Original ArticleLow serum vitamin A mothers breastfeed daughters more often than sons in drought-ridden northern Kenya: a test of the Trivers–Willard hypothesis
Introduction
The Trivers–Willard hypothesis predicts that, under certain conditions of the mother and the reproductive potential of her offspring, natural selection should favor unequal parental investment between daughters and sons. Specifically, the hypothesis predicts that mothers in good condition should increase investment toward sons, while mothers in poor condition should favor daughters (Trivers & Willard, 1973). This prediction is based on the observation that males tend to have higher reproductive variance than females in polygynous mammalian species, including humans. Different variances arise due to the biologically based differential parental investment of mothers and fathers, with females spending disproportionately more time and energy for each offspring from conception to birth and beyond (Roth, 2004). This difference may be amplified in polygynous societies where males can have a large number of mates/spouses, further increasing their reproductive success relative to males with an exclusive single partner (Roth, 2004).
Given this assumption of sex-based differential in reproductive potential, when offspring reproduction is affected by maternal condition, mothers in good condition who can provide more resources to the offspring potentially achieve higher reproductive fitness through preferential investment in sons, while parents in poor condition, who can provide fewer resources, attain potentially higher fitness through investment in daughters (Trivers and Willard, 1973, Quinlan et al., 2005).
In his autobiographical book, Natural Selection and Social Behavior (Trivers, 2002), Robert Trivers outlines the origin and construction of what he terms the “Trivers–Willard effect,” noting that the stimulus for the subsequent model arose when a graduate student, Dan Willard, heard Trivers' lecture on Fisher's (1930) theory of sex ratio adaptiveness. The theory posited that deviations from a 1:1 sex ratio were chronologically unstable because any sex produced in higher numbers has less chance of reproducing, thereby lowering its reproductive fitness value to its parents. Willard applied this logic to human marriage schemes featuring female hypergyny, or marrying someone of higher socioeconomic status (cf. Dickemann, 1979a, Dickemann, 1979b), arguing that men would be favored in higher socioeconomic strata or castes because they could readily find lower-status females eager to be their mate. Conversely, parents from lower socioeconomic strata and castes would benefit from having daughters who could marry above their parental position. Trivers subsequently applied this logic to include all mammalian species by substituting the concept of parental investment for socioeconomic value. While the underlying logic remained the same, substituting parental investment meant that the hypothesis could now be tested with reference to other variables, most notably “maternal condition.”
Previous tests of the Trivers–Willard hypothesis in human populations, particularly those examining postnatal parental investment, overwhelmingly focused on maternal economic resources as indicators of maternal condition. For example, variables such as education (Almond and Edlund, 2007, Freese and Powell, 1999, Koziel and Ulijaszek, 2001), income (Freese and Powell, 1999, Gaulin and Robbins, 1991, Gibson, 2008, Keller et al., 2001), wealth, and socioeconomic and/or occupational status (Hopcroft, 2005, Koziel and Ulijaszek, 2001, Kushnick, 2010, Quinlan et al., 2003) are prevalent in our systematic survey of maternal condition markers used in previous studies testing sex-biased postpartum parental investment (Table 1). This literature survey used the key words “Trivers–Willard” and “human parental investment” to search scientific journal databases (ScienceDirect, Springerlink, Wiley-Blackwell, Web of Science, and JSTOR). Relevant scholarly articles tested the applicability of the Trivers–Willard hypothesis to human populations in both industrialized and nonindustrialized settings. The resulting 35 articles were then summarized for the following features: study population, the hypothesis, the method used, variable(s) used for maternal condition, variable(s) used for parental investment, control variable(s), results, conclusions, and limitations of the research. Table 1 focuses on a subset of 13 articles that explicitly tested postpartum parental investment (as opposed to prenatal investment).
