3.1. Resources and women's life histories

In most societies, women's attractiveness to men as mates centers on their health and their reproductive value. Thus, cross-culturally, in societies in which the groom or his family transfers goods to the bride or her family, younger women tend to command a higher bride wealth or service than others do. High reproductive value is desirable, and men with greater economic resources are able to marry women with higher reproductive value in such marriage markets. Among the agricultural and pastoral Kipsigis, for example, the required bride wealth was directly related to a woman's reproductive value. With the introduction of western technology and medicine, differentials have been reduced (Borgerhoff Mulder, 1995, Borgerhoff Mulder, 1998a). Borgerhoff Mulder (personal communication) also notes that, in traditional societies, daughters of wealthier families may reach menarche, and possibly the mate market, at younger ages than do daughters of poorer families. Because wealthy men could typically marry younger women than poor men can, in some societies (e.g., Irons, 1979, Irons, 1983), the relationship between wealth and fertility for women mimicked that for men, but was seldom significant (an exception, in a historical society: Low & Clarke, 1991, Low & Clarke, 1992, Low et al., 1992).

In historical Germany, younger, higher reproductive value women tended to marry wealthy men, and parents skewed their investment in children along economic cost–benefit lines (e.g., Voland, 1990, Voland & Dunbar, 1995, Voland & Engel, 1989, Voland et al., 1990). Hughes (1986) found similar patterns in historical England with men's wealth and women's reproductive value. In 19th century Sweden, daughters of wealthy men tended to marry earlier than poorer women do, experience similar interbirth intervals, and have significantly greater fertility (Low & Clarke, 1992). Historical records (Drake, 1969) suggest that poor men sought older women with greater resources when they could, explicitly trading reproductive value for resource value. In 18th and 19th century Scandinavia, daughters of upper–middle class women (who would marry richer men) were considered marriageable at 18 years, while daughters of poorer men, who would marry poorer men, were not considered marriageable until their mid-to-late 20s (Drake, 1969, Low, 1991). Poorer women also were likelier to move from their home parish (Clarke, 1993, Low & Clarke, 1991).

Resources were thus relevant to women's age-specific fertility and mortality, likelihood of migration, and lifetime reproductive success. The direct effects of resource availability on women's fertility tend to be concentrated (above) at the low end of variation: Women with poor nutrition suffer infecundability and loss of fetuses and children. More subtle effects could arise, depending on whether resources were husband's resources, or whether the woman herself garnered them. In the German, English, and Swedish cases above, resources were almost entirely controlled by men. When women obtain their own reproductively important resources, there is a conflict: Effort spent garnering resources cannot be spent allocating the resource in care of offspring. Thus, among the Ache and the Hiwi (Hurtado, Hill, Kaplan, & Hurtado, 1992) and the Ye'kwana (Hames, 1988), for example, women who are nursing cannot gather as much as other women can.

Nonadaptive cultural preferences can affect women's fertility in wealth- or class-specific ways. In 19th century northern Sweden, for example, people adopted the practice of feeding infants cow's milk (a common practice in nearby northern Finland). Refrigeration was lacking, and infant mortality was high. The central Swedish government sent a medical doctor, along with civil servants, to the northern town of Haparanda. He spoke mainly with upper class diplomats and civil servants—but the household staffs in these houses also gained the information. At the upper and lower ends of the class spectrum, health information and resulting infant survivorship were better than for tradespeople and small holders. Similarly, in some parts of 17th century Europe, women who could afford it hired wet-nurses to nurse their children. The reproductive impacts of this practice varied for women of different statuses (see Hrdy, 1999). Wet-nurses fared worst (in terms both of conceptions and loss of infants). Wealthy women had short interbirth intervals and high conception rates, but many children whose care was delegated to other women died. The women whose reproductive lives were least disrupted, and most successful, were lower–middle class women who neither could afford to delegate their nursing activities nor were forced to take on the care of other women's children for pay.

3.2. Marriage system and women's options

A woman's options are constrained by the marriage system. Marriage systems also (below) give rise to further patterns that, in turn, affect women's possible options in life. For humans, as for other mammals, polygyny is the “default” mating system (even when marriage systems are specified as other than polygynous): When some males can dominate mating opportunities, they will. Over 85% of human societies for which data exist are polygynous.

Ecological factors (e.g., availability of large game, making hunting likely; ungulate presence, making animal husbandry likely; e.g., White & Burton, 1988) correlate with the likelihood of polygyny. There is a less well-known correlate, degree of pathogen stress, which may exacerbate sexual selection, increasing the degree of polygyny (Hamilton, 1980, Hamilton & Zuk, 1982). Polygyny (Low, 1988b, Low, 1990a, Low, 1990b, Low, 1990c) is common in many conditions, but in areas of high pathogen stress, monogamy is absent. Furthermore, although sororal polygyny is preferred cross-culturally, in areas of high polygyny, nonsororal polygyny is more common than expected, as is the capture of women from other groups (both predicted by the polygyny hypothesis and no other; Low, 1988b, Low, 1990c). Pathogen stress may be a real environmental uncertainty that renders fewer men acceptable as mates and further favors the production of genetically variable offspring.

