Paternal Provisioning versus Mate Seeking in Human Populations

Personality and Individual Differences, Vol. 17, August 1994, No. 2, 227-255.

Abstract

Paternal investment theory suggests that in cold climates males were selected for provisioning, rather than for mating success. In warm climates, where female gathering made male provisioning unessential, selection was for mating success. Male hunted meat was essential for female winter survival. Genes that encouraged mating success were selected for in warm (was cold) climates. Negroids (blacks) evolved in warm cold climates, while Caucasians (whites) and Mongoloids (Asians) evolved in colder climates. Mating is assisted by a strong sex drive, aggression, dominance, sociability, extraversion, impulsiveness, sensation seeking, and high testosterone. Provisioning is assisted by anxiety, altruism, empathy, behavioral restraint, gratification delay, and a long life span. Explanations are offered for racial differences in many personality characteristics, hormone levels, monamine oxidase levels, testosterone levels, lactase dehydrogenase metabolic paths, life spans, prostate cancer rates, hypertension, genital (penis and testes) size, vocal frequencies, liver size, muscle structure, mesomorphy, bone density, sports performance, crime rates, rape, child abuse, earnings, age at first sexual activity, AIDs, illegitimacy, divorce, marriage, and polygyny rates. Eye color correlations are discussed. Negro family structure in the Caribbean and the U.S. may reflect selection in Africa during hunter-gather times. Comparison is made with differential K theory and father absence theories.

Key words: race, climate, evolution, personality, polygyny, aggression, provisioning, mating, dominance seeking, paternal investment

TABLE OF CONTENTS

MALE COMPETITION FOR MATING OPPORTUNITIES 2
HUNTING AS A FUNCTION OF LATITUDE 4
THE COLD CLIMATE SITUATION 6
PREDICTIONS OF THE HYPOTHESIS 8
USING RACIAL DATA TO TEST CLINAL THEORIES 8
EXPLAINING RACIAL VARIATION BY MATING COMPETITION VERSUS PROVISIONING 10
Aggression 10
Dominance 11
Anxiety 11
Impulsivity 11
Delay of gratification 12
Sociability 13
Extraversion 13
Behavioral restraint 13
Criminal activity 13
Sexual restraint 14
Sexual behavior 15
Size of sex organs 15
Body build 16
Muscle characteristics 17
Life Span Variations 19
Hormonal Levels Versus Timing of Sexual Maturity 20
IMPLICATIONS FOR FEMALES 2
Why the Emphasis on Hunter-Gatherers? 23
THE DISTRIBUTION OF POLYGYNY 25
New World African Origin Populations 29
Father Absent Societies 31
TESTING THE THEORY WITH OTHER POPULATIONS 32
CONCLUSIONS 32
REFERENCES 33

Rushton (1985, 1987, 1988, 1989b, in press) and Ellis (1987) have drawn attention to the existence of many racial differences, including behavioral ones, and shown that they were frequently arranged in the order Mongoloid, Caucasoid, Negroid (or Negroid, Caucasoid, Mongoloid, depending on the trait). The differences Rushton drew attention to include twinning rates, (Negroids first, Caucasoids second, Mongoloids third), intelligence (Mongoloids first, Caucasoids second, Negroids third), various differences in aggression, dominance seeking, impulsivity, extraversion, sexual behavior, genital size (Negroids first, Caucasoids second, Mongoloids last), extent of altruism, and behavioral restraint (Mongoloids first, Caucasoids second, and Negroids last), and timing of birth, menarche, birth of first child, and death (Negroids earliest, Caucasoids second, Mongoloids last).

Rushton explains his observations by a version of r versus K selection theory. The terms r and K come from population biology, where r is a population's maximum growth rate, and K is the environment's carrying capacity. K selected species have been selected for success at competition with conspecifics. Species selected for fast reproduction with low ability to compete are called r selected. Rushton extends the idea to human populations. He argues that Negroids are the least K selected among human populations, Mongoloids the most K selected, and Caucasoids in between. This idea has produced considerable scientific (J. Anderson, 1991; Flynn, 1989; Leslie, 1990; Lynn, 1989; Roberts & Gabor, 1990; Silverman, 1990; Weizmann, Wiener, Wiesenthal, & Ziegler, 1990; Zuckerman, 1991; Miller, 1993) and popular controversy (Gross, 1990; Pearson, 1991, Chap. 5; Lerner, 1992, pp. 139-147), to which Rushton (1989a, 1990a, 1990b, Rushton & Ankney, 1993) has responded. Chisholm (1993) has also applied life cycle theory to humans, arguing that early experiences with the correlates of high death rates affects the allocation between mating and parenting.

This paper will propose an alternative explanation to differential K theory. Like differential K theory, it will be derived from a standard biological theory, parental investment theory. In some species, males devote more effort to seeking mating opportunities. In other species, they devote more effort to assisting their offspring. In each species, males evolve to use the strategy that most promotes their fitness. Which strategy most promotes their fitness depends on the effect of paternal investment on offspring survival and fitness, and on the opportunities for obtaining reproductively successful additional matings (Katz & Konner, 1981; Clutton-Brock, 1991). Likewise, within the same species different populations have been selected for different positions on the paternal investment versus mating effort continuum.

In humans, an especially important form of paternal investment is provisioning, bringing the mother and child food. Offspring's benefit from provisioning depends on climate. In warm climates, females typically can gather enough food for themselves and their children. In cold climates, hunting is required to survive winter, and females typically do not hunt (other than for easily captured small game). Hence, offspring survival requires male provisioning in cold climates. Thus, cold climate males were selected to devote more efforts to provisioning, and less to seeking matings. In warm climates, such male provisioning was not essential, even if desirable. Thus, warm climate males who devoted more efforts to seeking mating opportunities, and less to provisioning, left more offspring. This theory can explain many of the known racial differences.

Paternal investment theory's chief advantage is its specificity as to the traits populations should exhibit. It makes specific testable predictions (which are generally sustained) as to how cold-adapted populations and tropical populations should behave. In contrast, Rushton's and Ellis's differential K theories, after stating that Mongoloids are most K selected, and Negroids least, are very vague as to why this is. They fail to predict how other races (Malays, Australian aborigines, Polynesians, etc), or populations within the major races, should differ (J. Anderson, 1991; Miller, 1993).

To limit this paper's length, a detailed treatment of the evidence for the racial differences discussed will not be attempted. The reader can find relevant references in Ellis (1987, p. 159) and in Rushton's papers (especially 1987, p. 1019, 1989a, p. 9) and in Rushton's forthcoming book (in press). The author has checked most of these. Although many of the individual studies could be improved on, the racial differences do appear to be as Rushton and Ellis depict them. Strong evidence as to whether most racial differences in behavior are of environmental or genetic origin is lacking. While cultural explanations have been proposed for many of the behavioral differences, (but not for the morphological, or biochemical ones), there is not space to discuss them here. Occasionally, when new evidence has appeared since Rushton and Ellis wrote, it will be noted.

This paper combines several generally accepted ideas from different disciplines. It accepts the evidence from human behavior genetics that most personality traits have an inherited component (for instance Eaves, Eysenck & Martin, 1989; or Bouchard, Lykken, McGue, Segal, & Tellegen, 1990, or Rushton, in press, chap. 4). From biology and population genetics, it takes natural selection. From physical anthropology, it takes the theory that humans have been separated into races long enough to have evolved somewhat different appearances, many of which can be traced to climatic differences. From sociobiology, it takes the idea that males differ in the extent to which they devote their efforts to mating versus parenting, and that basic human behavior traits evolved during the hunter-gatherer 99% of human history (Barash, 1977; E. Wilson, 1975). It extends this by arguing that cold weather hunter-gatherers evolved into Mongoloids and Caucasoids, and tropical African hunter-gatherers into Negroids, and that differences in morphological and behavioral gene frequencies can be explained by the climatic differences during hunter-gatherer times.

In many species, much male behavior consists of competition for females (Barash, 1977; Wilson, 1975; Trivers, 1972). This seems to be true of humans. Standard sociobiology explains differences between human male and female behavior by contrasting the male's ability to father children by several females with the female's inability to have children by more than one male at a time (Symons, 1979; Hrdy, 1981). Thus, men evolved to exploit any impregnation opportunities that arise, while women direct their copulations to males who are willing and able to invest in them and their children.

Thus, in discussing male behavior across species, biologists argue that males evolved a trade-off between paternal investment and mating efforts that maximizes their inclusive fitness for that environment (Draper, 1989, pp. 149-150). The argument generalizes easily to explain differences in male mating and paternal investment within a single species, such as humans.

Under some conditions males leave more descendants by devoting more energy to seeking copulations (an endeavor that often includes prestige seeking), and relatively less to provisioning their offspring. In these conditions, selection will be for such characteristics as high sex drive, aggression, a mesomorphic body build, and large testes. In these conditions females can raise children with only limited male provisioning.

In other circumstances, males are selected for extensive provisioning of their children. This normally implies less energy being devoted to seeking matings for two reasons. First, energy and resources devoted to seeking additional matings would be diverted from provisioning their mates and children, thus reducing the number of surviving children. Secondly, added matings would frequently produce children able to survive only if resources were diverted from the father's other children. The net gain in surviving offspring would be small, or even negative. The first effect of devoting more energy to mating is to reduce total male investment in offspring, while the second spreads it among more offspring, reducing per capita investment.

I will argue that selection for male provisioning is especially common in climates with cold winters, where large game hunting is required for survival. Children of males who failed to provision would often not survive the winter.

Physical anthropologists can explain many racial variations as climatic adaptations (Baker, 1974; Coon, 1965, 1982; Krantz, 1980; Roberts, 1978). For instance, in northern latitudes, winter ultraviolet radiation intensity is low, and pale skin permits maximum vitamin D production. In low latitudes, there is a risk of excess vitamin D production and sunburn. Here dark skin is optimal (see Robins, 1991). Likewise, variations in adult ability to tolerate lactose has been interpreted as partially an adaption to low levels of ultraviolet radiation (Durham, 1991). Lynn (1991b) has contrasted the hunting required for survival in cold climates, in which Mongoloids and Caucasoids evolved, with the tropical gathering. He used this to explain the racial intelligence differences he had earlier documented (Lynn, 1991a). Here, this difference in reliance on hunting will be used to explain many other behavioral differences.

The reader will probably have little difficulty in accepting that Negroids evolved in the tropics, and Caucasoids and Mongoloids in colder climates. However, some may wonder why the Mongoloids are considered to have evolved in a colder climate than did the Caucasoids. One reason is that certain Mongoloid features (a stocky body build, and distinctive eye folds) appear to be adaptations to arctic conditions (Baker, 1974; Coon, 1965, 1982; Krantz, 1980; Roberts, 1978).

