds and 1.264, 1.425, and 1.837 (Wald χ2=0.777, ns) for the richerhouseholds, yielding a significant interaction effect (Wald χ2=6.681,.05bpb.10). For the multivariate model statistics, see Table 3.

holds in both the richer and the poorer (at marginalsignificance) halves of the population (Table 3, Model E).There is a marginally significant interaction between motherand wealth, indicating that mothers' survival is morestrongly associated with child survival in rich than inpoorer households, an unpredicted trend we return to inSection 4.

3.2. Paternal kin effects

The effects of paternal kin on survival (Fig. 1) are shownin the second panel of Table 2. Single-variate analyses(Model C) show that children with a living FM and childrenwith a living FF are significantly less likely to die than thosewithout these paternal kin.

The principal paternal grandparental effect lies betweenthose who are deceased and those who are alive and liveeither locally or distant. This indicates that for this categoryof kin, the distance at which they live is not a major factor inproviding support, conforming to ethnographic observations:


paternal kin amass at times of conflict over land, interlineageconflicts, or family crises in current times, as they did in theearly years of the 20th century (Peristiany, 1939).

Examining paternal kin effects in a full model in which alldemographic and economic variables are entered (Table 3,Model D) shows that both a surviving FM and the number ofFB are retained as significant factors reducing childmortality, although the latter is only marginally statisticallysignificant. In other words, paternal kin raise survivalchances, independent of the positive effect associated withthe number of MB (Panel 3, and see below) and of controlvariables (see Table 3, Panel 4), including wealth.

Splitting the sample by wealth shows clear wealthinteractions. As expected, the positive association betweenFB (FBCAT) and child survival is stronger in the richerthan in the poorer section of the population (Fig. 2).Children with a large number of FB show reducedmortality in the wealthier but not in the poorer half ofthe population. These results hold in the fully controlledmodel (Table 3, Model E), and there is a significantinteraction effect between number of FB and wealth(Model F). The reasons why the FB might enhancesurvival in wealthy but not in poor households areexamined in Section 4. A further wealth interactionindicates that the association between the FM and childsurvival is stronger in the poorer than in the richer half ofthe population, another unexpected result addressed inSection 4.

3.3. Maternal kin effects

The effects of maternal kin on survival (Fig. 3) are shownin the third panel of Table 2. Single-variate analyses (ModelC) reveal that the status of the MM has no overall effect onchild survival and that there is only a marginally significanttrend for the MF and MB.

These marginally statistically significant effects reflectlow mortality (high proportional hazard ratio) in cases wherethe MM and MF are alive but living beyond the neighbor-hood compared to those who are either deceased or alive andpresent. This suggests an unusual but ethnographically quiteplausible maternal kin effect—a woman benefits if herparents live at some distance from her marital home. This“refuge” effect is explored in Section 4.

Examining the effects of maternal kin in the full analysis(Table 3, Model D) reveals that neither the effect of MMnor the MF retains significance. There is, however, a strongindependent and positive effect of the number of MB onchild survival. In other words, children whose mothershave three or more brothers experience a significantlyreduced mortality risk, independent of whether the motherherself is alive (Panel 1), of the effects of the FM andFB (discussed above, Panel 2), and of other controlvariables (Panel 4).

Splitting the sample by wealth shows that the apparentlybeneficial effects of the MB are, in fact, most marked in the

poorer half of the population (Fig. 4). Statistically, this ispreserved in the full model (Table 3, Model E) where thenumber of MB is marginally significantly associated withchild survival only in poorer households, yielding amarginally significant interaction between MB and wealth(Table 3, Model F). Adding to the model a further interactionterm that identified whether the MB lived locally or not (asgauged by the location of the mother's parents) shows thatthe number of MB is associated with higher survival in poorhouseholds when these brothers live at a distance (Waldχ2=9.17, df=3, pb.05).

3.4. Control variables and the role of competitionamong kin

Analyses showing that nepotism toward grandchildren,nephews, and nieces within patrilineages varies by wealthprovide only indirect evidence that kin are motivated tohelp their relatives as a function of potential competitionover resources within the patrilineage. The most convin-cing evidence for this interpretation comes, unsurprisingly,from ethnographic observations (see Sections 1 and 4).Further indication of competition contributing to offspringsurvival in resource-stressed situations is found in thepatterning of the effects of the control variables in theseanalyses. First, consistent with previous analyses, thenumber of a child's mother's cowives is negativelyassociated with child survival (Borgerhoff Mulder, 1990,1997), an effect that is observed in poorer but not in richerhouseholds (although the interaction term is not signifi-cant). Second, again consistent with previous analyses(Borgerhoff Mulder, 1998, Table B1, and here justfocusing on male siblings—the most likely competitors),survival chances are negatively associated with the numberof a child's full or half brothers. As with the analysis ofmother's cowives, this effect is stronger in poorer than inricher households (Model F).

