Anth235 Midterm

Other Menstrual Cycle Effects

As short term mates, fertile women prefer: Symmetrical men The scent of symmetrical men Deeper male voices Creativity over wealth Differences are not seen for long-term mate preferences

Gregor Mendel (1822-1884)

Austrian Monk Pea Plant Experiments Mechanism of Inheritance over Generations

Pelvis

Basin-shaped Ilium—short, broad, flared Large sacrum **Remember: Functional Compromises

Costly Behaviors

Behaviors can be costly signals like ornaments Costly behaviors include: mating dances stotting risk taking

Men and Long-term Mating

Benefits Cooperation / division of labor Increase in mate quality Combat problem of concealed ovulation High quality offspring Costs Lost opportunity for multiple matings

Ritual

Ceremonial burial Grave goods Decorative beads Pigments (red ochre) Variation

Why Study Primates?

Closest living relatives (humans as primates) Reconstruct hominid behavior and ecology Evolutionary processes Diversity of primate adaptations Urgency

Intra-sexual Selection

Competition between members of the same sex (usually males) for access to mates Leads to the evolution of armaments

Punnett Square

Different allele combinations Calculate phenotype and genotype ratios

Parent Offspring Conflict (Trivers)

Fitness interests of parent and offspring are not identical Parental perspective: offspring increase in foraging allows increased investment in additional offspring, at some cost to the foraging offspring Child's perspective: own increase in foraging reduces investment and increases costs for self Decision: reflect a compromise between parental and offspring interests, although the actual result will depend on local ecology, foraging techniques, fertility levels

symmetry as a fitness indicator

Fluctuating asymmetry is a measure of developmental instability (Symmetry is a measure of stability) Both mutations and environment can cause slight deviations in development

Reproductive Value and Mating Context

For a long term mate, men should have a preference for high residual reproductive value = youth For a short term mate, men should be more sensitive to high fecundity

Prosimians

Found in Madagascar and Old World The greatest diversity is in Madagascar Many nocturnal Greater reliance on smell

"fitness" indicators

Genetic fitness is the inherited part of fitness - the part of fitness that can be passed on to offspring Complex traits make the best fitness indicators because they take many genes to produce, so will be affected by any of a number of mutations and are more difficult to fake

The La Cerra Effect

La Cerra (1994) wanted to see whether women tended to find men who were willing to invest in children more attractive than men who did not Women may select for cues that indicate a man will be a good father and a heavy investor in his offspring In order to test her hypothesis, La Cerra showed slides of men and women in the following conditions Standing alone Interacting positively (smiling, maintaining eye contact, reaching for the child) with an eighteen-month-old child Ignoring the child who was crying In the room neutrally facing the child Vacuuming a living room rug Male and female undergraduates were then asked about the individual's attractiveness as a) a date, b) a sexual partner, c) a marriage partner, d) a friend, e) and as a neighbor. Do you expect that males and females responded differently to the different photographs? Which pictures would you expect that females would be most attracted to for a marriage partner? Which pictures would you expect that males would be most attracted to for a marriage partner?

An individual's genetic fitness should affect their mating strategy

Low FA men have more mating opportunities Low FA men invest less in relationships Men with low FA have more lifetime partners Men with low FA have more extra-pair sex (short term matings)

The Red Queen Hypothesis

Males and females may evolve strategies and counter-strategies and counter-counter-strategies

Infanticide

Males may kill the offspring of other males in order to bring females into oestrus and/or avoid investing in other males children

Homosexuality: Conclusions

Many more studies are needed There is limited support for the kin selection hypothesis Reasonable support for the alliance hypothesis Most likely there are multiple causes/forms or homosexuality. So this needs to be taken into consideration We are far from understanding the genetic/environmental/hormonal influences

Old World Monkeys

Medium to Big Examples: Baboons Macaques Colobus Guenons

Of Gay Sheep, Modern Science, and Bad Publicity

OHSU researcher Charles Roselli was studying hormonal influences on homosexuality in sheep Protests erupted over the idea that he was trying to find a cure for homosexuality

Common Chimpanzee

Pan troglodytes Tropical Africa Knuckle-Walking

Growth & Development

Pronouncedencephalization Childrenbornprogressively"younger"(secondaryaltriciality) Extendedlifehistory Childhood&adolescence

preferences based on WHR

Psychology for WHR assessment and male preference for low WHR are hypothesized to have been selected for because WHR is an indicator of reproductive value, including sex life-stage pregnancy Parity-refers to the number of times a woman or female animal has given birth. health problems (e.g., diabetes, heart disease) fat stores available for fetal cognitive development

But, in real life ...

Ranking known peers on traits and desirability Female rankings much more correlated with resource linked traits Physical attractiveness much more important for both

Circumcision

Reduces penile sensitivity Increases sperm displacement Circumcised men thrust deeper and harder Why do men have a foreskin?

Reproductive Strategies

Reproduction requires survival, mating, rearing offspring Male vs. Female Strategies

The Apes

Separated into two groups Lesser Great Live in Africa and Asia No tails Arms can rotate for swinging through trees (monkeys can't swing on monkey bars) Larger Bigger brains

The Problem of Paternity Certainty

Sexual jealousy 60%ofmenpicked sexual jealousy as more distressing than emotional 85%ofwomenpicked emotional infidelity as more distressing Preferenceforfaithfulness and chastity

Post-copulatory Sexual Selection

Sexual selection can occur both before and after copulation (sexual intercourse) When two or more males mate with one female, male sperm compete to fertilize a female egg (sperm competition) Sometimes females can influence which male's sperm wins (cryptic female choice)

Bonobos

Sexually bonded females can form coalitions to mitigate male dominance Sex is a form of communication that reduced friction between individuals

Marriage

Shared interest in genetic offspring requires expectation of exclusive sexual access Not just about reproduction (directly): Marriages bind alliances, economic relations between groups and families Allows for division of labor

The Sexy Son Hypothesis

Similar to runaway selection Females choose males with a preferred trait They have sons who also have the preferred trait - thus, their sons are sexy Their sexy sons get more mates, thus passing on the trait even further

Male Strategies

Single male groups Become dominant male Recruit new group Infanticide Multi-male groups Monitor estrus cycles (timing is everything!) Friendship/consortships Peripheral mating

New World Monkeys

Smallest: Marmosets and Tamarins Middle sized: includes Capuchins and Squirrel monkeys Biggest: Atelines or Spiders and Howlers

cultural variation

Some preferences will be consistent across cultures if they signal things that are valuable across cultures E.g., every culture values health as signaled by unblemished skin Others will be culture specific - body weight is differentially related to health in resource poor and resource rich environments

Local Resource Competition (LRC)

Summary: if food is limited and rank inherited, low ranking females do better producing male offspring who leave and don't inherit low rank, but can potentially reproduce elsewhere If high ranking, produce females who can help maintain their matriline's dominance.

Mating Context

The problems faced by men and women should depend on mating context Long term mating: selecting a partner for marriage or shared investment in children Short term mating: selecting a partner for a single sexual encounter, a summer fling, etc. Because of PI asymmetry, men are expected to devote more time to short-term mating

Mating Systems and Parental Investment

The variance in male reproductive success is closely tied to the mating system Monogamous males have about the same number of children as females Polygynous males can have many more children Thus, male-male competition should be greater in polygynous mating systems Males should be choosier in picking monogamous mates

waist-to-hip ratio (WHR)

The waist is measured as the thinnest part between the belly button and ribs The hips are measured at the widest point around the buttocks

Most Traits Probably have Elements of All

Traits selected for through runaway selection tend to become correlated with genetic quality since only high quality males can produce them Sexy tends to be the same as genetic quality for the same reason, so males with good genes have sexy sons with good genes

Heterosexual boyfriends

Travesti's boyfriends consider themselves to be heterosexual The boyfriends penetrate and are not penetrated

Alliances

Typically the predominant same-sex behavior goes with which sex has more social bonding Male-bonded species (chimpanzees, geladas, baboons, macaques, dolphins) show male-male same sex behavior Female-bonded (squirrel monkeys, macaques, bonobos, elephants) show female-female same sex behavior.

FA and sexual behavior

Women prefer men with low FA as short-term mates Low FA men are rated as more attractive by females Women with more resources prefer symmetrical men

O'Hara and O'Hara 1999

Women were more than four times less likely to have an orgasm with a circumcised male Circumcised males were almost twice as likely to have a premature ejaculation Twenty times more likely to report vaginal discomfort with a circumcised partner

Other Ways to Advertise Ovulation

Women who are ovulating wear skimpier clothing to nightclubs Women are more likely to engage in EPCs when ovulating

Sexual Strategies Theory

Human mating is strategic Humans seek particular mates to solve specific adaptive problems that their ancestors confronted during the course of human evolution Drawing on Parental Investment Theory, sexual strategies theory predicts: Over evolutionary history, men and women have faced different mating problems Men and women should have evolved different solutions

Human Penises

Humans have relatively simple but very large penises Humans have relatively large testes

The Well Hung Humans

Hypotheses for the large penis: Female choice Visual Tactile Male-male competition Sperm competition

Primate Social Systems

In general, primate social systems evolve based on food resources Females group where food is Males group where females group Semi-Solitary For example: orangutans & nocturnal strepsirhines Single male polygyny (one male/multi- female) e.g., gorillas Multi-male/multi-female groups e.g., baboons Fission-Fusion Type of multi-male/multi-female groups e.g., Chimpanzees Monogamous (Pair-bonded) e.g., gibbons Polyandry (Multi-male/one-female) e.g., marmosets & tamarins

The Coolidge Effect

In nearly every species, males show renewed sexual arousal when introduced to a new female

Comparative Data from Other Animals

In non-human primates, homosexual behavior is primarily used in two ways Alliance formation Dominance interactions

Dominance and Inter-sexual Selection

In primate systems, dominant males typically obtain the most mates The same principle applies in humans Human "dominance" can take many forms - often not related to physical dominance

Australopithecus

"Southern Apes" Adaptive Radiation: >4mya to ~1mya

Patriarchy and Sex Ratio

In the common patriarchal society, a well-fed son can produce more grandchildren than a well-fed daughter. BUT if you stop having children as soon as you have a son, this could also produce a male-biased ratio. Selective abortion Killing infants of one sex Investing more in one sex

Male Hunting

In traditional societies, males hunt large, difficult to kill game Meat is usually shared communally, not just with the family But better hunters get more/better wives