3.3. Marriage system, choice, and voice

Humans are unusual among mammals in that women, in most societies, do not simply exert mate choice, as do most female mammals—although women may informally make their views known. Marriage (within which much reproduction occurs, of course) serves functions of alliance to strengthen a lineage's resource control and power, both within and between groups (cf. Alexander, 1979).

One anthropological measure that attempts to capture the informal influence of women (e.g., Whyte, 1978) is women's “voice” or “say” in important alliances, such as marriage arrangements. We might imagine that women would have more informal and formal influence under two conditions: in monogamous societies (because men's and women's interests are likely to converge there more than in many polygynous conditions, and because women may gain more paternal care under monogamy) and when women contribute more than men do to the subsistence base. Neither of these predicted associations is as simple as they might at first seem.

3.4. Marriage system and conflicts

In polygynous systems, there is clear potential for sexual conflict of interest: Are women's interests suppressed or in conflict with men's in these systems (e.g., Davies, 1989, Hrdy, 1979, Irons, 1983, Smuts, 1992, Smuts, 1995)? In many cases, yes. In about 40% (65/157) of traditional societies for which data exist, for example, women are captured as brides or concubines. Captured women live far from kin and other possible allies, while their husbands remain embedded in a male kin network. [In a few other primates, such as red tail (Cercopithecus ascanius) and patas (Erythrocebus patas) monkeys, males may form group alliances and capture females; e.g., Cords, 1986.]

Among the Dogon people of Mali (Strassmann, 2000), the more co-wives a woman has, the greater the likelihood that her child/children will die; the nonrandom mortality rate arises from conflicts of interest among those who share both agricultural work and food (“work–eat group members”; Strassmann, 1997, Strassmann, 2000). Dogon men who can be polygynous gain reproductively, but their wives lose. In polygyny, generally, women's interests do not converge with either their husbands' or their co-wives' interests. If the system is sororally polygynous, we do expect the interests of sisters/co-wives to converge more than that of those in nonsororally polygynous systems.

The costs and benefits of polygyny for women are not even, nor necessarily long lasting. For the wives of polygynous men in traditional societies, fertility may be uneven; in a number of societies, second and subsequent wives, for whatever reasons, have fewer children than first wives do (e.g., Daly & Wilson, 1983, Garenne & van de Walle, 1989, Pebley & Mbugua, 1989, Sichona, 1993). The fertility effects are sometimes small (e.g., Timaeus & Reynar, 1988) and sometimes ephemeral in terms of number of grandchildren (Josephson, 1993). There are many covariants of fertility (e.g., women who remain unmarried until their reproductive value is low are unlikely to become first wives and also likely to have few children in their remaining lifetimes); the most detailed and methodologically elegant study so far is Strassmann (2000).

Polygyny is not the only marriage system in which wife–husband conflicts of interest are likely. Because of the socioecological complexities of human systems, women's interests often do not converge completely with their husbands' in socially monogamous (Alexander et al., 1979, Reichard, 2004) situations, as they would in genetically monogamous systems (Low, 2004). The reproductive interests of parents converge to a degree—but this does not necessarily protect women's reproductive interests, as when divorce is common. Human life histories are unusual among mammals; offspring are extremely altricial and require considerable care for extended periods of time (reviews by Kaplan, 1997, Low, 1998). Social monogamy with intense parental investment by both parents does not preclude extra-pair copulations (EPCs; and fertilizations), but any resulting children are likely to command less investment than will children born within socially recognized unions.

While women's interests converge imperfectly with their husbands in socially monogamous systems, women are likely to see monogamy as serving their interests. It is important that the conditions that favor monogamy can arise from either ecological or social forces. The origins of socially imposed monogamy appear to arise from the pressures of group living, interacting with historical particulars (e.g., Alexander, 1979). As societies increase in size, under many conditions, it becomes more difficult for powerful (and thus polygynous) men to maintain their reproductive advantages openly. Social monogamy typically follows. In these circumstances, EPCs may be reproductively profitable, and husbands' and wives' interests often conflict. A number of sociocultural conventions (legitimacy, inheritance rules) arise in most societies to mediate these conflicts, and men and women are likely to use them differently; for example, men and women tend to distribute their bequests differently, in ways that reflect these conflicts of interest (Gaulin et al., 1997, Judge, 1995, Judge & Hrdy, 1992).