Admittedly, Europe is also cold. However, it is believed that farming started in the Middle East and then spread with the farmers' migration into Europe. This conclusion comes from archaeology and a northwest to southeast distribution of certain genes (Menozzi, Piazza, & Cavalli-Sforza, 1978; Piazza,1993; Sokal, Oden, & Wilson, 1991). The latter suggests a population movement, rather than merely technology diffusion. The result is that some European ancestors were Middle Eastern hunter-gatherers, rather than hunter-gatherers who had evolved where the populations now live.

Richard Lee (1968, pp. 42-43) in the widely cited Man the Hunter book emphasized that most calories eaten among typical (tropical) hunter-gatherers come from gathering, leaving the impression that gathering was the primary subsistence mode for the ancestors of most peoples. However, this really held only for the well studied inhabitants of warm and tropical areas, which are the majority of surviving hunter-gatherers. He reports that "In warm-temperature, sub-tropical, and tropical latitudes, zero to thirty-nine degrees from the equator, gathering is by far the dominant mode of subsistence, appearing as primary in 25 of the 28 cases." He found that 6 of the 8 societies whose latitude exceeded 60 degrees relied primarily on hunting. Eskimos (Inuits) are classic examples. In cool to cold temperature latitudes, from 40-59, degrees fishing was the most common subsistence mode. Temperature tells the same story, "Hunting is primary in three of the five societies in very cold climates (annual temperatures less than 32 F0.), fishing is primary in 10 of the 17 societies in cold climates (32-50 F0.): and gathering is primary in 27 of 36 societies in mild to hot climates (over 50 F0.)."

Fishing (usually male) appears to be a relatively recent development (Washburn & Lancaster, 1968, p. 294). In earlier times, before the most fertile mid-latitude lands were cleared for farming, there was probably more mid-latitude hunting. The hunting of large sea mammals from boats, so important to Eskimos, developed relatively recently.

Especially significant in the Lee tabulation (p. 43) is the absence of any predominantly gathering society above 500 latitude. Gathering is the primary way a single mother can feed her family.

The above tabulations suggest that ancestral Negroids relied on gathered vegetable material and ancestral Caucasoids and Mongoloids relied on animal matter from hunting and fishing. Supporting evidence is provided by blacks' greater current salt retention compared to whites (Luft et al., 1991). Presumably, the lower salt content of the prehistoric low meat African diet, combined with greater sweat loss, selected for salt retention.

Another piece of evidence consistent with the environment of evolutionary adaptation for Caucasoids involving more meat eating than Negroids' original environment is that the Negroids' livers are smaller (Lewis, 1942, p. 276). The liver' s function is to produce bile, which is used in fat digestion. A diet lower in fats (which are much more common in animal foods) would requires less bile for digestion.

The obvious problems of hunting while pregnant or with a small child assure that males do the hunting (for other than small and slow game) in virtually all societies (Friedl, 1975, p. 16-18; Watanabe, 1968).*(1) In climates typical of those now prevailing north of 500 a woman would have difficulty raising children without male supplied meat.

An example of women hunting small game is provided by the Twi of Melville Island, Australia (Goodale, 1971, pp. 151-169). They commonly hunt lizards, snakes, crabs, rats, and opposum and bandicoot. The last two are small animals found sleeping in logs or hollow trees, and typically are easily killed where found (i.e. without pursuit). In northern climates most such small animals would be hibernating during winter, or would be rare then.

The Hadza of Tanzania are typical tropical hunter-gatherers. Woodburn (1968, p. 52) describes the abundance of game and other food, stating, "For a Hadza to die of hunger, or even to fail to satisfy his hunger for more than a day or two, is almost inconceivable."

Barnard and Woodburn (1988) noted that, "In all known hunter-gatherer societies, with immediate return systems, and in many but not all, hunter-gatherer societies with delayed return systems, people are almost always able to meet their nutritional needs very adequately without working long hours." If gathering permits this, one would expect that women could adequately feed themselves and their children without male provisioning. Most tropical hunter-gatherer societies are immediate return ones.

Gardner (1972, p. 414), in describing the Paliyans, a foraging people of India, has pointed out that, "In normal times Paliyan men and women spend a bare three to four hours a day obtaining food and evidence no anxiety whatsoever about its supply." Single individuals are able to feed themselves easily, and married couples may not feed each other. Male provisioning is clearly not necessary. Not surprisingly, with the parties not needing each other economically, "the usual situation is one of fragile, often serial, unions, terminated quickly by the offended party when conflict arises" (p. 419).

A similar impression is left by descriptions of other tropical hunter-gatherer societies. Lee & DeVore's famous Man the Hunter is often summarized as showing that most calories come from gathering, not hunting, that most gathering is done by females, and that hunter-gatherers need spend only a relatively small part of their time in gathering. Taken together, these facts imply that a woman can feed her family with little male assistance. This suggests that males would leave more descendants by focusing their efforts on mating rather than on provisioning.*(2)

Now consider a cold climate, such as prevailed where the northeastern Mongoloids (Chinese, Japanese, Koreans) are believed to have evolved. Even now these areas have cold winters. During the last ice age they were much colder. Their people's physical features evidence numerous cold adaptations (Coon, 1965, 1982; Krantz, 1980; Roberts, 1978).

Although winter is only part of the year, it is the season animals have the greatest difficulty surviving. Most plant foods disappear (fruits, berries, edible leaves). Hunting becomes more difficult. Most animals time their reproductive cycles so there are no winter young or eggs. Eggs and young were the easiest animal protein for primitive men and women to obtain, since they were immobile or moved too slowly to escape. Many adult animals migrate (birds) or hibernate to escape the winter food shortage. Frozen ground prevents humans from digging for tubers and ice prevents or complicates fishing. Snow cover makes it hard to find fallen nuts or tuber locations, and makes walking much more difficult (Jochim, 1976, p. 138). Winter has severe weather episodes in which it is unsafe to leave shelter to hunt. Also, winter cold increases the body's food requirements (Kleiber, 1961, p. 164, p. 239). Finally, winter is a time of short days (Torrence, 1983, emphasizes this). Thus, the very period when food is needed most is also when it is scarcest, hardest to acquire, and the time for gathering it is least.

Admittedly, there could be geographical circumstances where winter is easier. Many ungulates (such as elk) in mountainous areas winter in the lowlands. Possibly this concentration, aided by ease of tracking in snow, could make winter an easier time for survival. However, males would still be expected to be the sex that took advantage of this situation.

Jochim (1976, pp. 141-143) has estimated food consumption for German mesolithic foragers by seasons. Plants (female gatherable) are estimated to be 30% for spring and summer, 23% for fall, but zero in winter.

In northern climates a female cannot be self supporting. A female would avoid marriage to a hunter already supporting another's family. Even if married to an excellent hunter, a second wife (receiving half of his support) would probably be poorly provisioned. Cold climates lead to environmental monogamy (Alexander, Hoogland, Howard, Noonan, and Sherman, 1979). Males evolve drives that encourage family provisioning, and discourage competing for mates.

Large game hunting often requires cooperation by groups of males. A male who doesn't have the cooperation of others is likely to bring home less meat, and to leave fewer descendants. Variability in hunting success makes it desirable to hunt in groups that share their kills. Withdrawal of cooperation by other males is a likely penalty for trying to impregnate another's wife. In such circumstances, the drives that lead to seeking multiple mates are selected against.

Of course, in a warm climate hunting groups are likely to exist and a philandering male may be penalized by exclusion from the group, or less cooperation. However, with opportunities for gathering and hunting small game abundant at all times, the penalty of loss of participation in the large hunts is less severe, and the selection against mating drives weaker.

Why presume that primitive hunters could not kill enough food to carry multiple females through the winter? After all, there is a lot of meat on even a single ungulate carcass. If game were abundant enough for a typical hunter to support more than one wife, population would grow. It would grow until the pressure on the game resource had reduced the yield from hunting to where one hunter could typically support only a single wife and offspring. The argument is the standard Malthusian one.

Although Man The Hunter is usually cited as showing that hunter-gatherers can feed themselves with a short work day, this appears true only of the tropical peoples discussed. That book also contains Balikci's (1968, p. 82) discussion of the Netsilik Eskimos, who had a 10% loss of life from starvation in two years. The inland Eskimos appear to be able to support only one family per male (Alexander et al., 1979). Other northern hunter-gatherers such as the Ainu appear to have monogamy as the typical pattern, even if a few males have more than one wife.

Other descriptions of northern hunting peoples emphasize the difficulty of the life and the risks of starvation. For instance, Rogers (1972, p. 104) in his discussion of the Mistassini Cree, American sub-Arctic natives, states that, "Securing sufficient food is a constant problem and a never ending concern. Times of starvation are vividly remembered." He also states that the gathering of vegetable foods is minimal. In such an environment a typical male could not support more than a single female. Any inadequate provisioning of her offspring could reduce his reproductive success.

Likewise, Nelson (1973) discusses how those Kutchin (north Alaska) who remembered the old ways emphasized hardships and lack of food. He reports that similar attitudes to the past are found among other Athapaskan people. Emphasis is placed on far north people because Ice Age Europe and Asia had climates similar to these peoples' current homes (Soffer & Gamble, 1989; Scott, 1984).

Admittedly, a female without a "husband" would probably receive some meat from brothers, other family members, and other band members. In contemporary hunting bands game is usually shared with other band members (Hawkes, 1993). Even unmated females get some. Such a meat sharing system is very useful in spreading the risks of the hunt among families (Hames, 1990). In the short run, this clearly supplies some meat to women without husbands. However, it is likely that in the long run a female unattached to a male hunter would suffer. Adverse effects are most likely during occasional famines, when social traditions of sharing are likely to break down. Then band members are likely to give priority to their own offspring. One possible mechanism is through the more successful hunters joining bands with fewer non-related dependents, where their own families would be better provisioned. Thus, in time of famine, poor provisioning by a father would affect his children's survival. The sharing systems observed among current hunters probably evolved relatively late, after an earlier system in which males gave priority to themselves and their families. While sharing may moderate regional differences in the consequences for offspring survival of male provisioning, they are unlikely to eliminate them.

There appear to be several traits that contribute to success in mating competition. A strong sex drive would lead to more matings. A strong Coolidge effect (in which women other than regular sex partners were considered unusually attractive, see Symons, 1979, p. 209) would encourage taking additional wives, matings with other men's wives, and rape. Efforts to mate with other men's wives involves risks of retaliation. Thus, aggressiveness, impulsiveness, and lack of fear would contribute to leaving more descendants. If one were to have only an occasional mating with certain women, greater sperm production would lead to leaving more children. If males frequently succeeded in mating another's wife, a high sperm production and strong sex drive would lead to leaving more offspring.