Embodied Capital Theory

Kaplan et al. (2000) argue that the extended juvenile period and long lifespan of humans co- evolved with: Dietary transition to high- quality, difficult-to-acquire foods

Care for Sister's Offspring

Kin selection and male androphilia in Samoan fa'afafine Paul L. Vasey, David S. Pocock, Doug P. VanderLaan

The Discovery of DNA

1953 Watson and Crick Franklin & Wilkins

Gorilla (Gorilla gorilla)

3 Subspecies: Mountain, Western, & Eastern Lowland Gorilla Tropical Africa Knuckle Walking

Smith, Cornellisen and Tovee 2007

3D bodies Body fat was a better predictor of attractiveness than WHR Physical fitness was not a good predictor

Female Strategies

Age of first reproduction Age at weaning of offspring Interbirth interval Alloparenting Mate choice

Frequency in the USA

Alfred Kinsey report (1953): Sliding scale 7 levels from homo to hetero Men 4% exclusively homosexual 37% had orgasmic sex with other men 10% were primarily homosexual for at least a three-year period (average rate 4 to 10%) Women 20% had homosexual sex 2% were exclusively homosexual. In a representative sample of U.S. men in 1970 (n = 1,450), 20% had orgasmic sex with other men, 7% after age 19 In a sample of contemporary U.S. women (n = 1,384), (Janus and Janus 1993) 17% had homosexual sex Only 5% defined themselves as gay or bisexual For men (n = 1,335) 22% had homosexual sex 9% defined themselves as gay or bisexual

WHR and Parity

Although in Western cultures women get larger as they have more kids, relative to those who have no kids they decrease in hip fat