Polyandrous systems are rare and associated with either resource constraints (as among the Lepcha in northern India, where it appears to take the work of two men to produce enough to support the children of one woman) or conservation of familial resources within some lineages (as among some Tibetan communities). There are too few data to tell whether women's interests thrive in these systems. Thus, with current information, it is not clear that any particular marriage system automatically favors women's resource-accumulation or reproductive interests.

3.5. Other factors influencing conflicts of interest

In societies (regardless of marriage system) in which men's sources of power are unpredictable, and women have sufficient resources to be independent, men cannot always control women. Resource distribution, coalitions for mate seeking, inbreeding avoidance, and nepotism (familial coalitions)—all are important in marriage patterns. In such societies, either straightforward polygyny or serial monogamy (really serial polygyny) results. Serial polygyny occurs, for example, among societies as disparate as the Ache of Paraguay and the contemporary United States. Ache men and women have perhaps 10 spouses in a lifetime (most occur early in their reproductive careers; Hill & Hurtado, 1996), U.S. women roughly 1–4.

4.1. Female coalitions

Female–female coalitions in primates, including humans, tend to center on parental effort (e.g., Irons, 1983) and on cooperation to protect females from male–female conflicts of interest (e.g., Smuts, 1992). Implicit in the contrast between the “mating effort” emphasis of male coalitions and the “parental effort” emphasis of female coalitions is that the amount of resources under consideration differs—mating effort can turn huge amounts of resources to reproductive profit; typically, parental effort cannot (Low, 2000b, Chap. 3). Female parental effort coalitions tend to be among kin or (in humans) co-wives. Both “positive” (food gathering) and “negative” (e.g., reproductive suppression) parental effort coalitions exist. Reproductive suppression can have bloody outcomes. Marmoset females may kill other females' offspring (Digby, 1995, Lazaro-Perea et al., 2000). In several baboons, females in dominant matrilines harass subordinates (e.g., Wasser, 1983) and sometimes kill their offspring.

In many societies, co-wives gather food together and share information (e.g., Irons, 1983). Women may have considerable informal say not only in family alliance and marriage matters but also behind-the-scenes influence in larger political matters, through their spouses and male relatives; this, however, is difficult to document. Because reproductive return curves differ for mating and parental efforts, it should be no surprise that the payoffs differ for mating-effort versus parental-effort coalitions. Parental-effort coalitions are more likely to be among related women or co-wives and to be longer term, to be relatively more stable, and to involve lower risk than mating-effort coalitions do (more typical of males).

Despite the inherent potential conflicts of interest in polygyny, women do well reproductively in at least some polygynous systems. In part, this appears to arise from the differential ability of women to find allies and form coalitions in different systems. Sellen, Borgerhoff Mulder, and Seiff (2000) found that among Datoga pastoralists of East Africa, women in marriages with a single co-wife produced fewer surviving offspring than did women with more co-wives; furthermore, their children grew more poorly during their first 3 years. Resources play a part in this pattern; Sellen et al. noted that, among the rich, the number of co-wives is irrelevant, but that in other marriages, women's direct access to critical resources is crucial. Among the swidden agriculturalist Yanomamö of the Orinoco valley in South America, there is little in the way of wealth or heritable resources. Here, a man of high status (who also is likely to have more wives and more children than other men will) received more food gifts from others in the village than poorer men did. His wives thus have more food available (Hames, 1996).

Cross-culturally, women sometimes have significant allies in polygynous systems (Yanca & Low, 2004), in part, depending on whether the polygyny is sororal (wives are sisters) or nonsororal. When co-wives are sisters, female coalitions can exert considerable influence in the kin group; when women live close to their natal families, as well, they have nearby allies. Hence, the relationship between women's interests and polygynous/nonpolygynous systems is not an obvious one; as Borgerhoff Mulder (1992) has noted, women adopt strategies in the face of conflicts of interest and female–female competition, and there is a hierarchy of questions that need to be addressed.

4.2. When do women control significant resources?

Cross-culturally, the major influence on women's ability to control resources was the level of female contribution to subsistence; it showed a linear pattern (greater female contribution, greater female resource control; Low, 1990a, Low, 1990d, Low, 1992). Some proposed that broad ecological correlates of female resource control, such as protein deficiency and risk of starvation, show no correlation. Subsistence modes show patterns: that women in fishing and hunting societies were likely to have certain kinds of resource control, and women in irrigation, agricultural, and animal husbandry societies were likely to have little resource control. However, the subsistence mode is related to relative male and female contribution to subsistence (which independently influences women's resource control), residence, and other social phenomena, which renders any proposed causality less clear (cf. Yanca & Low, 2004).