Examining control variables in this way providesstatistical support for the claim that competition withinpoorer patrilineages is modulated by wealth availability. Onthe other hand, it is important to stress that these imputedpatterns of competition are not entirely accounted for bywealth since both the main effects and the wealth interactionterms are independent of the wealth covariate (acres).Furthermore, an anticipated interaction between birth date,wealth, and child survival (reflecting increased competitionover time with escalating land shortages) was not found,probably because, in recent years, child mortality hasdeclined substantially and wealth nowadays shows highercorrelations with fertility than with offspring survival(Borgerhoff Mulder, 1987b).

The motivations for maternal kin to assist with the raisingof grandchildren, nephews, and nieces are more difficult toevaluate since we know less about their circumstances; this isbecause many of them reside outside of the study area. Thepresent analyses do, however, indicate that these lineages


help out their daughters when their daughters find them-selves married to men with limited resources.

4. Discussion

In this study, demographic correlations between theavailability of kin and the survival of children prior to theirfifth birthday are presented to examine how kin cooperationaffects reproduction in human societies. The principalfinding is that paternal kin play a larger role in affectingchild survival than maternal kin, no doubt reflectingpatrilineal social institutions in the Kipsigis. As predictedby the local resource competition model, there is asignificant interaction between wealth and number of FB,suggesting that the apparently protective role of paternaluncles is stronger in richer than in poorer lineages.An unanticipated finding is the apparent buffering effectof maternal kin in situations where paternal kin controllimited resources.

The mechanisms whereby kin do, or do not, improvechild survival are not identified in this study but are likelyto entail provision of food, labor, and cash at critical familycrises—food shortages, crop loss, cattle disease, politicaldisputes, and acute illness. As mentioned in Section 1,public health scholars are increasingly impressed with theimportance of kin networks in promoting positive healthoutcomes (Cohen et al., 2000). My ethnographic observa-tions with Kipsigis provide abundant evidence of the rolekin often play in supporting one another in child rearing,whether this involves taking a child to a traditional orWestern health practitioner, providing money for care,slaughtering livestock in times of drought or temporaryfood shortage, helping with the tasks of firewood and watercollection, assisting with agricultural work, or providingdirect childcare (Borgerhoff Mulder & Milton, 1985). Morediffusely, and as noted by Peristiany and myself, patrilinealkin amass at times of crisis and provide critical assistancewith events surrounding circumcision, weddings, and bloodpayment negotiations (Peristiany, 1939, p. 98). Althoughneighborhoods (kokwetinwek) have strong cooperativeinstitutions for coordinating daily tasks (Peristiany, 1939),relatives are usually the most reliable partners even withina neighborhood.

Although much of what we know about cooperativebreeding in humans is based on correlational data (asreviewed in Sear & Mace, 2006), we should caution thatcorrelation is not causation. Since relatives share not onlygenes but many environmental factors as well, underlyingheterogeneities between families may account for some ofthese correlations. To some extent, this problem is mitigatedhere by the examination of the shared frailty parameters,which, perhaps unsurprisingly, proved to be nonsignificantgiven the number of independent variables available in thisstudy. Furthermore, if phenotypic correlations were impor-tant, we would expect positive associations between child

survival and all categories of equally related relatives, butthis is not the case. A further problem with analyses of thetype reported in this article is that contingencies are notexplored. In other words, no attempt is made to determinehow the loss of one relative, for example, a mother, mightrender other relatives more, or less, important in terms ofaffecting child survival; this could be the subject of furtheranalysis and would undoubtedly add texture to our under-standing of the facultative aspects of kinship behavior.

4.1. Evidence for local resource competition

Why do paternal kin, and specifically FB, enhancesurvival in wealthy households but not in poor households?A man and his brothers and half brothers are directcompetitors over their father's land and livestock. Kipsigiscustom dictates an equal division of resources among sons, apractice that worked well in a traditional system in whichlivestock was the principal wealth source. During thecolonial period, the ownership of cultivable land becameimportant. Unlike cattle, land is an inelastic resource, andplot sizes are shrinking across generations as populationgrows. Violent cases of strife over land among brothers,among brothers' wives, and among the cowives married topolygynously married men were observed during fieldwork,including the following: a man burning down the house ofhis brother, with wife and children inside; a womanattempting to poison the children of her husband's elderbrother; attempts to “move” fences (effectively brush piles)over night prior to the formal delineation of fields; witchcraftaccusations; and local court cases. Equally notable wereinstances of supreme altruism and fraternal cooperation(such as shouldering the medical expenses of a brother'schildren or providing secondary school fees for a nephew orniece). The unexpected observation that living paternalgrandmothers improve child survival in poor but not in richhouseholds suggests that these grannies can perhaps bufferyoung children from mortality in patrilineages wrought withresource conflict.