Empathy with others is not conducive to extra-marital relationships or even to taking additional wives. For instance, strong sympathy for one's children is likely to lead to devoting spare resources to them, rather than using the resources to purchase sexual access through prostitutes, concubines, or extra wives. Likewise, if extramarital intercourse violates social mores, a strong tendency to follow social mores is not conducive to extra-marital access.

The above paternal investment theory makes a number of interesting predictions about the mating patterns of hunter-gatherer populations in warm versus cold climates.

USING RACIAL DATA TO TEST CLINAL THEORIES

Because agriculture was adopted relatively recently, differences that emerged during the hunter-gatherer stage should have survived into the ethnographic present. The above suggests that personality and behavior differences among modern populations should be correlated with the winter temperatures where they evolved. Tropical populations would be selected for greater mating efforts and lower paternal investment. In cold climates, the opposite would be true.

Ideally, analysis would use data expressed as behavioral clines drawn from data on many different populations. (A cline is a line connecting points of equal values for observations, such as lines on a weather map.) Unfortunately, such data is rare. Admittedly, some data is available from physical anthropologists' studies of specific populations, and from the ethnographic record, coded in the Ethnographic Atlas (Murdock, 1967).

However, some population level studies do exist using such data. Wolfe and Gray (1982) found a correlation between the extent of polygyny and the height of males and females. This is easily explained by the taller males having an advantage in acquiring mates, which leads to greater selection for height in polygynous societies. Wolfe and Gray were surprized not to find clear evidence of greater sexual dimorphism is such societies, which they expected from the animal evidence. However, in humans it is known that most genes that affect height appear to affect both males and females equally (excluding the obvious exception of those carried on the X or Y chromosomes). What they regarded as a puzzling correlation between polygyny and female height is easily explained. If taller males leave more offspring, the mean height of both males and females will be raised, leaving sexual dimorphism little changed.

Wolfe and Gray used a code for the extent of father-infant proximity in the first year of life as a measure of male parental investment. They found that this correlated to a statistically significant degree with measures of polygyny (Table 8.1), and with male and female heights. Since it is unlikely that height causes differences in marriage patterns, it is more plausible that sexual selection has affected the frequencies of genes for height, and possibly for a measure of paternal investment (if that is subject to genetic variability). For these effects to have appeared, the differences in marriage patterns must have persisted long enough for natural selection to have acted. Although Wolfe and Gray did not notice this, this is the first clear evidence that sexual selection has played a role in the evolution of differences in gene frequencies among human populations.

Studies of the above type can be done for only a few variables. As Wolfe and Gray (1982, pp. 206-207) report, "Our search of the literature of cross-cultural codes revealed no codes that directly rate societies on the degree of male parental investment. This lack of codes is partly due to the fact that ethnographers rarely had a theoretical orientation in which the problem of male parental investment was of central concern and therefore rarely collected data on this problem."

However, various physical variables do correlate with climate, among which are skin color, shape of nose, body mass and shape, etc. These highly visible surface features include the variables usually used for racial classification. It appears that Negroids evolved in the warmest climate (tropical Africa), Mongoloids in the coldest (the North China-Siberia area), and Caucasoids in intermediate climates (Europe and the Middle East). These areas are separated from each other by barriers to gene flow. The Sahara partially isolates tropical Africa. During the ice ages, the Himalayan ice sheet separated the Caucasoids from the Mongoloids.

Limits on gene flow between different areas (although not complete) permitted populations in each region to develop somewhat different morphology and behaviors. Each evolved in their respective environments so as to produce the largest numbers of descendants. Each of the major races of mankind, Negroid, Caucasoid, and Mongoloid, is itself composed of numerous separate populations. In the absence of detailed information on personality in a large number of localities around the world, the best way to look for evidence of personality varying with climate is through examining racial variability.

Rushton and Ellis have assembled considerable racial information. (These differences of course are related only to central tendencies, and any one individual need not resemble his race with regard to any single trait.) Theirs are the best summaries of non-morphological racial differences. They deserve considerable credit for assembling this data. While they interpreted their findings in differential K selection terms, their data can also be used to test the male mating effort versus paternal investment hypothesis.

Let us start by taking the list of personality characteristics Rushton assembled (1987, p. 1019; 1988, p. 1010; 1989a, p. 9; Rushton & Bogaert, 1989) and see whether a mating effort versus parental investment framework can explain them. This can be done for most items.

Aggression

Negroids are reported (by Rushton) to be the most aggressive and Mongoloids least aggressive (Caucasoids intermediate). From a reproductive viewpoint, aggression's benefits include leadership positions. These assist in obtaining multiple wives, and frequently provide opportunities for extramarital relationships. Aggression also produces children through rape. In warm climates, where extra wives can be self-supporting, there are clear reproductive benefits to additional wives. In cold climates, lower survival of children by the first wife will provide a partial or even complete offset.

The disadvantage to high levels of aggression is that it leads to injury or even death, either in the course of the conflict caused by the aggression, or through retaliation by victims or society. In all climates death eliminates, and disability reduces, the chances of fathering additional children. However, in cold climates, death and disability are more likely to lead to the death of existing children, while in warm climates the mother is more likely to be able to rear her children alone. This effect alone would make the optimal position on an aggression-fearfulness continuum climate dependent.

Aggression also leads to interband raiding and warfare. These increase sexual access to other bands' females. Direct access may be through rape or wife capture. Indirectly, killing or defeating competing males reduces mating competition. Wives are later obtained by courtship or exchange.

However, additional wives only contribute to reproductive success if there is enough food to rear the resulting offspring. Where women supply their own subsistence, warfare and wife capture can produce reproductive gains. This is likely to be true in tropical areas. In cold areas, where wives must be provisioned by hunting, additional children from captured wives would divert scarce resources from other children, limiting the net gain.

A defeat in interband conflict leads to deaths and injuries. In northern climates, where the gains from success are small, and the losses large, the relatively non-aggressive will leave more descendants. In tropical climates, where the benefits are larger, selection will be for higher aggression.

While Rushton interprets aggression as a r characteristic, this is implausible. Gould (1982, p. 367) argues, "Since member of K-selected species inhabit the same niche and compete for population-limiting resources, it should not be surprising that these animals regularly fight with each other for control of those resources. Among r-selected species, on the other hand, fighting would be a waste of their most precious commodity, time." In essence, if resources are abundant, organisms will not benefit from fighting. Destroying other individuals is only beneficial when it eliminates competitors, which is to say in K conditions. In contrast, aggression appears unambiguously useful for obtaining numerous matings, even if not for provisioning.

Dominance

Very similar comments can be made regarding dominance, where Negroids are reported to be the strongest seekers of dominance and Mongoloids least (Caucasoids intermediate). Dominance seeking leads to more mating opportunities, but also to death and injury, which reduces the survival of already born offspring, especially where male provisioning is essential. If the extra wives that dominance and prestige provide cannot be supported, the ability to attract them gives little reproductive benefit.

Anxiety

Mongoloids are reported to be the most anxious, and Negroids least (with Caucasoids in between). This is closely related to dominance seeking and aggression, in that high anxiety deters dominance seeking and aggression. The more prone an individual is to anxiety, the less likely he is to seek additional matings beyond his first wife. It is usually not in the wife's reproductive interest for him to either take additional wives, or to seek extramarital relationships. She can be expected to have evolved to threaten retaliation, and occasionally retaliate by leaving, attacking the offending male, or enlisting the aid of her relatives or society against him. Conducting an affair with another man's wife, or an unmarried chaste female, or rape, all involve risk taking. Thus, where the optimal male strategy is to devote less efforts to mating than to provisioning existing children, high anxiety is selected for .

There are additional reasons for selection for high anxiety in cold climates. One strategy for surviving the winter is food storage (see Miller, 1991). Food storage is practiced only (with exceptions) in societies whose growing seasons are less than about 200 days (Binford, 1980). Anxiety about food supply encourages storage, and discourages their too rapid consumption. Where storage makes a difference, high anxiety is selected for.

Coon's descriptions (1971, p. 26-39) of hunter-gatherer housing shows that simple domed houses and lean-tos are used by southern people, while igloos, plank houses, and pit houses are used further north. Pit houses, roofed holes, are common among northern hunters because they protect well against intense cold. Houses in snowy areas can collapse with a heavy snowfall, and cause loss of life, as well as leaving the inhabitants exposed to the cold. Collapse of a tropical conical hut, or lean to, is only an inconvenience. Anxiety would appear to encourage the construction of houses with adequate safety margins, and possibly an early start to such construction. Thus, anxiety would appear to promote reproductive success more in areas with snow than in tropical areas.

Anxiety about being caught in a freezing storm, or away from a warm campfire overnight is likely to promote survival in cold climates. Thus, stronger cold climates selection for anxiety is predicted. (Nelson [1969] mentions the caution and absence of thrill seeking in Eskimos).

Impulsivity

Frequently, short term pleasures can be obtained by acting, but acting requires risking adverse consequences. One who frequently takes advantage of short term opportunities displays impulsivity. Those with high anxiety levels are less likely to seek immediate pleasures. Thus, it is not surprising that the ordering on impulsivity, Negroids highest, Mongoloids lowest (Caucasoids in between), is opposite to the ordering on anxiety.

In particular, males frequently have the opportunity to father children through extramarital relationships or rape. High impulsivity individuals would be expected to more often act on these opportunities than would low impulsivity individuals. Thus, impulsivity would be selected for where a high mating effort contributed to fitness.

Delay of gratification

Closely related to impulsivity is the ability to delay gratification. Mischel (1958, 1961c, 1971, p. 127) found a racial difference in preference for delayed gratification. Trinidad Indians (i.e. India origin) children would wait longer for a reward than Negro children, although he interpreted this as reflecting the greater absence of fathers among the Negro children. As is common in Negroid populations (see below), many of the Negroes lacked a father in the home, while few Indians lacked a father in the home. Cultural factors probably also play a role, since Grenada Negroes were similar to Trinidad Indians (Mischel, 1961c). Delay of gratification was less in a Trinidad industrial school for juvenile deliquents than in an ordinary Trinidad school, supporting the theory that difficulty in deferring gratification (such as choosing the immediate gratification obtainable from stealing, risking a future punishment) contributes to criminal activity (Mischel, 1961a). Gratification delay in Trinidad Negroes was found to be positively related to Harris's Social Responsibility Scale, which conceptualized responsibility "as a composite of attitude elements reflecting behavior classifiable as reliable, accountable, loyal, or doing an effective job" (Mischel, 1961a, b). Interestingly, the Trinidad Negroes scored much lower than a similar aged (presumably predominantly white) US group, which Mischel (1961a, p. 6) notes is "fully consistent with anthropological observations."