Women compete directly for resources primarily in two conditions, both more prevalent in Africa than elsewhere. In complex matrilineal or dual-inheritance societies, a powerful woman, while not increasing her own number of children, may increase her number of grandchildren by passing wealth and power to her son, as in the Ashanti people of Ghana (Low, 2000b, Chap. 12). Even when such systems are changed by contact with industrialized societies, traditions of women's independence and power, although they no longer yield reproductive gains, may persist and thrive, as in West Africa today.

Within polygynous societies, the greater the percent of polygynous marriages (degree of polygyny), the more women and their children tended to become independent resource units. Cross-culturally, the more women contributed to the subsistence base, the more say they had over the distribution of the fruits of their own labor (Low, 1990a, Whyte, 1978).

4.3. When do women hold significant political power?

We normally think of nation-states and formal political arrangements when we hear the word politics. Yet, of many definitions, the definition of politics of primatologist Frans de Waal (1982) as “social manipulation to secure and maintain influential positions” is a biologically useful one and applicable to women as well as men. The art of politics, defined this inclusively, is older than humanity itself. De Waal commented (p. 19) that “whole passages of Machiavelli seem to be directly applicable to chimpanzee behavior.” But what males, who specialize in mating effort, versus females, who typically invest in offspring-specific maternal effort, stand to gain from political maneuvering differs greatly (above). If coalitions had reproductive ends in our evolutionary history, and if men and women succeeded reproductively through different strategies, then men's and women's coalitions should differ (e.g., Low, 2000a, Low, 2000b, Chap. 11). Not surprisingly, human sex differences, similar to those of other primates, reflect these divergent evolutionary ends.

Cross-culturally, although women's participation varies in many domains, public political power is almost exclusively male (Low, 1990d, Low, 2000b, Chap. 12). About 70% of societies cross-culturally have only male political leaders; just over 7% have both sexes, although men are more numerous and more powerful (Whyte, 1978). In very few societies are women coded as equally powerful as men are (and men are always more numerous). Women are more likely to have influence within their kin groups than in community affairs. Women can hold kin leadership posts and have more say and influence in kinship matters in 6 of 93 societies. Both sexes have roughly equal influence in 3 other societies. In a few, women can attend and participate equally in community gatherings, although they may be segregated from men. While women may widely have an informal voice in community affairs, formal political influence is rare for them, both within the kin group and in the community at large.

The very rarity of women's public political power (9/93 societies; Whyte, 1978) makes statistical analysis difficult. Four of the nine societies are African, suggesting a possible geographic bias, but in a study by Whyte (1978), African societies comprise over a fourth of the odd-numbered societies of the Standard Cross-Cultural Sample (a sample chosen to be representative: stratified by geographic region and within region by language group, and with data from trained ethnographers; Murdock, 1981, Murdock & White, 1969). In fact, the geographic association appears weak (Low, 1992).

Descent rules comprise the one strong pattern with women's large-scale political activity (Low, 1992). Five of the nine societies in which women are said to have such political power are societies with matrilineal or double descent, precisely those societies in which women's power is predicted to enhance their son's (not their own) reproductive success. Over half of the cases of women having political power occur in one fifth of the (complete) Standard Cross-Cultural Sample: the matrilineal and duolineal societies. Of the 18 matrilineal and double-descent societies, 5 have women in political office (2 expected); of the 56 patrilineal, ambilineal, and bilateral-descent societies, 4 have women in political office (7 expected). Thus, despite the difficulties of recording descent systems in great detail, there does appear to be an association between descent systems and women's ability to hold public office.

5.1. Inheritance

Hartung (1982) found that inheritance tends to be male biased in polygynous systems because resources influence male reproductive success more than female reproductive success in such systems. In bride wealth systems (above), young, high-reproductive-value women cost more than older women do so that parental gains through sons are more affected by material investment than parental gains through daughters. Cowlishaw and Mace (1996), using a phylogenetic approach, confirmed this pattern, noting that marriage patterns and inheritance “evolve together in a way that is adaptive” that strongly polygynous systems have male-biased inheritance and monogamous ones are more characterized by an absence of sex-biased inheritance.

5.2. Status and the treatment of daughters

Sex-differential investment can take more subtle forms than inheritance biases do and can operate both as physiological and as cultural biases. The evolutionary framework leads us to expect some bias in the treatment of dependent sons and daughters, because in many polygynous mammals, only the best-invested males have much chance of reproducing, but any daughter is likely to have some reproductive success. The guiding selective principle appears to be the following: When parental effort is effective and variance in the reproductive success of one sex exceeds that of the other sex (as in elephant seals), or when resources or parental investment can influence the reproductive success of one sex more than the other, there should be a correlation between parental condition and investment in that sex (Low, 2000b, Trivers & Willard, 1973, Clutton-Brock, 1991).