men and baby faces

1. Introduction Kin discrimination mechanisms allow individuals to modify their behaviour with respect to genetic relatedness (Lehman & Perrin, 2002) and have been shown to affect social (DeBruine, 2002, 2004b; Hauber & Sherman, 2001), sexual (Lacy & Sherman, 1983; Neff & Sherman, 2002), and parental behavior (Daly & Wilson, 1982; Platek, Burch, Panyavin, Wasserman, & Gallup, 2002, 2003; Regalski & Gaulin, 1993; Volk & Quinsey, 2002). In humans, facial resemblance has been suggested as a kin identification mechanism (e.g., Daly & Wilson, 1998). Because of concealed ovulation, internal fertilization, and female infidelity, parental cer- tainty is asymmetrical: Only males are susceptible to error in identifying their young. Current estimates of extra-pair paternity in humans vary between 1% and 20%, with most around 10% (Baker & Bellis, 1995; Cerda-Flores, Barton, Marty-Gonzalez, Rivas, & Chakrborty, 1999; Neale, Neale, & Sullivan, 2002; Sasse, Myller, Chakraborty, & Ott, 1994; Sykes & Irven, 2000), and the human male has apparently evolved various paternal assurance tactics to counter female infidelity (Gallup et al., 2003; Goetz et al., in press; Platek & Shackelford, in press; Shackelford, 2003; Shackelford, Goetz, & Pound, in press), thereby increasing the likelihood that paternal investment will be bestowed on one's own genetic offspring. Men may use assessment of facial resemblance as a paternal investment strategy. Using a hypothetical investment paradigm and computerized facial morphing, Platek et al. (2002, 2003) found that males were more likely than females to utilize facial resemblance when asked to make hypothetical parental investments. In a recent attempt to replicate Platek et al., however, DeBruine (2004a) found that both sexes responded positively to self- resemblance in children's faces, with no apparent sex difference. The methods of DeBruine differed from the original studies of Platek et al. by using color photographs of faces, by using a different set of faces for each rating task, by using an algorithm that neotonized participants' faces (i.e., warped participants' faces to appear childlike prior to morphing) and by using images of infants aged 15-27 weeks, rather than 2 years. These modifications in methodology may account for DeBruine's failure to replicate Platek et al. and suggest that multiple factors, requiring further research, may influence the impact of facial resemblance on male parental investment decision making. Here, we conducted two experiments to address the neurocognitive factors involved in reactions to children's faces. In Experiment 1, we replicated the procedures and slightly 396 S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 modified the parameters used in Platek et al. (2002). First, we used high-resolution color photographs of children and participants comparable with those in DeBruine (2004a). Second, we did not warp faces to a standardized face space, as was done previously (see Platek et al., 2002, 2003). In Experiment 2, we employed functional magnetic resonance imaging (fMRI) to investigate neural substrates activated by children's faces. 2. Experiment 1: Replication of Platek et al. (2002) in color 2.1. Methods 2.1.1. Participants Forty-one right-handed college students (22 males, 19 females) volunteered for this study and received course credit for their participation. Participants were recruited to participate in a self-face recognition experiment and a childcare experiment. Males who were not clean- shaven were excluded from participation. Participants had their pictures taken and were instructed that the initial phase of their participation was a self-face recognition task. In brief, this task entailed responding to self-, famous, and several novel, unknown faces and lasted approximately 10-15 min. This procedure was used for two reasons: first, because of our ongoing interest in cognitive aspects of self-face recognition, and second, to make it less obvious to the participants that their pictures were being used in the bchildcareQ experiment. At the end of the study, all participants were debriefed: They were told that their pictures were used in both parts (self-face and childcare) of the study and were given the opportunity to ask any questions. 2.1.2. Stimuli and procedure High-resolution pictures of the participants were taken using a Hewlett Packard (Model 315) 2.1-megapixel digital camera under uniform lighting conditions. The participants were asked not to smile or frown, and to try to maintain a neutral unexpressive face; if a participant blinked or made a facial expression (e.g., smiled), the photo was retaken. Images were processed using a 1.9-GHz laptop computer (Dell), Adobe Photoshop Elements (Version 2.0), and Ulead MorphEditor (Version 1.0) software and were presented in color and were matched for luminance. Images were cropped (using the magnetic lasso tool in Photoshop) just under the chin, from ear to ear, and just below the hairline so that only the face was cropped. Images were then feathered (10 pt) onto a black background and mounted on a canvas of consistent size; image aspect ratio was maintained to eliminate distortion via forced warping to a standardized bface space.Q Each participant's picture was morphed (Ulead Morph Editor, Version 1.0) with three children's faces (two males, one female; age range 1.5-2.25 years), so that each stimulus image combined 50% of the participant's face and 50% of the child's face (see Fig. 1). Images were presented using SuperLab (Cedrus, Version 2.0) in an experimental design following Platek et al. (2002): Participants were asked to select one child's face out of an array of five faces in response to 10 hypothetical investment questions (see Table 1) by depressing a key S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 397 Fig. 1. Example facial morphs. (Color version of figure located at http://psychology.drexel.edu/ECNL/ Color_morphs.tif ). on a standard computer keyboard (all responses were forced-choice). Questions were presented in random order across participants, and face positions changed across questions and participants. 2.2. Results There was no effect of sex of the child. We created composite positive (Questions 1, 2, 4, 6, 7, and 10; see Table 1) and negative (Questions 3, 5, 8, and 9) investment scores that consisted of averaging responses (selected self-child morph = 1; did not = 0) across faces. A Mann-Whitney U test revealed that males (mean rank = 22.70) were more likely than females (13.24) to select self-morphs in response to the positive questions ( p b .001). There was no sex difference in response to negative questions (male mean rank=21.73, female=20.16; p = .6). The difference between the composite positive and the composite negative scores also differed between the sexes (male mean rank = 26.20, female = 14.97; p b .01). Considering individual questions, males chose self-morphs significantly more often than females in response to bwhich one of these children would you adopt?Q (v2=11.27, p b .01), bwhich one of these children would you resent least paying child support for?Q (v2 = 11.80, p b .01), and bif one of these children damaged something valuable of yours, which one would you punish the least?Q (v2 = 18.34, p b .01), and the sex difference in response to bwhich one of these children would you spend the most time with?Q (v2 = 7.32, p = .06) and bwhich one of 398 S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 Table 1Ten hypothetical parental investment questions and the number of participants who selected 0, 1, 2, or 3 of the self-child morphs for each question, by sex Sex Male = Questions Female 1. Which one of these Male children would you adopt? Female 2. Which one of these children would Male you spend the most time with? Female 3. Which one of these children Male would spend the least time with? Female 4. Which one of these children Male would spend US$50 on first? Female 5. Which one of these children Male would spend US$50 on last? Female 6. Which one of these children Male do you think is the cutest? Female 7. Which one of these children would Male you resent least having to pay Female child support for? 8. Which one of these children would Male you resent most having to pay child Female support for? 9. If one of these children damaged Male something valuable of yours, which Female one would you punish most? 10. If one of these children damaged Male something valuable of yours, which Female one would you punish least? 22 =19 Number of participants who selected self-morphs (out of three) 01 2 3 4 5 7 6 11 6 1 1 5 7 6 4 12 4 2 1 17 4 1 0 14 4 1 0 6 10 4 2 8 10 1 0 13 8 1 0 12 5 2 0 7 9 5 1 13 4 1 1 2 9 7 4 11 5 2 1 15 5 2 0 14 5 0 0 19 3 0 0 14 4 1 0 3 7 7 5 15 2 2 0 these children do you think is the cutest?Q (v2 = 6.20, p b .1) approached significance (see Table 1). These data on individual questions need to be interpreted with caution because no statistical correction for Type I error (e.g., Bonferroni) was applied. 3. Experiment 2: Sex difference in neural substrates that process child facial resemblance 3.1. Methods 3.1.1. Participants Nine right-handed students (four males, five females), none of whom had participated in Experiment 1, volunteered for participation. Handedness was assessed using a modified version of the Edinburgh Handedness Inventory (Oldfield, 1971). 3.1.2. Pictures S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 399 Each participant's picture was taken and morphed with a child's face (50%, Ulead Morph Editor version 1.0), as in Experiment 1. Additionally, one novel and one famous face were also morphed with the same child's face and used as control stimuli. At the completion of the study, participants were asked whether they recognized any of the faces, and none did. Each of the three stimulus classes was created independently for each participant. The stimuli were of the same visual quality as in Experiment 1, that is, presented in color and at the same size and resolution. 3.1.3. fMRI imaging parameters Images were collected by a Siemens Magnetom Vision 1.5 Tesla scanner with echoplanar capability (25 mT/m, rapid switching gradients). The participants were instructed to lie still throughout the scanning procedure and focus on the center of the field of view while looking into the goggles (see below). Additionally, foam pads within the head coil helped secure head fixation and prevent motion during scanning. Scanning began with the collection of high-resolution T1-weighted imaging sequence acquired in the axial plane to locate the positions for in-plane structural images. Twenty-six (whole brain) contiguous (no gap) 5-mm axial high-resolution T1-weighted structural images (matrix size=256256; TR=600; TE=15 ms; FOV=21 cm; NEX=1; and slice thick- ness = 5 mm) were collected for spatial normalization procedures and overlay of functional data. Precise localization-based standard anatomic markers (AC-PC line) were used for all participants (Talairach and Tournoux, 1988). Next, functional images were acquired with a gradient-echo echo planar free induction decay (EPI-FID) sequence (T2* weighted: 128128 matrix; FOV=21 cm; slice thickness=5 mm; TR=4 s; and TE=54 ms) in the same plane as the structural images. The size of the imaging voxel was 1.721.725 mm. 3.1.4. fMRI experimental design The study was designed to measure blood oxygenation level dependent (BOLD) responses to self- versus nonself morphs. A boxcar design was used, in which participants were shown a stimulus for 1500 ms every 2 s during the activation period. During the rest period, a radial checkerboard pattern was presented in the same fashion to collect the baseline MRI signal. The checkerboard has been used as a rest stimulus in other face-processing studies and serves to control for visual cortex activation (Henson, Shallice, Gorno-Tempini, & Dolan, 2002). Participants were asked to look at the images and think about whether the image made them feel positive, neutral, or negative. Each of the three stimuli (self-child, stranger-child, and famous-child morphs) was presented in blocks of six activation and six rest epochs (20 s) to maximize BOLD signal response and minimize habituation. The stimuli were delivered through stereoscopic goggles designed for use within the fMRI environment (Resonance Technology, Inc) and Neurobehavioral Systems Presentation software (www.neurobs.com). 3.1.5. Analysis of fMRI data The postacquisition image preprocessing and statistical analysis were performed using SPM'2 (Statistical Parametric Mapping, Wellcome Department of Cognitive Neurology, 400 S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 University College of London, UK), run under the MatlabR (The Mathworks, Natick, MA) environment. Images were converted from the Siemens format into the ANALYZE (AnalyzeDirect, Lenexa, KY) format adopted in the SPM package. Slice timing correction was performed to compensate for delays associated with acquisition time differences among slices during the sequential imaging. A 3D automated image registration routine (six- parameter rigid body, sinc interpolation; second-order adjustment for movement) was applied to the volumes to realign them with the first volume of the first series used as a spatial reference. All functional and anatomical volumes were then transformed into the standard anatomical space using the T2 EPI template and the SPM normalization procedure (Ashburber & Friston, 1999). This procedure uses a sinc interpolation algorithm to account for brain size and position with a 12-parameter affine transformation, followed by a series of nonlinear basic function transformations seven, eight, and seven nonlinear basis functions for the x, y, and z directions, respectively, with 12 nonlinear iterations to correct for mor- phological differences between the template and given brain volume. Next, all volumes underwent spatial smoothing by convolution with a Gaussian kernel of 3.443.4410 mm3 (two times the voxel size) full width at half maximum (FWHM), to increase the signal-to- noise ratio (SNR) and account for residual intersession differences. SPM'2 General Linear Model (GLM) and random effects analysis (RFX) procedures were used to identify voxels associated with significant activation. To test for sex differences in activation to self-morphs, we performed two primary contrasts: overall child face versus null (nonface stimulus) and self-child morph versus non-self-child morph (by sex: independent samples t test in SPM'2). Statistical parametric maps (SPM{t}) were obtained reflecting significantly activated voxels for each contrast and the model used (pb.005; cluster=four voxels). 3.2. Results When responding to children's faces, independent of whether the child face was a self- or non-self-morph, we observed significant ( p b .005, four-voxel cluster detection) activation in right middle (BA 11) and left inferior frontal (BA 45) gyri (see Fig. 2a and Table 2) for both sexes and no significant sex difference. There were no significant patterns of activation when comparing morphs made with famous faces and those made with stranger faces, hence, these conditions were collapsed for comparison to the self-morphs. When comparing activation associated with the self- morphs minus the non-self-morphs, males showed different activation when viewing self-child morphs than did females. In males, cluster analysis for self-morphs minus non- self-morphs pointed to significant activation (pb.005, cluster=4) in the left superior, middle, and medial frontal gyri (BA 10,9; see Fig. 2b and Table 2), whereas in females, cluster analysis for self-child morph minus non-self-child morph revealed significant activation in the right superior and medial frontal gyri (BA 10) and insula, as well as the left medial superior frontal gyrus (see Fig. 2c and Table 2). Males and females did not differ in patterns of activation associated with viewing non-self morphs. However, the right hemisphere frontal lobe activation observed in females to self-morphs minus S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 401 Fig. 2. (a) Significant activation for overall effect of looking at children's faces. (b) Significant activation in males when contrasting self-child morphs with non-self-child morphs. (c) Significant activation in females when con- trasting self-child morphs with non-self-child morphs. All contrasts are set to a threshold of pb.01 with a spatial detection cluster threshold set at four voxels. (Color version of figure located at http://psychology.drexel.edu/ ECNL/Color_Figs1-3.tif ). 402 S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 Table 2Local maxima of cerebral blood flow (CBF) minimum spatial extent = 4 voxels) change for each experimental contrast (height threshold p b .01, Region All child facesMiddle frontal gyrus Inferior frontal gyrus BA 11 HEM x R 38 yz 48 7 21 3 44 18 54 14 40 26 Z score 2.68 2.65 2.93 2.39 3.09 4.64 3.61 3.53 2.47 p (nmax N k) b .01 b .01 b .01 b .01 b .01 b .01 b.01 b .01 b .01 45Males (self-child morph-non-self-child morph) Middle frontal gyrus Superior frontal gyrus Medial frontal gyrus Females (self-child morph-non-self-child morph) Superior frontal gyrusInsulaMedial frontal gyrusMedial superior frontal gyrus 4 16 36 31 10 8 L 20 10 L 28 10 L 30 9 L 2 10 R 40- R 3610 R 12 50 10 L -18 60 BA = Brodmann's area; HEM = cerebral hemisphere; coordinates in Talairach space (Talairach & Tournoux, 1988). non-self-morphs was in the same region as that observed in the overall child face condition, which was not true for males. 4. Discussion These data show that the sex difference in reactions to facial resemblance of children (Burch & Gallup, 2000; Platek, 2002; Platek et al., 2002, 2003; Volk & Quinsey, 2002) is replicable using high-resolution color images and may be associated with a sex difference in neural processing. In Experiment 1, using software comparable to DeBruine (2004a), we replicated and improved upon Platek et al. (2002) by demonstrating that males respond favorably to children as a function of facial resemblance when making hypothetical investment decisions. In Experiment 2, we extended these results and found a sex difference in neural substrates activated by seeing children's faces that resembled the viewer. Unlike females, males showed significant neural activation in the left frontal cortex, which has been hypothesized to be involved in the inhibition of negative responses (Collette et al., 2001; Davidson, 1997; Harmon-Jones & Sigelman, 2001). In another recent investigation of the neural correlates of facial resemblance and face recognition in child and adult faces, using fast event-related fMRI, Platek, Keenan, & Mohamed (in preparation) also found a sex difference in activation associated with facial resemblance—specifically when responding to child faces that were morphed to resemble the participant. It would appear, therefore, that facial resemblance in children may activate neural substrates associated with the inhibition of negative responses in males. Thus, sex differences in reactions to facial resemblance may be driven by neurocognitive processes that are recruited when facing specific adaptive problems, such S.M. Platek et al. / Evolution and Human Behavior 25 (2004) 394-405 403 as making determinations about paternity, which supports the hypothesis put forth by Daly and Wilson (1998) that males may use self-resemblance in allocating paternal investment. Although not directly tested in this study, the right lateralized frontal and medial frontal activation observed in females may be part of a mentalizing (e.g., theory of mind) module. In our behavioral studies (e.g., Platek et al., 2002), participants were asked how they made their choices: Most males denied adopting any particular strategy, but females often reported attributing personality characteristics when responding to hypothetical investment questions (e.g., trying to find the dnicestT child to give money to, spend time with, or adopt). In other words, females may have consulted psychological mechanisms that infer personality characteristics to make their hypothetical investment decisions. This hypothesis is supported by recent neuropsychological and neuroimaging data on mentalizing (see Castelli, Happe, Frith, & Frith, 2000; Fletcher et al., 1995; Gallagher et al., 2000; Ishii et al., 2002; Klin, Jones, Schultz, & Volkmar, 2003; Platek, Keenan, Gallup, & Mohamed, 2004; Stuss, Gallup, Gallup, & Alexander, 2001; Vogeley et al., 2001). There are obvious limitations to these studies. Experiment 1 relies on responses to hypothetical questions about morphed child faces, which may not adequately model real- world scenarios. Additionally, participating in a self-face recognition task prior to making responses about self-child morphs may have introduced an unconscious self-face priming bias. However, there is no obvious reason why males should be more affected by self-face recognition priming than females are; indeed, one might expect that females would be more affected because they perform better on unconscious self-recognition tasks (Keenan et al., personal communication; Platek, Burch, & Gallup, 2001). Although we used a small sample in Experiment 2, large sample sizes are not typically needed to achieve adequate statistical power in functional imaging studies. The sample size was adequate to demonstrate group effects but insufficient to examine correlates of individual differences in the BOLD response to self-morph presentations. In addition, the boxcar experimental design may not be as sensitive to subtle changes in BOLD response as event-related fMRI is (D'Esposito, Zarahn, & Aguirre, 1999). However, to account for the possibility of low activation levels and habituation, we used six blocks that have been shown in the Drexel Functional Neuroimaging Laboratory to be a reliable number of epochs to produce maximum activation associated with stimulus exposure, while limiting habituation effects. Despite these limitations, this study demonstrates a sex difference in hypothetical investment decision making to high-resolution color child morphs, as well as in the underlying neural activation associated with viewing self-resembling child morphs. These findings suggest that selection produced a neurocognitive response module in males that is specific to detecting, responding, and favoring facial resemblance in children.