The suggestion that brothers cooperate more in richerthan in poorer patrilineages likely results from two specificfeatures of Kipsigis society and ecology. First, patrilineallyrelated kin tend to be equivalent in wealth status, reflectingthe traditional egalitarian inheritance system and limitedaccess to off-farm employment in this population. Second,the relationship between resource availability and fitnessshows diminishing marginal benefits as wealth increases(Borgerhoff Mulder, 1987a). Under such circumstances, apoor man might find that the inclusive fitness benefit (b) ofmaking a resource transfer to a sick nephew or niece iseclipsed by the direct fitness cost (c) of the lost resource;conversely, for a rich man, the marginal benefit oftransferring resources to save the life of an ailing nephewor niece could outweigh the small, direct fitness cost. If thisis the logic underlying these empirical patterns, nepotismwould indeed be more likely to trump direct fitness costs in


richer than in poorer lineages. This framework helps toexplain why, in a study of the Tanzanian Pimbwe, Hadley(2004) found exactly the opposite pattern to the onereported here. The presence of mother's sisters in thevillage was positively associated with child weight-for-agescores only in families of low socioeconomic status, a moreintuitively understood and possibly widespread pattern(e.g., Stack, 1997, for African Americans). Why is thePimbwe pattern so different from the Kipsigis? With littleto no inherited possessions, poorer families have fewmaterial assets to compete over and have much to gainfrom cooperation. In fact, because close relatives can varyenormously in wealth, it is the rich individuals who reportbeing pestered by their poorer relatives into providing acostly resource stream (Hadley, 2004).

An alternative but unlikely explanation for the Kipsigisresults is that it is not the b and c terms in Hamilton'sequation that vary with wealth but rather r. Thus, relatednessamong agnatic kin might be higher in richer than in poorerhouseholds, reflecting greater mate guarding potential on thepart of husbands or lower temptations to infidelity on the partof women. However, extramarital affairs are heavily sanc-tioned in this population and lead to acute punitive outcomesfor women, which (anecdotally at least) appear unrelated towealth. Insofar as richer people tend to be more educated andcosmopolitan, one might expect (if anything) richer house-holds to show lower r, although this is pure supposition.

These results can be placed in the broader framework offamily conflict (Emlen, 1995). Hill and Hurtado (1996), afterfailing to show effects on survival associated with individualrelatives, turn to the use of the ratio of related helpers topotential recipients in the Ache to indicate kin assistance inthe context of competing demands on their time. InDominica, Quinlan (2005) shows how the number ofbrothers in a community negatively affects reproductivesuccess, a pattern elsewhere linked to competition overinheritance and marriage payments (East Africa, Mace,1996) and land inheritance (agrarian Sweden, Low, 1994).Furthermore, in many parts of the world, strong negativeeffects of the number (Knodel, Havanon, & Sittitrai, 1990)and sex (Parish &Willis, 1994) of siblings are seen on healthand educational outcomes (Downey, 2001; Garg & Mor-duch, 1998) as parents strategize within offspring sets.Specific links can be drawn between parental strategies andresource shortages in a historical German peasant popula-tion, where differential investment in sons and daughters isshown to vary in relation to habitat saturation (Voland &Dunbar, 1995). The demographic evidence presented here onthe Kipsigis, a population with a strong ethos of equalinheritance to sons and an almost exclusive reliance oninherited resources of land and livestock, demonstrates howthe balance between apparent kin assistance and kincompetition pivots on the availability of resources. Onthese grounds, we might also expect that as rural populationsbecome integrated into a cash economy (with wealthultimately held individually in bank accounts), the impor-

tance of kin networks for resource transfers will decline, as isoccurring in India (Shenk, 2005); kin networks maynevertheless continue to play an important political role(Munshi & Rosenzweig, 2005).

The implication of these results is that for humans, aswith other species (Griffin et al., 2004; West et al., 2001),patterns of dispersal and resource distributions are key tothe development of more precise predictions regardingaltruism. This might explain why, in many domains ofhuman behavior, altruism is only loosely explained byrelatedness; for example, only one third of the variability incash remittances in South Africa is attributable to a modelof kin selection that includes reproductive value (Bowles &Posel, 2005). Furthermore, local competition among kinover resources may explain why, in some situations,kin of equal relatedness but different categories, suchas those related through maternal or paternal ties, mayevince very different patterns of nepotism (Alvard, 2003,Quinlan, 2005).