In nonhuman species, and in many preindustrial societies and developing countries, a mother's physiological condition (resources available to rear a successful offspring) typically declines with age. But whenever the nutritional condition of mothers does not decline with age, a male bias in sex ratio might be expected in older mothers (e.g., Low, 1991, Mace, 1990). If a female's condition is good as she nears the end of her reproduction, it may pay to invest ever more heavily, with a greater potential reproductive profit if successful; for example, older female gorillas tend to have sons (Mace, 1990). In nutritionally rich 19th century Sweden, mothers over age 35 showed a sex-ratio bias toward sons, and mothers under 25, toward daughters (Low, 1991, Low, 2000b, p. 105). It is interesting, in light of these physiological patterns, that parents shift their preference for sons versus daughters under different circumstances (Knodel, 1988).

5.3. Dowry

Whenever males vary greatly in their status and wealth, women married to wealthy, high-status men (and the women's children) benefit reproductively. Hence, it may pay would-be brides (usually, the brides' fathers) to compete, bargaining for wealthier men as mates. Dowry is 50 times more common in monogamous, stratified societies than in polygynous or nonstratified ones (Gaulin & Boster, 1990, Gaulin & Boster, 1997). In many of these societies, poorer women's families must pay more dowry than wealthy women's families do; dowry becomes a form of mate competition among women. Insofar as poor families cannot pay well, the stratification is intensified. Such competition can incur costs. Consider dowry in modern rural India. Since about 1950, demographic shifts have resulted in a decline in potential grooms for potential brides of marriageable ages—and dowries have risen steadily. By 1990, a dowry was likely to be over 50% of a household's assets. Wives from poor families, able to pay less dowry, are less likely to marry; if they marry, they have a high risk of being abused by their husbands. Furthermore, domestic violence and spousal abuse have been correlated with increases in dowry worth; “insufficient dowry” is an important recorded cause of spousal abuse (Rao, 1997).

5.4. Training sons versus daughters

We know a little about the genetic bases of behavioral differences, but equally important are sociocultural differences in how sons versus daughters are trained (Low, 1989, Low, 2000b, Chap. 6). And we know a lot about that. Across cultures, sons and daughters are trained differently in ways that relate to the evolutionary history of each sex's reproductive success. Cross-culturally, sons are more strongly trained than are daughters in behaviors useful in open competition, while daughters are more strongly trained in such values as sexual restraint, obedience, and responsibility—traits widely sought by men in their wives. Put bluntly, boys get more practice than girls do in negotiating their best interests, starting early in life. Furthermore, the greater the degree of polygyny in nonstratified societies (the higher the potential reproductive rewards), the more that boys are taught to strive.

Cross-culturally, the more women actually control important resources or exercise power, the less that daughters are taught to be submissive. For example, the more that women are able to inherit property, the less that daughters are taught to be obedient. The more formal power that women have within the kin group, the more that daughters are taught to be aggressive and the less they are trained to be industrious.

5.5. Sex-preferential infanticide

The extreme of differential parental allocation is sex-biased infanticide. Parents will do best to allocate effort into the reproductively most profitable sex (e.g., Alexander, 1979), thus, infanticide by genetic parents is relatively rare. Infanticide in most species is typically committed by reproductive competitors rather than parents; for example, in lions, langurs, and gorillas (Packer & Pusey, 1983, Watts, 1989; see Hrdy, 1979, for an overview), when a male takes over a harem, he is likely to kill all infants under a certain age. The mother becomes sexually receptive and the male profits both by eliminating an offspring with a competitor's genes and by gaining a mating. In humans, stepparents (whose reproductive interests do not coincide with the child) are more likely to abuse or neglect children than genetic parents do (Daly & Wilson, 1996).

But genetic parents can and do commit infanticide. Among the Innuit, infanticide patterns suggest that parents are attempting to manipulate the sex ratio of their children to maximize mate choice resulting from local sex ratios (keeping daughters when the local ratio is male biased, and vice versa: Smith & Smith, 1994, Smith, 1995). In some caste systems, women, but not men, can “marry up” (hypergyny). High-status men can marry women of all castes; low-status men can only marry low-status women. In these circumstances, sons and daughters differ in reproductive value to high- versus low-status parents. One result is female infanticide by high-status parents (Dickemann, 1979, Parry, 1979).

Cross-culturally, female infants are at risk under predictable socioecological conditions, including plow agriculture and herding, patrilineal inheritance, dowry, and low resource contributions by women (e.g., Bugos & McCarthy, 1984, Dickemann, 1979, Hill & Low, 1991, Low, 1993, Low, 1994, Torres & Forrest, 1988; see also Low, 2000b; Whyte, 1978, for the cross-cultural patterns in traditional societies). Broad historical patterns are also consistent with these findings (see review in Low, 2000b), just as sociologists note that infanticide can be biased according to cultural practices.