Most Link Bipedalism to Changing Environment

Cooling & Drying Breakup of Tropical Forests Expansion of Grasslands and Savannah Dispersed Resources

early hominin mating systems

Equality for the sexes in human evolution? Early hominid sexual dimorphism and implications for mating systems and social behavior Clark Spencer Larsen* Department of Anthropology, 244 Lord Hall, 124 West 17th Avenue, Ohio State University, Columbus, OH 43210-1364 Since the publication of Charles Darwin's The Descent of Man and Selection in Relation to Sex in 1871 (1), there has been a vigor- ous debate about the meaning of sexual dimorphism for a range of physical at- tributes in numerous animal species, including primates and humans, extinct and extant. Key points of discussion are how to interpret size dimorphism in past humans and human-like ancestors and what inferences can be drawn about the evolution of human mating systems and social organization. In this issue of PNAS, Reno et al. (2) report on their investigation of sexual dimorphism in the three-million-year-old Australopithe- cus afarensis, an important and well known hominid, ancestral to the genus Homo (3). Insight into dimorphism in this taxon has important implications for social behavior and organization in later and present-day humans. Body mass dimorphism varies dramat- ically among primate species, both present and past. For most anthropoids, males are bigger than females (4-8). Humans today display relatively limited sexual dimorphism (15%), whereas some of the other hominoids (gorillas and orangutans) are highly dimorphic (50%) (5, 9). Body mass is easily de- termined in living species. For past non- human primates and human ancestors, mostly represented by fragmentary fossil remains, body mass is far less accessible. Recently, the femur head (the ball of bone at the top of the femur that fits into the hip joint) has been invoked as a source for estimating body mass in early hominids, Homo, and its evolutionary predecessor, Australopithecus (10, 11). Comparisons of body mass in fossil hominids reveal that general levels of dimorphism have likely remained more or less the same for most of the evolu- tion of Homo, or most of the last two million years to the present (9). In hominids predating Homo, namely the multiple species of Australopithecus, the consensus among paleoanthropologists that has emerged over the last two de- cades is that pre-Homo species are char- acterized by high levels of sexual dimor- phism (4, 5, 12-15). Close scrutiny of the fossil record, however, suggests that this consensus is built on a data set re- Fig. 1. plete with limitations, especially in re- gard to reconstructing size dimorphism in Australopithecus. First, the sample used to estimate di- morphism is very small (fewer than six individuals for A. afarensis). Second, estimates of dimorphism are based on the assumption that sex identification in fragmentary fossil remains used to de- rive these estimates is accurate. Indeed, the secondary sex characteristics exhib- ited in the bony pelvis, by far the most reliable of the indicators for humans (16, 17), are largely missing. Thus, inves- tigators are left with size of skeletal ele- ments alone (males have big bones and females have small bones), a poor proxy for pelvic sex identification. Third, accu- racy in determination of sexual dimor- phism is predicated on correct taxo- nomic identification. This is especially problematic given that level of sexual dimorphism shows substantial intertaxa variation. Fourth, levels of dimorphism can shift over broad expanses of time (potentially hundreds of thousands of years) or even relatively narrow ex- panses of time involving hundreds or tens of years (18). Finally, sexual dimor- phism levels across broad geographic areas and ecological variation therein may be exaggerated in comparison with contemporary members of a species liv- ing in the same place (9). Reno et al. (2) draw on advances made in statistical modeling to circum- vent these limitations of the early homi- nid fossil record. They apply a new and robust method of simulating dimorphism to an assemblage of A. afarensis repre- senting the remains of individuals who likely died simultaneously in a single catastrophic event some 3.2 million years ago at site A.L. 333, Hadar, Ethio- pia. Using the 40% complete skeleton (''Lucy'') from site A.L. 288 as a mor- phometric template (she has a relatively well preser ved femur head and other long bones; Fig. 1), they calculated fem- oral head diameters from measurements for the postcranial elements from A.L. 333 and other A. afarensis remains. In contrast to the consensus, their analysis See companion article on page 9404. *E-mail: [email protected]. A.L. 288 -1 (''Lucy''), the most complete skeleton of Australopithecus afarensis, serves as a morphometric template for determination of sexual dimorphism in other members of the taxon. Original fossil skeleton housed at the Na- tional Museum of Ethiopia. Photograph copy- right 1985, David L. Brill. www.pnas.orgcgidoi10.1073pnas.1633678100 PNAS August 5, 2003 vol. 100 no. 16 9103-9104 revealed only slight to moderate levels of sexual dimorphism, more like Homo and chimpanzees than gorillas. How to interpret this interesting re- sult? By using models derived from study of living nonhuman primates and humans, analysis of sexual dimorphism provides a window onto behavior in earlier hominids and added perspective on the evolution of human social be- havior and mating systems. Monomor- phic species of living primates (those taxa exhibiting low levels of sexual di- morphism) tend to express minimal male-male competition, whereas di- morphic species tend to express rela- tively high levels of competition (19- 21). Baboon males, for example, are highly intolerant of one another and aggressively compete for access to fe- male mates; simply, success in fights results in greater access to females. For this and other dimorphic primates, sexual selection is only one explanation for high levels of dimorphism, and may not be the best one (22). However, new analyses indicate associations be- tween dimorphism and competition levels (6, 7): where dimorphism is high, male-male competition is common- place; conversely, where dimorphism is low, competition among males is less frequent. Although chimpanzee adult males ex- press aggressive behavior toward one another, they tolerate each other, live in multimale kin groups, and are collabo- rative. Chimpanzee males defend terri- Darwin, C. (1871) The Descent of Man and Selec- tion in Relation to Sex (D. Appleton, New York); reprinted (1898) (D. Appleton, New York). Reno, P. L., Meindl, R. S., McCollum, M. A. & Lovejoy, C. O. (2003) Proc. Natl. Acad. Sci. USA 100, 9404-9409. White, T. D. (2002) in The Primate Fossil Record, ed. Hartwig, W. C. (Cambridge Univ. Press, Cam- bridge, U.K.), pp. 407-417. Frayer, D. W. & Wolpoff, M. H. (1985) Annu. Rev. Anthropol. 14, 429-473. Plavcan, J. M. (2001) Yearbook Phys. Anthropol. 44, 25-53. Plavcan, J. M. & van Schaik, C. P. (1997) J. Hum. Evol. 32, 345-374. Plavcan, J. M. & van Schaik, C. P. (1997) Am. J. Phys. Anthropol. 103, 37-68. McHenry, H. M. & Coffing, K. (2000) Annu. Rev. Anthropol. 29, 125-146. Ruff, C. (2002) Annu. Rev. Anthropol. 31, 211-232. Ruff, C. B. (2000) Am. J. Phys. Anthropol. 113, 507-517. Ruff, C. B., Trinkaus, E. & Holliday, T. W. (1997) Nature 387, 173-176. tory and engage in cooperative, coali- tionary behavior (23-26). Perhaps, then, the social organiza- tion of A. afarensis might be best char- acterized as multimale, cooperating (generally noncompeting) kin groups. Based on these new reconstructions of relatively low skeletal dimorphism inA. afarensis, this would seem to be one possible conclusion. However, A. afa- rensis has lower canine dimorphism than chimpanzees (5-7, 12-14, 27), which suggests a different kind of so- cial organization for these early homin- ids altogether. The findings of Reno et al. (2) and interpretations based on a range of evidence suggest that A. afa- rensis had a monogamous and not a polygynous mating system with strong intermale competition as was implied from previous reconstructions of great body size dimorphism. However the data are interpreted, their findings do not contradict what would be expected in a monogamous mating system. In- deed, the relatively low amount of di- morphism is more consistent with pair bonding (and the behaviors associated with it), more so than with the higher levels of dimorphism in single- and multimale extant primate genera (28). We will never know what the social organization and mating systems were for early hominids; past behaviors do not preser ve. However, innovative doc- umentation of morphometric variation in the context of informed study of be- havior in living species provides essen- McHenry, H. M. (1992) Am. J. Phys. Anthropol. 87, 407- 431. McHenry, H. M. (1994) J. Hum. Evol. 27, 77- 87. McHenry, H. M. (1994) in Power, Sex, and Tradi- tion: The Archeology of Human Ancestry, ed. Shen- nan, S. & Steele, J. (Routledge, London), pp. 91-109. Richmond, B. G. & Jungers, W. L. (1995) J. Hum. Evol. 29, 229 -245. Krogman, W. M. & Iscan, M. Y. (1986) The Human Skeleton in Forensic Medicine (Thomas, Springfield, IL), 2nd Ed. Walker, A. & Ruff, C. B. (1993) in The Narioko- tome Homo erectus Skeleton, ed. Walker, A. & Leakey, R. (Harvard Univ. Press, Cambridge, MA), pp. 221-233. Ruff, C. B. & Larsen, C. S. (2001) in Bioarchae- ology of Spanish Florida: The Impact of Colonial- ism, ed. Larsen, C. S. (Univ. Press of Florida, Gainesville), pp. 113-145. Leigh, S. R. (1997) Am. J. Phys. Anthropol. 97, 339 -356. tial perspective on behavior in extinct species. In addition to charting new directions for future analysis, thesenew findings suggest that earlier be- havioral models based on supposedly highly dimorphic pre-Homo taxa are not the most appropriate, and that the Early hominids may have been more human-like in their basic social behavior. earlier consensus about body size di- morphism and its implications requires further discussion. Rather than imply- ing some form of unique behavior based on a combination of low sexual dimorphism in canine size and high sexual dimorphism in body size (5-7, 9), A. afarensis (and other early homin- ids) may have been more human-likein their basic social behavior. Thus, the roots of human behavior may go deep in time. The article by Reno et al. (2) and the discussion and debate pro- voked by it will move the field closer to deriving an increasingly informed understanding of sexual dimorphism and social behavior in the remote hu- man past, laying the groundwork for understanding the evolution of human social organization.