4.2. The role of other kin in contributing to child welfare

We now turn to the effects (and lack thereof) of parents,of the maternal grandmother, and of other maternallyrelated kin on child survival among the Kipsigis andimplications for understanding the role of kin as caretakersin human societies.

First, consider parents. Loss of a mother but not of a fatheris a strong determinant of offspring survival in the Kipsigis,as would indeed have been expected. Mother loss is animportant influence on mortality in the first years of life inmany populations (recently reviewed in Sear &Mace, 2006).While mother loss does not precipitate infanticide in theKipsigis (unlike in some populations, e.g., the ParaguayanAche, Hill & Hurtado, 1996), breast-feeding children areinevitably weaned rapidly, triggering potentially serioushealth complications that threaten survival, and weanedchildren move to the house of the deceased mother's cowifeor sister-in-law where they are in direct competition withother children. Father loss, in contrast, has no evidentsurvival costs. High mortality risks among children with adeceased father are most commonly attributed to themother's new marriage, specifically the stepparent (Daly &Wilson, 1985; Voland, 1988); indeed, the stress thataccompanies divorce may be more detrimental than thedeath of a father per se (Sear, Steele, McGreggor, & Mace,2002). In the Kipsigis, remarriage does not occur. Widowedmothers are subject to “widow inheritance,” a practicewhereby the woman is effectively passed on to one of herbrothers-in-law; the fact that widowed women do notremarry but cohabit with their late husband's brother likelycontributes to the finding that widowed Kipsigis women donot experience higher child mortality than nonwidowedwomen. Finally, the anomalous and unanticipated findingthat maternal loss is more prejudicial to child survival amongthe wealthy than among the poor most likely reflects the fact

Table 3Kin variables, control variables, and child survival (Model D), with data split by wealth (Model E), and with wealth interaction terms (Model F)

n

Model D

Model E

Model F aPoor Rich

Proportional hazardratio and Z-scoresignificance level

Wald χ2 andsignificancelevel






1. Parental survivalFatherDead b 19 1.49 1.39 0.01 2.40Alive 766 2.350 2.660 4.269MotherDead b 23 20.72 ⁎⁎⁎ 2.73 y 18.07 ⁎⁎⁎ 3.51 y

Alive 762 0.164 ⁎⁎⁎ 0.357 y 0.053 ⁎⁎⁎

2. Paternal kinFMDeceased b 117 19.50 ⁎⁎⁎ 10.58 ⁎⁎ 2.74 6.40 ⁎

Distant 214 0.294 ⁎⁎⁎ 0.173 ⁎⁎⁎ 0.537Local 454 0.337 ⁎⁎⁎ 0.481 y 0.227FFDeceased b 193 3.26 1.12 2.04 0.358Distant 219 0.574 y 0.589 0.455Local 373 0.713 0.629 1.304FBCAT0–1 b 215 7.35 y 3.29 10.11 ⁎⁎ 15.25 ⁎⁎

2 177 0.721 1.620 0.186 ⁎

3 202 0.502 ⁎ 1.683 0.079 ⁎⁎

4–6 191 0.494 ⁎ 0.747 0.162 ⁎⁎

3. Maternal kinMMDeceased b 84 3.59 0.30 3.65 1.39Distant 381 0.727 0.781 0.503Local 320 0.478 y 0.720 0.189 y

MFDeceased b 147 3.35 3.27 0.25 2.05Distant 342 0.873 0.664 0.698Local 296 1.544 1.485 0.963MBCAT0–1 b 112 10.82 ⁎ 7.14 y 4.75 6.77 y

2 320 1.790 1.314 2.189 y

3 292 0.838 0.561 1.3214–7 61 0.588 0.214 0.602

4. Control variablesNo. of cowives0 b 391 10.18 ⁎⁎ 4.58 y 2.25 1.211 317 2.012 ⁎⁎ 1.843 y 2.007N1 77 2.359 ⁎⁎ 2.332 1.481Birth year 785 0.940 ⁎⁎⁎ 26.35 0.948 7.56 ⁎⁎ 0.922 ⁎⁎ 9.87 ⁎⁎ 3.56 y

Single b 756 27.00 ⁎⁎⁎ 9.97 ⁎⁎ 18.70 ⁎⁎⁎ 2.46Twin 29 5.802 ⁎⁎⁎ 6.760 ⁎⁎ 9.314 ⁎⁎⁎

Acres 785 0.968 ⁎⁎⁎ 11.36 ⁎⁎⁎ 0.925 2.31 0.954 3.16 y NANo. of full or half brothers0–3 b 140 4.34 y 5.65 y 0.86 4.46 y

4–8 411 0.998 1.393 1.1649–15 234 1.813 3.318 2.078

Full model loglikelihood χ2

785 138.001, df=23, pb.001 78.34, df=23, pb.001 107.27, df=23, pb.001

a Model F is not strictly a single model. It results from a rerunning of Model D 12 times, each time with a wealth interaction term added for the specifiedvariable.

b Reference category.⁎ pb.05.⁎⁎ pb.01.⁎⁎⁎ pb.001.y pb.10.



that, in this small sample, children in the wealthy sector weresomewhat older at their mother's death.