Today, our best efforts may have unintended negative results. One modern thrust to combat female-preferential infanticide argues that improving women's education and employment will improve the status of women and result in better treatment of daughters. This reasonable argument may be stated too generally; recent demographic reports from India, China, and Korea suggest that when women's status improves, preferential allocation of resources to sons increases—the opposite of the desired outcome (Dasgupta & Visaria, 1996, Premi, 2001). In at least some populations, the education of women correlates positively with female infanticide (Dasgupta & Shuzhuo, 1999, Dasgupta & Visaria, 1996). It is not at all clear that raising women's status will protect female infants; discrimination against daughters may actually increase with women's status in these countries (Anderson & Romani, 1997, Dasgupta & Visaria, 1996).

5.6. When are daughters treated preferentially?

The biases arising from our mammalian background mean that successful sons in most circumstances cost more than successful daughters do (cf., e.g., Alexander, 1979). Thus, not only do we see biases in physiological investment (Trivers & Willard, 1973) but in behavioral care (Borgerhoff Mulder, 1998a, Mace, 1996), inheritance (Hartung, 1982, Hartung, 1997), and other social investments, even today (Gaulin & Robbins, 1991). Among the Kipsigis (Borgerhoff Mulder, 1998a), poorer families (who have little bride wealth for their sons, but whose daughters bring in bride wealth) invest preferentially in education for their daughters.

Similarly, the Mokogodo in central Kenya are a relatively poor people without power or prestige compared with neighboring groups, such as the Maasai, the Samburu, and the Mumonyot, whom they admire. The Mukogodo are somewhat stigmatized by their neighbors. This fact has had curious consequences. The Mukogodo treat their daughters better than they treat their sons (Cronk, 1991a, Cronk, 1991b, Cronk, 1991c, Cronk, 2000, Cronk, 2004), and this makes selective sense: Daughters are more able to marry and continue the family lineage.

Mukogodo mothers nurse their daughters longer than their they do their sons; Mukogodo caregivers (mothers and others) stay closer to girls than to boys and hold them more. Parents take their daughters more frequently than their sons to the dispensary and clinic for treatment and enroll their daughters more in the local Catholic mission's traveling baby clinic. Because of the biased care, girls show better growth than boys do: better height for age, weight for age, weight for height. The Mukogodo preference for daughters makes good behavioral ecological sense. Yet, the situation is nuanced, and, living among societies that have a strong son preference, the Mukogodo, if you ask them, claim (like the Maasai, whom they try to emulate) to prefer sons to daughters (Cronk, 1991a, Cronk, 1991b, Cronk, 2000, Cronk, 2004).

6.1. Variation and scale in modern societies

Today, of course, the influence of marriage systems on women's ability to gain resources independently and to wield power is even less obvious or direct than in traditional societies. In modern western democracies, serial polygyny (social monogamy with remarriage) is a widespread rule; women once again typically have free mate choice. Arranged marriages do exist in some democracies, although they appear less common than previously (e.g., in India, among some castes). Women in most nations may hold political office and are not restricted from particular jobs. Cultural factors beyond marriage rules may have significant impact. Religious dogma may affect women's lives; in some Muslim societies (particularly the less secular ones), women's dress codes may be severely mandated. They may be subject to clitoridectomy, may not have free mate choice, and may be restrained from daily activities we think of as common, such as voting or driving a car. Although these conditions may not prove conflicts of interest, they are suggestive.

6.2. Work, resources, and women's reproduction

Resource control clearly influences individual reproductive success in other mammals and in traditional and historical societies. Today, evolutionarily novel conditions complicate the picture. What is “wealth”? Men's earnings? Women's? Particularly in modern Western societies, women's tradeoffs are heightened: There are few professional (i.e., high-paying) jobs that women can perform while simultaneously doing childcare. We need appropriate repeatable measures of wealth, information on women's own wealth, separate from husband's or household wealth, and individual (rather than aggregate) data on age-specific fertility.

Taşiran (1995) carefully sorted out relevant covariants for Swedish, white U.S., and Black U.S. women. The results indicated that major general patterns remained, but that considerable complexity exists in transitory (and probably nonselective) economic and fertility preferences. Tasiran compared empirical studies and models for Swedish (pre-1935 to 1965 cohorts) and white and non-white U.S. women (pre-1945 to 1966 cohorts), using macro-, micro-, and integrated data (also Anderson & Low, 2003). For non-white U.S. women, 5–9 years of education made having a first child less likely; 10 or more years made having a first child more likely. For white U.S. women, significant negative effects of education existed for first and second births. For Swedish women, the effects were significant for first and third child (5–9 years; an additional effect was seen on first child for 10+ years of education).