Evolutionary Psychology

Evolutionary psychology (EP) attempts to explain mental and psychological traits such as memory, perception, or language as adaptations, that is, as the functional products of natural selection or sexual selection. Stinky T-shirts?

evolution of primate social systems

Evolution of Primate Social Systems Peter M. Kappeler1,3 and Carel P. van Schaik2 Received May 10, 2001; accepted August 8, 2001 We review evolutionary processes and mechanisms that gave rise to the diver- sity of primate social systems. We define social organization, social structure and mating system as distinct components of a social system. For each com- ponent, we summarize levels and patterns of variation among primates and discuss evolutionary determinants of this variation. We conclude that conclu- sive explanations for a solitary life and pair-living are still lacking. We then focus on interactions among the 3 components in order to identify main targets of selection and potential constraints for social evolution. Social organization and mating system are more closely linked to each other than either one is to social structure. Further, we conclude that it is important to seek a priori measures for the effects of presumed selective factors and that the genetic con- tribution to social systems is still poorly examined. Finally, we examine the role of primate socio-ecology in current evolutionary biology and conclude that primates are not prominently represented because the main questions asked in behavioral ecology are often irrelevant for primate behavior. For the future, we see a rapprochement of these areas as the role of disease and life-history theory are integrated more fully into primate socio-ecology. KEYWORDS: socialorganization;matingsystem;socialstructure;socialrelationships;sexual selection; life history; socio-ecology. 1 ̈Abteilung Verhaltensforschung/Okologie, Deutsches Primatenzentrum, Kellnerweg 4, 37077 Go ̈ttingen,Germany.2Department of Biological Anthropology and Anatomy, Duke University, Durham, North Carolina 27708.3To whom correspondence should be addressed; e-mail: [email protected]. 707 0164-0291/02/0800-0707/0 ⃝C 2002 Plenum Publishing Corporation 708 Kappeler and van Schaik INTRODUCTION The stunning diversity of primate social systems has been described and analyzed in reviews by Crook and Gartlan (1966), Eisenberg et al. (1972), Clutton-Brock (1974), Clutton-Brock and Harvey (1977), van Schaik and van Hooff (1983), Terborgh and Janson (1986), Wrangham (1987), Dunbar (1988), Janson (2000), and Strier (2000a). Aspects of this diversity include spacing, grouping and mating patterns, as well as variability in patterns and quality of social relationships. Moreover, diversity in social systems is not only evident among species but also exists within species (Sterck, 1999) and even within populations (Goldizen, 1987a; Richard, 1978), though in com- parison with other vertebrates (Lott, 1991) documentation is relatively poor. We do not attempt another review of this diversity, but instead systematically examine the evolutionary forces that have generated and shaped it. A prerequisite for this endeavor is clarification of how social systems or any of their component parts evolve. After all, definitions and characteristics of social systems focus on traits of groups and not on individuals, the latter of which are the targets of natural and sexual selection (Goss-Custard et al., 1972; cf. Rowell, 1993). We therefore need a theoretical framework that relates fitness-relevant behavior of individuals, such as foraging, predator avoidance, mating and parental care, to the defining characters of a social system. This link is provided by the socio-ecological model (Crook, 1970; Emlen and Oring, 1977; Terborgh and Janson, 1986), which recognizes that social systems represent emergent properties of individual behavioral interactions and strategies (Hinde, 1976). The underlying behavior of individuals towards conspecifics, in turn, is probably largely shaped by ecological factors, such as the distribution of risks and resources in the environment and their interac- tions (Elgar, 1986; Emlen, 1994; Mangel, 1990; Terborgh and Janson, 1986; van Schaik, 1983, 1989; Vehrencamp, 1983; Wrangham, 1980). However, the social organization and demographic conditions created by individual be- haviors also impose constraints on the behavioral options of these same individuals, leading to complex feedback loops (Janson, 1986; van Schaik, 1996). Males and females are treated separately in the model because their fit- ness is limited by different factors ultimately related to sexual differences in potential reproductive rates and the resulting intersexual conflict (Clutton- Brock and Parker, 1992, 1995). Accordingly, the model assumes that the dis- tribution of females is primarily determined by the distribution of risks and resources in the environment, whereas males distribute themselves primar- ily in response to the temporal and spatial distribution of receptive females. The resulting demographic categories broadly determine the kinds of social Evolution of Primate Social Systems 709 relationships possible, as well as the operational sex ratio, which is an im- portant predictor of sex roles, reproductive strategies and the intensity of matingcompetition(EmlenandOring,1977;KvarnemoandAhnesjo ̈,1996; Sterck et al., 1997). DEFINITIONS AND TARGETS OF SELECTION Despite considerable theoretical progress and increasing knowledge about the diversity of primate societies over the last two decades, confusion about targets of selection, combined with a lack of clear definitions, continue to hamper analyses of primate social systems. Accordingly, we begin by defining 3 different elements of social systems that had earlier been identified as fundamental and distinct components (Rowell, 1979, 1993; Struhsaker, 1969). The distinctions among social organization, social structure and mating system are important because they are not necessarily congruent. Each of them alone is insufficient to characterize adequately a social system com- prehensively because they are shaped by different selection pressures, even though they are not entirely independent of one another. For example, speaking of polygynous groups confuses two distinct aspects of the social system when one only refers to group composition (Clutton-Brock, 1989a; Davies, 1991; Mu ̈ller and Thalmann, 2000). Moreover, a particular social organization does not necessarily reflect the logically corresponding mating system (Cords, 1988, 2000). Because the 3 elements have been defined and named differently, sometimes by the same authors, and their labels have been used interchange- ably in the past (cf. Barton et al., 1996; Charles-Dominique, 1978; Eisenberg etal.,1972;HillandLee,1998;Janson,1988;Mu ̈llerandThalmann,2000; Rowell, 1988; Struhsaker, 1969; van Schaik and van Hooff, 1983; Whitehead, 1997; Wrangham, 1987), we propose consistent use of the following definitions. The society (= social unit; social system) is the set of conspecific ani- mals that interact regularly and more so with each other than with members of other such societies (Struhsaker, 1969). It is usually easy to recognize a society, but with animals that form parties of variable composition (fission- fusion) or are largely solitary, recognition is a challenging task. Primatolo- gists have recognized the following aspects of societies: social organization, mating system, and social structure. Social organization describes the size, sexual composition and spa- tiotemporal cohesion of a society. Traditionally, 5 fundamental social orga- nizations are recognizied: an animal either lives alone, or it is associated with 710 Kappeler and van Schaik either one or several members of the same and or opposite sex (Kappeler, 1999a). Obviously, this categorization only coincides with the definition of social organization if spatial dispersion of individuals corresponds with so- cial boundaries. Specifically, whether an animal is solitary or gregarious is just one aspect of social organization, and not a distinguishing feature, but for practical reasons we retain this categorization. The mating system has a social and a genetic component. The social mating system describes one subset of social interactions, namely mating couples, whereas the reproductive consequences of mating interactions, i.e., the genetic mating system, can only be determined via genetic studies. Social structure refers to the pattern of social interactions and the re- sulting relationships among the members of a society. These definitions and the resulting categories focus on adult males and females and do not consider infants and juveniles, presumably because of the historical focus on mating systems (Crook and Gartlan, 1966), even though they obviously also contribute importantly to a social structure. Together, these three components define the essentials of a society of a taxon. Social Organization The most basic characterization of primate societies has traditionally been based on social organization alone (Clutton-Brock and Harvey, 1977; Crook and Gartlan, 1966; Eisenberg et al., 1972). At this level, 3 fundamental types of social organization can be distinguished: neighborhood and solitary, pair-living, and group-living species. Solitary Primates: Neighborhood Systems The key aspect for the definition of solitary species is the fact that "the general activity, and particularly, the movements of different individ- uals about their habitat are not synchronised" (Charles-Dominique, 1978, p. 139). Thus, solitary individuals typically forage alone (Bearder, 1987), in contrast to gregarious animals, in which ≥2 individuals synchronize their ac- tivity in space and time (Boinski and Garber, 2000). Recognition of solitary foragers as a category of social organization is therefore justified if they also form a distinct social unit. The term solitary does not imply that they do not maintain social relationships or lack complex social networks; these traits simply describe other elements of the social system (Bearder, 1987, 1999; Charles-Dominique, 1978; Sterling and Richard, 1995; but see e.g., Radespiel, 2000; Thalmann, 2001). Similarly, it is neither necessary nor useful Evolution of Primate Social Systems 711 to introduce mating system terminology to characterize social organiza- tion, e.g., dispersed polygyny, and it also seems unnecessary to confine the term solitary to taxa without social relations outside the breeding season (cf.Mu ̈llerandThalmann,2000). The reason for the confusion about solitary species is that we know so little about their social life. With the exception of the orangutan, all solitary primate species are nocturnal. As a result, the pioneering field studies may have focused on what could be measured under these difficult circumstances, rather than on what should be measured to characterize the social organiza- tion of solitary foragers. Fortunately, a methodological consensus that could establish a standard for the study of solitary primates is emerging (Sterling et al., 2000). Until the 1990s, our conceptions about the diversity of their social systems were largely based on reports from a few pioneering studies, summarized by Bearder (1987). Since then, the number of long-term field studies of solitary primates, particularly lemurs, has grown almost exponen- tially, prompting several recent reviews (Bearder, 1999; Kappeler, 1997a; Mu ̈llerandThalmann,2000;Sterlingetal.,2000;vanSchaik,1999).Given the emerging picture of increasing diversity, we are still far from a complete overview, especially because the social systems of all Asian and most African strepsirhines remain poorly studied (Bearder, 1999) and those of 10 newly described species of cheirogaleids (Groves, 2000; Rasoloarison et al., 2000) are among those still completely unknown. Existing studies revealed that the major axes organizing the social diver- sity of solitary primates are (1) extent of range overlap with members of the same and opposite sex and (2) occurrence of sleeping groups and their sexual composition(Bearder,1987;Kappeler,1997a;Mu ̈llerandThalmann,2000). Home range overlap is highly variable, both within and between sexes, and malerangestendtoexceedthoseoffemales(Mu ̈llerandThalmann,2000), but in a few cases, the ranges of one male and one female coincide. Only when we know more about the nature and size of these still largely obscure basic units from exploratory studies in relatively small and randomly cho- sen study plots can systematic investigations of the more interesting level of social organization at the population level be designed. Hence, perhaps the currently safest generalization is that many solitary foragers live in neigh- borhoods (Richard, 1985), in which individuals do not live in distinct so- cial units but are decreasingly familiar with others that overlap increasingly less with their own home ranges. Unless populations