Second, we examine the absence of a maternal grand-mother effect. Generally, maternal grannies are thought to benatural caretakers of their daughters' children as opposed totheir sons' children because of the greater certainty ofrelatedness, and across many studies, their presence is foundto be associated with positive survival outcomes (Sear &Mace, 2006). In the Kipsigis, however, it is the paternalgrandmother whose presence has a beneficial effect on thesurvival of small children, no doubt reflecting the stronglypatrilineal organization of Kipsigis (also see Gibson &Mace,2005; Leonetti et al. 2005; Lycett, Dunbar, & Voland, 2000).This suggests a facultative component not only to paternalgrandparental behavior (as often surmised) but also, moresurprisingly, to maternal grandparental roles.

We turn now to other maternal kin. In the Kipsigis, thestatus of maternal kin, specifically maternal brothers, isassociated with child survival only in the poorer half of thepopulation. This suggests that maternal relatives provide helpprimarily in situations where limited nepotism from a child'sfather's lineage can be expected. This is the case despite thefact that there is relatively little cultural emphasis onmatrilineal as compared with patrilineal relatives in thispopulation. These results speak to the obvious but oftenignored fact that inheritance rules do not dictate behavior(Dickemann, 1982), as well as to the more specific point thatpatrilineal inheritance rules do not preclude an important rolefor maternal kin under certain circumstances, a phenomenonmuch appreciated by earlier ethnographers of African kinship(Radcliffe-Brown & Forde, 1950) and, indeed, by Peristianywho singled out a mother's true brother as a person “ofastounding importance” (Peristiany, 1939, p. 100) for theKipsigis. The data presented here do, indeed, suggest thatmaternal kin can play a key role in situations where theyobserve a patrilineage unable or unwilling to provide adequatecare to their grandchildren. A similar effect is found among thepatrilineal Ethiopian Oromo, where maternal grannies workhard in their daughters' households; accordingly, livingmaternal grannies tend to have taller granddaughters, althoughthey have no significant effects on the survival of theirdaughters' children (Gibson & Mace, 2005).

Interestingly, the role of maternal kin as indirect providersof child welfare is only in evidence when they live at adistance (over a day's walk) from the mother's maritalhousehold. This is opposite to the Ethiopian pattern wherechild survival is enhanced by the presence of unspecifiedmaternal kin in the village (Gibson &Mace, 2005). An effectsuperficially similar to the Kipsigis is found for non-coresiding maternal grandmothers among the NorthernIndian Khasi (Leonetti et al. 2005) but is attributed to self-selection, specifically the decisions of grannies to live withotherwise disadvantaged daughters. The explanation in theKipsigis is rather different (since women change residence atmarriage, not in grandmotherhood) and suggests a “refuge”hypothesis. In a cultural group where divorce is practically

nonexistent (Orchardson, 1961; Peristiany, 1939), womenare typically unhappy during at least some period of theirmarital life. In such circumstances, and particularly whenviolence ensues, women run away (or at least take theirchildren away) to the home of their parents, where theirbrothers also usually live. Several such instances wereobserved during fieldwork. When the natal homestead isnear to the marital homestead, the husband and his kin arrivein person to demand the wife and her children back. Whenthe natal homestead is at a distance, such delegationsdemanding the wife's return, to the extent that they occur, aregenerally less successful. This is primarily because apatrilineage typically lacks political allies beyond the areawhere it is localized. I therefore view this effect as evidenceof the wife finding safe haven, for herself and her children,with her natal family. To what extent absent kin play this rolein other populations is not known.

5. Implications

How has this analysis of the implications of local resourcecompetition for constraining the role of relatives ascooperators or competitors advanced our understanding ofthe evolution of the human family? Evolutionary anthro-pologists already know that across different societies,distinct categories of kin assist mothers in the arduoustasks of child rearing. Indeed, in many respects, the greaterrole of paternal than maternal kin in the patrilineallyorganized Kipsigis was to be expected. Evolutionaryanthropologists have speculated that this variation reflectsa facultative response to different ecological and socialconditions (e.g., Hames & Draper, 2004) and to the distinctobjectives of different sets of relatives (Sear, Mace, &McGreggor, 2003). However, there has been no applicationof general theory in regard to this variation.