It is not known how heritable the ability to defer gratification is, although most personality variables appear to have a significant degree of heritability. However, in one experiment, using Barbadian Negroes in the Cambridge area, the mother's delay of gratification (choice of large bottle of instant coffee in a week or a small bottle now) was found to be correlated with the child's violating or not violating a prohibition in a temptation situation (Mischel & Gilligan, 1964, p. 412). This suggests both behaviors are reflecting a heritable trait.

Difficulty in delaying gratification and impulsivity makes provisioning more difficult. Provisioning requires suppressing the desire to eat in order to bring food back to the family. In warm climates, not eating food when available would merely deny the male forager the nutrition required to compete with other males, since his children will normally be fed in any case.

Also, a food storage strategy for surviving the winter requires deferring gratification. The impulse to immediately eat available food must be resisted to store it, and later the impulse to eat the stores must be resisted in order to retain them for later use. Survival through the winter may require not only storing food, but also storing fuel, making clothing, and building shelters. These activities require resisting impulses to divert energy to activities with shorter term returns (gathering non-storable food, drinking, or even flirting). Thus, there are other reasons why seeking immediate gratification and impulsivity would be selected against in cold climates.

Banfield (1974) has proposed that seeking immediate gratification among the U. S. inner city poor (who tend to be Negroids) explains much of their poverty. While he carefully denies believing in genetic differences (like other writers of the time), it is plausible that this trait, like most personality traits, has a genetic component. Furthermore, tropical environments, such as that of Africa, are ones where hunter-gatherer populations are described by Woodburn (1980) as "immediate gratification" ones. He has described how for the Hadza of Tanzania, food is available in the bush at any time, and that as a result there is little need to plan ahead or to defer gratification. Thus, it is plausible that the immediate gratification behaviors that Banfield blames for so many inner city problems may be a continuation of tropical hunter-gatherer behavior.

Sociability

Sociability assists in learning of, and exploiting, mating opportunities. However, time spent socializing reduces the time available for gathering food and for other parental investments. Sociability often involves remaining near the camp where others are located, while provisioning requires leaving to forage. Thus, selection for provisioning will reduce sociability.

It should be noted that if sociability leads to talking it would be selected against in northern climates, where quiet is required for hunting.

Extraversion

Extraversion represents a combination of impulsivity and sociability (Eysenck & Eysenck, 1985). Thus, extraversion will be selected for where selection for seeking copulations occurs, and be selected against in other areas. Thus, it is not surprising that Negroids rank highest in extraversion, and Mongoloids rank lowest, with Caucasoids in between.

Behavioral restraint

Rushton (1988, p. 1013) combined many of these characteristics and described them in terms of behavioral restraint, with Mongoloids being highest in behavior restraint, and Negroids being lowest (Caucasoids intermediate). Behavioral restraint is not conducive to males seeking mating opportunities, but is conducive to making high paternal investment, which frequently requires resisting immediate gratification opportunities.

Criminal activity

Criminal activity is closely related to behavior restraint, for which the evidence is that Negroids are highest, Mongoloids lowest, and Caucasoids in between (for documentation see Wilson & Herrnstein, 1985; Ellis, 1988, p. 532; Jaynes & Williams, 1989, chap. 9; Rushton, 1990a). Paternal investment theory would explain high crime rates as resulting from high aggressivity and low empathy, altruism, and rule following behavior, traits that contributed to tropical reproductive success. A contributing factor is that the racial ordering for intelligence (Mongoloids first, Caucasoids next, and Negroids last [Jensen, 1987 on Negroids; see Rushton, in press, for other references]) is opposite to those for crime, and criminal behavior is known to be more common among the less intelligent. Intelligence differences have been offered as explaining the racial differences in crime (Gordon, 1987).

If those selected for mating effort have higher levels of aggression, lower behavioral restraint, and higher sex drives, it would be predicted that rape rates would be higher. They are known to be higher in Negroids than in Caucasoids (Brownmiller, 1975, pp. 213-216; Ellis, 1989, p. 94).

Child abuse is another form of crime. Abusing a child is the opposite of investing in it. If cold climate fathers were selected for paternal investment, their descendants should commit less child abuse. Caucasoids do have lower child abuse rates than Negroids (Ellis, 1987, p. 159), and Mongoloids the lowest (Ellis 1993, p. 171) .

Sexual restraint

The form of behavioral restraint most sensitive to natural selection is sexual restraint. With regard to a wide range of sex related variables, including marital instability, Rushton (1988) and Rushton & Bogaert (1987) show that Mongoloids are the most sexually restrained, and Negroids least, with Caucasoids intermediate. Sexual restraint is the ability to resist opportunities for copulation. Males that devote their energies to paternal investment have less energy left for seeking additional matings.

Two mechanisms could produce greater sexual restraint. The sex drives (including those for seeking variety in partners) could be weaker, or the inhibitions to sexual activity could be greater. That populations exhibiting greater sexual restraint are also more restrained in other ways suggests that part of the explanation is the greater inhibition (discussed above).

However, populations may also differ in the strength of sex drives. Selection for a stronger sex drive (and for a stronger desire for variety in partners) appears a more efficient mechanism for increasing mating effort than merely selection for less restraint. Generalized mechanisms, such as changing anxiety levels, might prove counterproductive in non-sexual spheres. For instance, less anxiety might encourage taking excessive hunting risks.

Simpson and Gangestad (1991) show that the strength of the desire for numerous sexual partners (what they call sociosexuality) varies independently of the strength of the desire for frequent sex. It is very plausible that both are genetically influenced. They (Gangestad & Simpson, 1990; Simpson & Gangestad, 1991) present evidence that sociosexuality varies with personality dimensions that have been shown to possess heritable components. Gangestad and Simpson (1990) argue that seeking separate partners, versus making a commitment to one partner represent different reproductive strategies, but fail to consider the possibility that the equilibrium percentage of individuals following the different strategies may vary with the environment (and hence with race).

A genetic influence on the drive for sexual variety is also suggested by the greater male desire for variety. This is probably caused by the effects of testosterone (or another sex related hormone) on some part of the brain. Genetically controlled variability in the number of receptors or the level of hormones could produce variability in the strength of the desire for sexual variety across populations.

Sexual behavior

Rushton and Bogaert (1987) document differences in sexual behavior. Besides a literature review, they reanalyzed the Kinsey data on sexual behavior in American whites and blacks. This showed greater sexual activity in blacks than in whites. For instance, the black frequency for coitus in their first marriage was 3.83 times per week for those aged 21-25 versus 3.11 for similar whites. The more frequent intercourse within marriage suggests differences in sex drive, rather than in a generalized restraint (which should not affect the frequency of marital relations). Measures of the extent and frequency of extra-marital sexual activity (p. 542) showed more activity outside of marriage among blacks than among whites, with all reported measures being statistically significant at the .001 level. Most black working class females believe that a wife should expect running around (Staples & Johnson, 1993, pp. 156-157). ". . .Black females have their first full sexual intercourse some years earlier than the typical White female." (Staples & Johnson, 1993, p. 77). Differences in either sex drive or in social restraint could explain these differences.

Since age at first intercourse (Martin, Eaves, & Eysenck, 1977), and age at first date, marriage, and first and second child appears to be genetically influenced (Mealey & Segal, 1993), it appears appropriate to consider hypotheses that population differences in sexual behavior are also genetically influenced.

Differences in sexual restraint and in the number of sexual partners predict racial differences in sexually transmitted diseases. Such differences exist in the distribution of AIDS (Rushton and Bogaert, 1989). They had earlier been reported for syphilis, where the negro rate (often approximated by the rate for non-whites) is a multiple of the white rate, both in civilians and in the military (Aral & Holmes, 1984, p. 130; Berg, 1984, p. 93; Lewis, 1942, p. 158). The reported gonorrhea rate is 21 times as high in blacks as in whites, although part of this difference may reflect better reporting from the public clinics frequented by blacks (Aral & Holmes, 1990, p. 22). For the common herpes simplex virus -2, the antibodies at ages 60-74 are found in over 80% of black females versus slightly over 20% of the white females (Aral & Holmes, 1990, p. 27). Pelvic inflammatory disease (caused by the spread of gonorrhea and/or chlamydia to the upper reproductive structures of women) is currently a major cause of female infertility in parts of Africa.

The most plausible proximate explanation for racial differences in sex drives is the possibility discussed below of differences in testosterone levels at the relevant ages. Testosterone promotes male sexual activity (Kemper, 1990, pp. 38-46).

Size of sex organs

One consequence of higher levels of puberty causing hormones could be greater development of the sex organs. Rushton and Bogaert (1987) use the Kinsey data to document longer penises and greater circumference of penises in blacks than in whites. From other sources they find Mongoloids to have shorter penises than Caucasoids. One condom manufacturer provides for larger size condoms for Africa than for other areas, and the smallest size for Asia (Wong, 1991). Mongoloids have smaller testes than Caucasoids (Short, 1984; Diamond, 1986).

It would be desirable to have good quality measurements of Negroid testes size, because the theory that they have been selected for high mating effort would predict larger testes in order to win at sperm competition. High levels of polygyny, and accompanying sperm competition, would select for large testes and high sperm production, especially allowing for the tendency for the largest number of wives to occur late in life. Among the large apes, the species that have multimale mating systems (notably the chimpanzee) have larger testes (Short, 1981; Smith, 1984).

The available evidence, while not of the highest quality, does not confirm this prediction. Ajmani, Jain, & Saxena (1985) found the scrotum diameters in 320 Nigerian males to be greater than had been reported for Europeans, as predicted. An American study of adolescents (Daniel, Feinstein, Howard-Peebles & Baxley, 1982) reported "There was no significant differences in testicular volumes between black and white adolescent boys. Any possible simple correlation with race was not significant against the general variability of testicular volume." No actual report is provided of the racial averages, leaving the possibility that some difference was found. In any case, a difference in adolescents might reflect only an earlier start to maturation among the blacks. In addition, one early autopsy study actually found lower testes weights in Negroids than in Caucasoids (Freeman, 1934).