are structured into dis- crete nuclei (Martin, 1972), recognition of discrete societies and hence social organization may be impossible. A third axis with potentially important relationships with social struc- ture and mating systems remains largely unexplored. It concerns the genetic structure of societies, which has sometimes been inferred from behavioral 712 Kappeler and van Schaik data. In particular, the existence of matriarchies, defined as clusters of closely-related females, has been postulated or assumed without genetic data for several species by Waser and Jones (1983), Clark (1985), Bearder (1987),andMu ̈llerandThalmann(2000),especiallywhenseveralfemales form sleeping groups. However, genetic analyses of mitochondrial DNA variability demonstrated the existence of spatially discrete clusters of fe- males sharing a particular mitotype in the absence of sleeping aggregations (Mirza coquereli: Kappeler, 1997b; Kappeler et al., 2002) and of matriarchies that include many more females than the largest sleeping groups (Microcebus murinus: Schmid, 1998; Wimmer et al., 2002). Future studies of solitary pri- mates that combine behavioral and genetic data to explore diversity along this axis and the consequences of a particular social organization for disper- sal, mating and social behavior should add exciting contributions to primate socioecology. Finally, few studies have addressed the fundamental question about the adaptive bases of a solitary life style during the active period, so that only the most obvious and general determinants and correlates have been discussed (Bearder, 1987; Charles-Dominique, 1978, 1995; Kappeler, 1997a). First, nocturnal activity is strongly correlated with a solitary life style, perhaps because group cohesion and coordination would be difficult and might attract predators, but some nocturnal primates are not solitary and the only solitary anthropoid is diurnal (Gursky, 2000a; Wright, 1989). Second, Clutton-Brock and Harvey (1977) suggested that small body size is an important deter- minant of solitary activity, but it is not strictly associated with a particular social organization. Third, there is no dietary specialization that correctly predicts solitary activity. van Schaik and van Hooff (1983), Bearder (1987), and Gursky (2000b) suggested that heavy reliance on animal prey and gum, which typically occur in small patches that cannot be shared, favors solitary foraging, but there are too many exceptions in both directions to make this a general determinant. Fourth, van Schaik and van Hooff (1983) suggested that some primates may be solitary because there is no predation pressure on them, and a solitary social organization is the optimal response to other selection pressures, such as intense feeding competition. But recent studies demonstrated that some solitary primates are subject to intense predation risk (Goodman et al., 1993; Rasoloarison et al., 1995). Thus, the optimal re- sponse to predation risk may differ among species, with solitariness being the optimal response for nocturnal species. A solitary way of life is also not a response by relatively immobile animals to the high risk of predation (Terborgh and Janson, 1986; van Schaik and van Hooff, 1983) because most solitary primates are not cryptic (Bearder, 1987). Thus, in contrast to group- living primates (Janson, 2000), there is no satisfactory explanation for why some primates do not live in groups. Evolution of Primate Social Systems 713 Primates in Pairs Permanent association of one adult male and female defines pair- living species (Fuentes, 2000; van Schaik and Dunbar, 1990). Thus, there is not only coincidence of male and female ranges but also synchronized spa- tial association between the pair. Recognition of this category of social organization is difficult in cases in which intersexual spatial association is relatively loose. For example, in several nocturnal strepsirhines, such as Galagoides zanzibaricus (Harcourt and Nash, 1986), Cheirogaleus medius (Fietz,1999;Mu ̈ller,1998),Phanerfurcifer(Charles-DominiqueandPetter, 1980; Schu ̈lke, 2003), Lepilemur edwardsi (Rasoloharijaona et al., 2000; Thalmann, 2001) and Lepilemur ruficaudatus (Hilgartner et al., unpublished data), male and female ranges overlap closely and are defended by both sexes against neighbors, but members of the pairs are neither consistently associated during the period of activity or necessarily always sleeping to- gether in the same shelter. These species are interesting because they may represent examples of independent transitions from a solitary to a pair-living social organization (Kappeler, 1999c), as also evidenced by the occasional occurrence of social units consisting of one male and two females. Phylogenic models revealed that pair-living has evolved independently in all major primate radiations, but it is nevertheless the rarest type of social organization (Kappeler, 1999c; Kappeler and Heymann, 1996; van Schaik and Dunbar, 1990). The rarity of pair-living among primates (and other mammals) is not surprising, given our current understanding of sexual dif- ferences in potential reproductive rates and the resulting conflict of interest between the sexes. Males should seek to maximize their reproductive suc- cess by mating with as many females as possible, and the defining characters of mammalian reproduction provide them with the opportunity to do so. Hence, we must ask why some males opt to live permanently with a sin- gle female, especially in primates with relatively long interbirth intervals. This has been the traditional formulation of this question, focusing on the potential consequences of pair-living for the mating system. Because these levels should be separated we should also seek explana- tions that account for pair-living as a form of social organization (Fuentes, 1999; van Schaik and Dunbar, 1990; Wright, 1986). Indeed, observations of extrapair copulations (EPC) in pair-living primates (Palombit, 1994; Reichard, 1995) and genetic evidence for the success of EPCs (Fietz et al., 2000) indicate that mating activities of neither sex are principally constrained by living in pairs. Why some primate males permanently bond with a single female has been explained with two arguments that distinguish between two fundamental types of monogamy in primates: systems with obligate male care and ones without paternal care (Clutton-Brock, 1989a). 714 Kappeler and van Schaik First, in species with male care, notably in the pair-living twinning cal- litrichids, Goldizen (1987a), Wright (1990), and Dunbar (1995a) invoked obligate paternal care via infant carrying as a crucial contribution of males to their own fitness. Interestingly, however, males that could not have fa- thered offspring also carry young (van Schaik and Paul, 1996). Pereira et al. (1987), Morland (1990), and Fietz (1999) suggested paternal care in the form of infant guarding as a mechanism to promote pair-living in some lemurs, but the empirical evidence for such an effect is weak and, in one case (Fietz et al., 2000), a large proportion of males was cuckolded, making evolutionary scenarios based on paternal certainty unlikely. Second, in species in which males provide no direct paternal care, no- tably most gibbons, there are two hypotheses for the permanent association between a male and a female. The first invokes limits of male monopolization potential (Emlen and Oring, 1977). Accordingly, males cannot successfully defend sexual access to >1 female because of the spatial distribution of females and their behavior (Brotherton et al., 1997; Komers, 1996). Accord- ing to the second hypothesis, permanent association as a pair is an evolved strategy that reduces the risk of infanticide by strange males (Palombit, 1999, 2000; van Schaik and Dunbar, 1990). Species with long interbirth intervals, long lactation in relation to gestation and infant carrying by the mother have a high risk of infanticide, which males can help to reduce via protecting their offspring (van Schaik, 2000a,b). The risk of infanticide and its gen- eral implications for primate social evolution have been recently examined (van Schaik et al., 1999; Nunn and van Schaik, 2000; van Schaik, 2000a,b). This hypothesis can be generalized to include all forms of permanent male- female association, i.e., include single males with >1 female (van Schaik and Kappeler, 1997). Explanations for why some males opt to associate with only one female therefore require an additional factor. The recognition that pair-living does not equate with a monogamous mating system raises questions about potential ecological causes for pair- living. Some obvious factors, such as range size, patch size and patch distri- bution have already been examined in some species (Robinson et al., 1987; van Schaik and Dunbar, 1990; Wright, 1986), but, given the taxonomic and ecological diversity of pair-living primates, different explanations may apply to different taxa so this possibility should be examined more comprehen- sively. An additional approach could be to turn the question around and ask why the well-known benefits of group-living are apparently not realized in pair-living species. Increased benefits or reduced costs of predator detec- tion and evasion or both factors, as well as resource defense, obviously have their largest per capita effects at small group sizes (Dunbar, 1988), which are clearly >2. There may be specific ecological reasons favoring extremely small group size, such as facilitated coordinated movements at night and Evolution of Primate Social Systems 715 improved hiding during the day in nocturnal pair-living species, but they are difficult to test post hoc. In brief, the adaptive bases for pair-living are far from being completely understood. Group-Living Primates The vast majority of primates lives in bisexual groups with ≥3 adults, which sets them apart from other mammals in which permanently bisexual groups are much less common (van Schaik and Kappeler, 1997). Moreover, an unusally large and detailed number of field studies of Primates versus other Mammalia, revealed an additional level of diversity with respect to the size, sex ratio and temporal stability of primate groups. The ratio of adult males and females provided the traditional criterion for further distinction among group-living species (Kappeler, 2000a). Accordingly, polyandrous, polygynous and multimale, multifemale groups have been distinguished (Clutton-Brock and Harvey, 1977; Crook and Gartlan, 1966; Eisenberg et al., 1972). However, these labels also confuse social organization and mating systems. Variation in group cohesion has been used for yet finer distinctions among groups with multiple males and females. Their members can be either relatively cohesive or exhibit one of two types of temporal variation in group cohesion. In fission-fusion groups, subgroups of varying size and composi- tion form temporarily (Nishida and Hiraiwa-Hasegawa, 1987; Rigamonti, 1993; Strier, 1992), whereas in multilevel societies, small social units, usually containing one male and several females, are also organized into higher-level sets (Stammbach, 1987). Variation in the number of adult males is the most striking feature of pri- mate group composition with far-reaching consequences for many aspects of male and female behavior (Hamilton and Bulger, 1992; Preuschoft and Paul, 2000; van Hooff, 2000). The most basic dichotomy is that between single- and multimale groups. During the early years of primate socio-ecology, re- searchers sought ecological explanations for this dichotomy, such as habitat type or predation risk (Kappeler, 2000a). Today, this dichotomy is no longer recognized as an invariant species-specific traits, but instead as a flexible response to variation among groups in ecological and demographic factors (Robbins, 1995; Steenbeek et al., 2000; Strier, 2000b; Struhsaker, 2000; Watts, 2000). Later analyses identified the number of females in a group and their degree of reproductive synchrony as important determinants of male mo- nopolization potential, which is an important determinant of the outcome of the dichotomy (Altmann, 1990; Andelman, 1986; Ridley, 1986). In 3 recent comparative studies, researchers re-examined the proposed key determinants of the number of group males. A comparison among 716 Kappeler and van Schaik arboreal folivores demonstrated that the presence of monkey-eating eagles tends to increase the number of males in howlers and colobus on average from one to two, whereas ecologically similar langurs (and some colobus), which are not exposed to such predators, tend to live in single-male groups (van Schaik and Ho ̈rstermann, 1994). Furthermore, Mitani et al. (1996a) found that the the qualitative difference between single- and multimale groups is positively associated with the number of females, and not with temporal distribution of their receptive periods, but Nunn (1999a) demon- strated that temporal overlap