To promote a more general model for which kin helpand why, this study explored the role of resourcecompetition in shaping kin effects, albeit presenting onlycorrelational evidence. The principal finding is that theroles of maternal and paternal kin are conditioned by thewealth of the household into which the child is born.Clearly, there are many factors likely to promote, or curtail,a facultative role for kin in promoting (or inhibiting) childsurvival. These include the availability of kin, thesensitivity of child survival to assistance from outside thenuclear family, and the potential for competition amongkin, all of which are at least indirectly assayed in this study(through kin survival and residence patterns, wealth, and acount of competing siblings). Another important factormight be transaction costs, here and elsewhere modeledthrough residential propinquity (although, in this study, itturned out that distance does not impede imputed patternsof patrilineal assistance and might actually enhancematernal kin contributions). Yet, other factors that shouldbe considered include competing demands on helpful kin,


something that is difficult to measure without a verydifferent sampling strategy.

While there are doubtless many factors promoting orinhibiting kin assistance, the importance of local resourcecompetition (and enhancement) is analytically tractable andhas potential generality, especially in systems whereresources are inherited across generations. While this studyhas restricted focus to the significance of local resourcecompetition for constraining the role of relatives ascooperators or competitors within a single population, thefindings suggest that the variation in kin roles acrosssocieties will be strongly influenced by the elasticity of theresource base, the pattern of inheritance, and the divisibilityof the inherited resources. We now need to shift scholarlyattention that has focused primarily on the factors respon-sible for the cohesiveness of kin groups (Hughes, 1986,Jones, 2000) to the source of their fragmentation, or in otherwords, to situations in which relatives transmute from friendto potential foe.

Acknowledgments

I am thankful to Mark Grote for advice on statisticalmodels, to Craig Hadley and Tim Caro for discussions, andMike Gurven, Richard McElreath, and an anonymousreviewer for helpful comments on the manuscript.

References

Alexander, R. D. (1979). Darwinism and human affairs. Seattle: Universityof Washington Press; 1979.

Allison, P. D. (1984).Event history analysis. Newbury Park, CA: Sage; 1984.Alvard, M. S. (2003). Kinship, lineage, and an evolutionary perspective on

cooperative hunting groups in Indonesia. Human Nature, 14, 129–163.Armelagos, G. J., Brown, P. J., & Turner, B. (2005). Evolutionary, historical

and political economic perspectives on health and disease. SocialScience & Medicine, 61, 755–765.

Beise, J. (2005). The helping grandmother and the helpful grandmother: Therole of maternal and paternal grandmothers in child mortality in the 17thand 18th century population of French settlers in Quebec, Canada.In E. Voland, A. Chasiotis, & W. Schiefenhoevel (Eds.), Grand-motherhood: The evolutionary significance of the second half ofthe female life (pp. 215–38). New Brunswick: Rutgers UniversityPress.

Black, R. E., Morris, S. S., & Bryce, J. (2003). Where and why are 10million children dying every year? Lancet, 361, 2226–2234.

Boone, J. L. (1986). Parental investment and elite family structure inpreindustrial states: A case study of late medieval–early modernPortuguese genealogies. American Anthropologist, 88, 859–878.

Borgerhoff Mulder, M. (1987a). On cultural and reproductive success:Kipsigis evidence. American Anthropologist, 89, 617–634.

Borgerhoff Mulder, M. (1987b). Resources and reproductive success inwomen, with an example from the Kipsigis. Journal of Zoology, 213,489–505.

Borgerhoff Mulder, M. (1989). Reproductive consequences of sex-biasedinheritance for the Kipsigis. In V. Standen, & R. A. Foley (Eds.),Comparative socioecology (pp. 405–427). Oxford: Blackwell ScientificPublications; 1989.

Borgerhoff Mulder, M. (1990). Kipsigis women's preferences for wealthymen: Evidence for female choice in mammals. Behavioral Ecology andSociobiology, 27, 255–264.

Borgerhoff Mulder, M. (1995). Bridewealth and its correlates: Quantifyingchanges over time. Current Anthropology, 36, 573–603.

Borgerhoff Mulder, M. (1997). Marrying a married man. In L. Betzig (Ed.),Human nature: A critical reader (pp. 115–117). Oxford: OxfordUniversity Press.

Borgerhoff Mulder, M. (1998). Brothers and sisters: How sibling interactionsaffect optimal parental allocations. Human Nature, 9, 119–162.

Borgerhoff Mulder, M., &Milton, M. (1985). Factors affecting infant care inthe Kipsigis. Journal of Anthropological Research, 41, 231–262.

Bowles, S., & Posel, D. (2005). Genetic relatedness predicts South Africanmigrant workers' remittances to their families. Nature, 434, 380–383.