Women's work has always appeared to present women with a fertility tradeoff harsher than men's, but today is it money, or time? The analysis of Taşiran (1995) showed how the tradeoffs work for these populations: When women's work experience was controlled for, women's wages had several positive effects on aspects of fertility. Women's work affects the likelihood of a first birth for Swedish and non-white U.S. women and second births for white U.S. women. This contrasts sharply with earlier studies, which confounded the tradeoffs of time and monetary gain for women's work.

Some differences appear to be simple cultural preferences, others seem to reflect the cultural costs and benefits of having a child. The effect of cohort was significant, reflecting the fact that both economic conditions and cultural preferences may vary over time. In Sweden, whether a woman was married or cohabiting (vs. alone) had no effect on her likelihood of having one or more children; social services for parental benefits had a positive influence only on the probability of having a third child. A man's wages increased the likelihood of having a first and (independently) a second child. In the study of Taşiran (1995), in the United States, a man's income had a negative relationship to a woman's probability of having a first, second, or third child. We have no information (e.g., lineage persistence) to help us understand whether this is a strategy shift (higher per capita investment per child, fewer children), a reproductively costly cultural preference, or simple co-variation with woman's age, since “husband” includes second or later husbands.

7.1. Work and family

The demographic transition, when family sizes fell from, for example, 8–10 to 2–3 children, is something of an evolutionary conundrum (e.g., Borgerhoff Mulder, 1998b), particularly in the context of women's resource control. It makes some ecological sense that, as competition to marrying and getting established increases (as in 19th century Europe and North America and in the developing nations today) and, thus, the cost of each successful child increases, two things might follow: (1) family sizes would fall; and (2) women might shift strategies from pure reproductive value (probable number of children that a woman will have in the remainder of her life, given prevailing age-specific fertility and mortality patters; Fisher, 1958) to a mixture of resource and reproductive value (Boone & Kessler, 1999, Kaplan & Lancaster, 2000, Kaplan & Lancaster, 1999, Low, 2000a, MacArthur & Wilson, 1967). That is, a woman's value might be not only the children that she might have but also the resources that she could contribute to those children's upbringing. Market factors are important in modern societies as diverse as New Mexican Latinos and Thais (Kaplan et al., 1995, Knodel et al., 1990). This represents an important teasing apart of influences; in early biological models, the “cost of getting established” was only influenced by conspecific density, clearly too simple a situation to describe human shifts. Mace, 1998, Mace, 2000 modeled the evolution of human reproductive rates when wealth is inherited and has shown that that the cost of children has a large effect. hence, women's shifts, in the face of shifting market competition and increasing per capita costs of children, make selective sense. But have we gone too far, in evolutionary terms (Borgerhoff Mulder, 1998b, Low, 2000a)?

Women today in western developed nations tend to marry later than in traditional and historical societies. Many women who marry relatively late work before (and often after) marriage—increasing their resource base. Family sizes are relatively small, and variance in family size is compressed, compared with traditional and historical societies. In general, women's resources and their reproductive success are positively correlated in studies that examine individual patterns and some census data. Resource effects are primarily at the very low end of the resource spectrum; for example, women on welfare have very low fertility (Acs, 1996, Rank, 1989). One study of census data (Daly & Wilson, 1983, p. 334) also found a positive relationship. Here, household income (not specified, but in that period largely male income) was part of the equation. Taşiran (1995) found that the impact of work, years at work, and income differed for white and Black American and Swedish women. In no case were the impacts of work on fertility wholly negative (above). Yet, it seems clear that today, even as among hunter–gatherers, women face tradeoffs in the expenditure of effort to get resources versus investment of existing resources into children.

7.2. Demographic transition in developing nations

Developing nations today are undergoing demographic transitions to smaller families, as the West did in the 19th century. Once, industrialization was seen (incorrectly: e.g., Schofield & Coleman, 1986) as the proximate cause of fertility decline. It is not a direct major force in modern transitions, but there is a deeper similarity of interest to evolutionary biologists: the degree of per capita investment required to produce highly competitive children, such as market forces (e.g., preferences in the job market; above). And within societies, even when wealthier families have fewer children than poorer families do, the income available for each child remains enormously higher for children in wealthy families (Birdsall, 1980, Birdsall & Griffin, 1988, Clarke & Low, 2001, Low, 2000a). In the past, the number of children was the best predictor of number of grandchildren; today, it is often not. Are wealth and offspring simply no longer related? This is clearly a question requiring further comparative work.