of female receptive periods predicts the num- ber of males after controlling for the number of females. A final assessment may therefore have to await the availability of additional data, in particular on group-living lemurs, which are characterized by the presence of several males, despite small female group size, and extremely short breeding seasons (Kappeler, 2000b). Variation in group size is another striking aspect of diversity in the social organization of group-living primates. Across species, primate group sizes vary by 2 orders of magnitude (Kappeler and Heymann, 1996), and varia- tion within some species is several-fold (Dunbar, 1988). Researchers have identified 4 main correlates, and hence possible determinants, of variation in primate group size. First, increasing foraging and travel costs set the upper limit of group size via increased intragroup feeding competition (Janson and Goldsmith, 1995; van Schaik, 1983). Second, living in larger groups is favored by decreasing predation risk and by intense intergroup feeding competition (van Schaik, 1983; van Schaik and van Hooff, 1983; Wrangham, 1980). Third, infanticide risk provides a selective force that reduces group size under cer- tain conditions (Crockett and Janson, 2000; Steenbeek and van Schaik, 2001). Finally, neocortical size may constrain group size because it determines the ability to process complex information about social relationships (Dunbar, 1992, 1995b, 1998). Birth and death rates are the most important proximate mechanisms regulating group size. Contrarily, immigration and emigration are constra- ined in effectiveness to modify group size by an inherent sexual bias in most species. Female philopatry is common among primates (Pusey and Packer, 1987), but female and bisexual group transfer also occur in various taxa (Goldizen and Terborgh, 1989; Isbell and van Vuren, 1996; Moore, 1984; Pope, 2000a; Sterck, 1998; Strier and Ziegler, 2000). Sexual biases in res- idency not only constrain the ability to adjust female group size but also have important consequences for the genetic structure of primate groups (Altmann et al., 1997; de Ruiter and Geffen, 1998; Gerloff et al., 1999; Melnick and Hoelzer, 1992; Pope, 1992, 1998), breeding patterns (Clutton- Brock, 1989b; Melnick et al., 1984; Moore and Ali, 1984; Packer, 1985), breed- ing success (Pope, 2000b) and sex-specific life history strategies (Alberts and Evolution of Primate Social Systems 717 Altmann, 1995a,b; Borries, 2000). Residency patterns also facilitate the for- mation of coalitions of same-sexed members of the resident sex through kin selection and structure other details of the social behavior of both sexes (Chapais et al., 1997; Moore, 1992; Silk, 2002). In sum, the social organization of group-living primates has been more intensely studied than that of other primates, so that variation in group size and composition, as well as the underlying mechanisms, are comparatively well-understood. Mating Systems The majority of mammalian mating systems that were characterized by Clutton-Brock (1989a) is found among primates. They include monogamy, in which males and females typically mate with only one member of the opposite sex and have roughly equal variances in reproductive success (MacKinnon and MacKinnon, 1980; Robinson et al., 1987; Rutberg, 1983; Sommer and Reichard, 2000); polyandry, in which one female mates with several males and each males mates only with her (Garber, 1997; Goldizen, 1987b; Tardif and Garber, 1994); and various forms of polygyny and polyg- ynandry. Spatial polygyny occurs among solitary species in which agonis- tically powerful males defend mating access to several females (Charles- Dominique, 1977). Scramble competition polygyny occurs when males roam widely in search of receptive females, which they desert soon after mating in search of additional mates, so that both sexes typically mate with sev- eral partners (Kappeler, 1997b). Moreover, several forms of female-defense polygyny exist among primates. In some species, coalitions of males de- fend a territory that contains several females with which most males mate (Gerloff et al., 1999; Watts, 1998). In the remaining species, groups of fe- males are defended directly by one or several males. Harem-polygyny occurs when a single male defends exclusive mating access to a group of females (Stammbach, 1987), whereas promiscuous mating occurs when several males defend groups of females (Altmann et al., 1997; Bercovitch, 1989; Brockman and Whitten, 1996; Paul et al., 1993; Sauther, 1991), often with pronounced skew in male mating success as a function of dominance rank (Cowlishaw and Dunbar, 1991; Johnstone et al., 1999; Paul, 1997). There is no case of lekking or resource-defense polygyny among primates. The ability of individual males to defend successfully exclusive mat- ing access to females is probably the major determinant of primate mating systems. When females are clumped in space and/or when their receptive periods are not synchronized, male monopolization is much greater than when females are solitary and/or when they have synchronized receptive 718 Kappeler and van Schaik periods (Dunbar, 2000; Emlen and Oring, 1977). However, females are not mere passive objects of male mating strategies. Their reproductive interests can be used to modify their spatial distribution and especially the degree of estrous synchrony (e.g., Zinner and Deschner, 2000; Zinner et al., 1994). Additional mechanisms, such as female choice, sexual swellings and modifi- cations of the fertile period of the cycle assure that sexual coercion of females is minimized and that females largely determine the identity and number of their mates (Nunn, 1999b; van Noordwijk and van Schaik, 2000; van Schaik et al., 1999; Zinner and Deschner, 2000). Many reproductive strategies of female primates appear to be aimed at confusing paternal certainty, thereby reducing the risk of infanticide (van Noordwijk and van Schaik, 2000). Fi- nally, because of the physiological constraints of gestation and lactation, most primates (and other mammals) are characterized by virtually exclusive female parental care, so that, compared to birds, polygynous mating systems are overwhelmingly common. With few puzzling exceptions among lemurs (Kappeler, 1993a), mor- phological correlates of different mating systems correspond well to theo- retical predictions and patterns that occur among other mammals (Clutton- Brock et al., 1977; Harcourt, 1997; Harvey et al., 1978; Kay et al., 1988; Mitani et al., 1996b; Plavcan and van Schaik, 1992). Accordingly, males of species in which females typically mate with several males have, on average, larger testes in relation to body mass than males of monandrously mating females. Similary, variation in male mate monopolization potential and intensity of direct male mating competition covaries positively with the degree of sexual dimorphism in body and canine size. Characterizations of primate mating systems continue to be importantly influenced by characteristics of social organization because observations of mating patterns, and, most importantly, genetic analyses of reproductive success are lacking for the majority of species. There is, of course, a rough correspondance between the social organization and mating system of a society or taxon but several facts argue against simple equation of them. Foremost, promiscuity outside the social unit appears to be widespread, as evidenced by copulations outside the pair-bond, the group or community (Cords, 2000; Digby, 1999; Launhardt et al., 2001; Richard, 1985; Sommer and Reichard, 2000). Intensified genetic sampling in the future may reveal that such discrepancies between social and mating partners are even more widespread than currently known. Moreover, temporal instability of social organization during the mating season contributes to a mismatch with mating systems (Borries, 2000; Cords, 2000). Finally, there is variation within species, and sometimes even within populations, in social organization and mating patterns, or both (Heymann, 2000; Steenbeek et al., 2000; Struhsaker, 2000). Therefore, ultimately analyses of individual female mating decisions may be Evolution of Primate Social Systems 719 more illuminating than those of system characteristics (Pereira and Weiss, 1989). Social Structure Relationships among individuals reflect behavioral strategies that have been selected because they maximize inclusive fitness (van Schaik, 1989). Differences in the patterning and nature of social interactions give rise to particular social relationships between pairs of individuals, and consistent features of dyadic relationships can be used to characterize social structure (Hinde, 1976). Variation among relationships is brought about by differences in the nature, frequency and intensity of affinitive, affiliative and agonis- tic interactions (de Waal, 1986, 1989). Sex is a major organizing principle in the analysis of social structure, with both ecological and social factors as ultimate determinants of the observed variation in social relationships. However, the relative importance attributed to these determinants or their components is still unresolved (Dunbar, 1988; Sterck et al., 1997; van Schaik, 1996; Wrangham, 1987). Interfemale Relationships Socioecological models of female social relationships are focussed on causes and consequences of feeding competition (Sterck et al., 1997; Koenig, 2002), primarily in Old World monkeys and apes. The nature of feeding competition is shaped by the distribution of resources and can occur within and between groups. When food patches are clumped, monopolizable, and of intermediate size relative to group size, contest competition among fe- males is expected, whereas scramble competition predominates over other types of patches (van Schaik, 1989). Whether related females form coali- tions to defend access to preferred food sources against other such coalitions (Wrangham, 1980) or intergroup feeding competition predomi- nates because groups form in response to predation risk (van Schaik and van Hooff, 1983), each female in a group-living species will experience a mix of contest and scramble competition within and between groups (Koenig, 2002). The consequences of a given competitive regime for social relation- ships with other females can be summarized by 4 interrelated variables: philopatry, nepotism, tolerance and despotism, according to which there are 4 main categories of female relationships (Sterck et al., 1997). Among them, 2 categories are most common. In resident-nepotistic groups, females 720 Kappeler and van Schaik are philopatric and establish stable, linear and nepotistic hierarchies with despotic dominance relations. Contrarily, in dispersal-egalitarian groups, fe- males regularly transfer between groups, forming neither stable linear hier- archies nor coalitions. The basic difference between them is the strength of intragroup contest competition, as confirmed by empirical field stud- ies (Barton et al., 1996; Koenig et al., 1998; Mitchell et al., 1991). There are other forms of female social structure. In resident-nepotistic-tolerant groups female philopatry is combined with decided relationships within a stable hierarchy—regular coalitions—and pronounced tolerance by domi- nants. The functional interpretation of this kind of social structure is that subordinates have enough leverage over dominants to reduce the extent to which they actually reap the benefits of dominance, though the course of the leverage is unidentified. It need not be ecological because, social fac- tors, especially mating conflict, may also affect female-female social rela- tionships (Preuschoft and van Schaik, 2001). An important goal for future research is to integrate the still poorly characterized competitive regimes of callitrichids and lemurs into the existing framework (Kappeler, 1999b; Pereira and Kappeler, 1997). Intermale Relationships Relationships among male primates in intra- and intergroup competi- tion are also highly variable across species (Kappeler, 1999a). They are pri- marily shaped by female distribution and the resulting nature of intrasexual selection. Because the contested fitness-limiting resource—fertilizations— cannot be shared, male relationships are typically characterized by com- petition, intolerance and clear dominance relations (Bercovitch, 1991; Cowlishaw and Dunbar, 1991; van Hooff, 2000; van Hooff and van Schaik, 1994). Therefore, agonistic interactions are common among males, while af- filiative behavior and alliances are mainly observed when large numbers of males find themselves within a single group (Noe ̈ and Sluijter, 1990; Plavcan and van Schaik, 1992; van Hooff and van Schaik, 1992). Contrarily, in a few species with male philopatry, grooming bonds and coalitionary behavior among males are well-developed (Strier, 1994; van Hooff, 2000; van Hooff and van Schaik, 1992, 1994). Intersexual Relationships