Boyd, R. (1982). Density-dependent mortality and the evolution of socialinteractions. Animal Behaviour, 30, 972–982.

Cesar, G. V., Wagstaff, A., Schellenberg, J. A., WGwatkin, D., Claeson, M.,& Habicht, J. P. (2003). Applying an equity lens to child health andmortality: More of the same is not enough. Lancet, 362, 233–241.

Chagnon, N. (1982). Sociodemographic attributes of nepotism in tribalpopulations: Man the rule-breaker. In K. C. S. Group (Ed.),Current problems in sociobiology (pp. 291–318). Cambridge: Cam-bridge University Press.

Chagnon, N. A. (1979). Mate competition, favoring close kin, and villagefissioning among the Yanomamo Indians. In N. A. Chagnon, &W. Irons(Eds.), Evolutionary biology and human social behavior (pp. 213–237).North Scituate: Duxbury Press.

Chagnon, N. A., & Bugos, P. (1979). Kin selection and conflict: An analysisof a Yanomamo ax fight. In N. A. Chagnon, & W. Irons (Eds.),Evolutionary biology and human social behavior (pp. 213–237).North Scituate: Duxbury Press.

Clarke, A. B. (1978). Sex ratio and local resource competition in a prosimianprimate. Science, 201, 163–165.

Cohen, S., Underwood, L., & Gottlieb, B. (2000). Social supportmeasurement and intervention: A guide for health and social scientists.Oxford: Oxford University Press.

Daly, M., & Wilson, M. (1985). Child abuse and other risks of not livingwith both parents. Ethology and sociobiology, 6, 197–210.

Daly, M., & Wilson, M. (1988). Evolutionary social psychology and familyhomicide. Science, 242, 519–524.

Dickemann, M. (1982). Commentary on John Hartung's “Polygyny andinheritance of wealth”. Current Anthropology, 23, 8.

Downey, D. B. (2001). Number of siblings and intellectual development: Theresource dilution explanation. American Psychologist, 56, 497–504.

Emlen, S. T. (1995). An evolutionary theory of the family. Proceedings ofthe National Academy of Sciences of the United States of America, 92,8092–8099.

Frank, S. A. (1998). Foundations of social evolution. Princeton: PrincetonUniversity Press.

Garg, A., & Morduch, J. (1998). Sibling rivalry and the gender gap:Evidence from child health outcomes in Ghana. Journal of PopulationEconomics, 11, 471–493.

Gibson, M., & Mace, R. (2005). Helpful grandmothers in rural Ethiopia.Evolution and Human Behavior, 26, 469–482.

Gowaty, P. A. (1993). Differential dispersal, local resource competition, andsex ratio variation in birds. American Naturalist, 141, 263–280.

Griffin, A. S., West, S. A., & Buckling, A. (2004). Cooperation andcompetition in pathogenic bacteria. Nature, 430, 1024–1027.

Gurven, M., Allen-Arave, W., Hill, K., & Hurtado, A. M. (2001).Reservation food sharing among the Ache of Paraguay. Human Nature,12, 273–297.

Gutierrez, R. G. (2002). Parametric frailty and shared frailty survivalmodels. Stata Journal, 2, 22–44.

Hadley, C. (2004). The costs and benefits of kin: Kin networks and children'shealth among the Pimbwe of Tanzania. Human Nature, 15, 377–395.

Hames, R., & Draper, P. (2004). Women's work, child care, and helpers-at-the-nest in a hunter–gatherer society. Human Nature, 15, 319–341.

Hamilton, W. D. (1964). The genetical evolution of social behavior, I and II.Journal of Theoretical Biology, 7, 1–52.

Hamilton, W. D. (1967). Extraordinary sex ratios. Science, 156:477–88.


Hertwig, R., Davis, J. N., & Sulloway, F. J. (2002). Parental investment:How an equity motive can produce inequality. Psychological Bulletin,128, 728–745.

Hill, K., & Hurtado, A. M. (1996). Aché life history: The ecology anddemography of a foraging people. Foundations of human behavior. NewYork: Aldine de Gruyter.

Hrdy, S. B. (2005). Cooperative breeders with an ace in the hole. InE. Voland, A. Chasiotis, & W. Schiefenhoevel (Eds.), Grandmother-hood: The evolutionary significance of the second half of female life (pp.295–317). New Brunswick: Rutgers University Press.

Hrdy, S. B., & Judge, D. S. (1993). Darwin and the puzzle ofprimogeniture—An essay on biases in parental investment afterdeathHuman Nature, 4, 145.

Hughes, A. (1986). Kin coalitions and social dominance. Journal ofTheoretical Biology, 123, 55–66.

Jones, D. (2000). Group nepotism and human kinship. Current Anthro-pology, 41, 779–809.