Rushton and his colleagues explain these sex organ differences with his differential K selection hypothesis. Yet it is not obvious that larger penises (or clitorises or vaginas) lead to more offspring. Thus, they should not be interpreted as r characteristics. More plausibly, they are a mere by product of selection for high levels of sex hormones. The testosterone surge at puberty enlarges the penis, and it is plausible that higher hormonal levels would produce a larger organ. Testosterone increases the penis size of castrated rats whether applied externally or injected (Wigodsky & Greene, 1940). This makes it more plausible that racial differences in penis size reflect hormonal differences.

Body build

There are racial differences in body build. Negroes have a more masculine body build than Caucasians (Laska-Mierzejewska, 1982). The masculine body build implies strong accentuation of such masculine characteristics as a large chest, and muscular body. Negro soldiers (males) have been found in two studies to be more mesomorphic (and less endomorphic) than white soldiers, with the difference being more than one standard deviation (Damon, Bleibtreu, Elliot, & Giles, 1962, Table 2).*(3). Simply put, mesomorphy is the amount of visible muscularity. Such a body appears to have been selected for conflict with other males. Notice, this observation is evidence for paternal investment theory, since other theories do not predict that the Negroid males will be more masculine.

However, the Japanese are relatively mesomorphic, both in Hawaii (Heath, Hopkins, & Miller, 1961), and in Japan (Kraus, 1951). Thus, the extent of mesomorphy appears to be one case where Ruston's generalization that the Caucasoids fall between the Negroids and Mongoloids breaks down.

Hudson & Holbrook (1982) found lower mean fundamental vocal frequencies in Negro males and females than others had found for whites. As is well known (and found by them), males display a lower frequency (deeper voice) than do females, and puberty deepens the male voice. This deepening is generally attributed to testosterone. The deeper Negro voice may reflect the influence of higher testosterone levels at puberty or prenatally.

Muscle characteristics

An interesting set of statistically significant differences in muscle characteristics has been found between black and white sedentary males (Ama, Simonau, Boulay, Serresse, Theriault, and Bouchard, 1986). African blacks were found to have less type I muscle fibers, more type IIa and lower activities in enzymes catalyzing reactions in phosphagenic and lactase dehydrogenase metabolic pathways. These were interpreted as likely to be inherited, and suggesting that blacks would exhibit better performance in sports requiring a short duration of exertion. Malina (1988) reviewed the literature on childhood performance, and found that blacks excelled in the dash, the standing long jump, the vertical jump, and the ball throw for distance. All of these involve a short burst of exertion. Tests with a bicycle ergocycle have shown Caucasians to have higher maximal O2 uptake, a trait adapted for endurance activities. Also, Ama, Lagasse, Bouchard, and Simonau (1990) reported better white performance (compared to black West Africans) in the last 30 seconds of a 90 second knee extension exercise, a result consistent with blacks making greater use of anaerobic energy metabolism than whites. What would have selected for racial differences in such traits?

Hunting is implausible both because Caucasoids are likely to have hunted more than Negroids, and because hunting often requires prolonged exertion to follow large animals. A likely explanation is male fighting, since fights often involve short periods of vigorous exertion (after which the opponent is hopefully defeated). Thus, Negroids appear to be selected more for fighting. This would be consistent with their more muscular body build and higher aggression. It is what would be expected if Negroids had been selected to win mating competitions.

Blacks have denser and stronger bones than whites (Himes, 1988; Pollitzer and Anderson, 1989). The disadvantage to higher density bones is higher weight (more energy required for movement) and greater need for calcium. The advantage is fewer fractures, and thus, lower mortality. The bone differences can be explained if black males engaged in more intermale conflicts, and those with stronger bones were less often injured. No other hypothesis comes immediately to mind that can explain these density differences.

Worthy (1977, chapter 2; Worthy & Markle, 1970; see also Jones & Hochner, 1973) has shown systematic differences in the sport positions blacks and whites play. He argues that where the positions require reacting properly to the opponents' actions, blacks are more successful, while whites do better where the player initiates his own motions, as in baseball pitching, marksmanship, or in shooting a basketball free goal. He reports a Negroid relative advantage in the defensive position of an experimental game where they had to react to their opponents' initiatives.

Worthy also correlated eye color with positions played. Dark eyed whites are overrepresented in the same positions in which blacks are overrepresented. In Worthy's theory, eye color plays a direct role. However, eye color also varies with ethnic origin, with north Europeans having more blue eyes. This suggests an Old World cline such that ability to react to opponents' actions increases to the south.

What circumstances would have selected for these different abilities? Survival in fights with other males would appear to depend on quick reactions to opponents' moves. In contrast, a hunter usually stalks his prey and then chooses the time to attack. Worthy's observations could be explained if male reproductive success in colder regions varied with hunting ability, while in the tropics it varied more with fighting ability. The eye color differences could be explained if hunting was even more important in northern Europe than in southern Europe, or if southern Europeans had interbred more with farmers of Middle Eastern origin.

Possibly relevant evidence is provided by Coleman (1980). Successful prone rifle shooters (who choose the moment of shooting) are the most introverted, while successful rapid fire pistol shooters (who have very little time to fire five shots, and have to move the pistol from target to target), are very extroverted. Apparently personality correlates with what looks like Worthy's reactive versus non-reactive distinction. Thus, an alternative explanation for these racial differences would rely on selection for different personality traits. Since tropical climates seem to select for both quick reactions (as in fighting) and for extraversion, and cold climates for the opposite, both theories predict a similar north-south behavioral gradient.

There are other intriguing reports of correlations of eye color with behavior. Rosenberg & Kagan, (1987, 1988) and Rubin & Both (1989) report that Caucasian children that are behaviorally inhibited ( a concept related to introversion) are disproportionately blue eyed. While Rosenberg & Kagan suggest several possible biological mechanisms for this effect, a very plausible explanation is that north European children (more likely to be blue eyed) are more behaviorally inhibited than South European children (who are more likely to have brown eyes). Both eye color and behavioral inhibition are believed to be genetically influenced. If the effect is really biological with both eye color and behavior having a common cause (a pleiotropic gene effect), it would be predicted that where one sibling was blue eyed and one brown eyed, that the blue eyed one was more likely to be behaviorally inhibited. If the genes for eye color and behavior were associated merely because the ancestors of different children came from different regions, there would be no sibling correlation. Unfortunately, such a study has not yet been done.

An argument against Negroids having evolved for fighting is that they show lower pain tolerance than other races (Worthy, 1977, pp. 123-124)

Life Span Variations

Negroids have shorter lives than Caucasoids, who have shorter lives than Mongoloids. For instance, U. S. whites have a life span estimated at 76.1 years versus 69.1 years for U. S. blacks (U. S. Department of Health and Human Services, 1993). If testosterone shortens life (Hamilton, 1948), as it appears to do (shown by the shorter life span of males than females, and of normal males compared to castrated males), differential testosterone levels could explain the life span ranking.

Part of the shorter Negroid life span reflects more violent and accidental deaths, which could result directly from higher aggression. However, disease causes most of the excess deaths. As a general rule, more polygamous species have higher male death rates (relative to female rates), with much of the difference being due to degenerative diseases (Daly and Wilson, 1988, p. 142). As discussed below, Negroids appear to be more polygamous than other races.

An evolutionary biology theory of aging (Rose, 1991) holds that many genes are pleiotropic (i.e. have more than one effect). The same genes that contribute to longer life often impose disadvantages earlier in life. In the simplest case, genes which delay degeneration (perhaps through directing more energy to cell repair) also imply early life disadvantages (perhaps through leaving less energy available for mating or for lactation). A longer life can be purchased only at the expense of an earlier reproductive disadvantage. Which genes are selected for depends on whether reproductive success is facilitated more by a long life, or by early life success. This may depend on whether selection is for high paternal investment or high mating effort.

Copulation precedes childbirth, while paternal investment follows childbirth. Since death destroys a man's ability to help his children, longevity facilitates paternal investing. Thus, if a father's death hurts his child's survival chances, selection for a long life will be stronger than if children can be raised without paternal assistance. Conversely, if an early death from mating competition is going to eliminate any reproductive benefits from slower degeneration later, the alleles that protect against degeneration in old age will be less beneficial (Diamond, 1992, Chapter 7, especially p. 132).

If a male does not obtain a mate when young, living longer will not add to his descendants. Suppose the number of matings determines how many descendants a man leaves. Then men are more likely to be selected for early vigor and competitive ability, even at the expense of an earlier death. Where males have been selected for mating competition, polygyny often emerges (see below). For many men to have multiple wives, others must have no wives. Thus, to perpetuate his genes in a polygynous society, a man must be ranked relatively highly by those who control sexual access. Dominance and prestige seeking would be selected for.

In a monogamous or nearly monogamous society, even undesirable men marry. Suppose genes that handicap men in mating also contribute to a longer life, and hence to greater offspring survival. Then carriers of these may leave more descendants than worse providers who are better at mating, but who still get only one wife. Genes for high testosterone probably shorten life but may contribute to mating success through stronger muscles, higher sex drive, and aggression. Many athletes shorten their lives by taking steroids (essentially synthetic testosterones) to promote competitive success.

Thus, strong male sexual competition leads to a shorter life span. This can be explained by paternal investment theory. In differential K theory, earlier deaths (in advanced countries, Negroids die earlier than Caucasoids, and Caucasoids earlier than Mongoloids) are interpreted as a result of going through the life cycle quicker. A long life span is thus a K characteristic.

Hormonal Levels Versus Timing of Sexual Maturity

Races differ in average age at sexual maturity (Rushton and Bogaert, 1987, p. 537). Negroids mature earlier than Caucasoids, and Caucasoids earlier than Mongoloids. Rushton interprets this as showing less K selection in Negroids.

Insert Figure 1 About Here

An alternative interpretation is that higher adult levels of sexual hormones contribute to mating success. To have high adult hormone levels, the levels must either start to rise earlier or the hormone levels must rise more rapidly (or the rise must end later). Thus, it is likely that in populations selected for high mating efforts, the young males at any given ages will have more sex hormones. The process is illustrated in Figure 1.

If various behavioral and physical characteristics appear only when the hormone levels reach a certain value, the stages will occur earlier in populations selected for intermale competition. For instance, this could explain the report (Westney, Jenkins, Butts, and Williams, 1984) that at 11 years, 60% of black males had reached the stage of accelerated penis growth, while in a white sample this stage was reached at an average of 12 1/2 years. This genital stage significantly predicted onset of sexual interest. The correlation between physical development and behavior is most plausibly interpreted as being due to there being a common genetic cause for both the penis development and the behavioral maturation (probably testosterone related). Udry, Billy, Morris, Groff, and Raj (1986) have shown that free testosterone predicts the onset of sexual motivation and behavior.