7.3. Modern adaptive and nonadaptive responses

Today, as in the past, individuals respond to both ecological and social challenges, and the major demographic shifts that we see are the accumulated effects of individual responses. Some are adaptive. For example, in life history theory (e.g., Roff, 2001, Stearns, 1992), a primary force influencing the age at first reproduction is the life expectancy of newly matured adults: When life expectancy is short, and adult mortality high, reproduction is early and intense. Early and late reproduction represent life-history tradeoffs that are affected by the environment, in particular, the probability of mortality. Early reproduction means a shorter exposure to mortality before reproduction (seen in dangerous environments); later reproduction (which occurs when adults are safe) means that a mother can have more, or larger, offspring in her first reproductive bout. Daly and Wilson (1997), comparing reproductive patterns in 30 Chicago neighborhoods differing in adult life expectancy, found this general rule to hold. Women in the neighborhoods with the shortest life expectancy had children earlier in life than did women in neighborhoods with the longest life expectancy.

Other shifts in modern patterns appear similarly adaptive; for example, fertility among poor urban Black women and women in ex-Soviet countries is relatively early (e.g., Geronimus, 1996, Geronimus et al., 1996). Such women face declines in their health and fecundability and in their support systems.

The least obviously adaptive shifts in Western nations today are seen among women who seek careers such as, CEOs, MDs, lawyers—occupations in which success is associated both with extremely high effort “on the job” and with some risk taking (e.g., Browne, 2002). Fortune magazine reviews the Top 50 women in American business annually. In each survey, these women had very few children of their own, a stark contrast to the Top 50 men in American business. Thus the question: Under what circumstances can a woman's delay of fertility compensate in increased resources enough to be reproductively profitable? Models by Low et al., 2002, Low et al., 2003 suggest that such circumstances are highly constrained. Women who are in the lowest socioeconomic strata in these models suffer very high infant and child mortality. When such women delay fertility enough to complete high school, their job opportunities are better, and they move up into strata with lower childhood mortality; they (or their daughters) then do better reproductively. There are essentially no conditions in these models under which delay of first birth past age 25–28 (except in declining populations: Fisher, 1958) can be reproductively profitable.

Surveys of men's and women's attitudes about sexual matters in modern Thailand (Knodel, Saingtienchair, Low, & Lucas, 1997) highlight sexual conflicts of interests (in a society in which resources are primarily male) and potentially adaptive responses under complex conditions. A long tradition exists of men, married as well as single, visiting sex workers in the context of men's social evenings out. In addition, some men have “minor wives” whom they support at least partially. Thai wives have historically not had the ability to constrain these behaviors and have held a long-standing preference that their husbands, if sexually active outside marriage, visit sex workers, because the transaction is brief and represents only a minor diversion of resources from the wife and children, in contrast to a minor wife, who can represent a real drain of resources. Interestingly, with the spread of HIV virus, an increasing proportion of women are responding to this sexual conflict of interests by preferring the husband to have a minor wife: The resource conflict is heightened, but the health risks are reduced.

7.4. Women, status striving, and coalitions

The very large scale and internal diversity of modern nation-states make generalizations premature. Certainly, women in many families continue, as in traditional societies, to work with female kin in childcare, for example. But both women's options for coalition building and the potential foci of such coalitions are far broader in most modern societies than in traditional and historical societies. Women may seek status and resources directly, not through association with powerful men, and we have probably underestimated the intensity of competitiveness among women (e.g., Campbell, 1999, Campbell, 2002, Cashdan, 1998). In mobile societies, such coalitions may also become more fluid; one's friends and allies change more over the years than in small-scale societies. In these regard, women's coalitions have perhaps converged toward men's: fluid membership, concerned with resource garnering (not simply allocation of resources to children). However, differences remain. The particular status or resource can wary widely, from money and status as a CEO, to violence to achieve rank in a girl's gang (e.g., Sikes, 1998). In most circumstances, women do not profit (either reproductively or in other currencies), in seeking status through coalitions, from the same sort of behaviors that work for men; they typically maintain more cross-generational ties than men do.

7.5. Women in politics

The inherent tradeoffs for women between resource garnering and reproduction have not only demographic but also political ramifications today. In most countries, women are rare in national and international politics (Norway, India, the United States, and England are recent exceptions), even when there are societal and legal mechanisms to mandate that women, as well as men, have access to power and control of resources. We think of men's and women's interests as converging most strongly in monogamous societies (even in socially monogamous societies as modern nation-states and “monogamous/mildly polygynous” traditional societies; Flinn & Low, 1986, Low, 1988a), but women are not more likely to hold political office in monogamous societies, societies in which men are absent, in those with biased sex ratios or in those with particular subsistence bases—women as leaders are too rare to see any pattern, except that of descent, discussed above.

Evolutionarily novel modern conditions mean that women have more direct control of physical resources and potentially more ability to form powerful coalitions than do women in traditional or historical societies. Yet, significant resources, status, and power, once proximate correlates of reproductive success (at least in men), do not appear in modern societies to have clear positive effects on women's lifetime reproductive success.