Johnstone, R. A. (2000). Models of reproductive skew: A review andsynthesis. Ethology, 106, 526.

Kaplan, H. S., & Lancaster, J. B. (2003). An evolutionary and ecologicalanalysis of human fertility, mating patterns, and parental investment.In K. W. Wachter, & R. A. Bulatao (Eds.), Offspring: Human fertilitybehavior in biodemographic perspective (pp. 170–223). Washington,DC: National Academies Press.

Keller, L., & Reeve, H. K. (1994). Partitioning reproduction in animalsocietiesTrends in Ecology & Evolution, 9, 98–102.

Knodel, J., Havanon, N., & Sittitrai, W. (1990). Family size and theeducation of children in the context of rapid fertility decline. Populationand Development Review, 16, 31–62.

Leonetti, D. L., Nath, D. C., Heman, N. S., & Neill, D. B. (2005). Kinshiporganization and grandmother's impact on reproductive success amongthematrilineal Khasi and patrilineal Bengali ofN.E. India. In E.Voland,A.Chasiotis, & W. Schiefenhoevel (Eds.), Grandmotherhood: The evolu-tionary significance of the second half of the female life (pp. 194–204).New Brunswick: Rutgers University Press.

Low, B. S. (1994). Men in the demographic transitionHuman Nature 5:223–53.Lycett, J. E., Dunbar, R. I. M., & Voland, E. (2000). Longevity and the costs

of reproduction in a historical human population. Proceedings of theRoyal Society of London Series B, 267, 31–35.

Mace, R. (1996). Biased parental investment and reproductive successin Gabbra pastoralists. Behavioral Ecology and Sociobiology, 38,75–81.

Manners, R. A. (1967). The Kipsigis of Kenya: Culture change in a “model”East African tribe. In J. Steward (Ed.), Contemporary change in

traditional societies, Vol. 1 (p. 207–359). Urbana: University of IllinoisPress.

Munshi, K., & Rosenzweig, M. (2005). Economic development and thedecline of rural and urban community-based networks. Economics ofTransition, 13, 427–443.

Mwanza, H. A. (1977). A history of the Kipsigis. Nairobi: East AfricanLiterature Bureau.

Orchardson, I. Q. (1961). The KipsigisNairobi: Kenya Literature Bureau.Parish, W., & Willis, R. (1994). Daughters, education and family budgets:

Taiwan experiences. Journal of Human Resources, 28, 862–898.Peristiany, J. G. (1939). The social institutions of the Kipsigis. Oxford:

Oxford University Press.Quinlan, R. J. (2005). Kinship, sex, and fitness in a Caribbean

communityHuman Nature, 16, 32–57.Radcliffe-Brown, A. R., & Forde, D. (Eds.) (1950). African systems of

kinship and marriage. London: Oxford University Press.Rutstein, S. O. (1984). Infant and child morality: Levels, trends and

demographic differentials. World Fertility Survey Comparative StudiesNo. 43. Voorburg, Netherlands: International Statistical Institute.

Rutstein, S. O. (2000). Factors associated with trends in infant and childmortality in developing countries during the 1990s. Bulletin of the WorldHealth Association, 78, 1256–1270.

Sear, R., & Mace, R. (2007). Evolution and the human family: A review ofthe evidence. Evolution and Human Behavior, in press.

Sear, R., Mace, R., & McGreggor, I. (2003). The effects of kin of femalefertility in rural Gambia. Evolution and Human Behavior, 24, 25–42.

Sear, R., Steele, F., McGreggor, I. A., & Mace, R. (2002). The effects of kinon child mortality in rural Gambia. Demography, 39, 43–63.

Shenk, M. K. (2005). Kin investment in wage-labor economies: Effects onchild and marriage outcomes. Human Nature, 16, 81–114.

Stack, C. (1997). All our kin. New York: Basic Books.Voland, E. (1988). Differential infant and child mortality in evolutionary

perspective; data from late 17th to 19th century Ostfriesland (Germany)L. Betzig, M. Borgerhoff Mulder, & P. W. Turke (Eds.), Humanreproductive behaviour: A Darwinian perspective (pp. 253–261).Cambridge: Cambridge University Press.

Voland, E., Dunbar, R. I. M. (1995). Resource competition and reproduction:The relationship between economic and parental strategies in theKrummhorn population (1720–1874). Human Nature, 6, 33–49.

West, S. A., Murray, M. G., Machado, C. A., Griffin, A. S., & Herre, E. A.(2001). Testing Hamilton's rule with competition between relatives.Nature, 409, 510–513.

Wilson, E. O. (1975). Sociobiology: The new synthesis. Cambridge, MA:Belknap Press.

hamilton's rule and kin competition: the kipsigis case

Documents