Lynn (1990) has argued that differences in testosterone could explain many of the observed racial differences, including the racial ordering in prostate cancer rates (Negroids highest, Caucasoids intermediate, Mongoloids lowest). Testosterone was 19% higher in black college students than in white students (Ross, Bernstein, Judd, Hanisch, Pike, & Henderson, 1986). Ellis and Nyborg (1992) have documented higher male testosterone levels in black veterans than in white veterans, although the magnitude of the difference appears too small to explain much of the behavioral difference. Prenatal or puberty differences have yet to be examined. Prenatal and pubertal testosterone play a major role in the emergence of masculine behavior (Gandleman, 1992, chapter 3). Olweus (1986) has shown in Swedish boys aged 15-17 that testosterone levels predict physical and verbal aggression. Testosterone levels have substantial heritability (Meikle, Bishop, Stringham, & West, 1987). High Negroid prenatal and postnatal testosterone could explain their more muscular body build, their deeper voices, their greater aggression, their greater dominance drive, their strong sex drive, and their shorter lives. Unfortunately, high testosterone correlates negatively with male occupational success (Dabbs, 1992) .

Thus, sex hormone differences could explain many racial differences. A strong point of differential K theory is its apparent parsimony. A large number of seemingly unrelated differences can be explained by a single evolutionary theory. If many of the differences trace to a single direct cause, sex hormone levels, or other ultimate causes that act through the sex hormones, then other theories are equally parsimonious.

Possibly, some other variable, such as monoamine oxidase levels, that affects multiple racially variable aspects of personality (Zuckerman, 1983, pp. 55-57), could explain many of the observed differences. Ellis (1991, p. 238) cites three studies showing that blacks have significantly lower monoamine oxidase levels than whites. Monoamine oxidase appears to be related to criminality, impulsiveness (Wilson, 1993), and sensation seeking (although the latter appears to be lower in Blacks [Ellis, 1991, p. 238]). Recently, further evidence of a link with aggression has been found through finding a family with a genetic defect in monoamine oxidase formation which lead to high levels of aggression in male carriers (Brunner et al., 1993).

Paternal investment theory deals with natural selection among males. The effects of such selection probably altered female traits also. Most human genes relevant to behavior have similar effects in both sexes (for instance, Eaves et al., 1989, p. 99). If the genes contribute greatly to male reproductive success and have relatively little effect on female reproductive success (as appears likely for many of the genes discussed above) male and female genotypes will both reflect selection for male reproductive success. Recall the earlier discussion of the effect of polygyny on height. This mechanism is consistent with Rohner's (1976) cross-cultural survey finding, that "the level of aggressive behavior displayed by children of one sex varies directly with the level of aggression among children of the other (r=.88, p<.01)." Similar results were found for adults.

However, behavior relevant genes can affect only one sex, especially if acting through testosterone, or other sex hormones. For instance, a gene acting only on the testes would affect only males. Races differ in some traits, such as in impulsivity and risk taking, that are affected by different genes in males and females (Eaves et al., 1989, p. 269).

Draper & Harpending (1988, p. 350) note that, "These father-absent females recognize that male parental effort is not crucial to reproduction and they are less coy and reticent, engage in sexual activities earlier and with less discrimination, and form less stable pair bonds."

All other things equal, females who start reproducing early leave more descendants. If the mother (or her female relatives) can rear the children, early opportunities for pregnancy should be accepted, especially if the genes are from a dominant and aggressive male. However, in cold climates other things are not equal, and waiting for a male who will provision her promotes female reproductive success. A female who has a child by a non-loyal male reduces her chance of catching a provisioning mate. If she does get one, he may be unwilling to support another man's child. Such male behavior reduces or eliminates the reproductive utility of having taken the first opportunity to be impregnated.

Females benefit reproductively from obtaining genes for success at polygyny or extra-marital matings. Females might benefit reproductively from mating a male who had exhibited his mating success, even if this implied less provisioning. Thus the model would predict tropical females would mate earlier than northern females, and would be more accepting of mates who were aggressive, dominant, self confident, non-empathetic, etc. Even if these traits appeared unlikely to lead to marital happiness, males exhibiting them probably had genes conducive to mating success. Northern females would prefer males who exhibited provisioning behavior, and traits conducive to it, such as altruism and empathy. Thus, cold climate females would be selected for the behavioral restraint or weak sex drive needed to resist taking the first mating opportunity, and for the intelligence to detect the courting male who will likely provision her. The earlier marriages and higher illegitimacy rates among Negroids are consistent with being less selective in accepting mating opportunities. The higher child abuse in Negroids (often by husbands or boyfriends) is consistent with less selectivity in mate choice, and possibly with more often choosing aggressive males for mating.

The prediction than Negroid females would select mates more for good genes than for a promise of provisioning is supported by Staples & Johnson's (1993, pp. 111-112) observation that, "Black women demonstrate a stronger preference for a physically attractive man than her White counterparts," although they complain that "When personal characteristics that are genetically influenced make such an important difference in a person's status, a genetic determinism emerges that is very similar to the operation of racial attitudes."

If securing male paternal investment is critical to female reproductive success, females should devote more effort to securing and keeping a provisioning husband than if such a husband is not critical. Notice that a low level of male mating effort is expected to correlate with a high level of female mating investment in this mode. In contrast, in a simple r versus K model, if one sex is high on mating effort the other sex would be expected to be also. Women evolved to attract and retain husbands would be expected to derive satisfaction from fulfilling the role of wife. Among whites (apparently college educated), 74% of women would choose the role of wife over that of mother, while only 50% of black college women, 24% of "middle" status black women (described as upper-lower class), and 16% of low status black women made that choice (Bell 1971, p. 254).

Males persuing a high mating effort strategy (low provisioning) would probably be less satisfactory as mates. Thus, it not surprising that 100% of married white women stated they would marry again if they could live their life over, while only 88% of black college educated women, 64% of the"middle status" blacks, and 36% of the low status blacks made that choice (Bell, 1971, p. 254). In considering these answers, it must be remembered that college educated women were a much lower percentage of the black population, so their answers were probably less typical of all blacks than the college educated were of all whites.

A similar lack of desire to live with ones lover is reported from Jamaica. Among Jamaican black females many actually prefer a "visiting" relationship with their man (i.e. one where the partners live separately, but have sexual relations leading to children) to a common law marriage where the man lives with them (Roberts & Sinclair, 1978). Of course, this may say more about the relevant men, than about the women.

Where males are more aggressive, one would expect females to need to be more aggressive in order to deal with their mates successfully. Where devoting energy to provisioning is not in the male's reproductive interests, females who aggressively assert their rights might be more successful than those who are more passive.

Where the males devoted much of their efforts to mating, leaving the females to provision themselves and their children, females would be strongly selected for traits conducive to successful gathering, such as initiative and a willingness to work hard. Males would be less strongly selected for willingness to do hard continuous work (although they might be selected for the quick burst of energy needed for fighting). Males might even benefit reproductively from laziness, if that led them to remain in camp where they could mate extra-maritally. Thus, the prediction is that tropical women would appear harder working than tropical males.

While direct evidence on this prediction is lacking, many have noted the greater occupational and educational success of U. S. black females relative to black males (see Taylor, 1992, p. 25). Black females earn 99.0% of white female annual earnings, versus 64.6% of white male earnings by black males (Jaynes & Williams, 1989, Table 6-5). The personality traits that were required to feed an African family, both in gathering and in agricultural times (notably consistent, continuous effort), are more consistent with occupational and educational success in an industrial society than those required for achieving multiple matings for males (aggression, for instance). African females without these traits probably left fewer descendants. A low level of male provisioning appears to have existed in sub-Saharan Africa both pre and post agriculture. Even in contemporary Africa, the women appear much more industrious than the men (Lamb, 1987, p. 38).

Why the Emphasis on Hunter-Gatherers? So far this study has focused on the selection for mating versus provisioning during the long hunter-gatherer period. This hunter-gatherer emphasis is traditional in sociobiology, since Homo sapiens have been hunter-gatherers for 99 percent of their time here on earth. Disagreement continues among experts as to just when modern humans emerged (Gee, 1992; Harpending, in press) and how long current races have existed. Mountain, Lin, Bowcock, & Cavalli-Sforza (1992) have shown the classifications that result from using several different methods. Their descent trees and those of others (Zhao and Lee, 1989; Nei & Roychoudhury 1993) agree that the largest genetic difference is between Africans and all other populations.

Recent interpretations of mitochondrial DNA mismatch distributions are consistent with the existence of racial differences. Harpending, Sherry, Rogers, and Stoneking, (1993, p. 494) report that "Given these caveats, our results show that human populations are derived from separate ancestral populations that were relatively isolated from each other before 50,000 years ago," and (p. 495), "The existence of between-group differences far older than within-group differences implies that the late Pleistocene expansion of our species occurred separately in populations that had been isolated from each other for several tens of thousands of years."

An alternative to the "out of Africa" hypothesis for human origins is the multi-regional hypothesis. Recently, two well preserved Homo erectus fossils in China were found that displayed many characteristics similar to those of current Mongoloid populations (Tianyuan & Etler, 1992). This discovery that possibly the early races were the ancestors of the current races. Under either hypothesis, there would have been time for racial differences in behavior to have evolved.

It is clear from the differences in skin color and other traits that human populations have been separate for long enough for major differences to have emerged. These adaptations to climate are believed to have emerged through small differences in survival rates among carriers of different genes. For instance long noses are believed to warm and humidify the air entering the lungs in dry or cold climates (Krantz, 1980, pp. 101-118), presumably reducing the death rate from respiratory diseases. Long periods are required for such small differences in survival to produce observable differences between populations.

In contrast, differences in mating success have the potential for comparatively rapid changes in gene frequencies. Just consider the difference in number of descendants between a male who has children with two wives, and one who has only a single one in areas where survival does not require male provisioning. Any personality traits that handicap a male in obtaining the second wife in such areas would be strongly selected against. For instance, Chagnon (1988) has reported that the Yanomamo males unokais (one who has killed another, usually in a raid) enjoyed higher marital success (1.63 wives per male versus .63 for non-unokais) and had more offspring (4.91 versus 1.59). If certain personality traits (aggressiveness, courage, sensation seeking) contribute to achieving unokais status, selection for these traits could be very rapid, much more rapid then selection that depends on merely slightly lower death rates for carriers of certain genes. In particular, population differences in personality relevant to marital success should emerge quicker than differences in noses. The observed differences in noses between populations imply that there has been time for mental differences to have emerged.