Although maternal biological condition, such as nutrition, may be considered a more proximate variable of maternal condition than economic resources, only one of 13 studies included in Table 1 (Tracer, 2009, using body weight and other anthropometric measurements) specifically adopted the nutritional status of the mother as an indicator of maternal condition. This trend holds true in studies of offspring sex ratio as a generalized extension of the Trivers–Willard hypothesis; only a few studies use maternal nutritional status as an indicator for maternal condition (e.g., Kanazawa, 2005, using height and weight; Gibson & Mace, 2003, using body mass and arm muscle indices), while the majority rely on wealth/status-related conditions (e.g., Almond et al., 2007, Bereczkei and Dunbar, 1997, Mackey and Immerman, 2008).
The Trivers–Willard hypothesis was originally considered the link between the biological condition (e.g., “health” or “body weight”) of the mother and the reproductive potential of her offspring (Roth, 2004, Trivers and Willard, 1973). As such, maternal nutrition is an important yet underutilized condition indicator. Sex-biased parental investment, particularly breastfeeding, has many implications for differences between growth, health, and overall wellness of boys compared to girls. Slower growth and decreased overall health among children have been associated with gender-biased infant feeding (e.g., Sellen, 2010). A study of sex bias in infant feeding patterns in relation to maternal condition has both theoretical and practical importance given the great magnitude of poverty and food insecurity contributing to maternal malnutrition. Micronutrient deficiencies, in particular, including vitamin and mineral deficiencies that can occur in the absence of protein–energy malnutrition, are widespread and persistent in low- and middle-income countries (Black, 2003), affecting upwards of two billion people worldwide (Micronutrient Initiative, 2009).
The study presented here examines the Trivers–Willard hypothesis using macro- and micronutrition status, represented by energy and vitamin A levels, respectively, as indicators of maternal condition and 24-h breastfeeding frequency recalls as the indicator for parental investment. Maternal vitamin A status is an important contributor to maternal condition since low vitamin A levels have functional consequences such as visual impairment and compromised immune responses (Black, 2003, Christian et al., 1998, Micronutrient Initiative, 2009). Severe deficiencies can result in night blindness (Christian et al., 1998) and maternal mortality (Black, 2003, Micronutrient Initiative, 2009). In animal models, vitamin A deficiency in females contributes to a variety of reproductive failures such as unsuccessful implantation, fetal resorption, prolonged gestation with or without fetal death, and congenital malformation of the offspring's eyes and kidneys (Clagett-Dame & DeLuca, 2002).
Section snippets
Study population
The study population features rural mothers in Ariaal agropastoral villages in Marsabit District, Kenya. The Ariaal traditionally practiced seminomadic pastoralism in northern Kenya, subsisting on camel and cattle milk (Fratkin, 1998, Fratkin et al., 1999). In the past 40 years, many northern Kenyan nomadic pastoralists have settled in response to dwindling access to range land, human population growth, and recurring droughts and famine (Roth & Fratkin, 2005). Today, about half of approximately
Sample size and data collection
We use cross-sectional data from mothers who exclusively breastfed their offspring (n=86). This data set is a subset of data collected in 2006 as part of a larger study of mother–offspring vitamin A transfer (Fujita, 2008, Fujita et al., 2011), which included breastfeeding frequency recalls, socioeconomic/demographic information, body mass index (BMI), serum C-reactive protein (CRP), and serum vitamin A (retinol) concentrations from 197 nonpregnant Ariaal mothers in Hulahula, Karare, Kituruni,
Results
The general characteristics of the mothers are summarized in Table 2. The mean (±S.D.) maternal age was 28±7 years old, the mean parity was 3.6±2 (live births), and the mean duration of time since birth was 125±75 days (N=83). Mothers with a son had lower parity (3±2 births, N=46) than those with a daughter (4±3 births, N=37; t81=2.09, p =.04), and more mothers with a son owned livestock (72%) than mothers with a daughter (51%; χ21=4.52, p =.04). Otherwise, these two groups did not differ in
Discussion
This study tested the Trivers–Willard hypothesis on breastfeeding frequency using maternal nutrition variables as the maternal condition indicator. In doing so, it differed from previous research that used socioeconomic variables, rather than biological factors, to determine “maternal condition.” The vitamin A status marker of maternal condition provided results that are consistent with the prediction from the Trivers–Willard hypothesis, while maternal energy or the relative wealth indicator of
Acknowledgments
This study was supported by the NSF Dissertation Improvement Grant #0622358, the Wenner-Gren Foundation, and the Micronutrient Initiative. Logistical and methodological support was provided by the Center for Studies in Demography and Ecology (University of Washington), the Centre for Public Health Research (Kenya Medical Research Institute), and the Center for Statistical Training and Consulting (Michigan State University). The authors are grateful to Ariaal mothers and research staff in
References (57)
- et al.