Recently, the a behaviorally relevant alle was shown to vary with race, being more common in blacks than whites (Blum, et al. 1991) The relevant gene (the D2 dopamine receptor gene) is associated with severe alcoholism. It is easy to imagine that alcoholics would be avoided as mates, and that once alcohol was abundant in a society that the genes contributing to it would be rapidly selected against.

Thus, from the observation that populations differ in skin color and related traits whose frequency is determined by differential survival, we can deduce that northern and southern populations have been separated long enough for differences in personality variables that affect mating versus provisioning to have emerged.

Once a population had evolved drives for low paternal investment and high mating effort, these genes would survive the adoption of agriculture. Men would devote their efforts to acquiring additional wives, and to mating with other available females. This would be especially likely if technology was such that a female could raise enough food to provision herself and her family, as is frequently the case in tropical horticulture. Thus, the paternal investment versus mating theory predicts that polygyny would be more common among tropical origin populations.

The model predicts the pattern of current polygyny. In White's (1988, p. 553) map of what he calls male stratified polygyny with autonomous cowives ("autonomous cowives" means that the wives basically support themselves) Africa is conspicuous. Interestingly, several New World examples are among Negroid or mixed populations (Saramacca, Goajiro). Pebley and Mbugua (1989, p. 338) report that "throughout most of southern West Africa and western Central Africa, as many as 20 to 50% of married men have more than one wife," and that "The frequency is somewhat lower in East and South Africa, although 15 to 30% of husbands are reported to be polygynists in Kenya and Tanzania," confirming Welch & Glick's (1981) summary of official statistics. In contrast, in Arab Muslim countries only 5 to 12% of men have more than one wife. While polygyny occurred in Europe and Asia, the rates do not appear to have been as high. Much of their polygyny was in ruling class harems.

The average polygyny rate in each of several "culture areas" was calculated from a tabulation provided by Hartung (1982), whose measure correlates with other measures of polygyny (Low, 1987). The highest was 39% for sub-Saharan Africa. The second highest was 21% for the Island area (including Australia, Indonesia, Polynesia, and Madagascar) which is populated by people that appear to have evolved in warm areas of Southeast Asia. The rates for societies containing Caucasoid (circum-Mediterranean, 15%) and Mongoloid (Eastern Eurasian, 10%) peoples are much lower, with the rates for North (11%) and South American (11%) natives resembling those of other Mongoloid peoples. These rates seem low, even where polygyny is culturally acceptable.

It was argued that during the hunter-gatherer period stronger mating drives emerged in warm areas where winter gathering was possible, and these survived the coming of agriculture. To test this theory, a regression of polygyny on latitude was computed giving (standard errors in parenthesis):

Percentage Polygyny= 32.1 - 0.481 latitude R2=0.098
(2.5) (0.098)

A highly significant effect exists for the Old World considered alone:

Percentage Polygyny= 35.0 - 0.56 latitude R2=0.11

(2.8) (.13)

However, the New World lacks a statistically significant correlation of polygyny with latitude. This lack is striking, and inconsistent with the Old World correlation being caused by the environment somehow affecting culture. Africa's high polygyny rate has been associated with its lack of animal drawn plows (possibly due to tsetse fly infection eliminating draft animals or due to failure to invent or to adopt the device) (see Burton & Reitz, 1981; Goody, 1976; White & Burton, 1988). Animal drawn plows are argued to require male operation, and a male cannot plow enough land to support several families. However, the low New World polygyny (prior to the European arrival), where the lack of suitable draft animals also prevented plow agriculture, presents a problem for this theory. The New World uniformity across latitudes can be explained by inadequate time for climate to have affected gene frequencies related to polygyny. This explains why New World polygyny rates resemble those of other Mongoloids.

Admittedly, African conditions were such that a woman engaged in horticulture could continue to support herself and her offspring. There was a low population density (possibly due to disease) which made shifting cultivation possible. Such slash and burn horticulture has a high per unit labor output (Boserup, 1970). This is because newly cleared land is relatively fertile. In such conditions females can grow enough food for themselves and their children, making it possible for the males to continue their old high mating effort strategy. Males who diverted effort from mating to tending crops would have increased their offspring's survival too little to offset the reduction in the number of children fathered. There are other theories.

Low (1988, 1990) argues that pathogen stress leads to polygyny, through increasing the female incentive to seek genetically superior males. Her data does show a statistically significant relationship between polygyny and pathogen exposure, although the relationship may reflect only the above noted tendency for polygyny to be more common in tropical areas, and the higher disease rates within these areas. Besides Low's mechanism, high pathogen levels could lead to polygyny through lowering the population density to where land productivity was high enough so that females did not require male provisioning. The story then becomes very similar to the argument of this paper.

Hrdy ( 1992, p. 434) has concluded that in horticultural Africa that "Matrilineal social organization combined with female-centered horticultural practices mean that by and large male investment is not critical for child survival and well-being. . ." These are presumedly the conditions in which the optimal male strategy is to emphasis mating rather than paternal investment.

The theory here is that polygyny is a least partially a response to an evolved male desire for sexual variety. This theory has appeared before, only to be quickly dismissed. G. Lee (1979, p. 702) claimed there was no biological evidence in favor of the view that males had a predisposition for variety in sexual partners, and that "there is no reason to expect the properties of the male sex drive to differ across cultures." Male/female differences in the nature of sex drives are now a stable in sociobiology (Symons, 1979), while this paper provides a plausible reason for the properties of the male sex drive to vary across cultures. It should be noted that in a society where polygyny is both accepted and practical (i.e. a typical man can support more than one wife and her offspring), males who lack the drives which lead to polygyny will leave fewer descendants, and such drives will be selected for, or more strongly selected for, than in other societies. Where polygyny is either forbidden, or not usually practical (i.e. multiple wives can not be successfully provisioned), the drives may be selected against, especially if they reduce the provisioning of offspring, or lead to conflicts with other males. Thus, conditions that lead to polygyny are likely to select for male characteristics that will lead to its continuation. However, in the proposed theory, the drives had been earlier selected for in the hunter-gatherer stage, and favorable conditions in many tropical lands (low enough population densities so that women could grow enough food to support themselves and their offspring) merely permitted polygyny to become widespread, and continued the selection for drives leading to mating success.

Thus, African polygyny is probably not recent, but has existed since the early days of African agriculture. Notice how evolutionary reasoning can provide evidence about the period before a written history.

Once males had multiple wives, they rationalized their behavior. Many will protest that culture is independent of biological drives, and that it has entirely separate origins. Here mores are argued to be affected by actions, as well as to affect actions. If someone doubts man's power to rationalize what his drives lead him to do, he might consider homosexuality. There is a remarkable correspondence between acceptance of homosexual acts and homosexuality (defined as a type of sexual orientation). Relatively few homosexuals (i.e., those with strong attractions to those of the same sex) strongly condemn homosexual acts. Current scientific evidence is that homosexuality has biological correlates (Allen & Gorski, 1992; Bailey & Pillard, 1991; Bailey, Pillard, Neale, & Agyei, 1993; LeVay, 1991), and probably a biological basis. If biology does not influence mores (a cultural variable), why the strong correlation between homosexual drives and having mores permitting acting on such drives?

As a thought experiment, it may be useful to imagine the sexual culture and mores that would emerge on an island occupied only by homosexuals (perhaps as a result of exile). As an even stronger thought experiment imagine an island inhabitant by Lesbians who were supplied with sperm from males conducive to offspring being Lesbian. Very likely, in a few generations the culture's sexual mores would be quite different from those of a heterosexual population.

As another thought experiment, imagine a good family man who becomes involved with an attractive female worker at a Christmas party. Surely, he would be less condemning of extramarital affairs afterwards.

Work with bisected brain patients and other sources suggests the left hemisphere of the brain has an "interpreter" which provides interpretation for actions undertaken for reasons it is not conscious of (Gazzaniga 1992, pp. 121-137). While culture clearly influences behavior, behavior and the drives leading to it may also influence culture.

It should be noted that the African polygyny with autonomous cowives would select for a high level of mating effort relative to provisioning effort, and continue a pattern that had probably emerged earlier in hunter-gatherer times. If the argument regarding hunter-gatherer conditions is rejected, African patterns of horticulture with the potential for wives to be self supporting may have existed long enough to select for some of the same traits (especially allowing for the difference in reproductive success among males who differ in their ability to acquire self-supporting wives).

Where males are following a high mating effort strategy, fathers will frequently leave their mates while the children are young and still dependent on their mothers. In such circumstances, the link with the mother will appear more important than those with the father. It then becomes logical to trace descent through the mother rather than the father. This is a plausible explanation for the observations that matrilineal systems are most frequent in simple horticultural systems (Aberle, 1962, Table 17-4), where females supply most of the agricultural labor (i..e. male provisioning is not necessary). Indeed, Schneider (1962, p. 16) asserts that, "The institutionalization of very strong, lasting, or intense solidarities between husband and wife is not compatible with the maintenance of matrilineal descent groups.S and that (p. 22) RIn matrilineal descent groups the emotional interest of the father in his own children constitutes a source of strain . . ."

To argue that cultural differences led to appreciable differences in gene frequencies one must argue that the relevant cultural differences have persisted for a very long period, given the slow speed at which natural selection operates. One should provide a plausible explanation for the persistence of the cultural differences. Cultural differences due to random drift probably do not last long enough to produce appreciable differences in gene frequencies. However, cultural differences due to environmental features can persist long enough for natural selection to change gene frequencies. Suppose African polygyny can indeed to be traced to a low population density (perhaps disease caused) which makes fertile land abundant enough so that wives can support themselves and their children. Then such a pattern could persist for long enough to shift gene frequencies in directions conducive to mating success.

Of course differences between populations in the frequency of behavior relevant genes could arise without being related to differences in appearance, or "race." One plausible place to look for differences would be between populations dependent on herding and those dependent on agriculture. Several observers have noted higher levels of aggression in herding populations. Edgerton (1971) discovered that herding populations are more aggressive than agricultural populations in four East African cultures. Livestock is easily stolen since it can merely be driven away, while crops usually must be harvested and then carried away. Already harvested crops are likely to be in defended storehouses. Thus, in herding areas a pattern of raiding for livestock emerges. Herders appear to be very fierce and willing to fight. Herding is a subsistence method that emerges in certain geographical areas (notably those too dry for agriculture and not easily irrigated). Herding may have been used by certain populations for a very long period, and populations may have been selected for the personalities suitable for raiding or for defending against raids. Higher levels of lactose tolerance in certain milk drinking herders (Bedouins for instance) suggest that they have been dependent on milk from their herds for long enough for gene frequencies to have been altered (Durham, 1991)

Of course, once gene frequencies conducive to devoting efforts to mating at the expense of provisioning had emerged for any reason, a transfer of the populations to other environments would result in a continuation of high mating effort and low provisioning. The forced transference of Africans to the New World provides an opportunity to test this prediction in a manner analogous to the adoption study in behavior genetics.

Interestingly, a pattern of weak husband-wife attachments with a succession of liaisons exists in New World populations of African descent, such as in the Caribbean, with its high percentage of female headed households (see Otterbein, 1965, Table 1, column 4 for the percentage in many societies, and Weinstein [1962, chap. 3] for a description of mating in the U. S. Virgin Islands). Smith (1962, p. 263), after describing several Caribbean systems with frequent changes of mate, notes that European societies have maintained monogamous systems since Tacitus, while West Indians rejected it. While he does not mention genetics, he does describe the extremely brittle marriage systems of the polygynous Hausa of Nigeria, in which the typical woman marries three or four times between menarche and menopause (p. 257). He describes how formal polygyny is not accepted on Carriacou Island (presumably because of church and government opposition), but describes a system where men frequently mate simultaneously with two women, each living separately (p. 29-30).

Roberts & Sinclair (1978) document the Jamaican system, which they report closely resembles the system among Trinidad Negroes (but not Trinidad Indians), with widespread "visiting" unions which produce children without the parents living together. They point out that the Jamaican marriage pattern appears to have been stable since the last century in spite of many political, economic, and social changes, an observation that appears inconsistent with it being a cultural holdover from African days, and with it being a result of slavery. Both cultural models would have predicted some response to the government and church pressures to adopt a more European mating system.

These Caribbean family systems appear to be low paternal investment ones not only because the fathers frequently do not live with the families, and there are multiple partners during their lifetime, but also because the fathers appear to make relatively small financial contributions to the support of their mates and children, even when they live together (Roberts & Sinclair, 1978, p. 64 and their case studies)

The New World Negroid pattern has frequently been explained as a continuation of African culture. "Retention of African polygamous practices was observed by nineteenth-century abolitionists working in the Sea Islands off the coast of South Carolina" (Staples & Johnson, 1993, p. 142). Sudarkasa (1988, p. 31), after noting that "African conjugal families normally involved polygynous marriages at some stage in their development," suggests that polygyny is reemerging in the American black community. Herskovits, who emphasizes the continuity of African traditions, comments that (1947, p. 299), "the father, as in Africa, remains on the periphery of the nucleus constituting the household, whose center is the mother, a grandmother, an aunt." Alternatively, the matrifocal ex-African family can be viewed as a social adaptation (possibly with a genetic component) to a male genetic strategy of low paternal investment, and high mating effort. Among American blacks, very high illegitimacy rates (66.7% of black women who bore a child in the last year were unmarried, versus 16.9% of whites, and 6.9% of Asians and Pacific Islanders [Bachu, 1993, table B]) show the continuation of a pattern of weak paternal investment. Although the percentage of black children living in families headed by women has been increasing (as has the white rate), it appears to have been consistently higher than the white percentage (Morgan, McDaniel, Miller, & Preston, 1993). Of course, the two parent black family has been common in the US, and Gutman (1976) has shown that in slavery and afterwards it was the most common pattern, although he does not claim that the black pattern was the same as the white pattern.`Recent historical research using a sample of Census returns shows that many never married black women with children apparently reported themselves as widows in the Census of 1910 (and presumably in other censuses), making the true rates of illegitimacy appear lower than they really were (Preston, Lim, & Morgan, 1992).

Bulcroft & Bulcroft (1993) found a relatively minor difference between white males and white females in the desire to marry, but a much larger differences between black males and black females. For instance, the percentage of white males not desiring marriage aged 19-25 was 12.6% versus 11.2% for white females. Black females were similar at 12.7%, but 22.8% of black males did not desire marriage. The authors present evidence that the most important explanation for the black male's lower interest in marriage is a belief that it will not have a positive effect on their sex lives. While the authors interpret this as being a result of the favorable sex ratio black males enjoy, the pattern of results could be explained by the hypothesis of this paper. It reflects the blacks male's low paternal investment strategy, with a strong desire for sexual variety (inconsistent with American monogamous marriages) being one of the mechanisms that evolved to direct his primary efforts to mating.

Ethnographic accounts of life among American lower class blacks also report that black females desire traditional marriages, while black males are reluctant (E. Anderson, 1989). Notice that ethnic culture cannot explain differences between black males and females since both share the same ethnicity, while the observed difference is well explained by the black males having been selected in Africa to use a low paternal investment strategy. The reproductive interests of black females both in Africa and in the US is served by being mated to a male who will make high paternal investments.

These patterns of weak pair-bonding, and low levels of male provisioning, are sometimes explained as a residue of plantation slavery. Testing this theory requires finding a New World African origin population that lacks a history of plantation slavery. Such a population exists. It is the Black Carib (coastal Belize, Guatemala, and Honduras), who are the descendants of survivors of slave ships wrecked before reaching the plantations, and who intermarried with the Carib natives (and later with other blacks). Still there emerged a pattern of multiple marriage partners during life, with some having simultaneous wives (Gonzalez, 1969, p. 72). This mating pattern appears quite similar to the pattern among other New World African origin populations.

The evidence provided by ex-Africans in the New World is important because, like the adoption study in behavior genetics, it alters the environment while leaving the genes unchanged (although often mixed). Since culture seems to change rapidly among immigrants to a new land, it is hard to imagine much of the original African culture having survived the voyage to the New World, and the many generations of influence by Christianity and other aspects of Western cultures.

Differences in selection for provisioning versus mating may explain other cultural regularities, such as the association between father-absence and aggression or crime (Whiting, 1965; Bacon, Child, & Barry, 1963). Draper & Harpending (1987, p. 349) have stated "father present societies are those where most males act like dads and father absent societies are those where most males act like cads," and have described other characteristics of the two types of societies. For instance, father absent societies are associated with local raiding and warfare, hostile relations between men and women, higher level of male violence, male public bombast oratory and rhetoric, transient bonding between males and females, less male direct provisioning, and women devaluing the male paternal role.

Draper and Harpending recognize the relevance of paternal investment theory (their theory involves early childhood experiences), but do not predict which societies will be father-absent ones (but see Belsky, Steinberg, & Draper, 1991, for an update). The alternative proposed here is that father-absence exists in descendants of tropical hunter-gatherers (which include most so-called middle-range societies practicing non-intensive agriculture), while the father-present pattern exists in descendants of northern hunter-gatherers (including the wet rice and other plow-using grain growers).

Whiting & Whiting (1975) use the label, the "aloof societies." They document an association between spouses sleeping apart and male aggressivity. In their theory, male children deprived of father contact compensate by becoming hyper-masculine and aggressive (notice the effect contradicts the simple hypothesis that male children learn aggression from frequent contact with their fathers, since father absence would then reduce aggressivity). The provisioning versus mating model suggests that tropical selection for mating success produces both polygyny (which leaves males with less opportunity to interact with their children) and male aggression.

As an additional factor, males not selected for provisioning would be less nurturing, and would not desire to spend time with small children. This would make them more likely to live away from their wives. In addition, a hyper-masculine male would not be very pleasant to live with, and the wives would probably be content to live apart from him. Part of the individual level correlation between criminality and father-absence (R. Anderson, 1968) may be due to "hyper-masculinity" leading to both.

Likewise, White and Burton (1988) have documented a relationship between warfare and polygyny. While the mechanisms they describe (such as marrying captives) are plausible, an alternative is that both result from a complex of traits, including aggressiveness, that reflect selection for mating success. White and Burton end their paper by noting that their model works less well for the New World, where the polygyny is different from the general polygyny they describe. They note (p. 884) that, "Much of New World polygyny appears to be of a different pattern, in which wives tend to be related to one another and to live in the same house." As noted, the genetic theory proposed here predicts lower rates of general polygyny in the New World, whose natives descended from relatively recent North Asian immigrants.

The above sexual selection theory was developed by considering the data assembled by Rushton on the major races, Mongoloid, Caucasoid, and Negroid. As seen, it seems to fit well.

However, there are other groups which are often considered distinct from the major races, Australian aborigines, the "brown" people of Southeast Asia, Polynesians, Micronesians, etc. Other populations appear to be mixtures of original populations. A climatic adaptation theory predicts that populations evolving in the tropics will resemble Negroids in their behavior and life history. Differential K theory makes no such prediction until it is specified how variable or predictable the area they evolved in is. If it is tropical rain forest, it would be of low variability, and one would expect them to be K selected.

While the writer's impression is that these other groups do resemble Negroids more than they do Mongoloids in their behavior, no systematic investigation has been undertaken. These peoples provide a holdout population that others can use to test the hypothesis proposed here.

An interesting population to examine would be the pygmies of various tropical rain forest groups. Their diminutive stature shows that they have been separated from their neighbors long enough to be physically different. These are predictable climatic areas, for which differential K theory would predict K traits. It is here hypothesized that they will have most personality and life history traits similar to those of their tropical neighbors.

Climatic theories predict gradual shifts in gene frequencies. As one moves towards the Arctic, isolated populations should evolve more in the Mongoloid direction. The theory here predicts that the more cold adapted groups among the major races, such as the Eskimos among the Mongoloids, will exhibit an even greater contrast with the tropical peoples than typical Mongoloids. Likewise, as one moves south, the behavior should shift in the tropical direction. Thus the northern European Caucasoids should resemble the Mongoloids more than those further south, and the Mediterranean Caucasoids should resemble the Negroids more. While formal studies appear lacking, the reported greater sociability, macho complex (Peristiany, 1965), and acceptability of mistress keeping among Mediterranean peoples, and the higher levels of behavioral restraint further north, appear to be as predicted by the existence of a male provisioning versus mating cline.

Thus, the provisioning versus mating theory is testable. Hopefully, further research will test it.

Offspring survival in cold climates requires provisioning by male hunters, while it is not critical in warm climates. Thus, the optimal male tradeoff between seeking copulations and provisioning depends on the climate. Hence, the colder the climate a population evolved in, the more they should have evolved drives that lead to provisioning (altruism, sexual restraint, rule following behavior) while in tropical areas the drives should have evolved towards competing for mating opportunities (which implies dominance seeking, aggression, high masculinity, extraversion etc.). This can explain many of the observed differences between the major races. While cultural explanations exist for many of the behavioral differences, they are unable to explain such differences as body build, genital length, muscle structure, bone structure, the size of the liver, testosterone levels, and monoamine oxidase levels, all of which are explained by the paternal investment versus mating success theory.