Simultaneous determination of alpha-tocopherol and retinol in plasma or red cells by high pressure liquid chromatography
American Journal of Clinical Nutrition
(1979) - et al.
Serum, plasma, and dried blood spot high sensitivity C-reactive protein enzyme immunoassay for population research
Journal of Immunological Methods
(2010) The reproductive success of wealthy Americans
Ethology and Sociobiology
(1984)- et al.
Dietary variety increases the probability of nutrient adequacy among adults
Journal of Nutrition
(2004) Big and tall parents have more sons: further generalizations of the Trivers–Willard hypothesis
Journal of Theoretical Biology
(2005)- et al.
The Trivers–Willard hypothesis of parental investment: no effect in the contemporary United States
Evolution and Human Behavior
(2001) - et al.
Parental investment and age at weaning in a Caribbean village
Evolution and Human Behavior
(2003) - et al.
Effects of acute inflammation on plasma retinol, retinol-binding protein, and its mRNA in the liver and kidneys of vitamin A sufficient rats
Journal of Lipid Research
(1996) - et al.
Effects of subclinical infection on plasma retinol concentrations and assessment of prevalence of vitamin A deficiency: meta-analysis
Lancet
(2003) - et al.
Trivers–Willard at birth and one year: evidence from UW natality data 1983–2001
Proceedings of the Royal Society B: Biological Sciences
(2007)
Long-term effects of the 1959-1961 China famine: mainland China and Hong Kong
National Bureau of Economic Research working Paper No. 13384
Results of research into the frequency of exclusive breastfeeding vary depending on the approach taken in the interview
Journal de Pediatria (Rio J)
Female-biased reproductive strategies in a Hungarian Gypsy population
Proceedings of the Royal Society of London Series B: Biological Sciences
Micronutrient deficiency—an underlying cause of morbidity and mortality
Bulletin of the World Health Organization
Women's autonomy and its relationship to children's nutrition among the Rendille of northern Kenya
American Journal of Human Biology
Night blindness of pregnancy in rural Nepal—nutritional and health risks
International Journal of Epidemiology
The role of VA in mammalian reproduction and embryonic development
Annual Review of Nutrition
Low socioeconomic status and female-biased parental investment: the Mukogodo example
American Anthropologist
Female-biased parental investment and growth performance among the Mukogodo
Female infanticide, reproductive strategies, and social stratification: a preliminary model
The ecology of mating systems in hypergynous dowry societies
Social Science Information
A simplified approach of assessing adult energy deficiency
European Journal of Clinical Nutrition
The genetical theory of natural selection
Ariaal pastoralists of Kenya: surviving drought and development in Africa's arid lands
Women's changing economic roles with pastoral sedentarization: varying strategies in alternate Rendille communities
When nomads settle: the effects of commoditization, nutritional change, and formal education on Ariaal and Rendille pastoralists
Current Anthropology
Sociobiology, status, and parental investment in sons and daughters: testing the Trivers–Willard hypothesis
American Journal of Sociology
Anthropometric standards for the assessment of growth and nutritional status
Cited by (8)
Evolutionary psychology: The new science of the mind
2019, Evolutionary Psychology: The New Science of the MindBreastfeeding over two years is associated with longer birth intervals, but not measures of growth or health, among children in Kilimanjaro, TZ
2015, American Journal of Human BiologyLetter to the editor: How to test the Trivers-Willard hypothesis in a regression model: A comment on Fujita 2012
2015, American Journal of Physical AnthropologyMilk immunity and reproductive status among Ariaal women of northern Kenya
2015, Annals of Human BiologyThe evolutionary ecology of early weaning in Kilimanjaro, Tanzania
2013, Proceedings of the Royal Society B: Biological Sciences