Male-female relationships are ultimately shaped by sexual selection and sexual conflict (Smuts, 1987; Smuts and Smuts, 1993; van Schaik, 1996). Intersexual relations among primates are highly variable, both within and among species. They are influenced, among other things, by the duration of Evolution of Primate Social Systems 721 male residence in a group, the respective rank in the same-sex dominance hierarchy, the degree of paternal certainty, the risk of infanticide and the degree of sexual dimorphism (Hamilton and Bulger, 1992; Kappeler, 1999c; Manson, 1994; Paul et al., 2000; Perry, 1997; Sicotte, 1994; Takahata, 1982; van Noordwijk and van Schaik, 1988; Watts, 1992; Weingrill, 2000). If males and females form affiliative bonds, females often provide most of the grooming, whereas males provide agonistic protection for the female and her infants and additional vigilance against predators (de Ruiter, 1986; Hemelrijk and Ek, 1991; Kappeler, 1993c; Koenig, 1998; Noe ̈ and Hammerstein, 1994a,b; Palombit et al., 1997; Smuts, 1985). Except for most lemurs (Kappeler, 1993b; Pereira and McGlynn, 1997; Richard, 1987), adult primate males dominate females in dyadic interactions. Interrelationships The 3 elements of a social system are discrete, interrelated entities. The links may be of 3 possible kinds: (i) causal—one aspect imposes a di- rect constraining effect on the values of the other; for instance, when so- cial organization determines which animals are available for relationships; (ii) evolutionary—adaptive values of one aspect have led over time to a particular range of values in another aspect, such as when the benefits of the formation of particular kinds of alliances has led to a particular social organization; and (iii) correlational—the values of the 2 elements reflect the effects of another factor, for instance sex-biased dispersal, without strongly affecting each other. However, it is often impossible to say which relation- ships actually hold in a particular example, especially since they are usually not exclusive. Moreover, there is a remarkable degree of freedom for each element to vary independently, which is the raison d'eˆtre for distinguishing the separate elements in the first place. The most pronounced relationships exist between social organization on the one hand, and mating system and social structure, on the other. First, in many cases the spatial dispersion of individuals, and, in the case of fe- males, their number and reproductive synchrony, impose a direct constraint on the mating system. Thus, polyandrous mating by females is less likely and indeed less common in single-male groups (pairs or female groups) than in multimale groups, though many exceptions occur to this rule. However, it is difficult or impossible to model the polarity of the evolutionary relationships between the 2 aspects. For instance, did multimale, multifemale groups form for ecological reasons after which promiscuous mating developed by default, or did selection for polyandrous mating by females, e.g., to reduce the risk of infanticide (Hrdy, 1979), lead to the formation of multimale groups? Thus, 722 Kappeler and van Schaik at the proximate level, social organization determines the mating system in some cases, rather than vice versa, but the relationship is neither strict nor predictable, and the evolutionary relationships are difficult to disentangle. Second, social organization and social structure are trivially linked at the proximate level in that certain demographic conditions must exist for the occurrence of particular classes of relationships; for example, female social relationships cannot develop in pair-living species. However, demography may also have more direct effects on social options (Strier, 2000b). Whether social organization is a given and social relationships develop within these constraints (van Schaik, 1996) or social organization emerged as a byproduct of the establishment of particular social relationships (Wrangham, 1980) is unresolved. The diversity of female social relationships among group-living species (Sterck et al., 1997) argues against the notion that a particular kind of social organization predisposes for a particular social structure. For exam- ple, demographically similar groups of gelada and hamadryas baboons are characterized by very different social structures (Stammbach, 1987). Thus, the two aspects of a social system are relatively free to vary independently of one another, both synchronically and evolutionarily. The link between social structure and mating system is clearly more variable. However, social relationships can affect mating relations when male-female friendships exist (Smuts, 1985), which could lead to the forma- tion of multiple pairs within larger groups (Pereira and McGlynn, 1997; van Schaik and Kappeler, 1993). Similarly, mating history predicts male-female and male-infant relations in some species (Palombit et al., 1997; Paul et al., 2000). Male-female association may have evolved as a result of paternity defense (Palombit, 1999, 2000) or to reduce risk of infanticide (van Schaik and Kappeler, 1997). Unlike most birds, in most primates obligate paternal care is absent, so that parental care has presumably played only a minor role in shaping primate social structure evolutionarily. Dispersal pattern may determine social organization and structure, which would be correlated due to this effect. Thus, female philopatry pro- duces clusters of females, which also happen to be relatives, facilitating coop- erative behavior among them (Moore, 1992; Waser and Jones, 1983). How- ever, although female philopatry may historically have been the pathway to gregariousness, group-living is adaptive in itself and no longer necessarily linked with female philopatry. Likewise, given that female dispersal occurs where the need for female cooperation is reduced, the benefits of cooper- ation with relatives may have selected for female philopatry (van Schaik, 1989; Wrangham, 1980), which suggests that the dispersal pattern is the evolutionary product of social processes rather than its evolutionary cause. In other cases, too, it is plausible to regard dispersal as the evolutionary prod- uct of social processes. Eviction of particular targets or group fission occur Evolution of Primate Social Systems 723 in groups that exceed a particular size (Henzi et al., 1997; Vick and Pereira, 1989). In brief, whether dispersal is the evolutionary cause or consequence of particular social processes is unresolved. DETERMINANTS AND MECHANISMS For each of the three components of social systems, we have already summarized important ultimate determinants of inter- and intra-specific variability, as well as the mechanisms that implement them, in previous sec- tions. An important remaining question is whether all important factors have already been identified. Might we have identified and quantified factors that are easy to measure at the expense of more important ones that are more difficult to measure? a priori Measures Recent progress in understanding the difference between predation rate and risk provide an illustrative example of the problem. The fact that predation risk exerts a major influence on primate behavior is uncontested (Stanford, 2002). Previous researchers who examined specific predictions about the effects of predation on primate social systems used predation rates to operationalize this independent variable (Anderson, 1986; Boinski and Chapman, 1995; Cheney and Wrangham, 1987). However, now there is consensus that predation rate is not suitable for such analyses because it ignores the effects of various countermeasures already in place to reduce the risk (Hill and Dunbar, 1998a; Hill and Lee, 1998; Janson, 1998). Realistic estimates of the underlying predation risk are much more difficult to obtain, so definitive tests are difficult (Janson, 1998). The same logic may explain why some authors refuse to acknowledge the importance of the threat of male infanticide in the evolution of (primate) social systems (Bartlett et al., 1993; Dagg, 1999; Sussman et al., 1995). For them, rates of male infanticide appear too low to qualify as a major force in social evolution, but the rates may be low precisely because effective countermeasures are already in place to minimize the risk: the white knight rejoinder (van Schaik, 2000a). The same problem may apply to studies of feeding competition and its social consequences (Koenig, 2002). Ideally we need independent a priori measures of risk, such as the ratio of lactation and gestation length for risk of infanticide (van Schaik, 2000b; van Schaik and Kappeler, 1997), and longevity (Janson, 2003) or predator assemblages (Nunn and van Schaik, 2001) for intrinsic predation risk. Such logically and independently derived variables can provide much more powerful explana- tions than post hoc arguments, which are often difficult to falsify. 724 Kappeler and van Schaik Genetic Aspects Social behavior also has a genetic basis, which has received little at- tention in recent socioecological analyses. Studies of hybrids and controlled experiments help to illuminate the relative importance of genetic predispo- sitions for a particular social organization, social structure or mating system. First, some sister taxa with radically different social systems produce hybrids under natural conditions. For example, hybrids between hamadryas and anu- bis baboons express certain taxon-typical traits of their respective parental specific social system, such as male herding of females (Colmenares, 1992; Nagel, 1973; Sugawara, 1988), indicating a genetic basis for the behavior. Studies of the behavior of other hybrids could help to define the nature and extent of genetic predispositions for other aspects of social behavior. Second, several experimental approaches have been successfully used to determine social reaction norms of individual species. Demographic manip- ulations, such as captive housing under different social conditions, can reveal limits of plasticity in social organization. A classic example, which is well un- derstood in retrospect (Kummer, 1968), concerns the formation of a large multimale multifemale group of hamadryas baboons at London Zoo, fol- lowed by an immediate violent response of adult males (Zuckerman, 1932). A similar approach was used by Mendoza and Manson (1986), Harrison and Tardif (1989), and French et al. (1995) to test the strength of pair bonds in captive callitrichids by presenting strange adults of both sexes to them. Understanding the limits of social plasticity also has obvious practi- cal applications for captive housing and propagation (Berger, 1996; Tilson, 1986). The flexibility of some aspects of the social structure can also be studied experimentally. For example, de Waal and Johanowicz (1993) showed that juvenile rhesus macaques housed for several months with juvenile stump- tailed macaques subsequently showed elevated levels of reconciliation, but not of affiliative and grooming behavior. Thus, some components of the species-specific social structure (de Waal and Luttrell, 1989) are apparently more flexible than others. Additional cross-fostering studies would clearly be of great interest in this context. Using a different experimental approach, Gore (1993) manipulated food distribution of captive rhesus macaques and hamadryas baboons with the goal of changing the quality of female social relationships. The food ma- nipulations had no detectable effect, which demonstrates that the degree of female bonding is inert to short-term fluctuations in the selective force that is thought to have generated it. Gore's experiment serves as an important reminder that immediate and evolutionary levels of responses should not be confused. Evolution of Primate Social Systems 725 Finally, the study of intraspecific variation in aspects of social systems can help to identify the breadth and limits of social reaction norms. Variation among subspecies or populations in social organization, social structure or mating systems can be used to separate relative contributions of genetic and ecological determinants, but this information is available for only a few primate species in sufficient detail (Barton et al., 1996; Castles et al., 1996). Therefore, there are results from few carefully controlled studies of the exact nature and causes of intraspecific social variation (Barton, 2000; Srivastava and Dunbar, 1996; Sterck, 1999). A logically separate issue is the degree of phylogenetic inertia in as- pects of societies. Clearly, closely related taxa tend to share more similarities in their social systems than more distantly related ones do. This holds not only for sister species but also at higher taxonomic levels. For example, Old World cercopithecoids tend to live in multimale multifemale groups with female philopatry and very similar social structures (di Fiore and Rendall, 1994) and female dominance is limited to lemuriformes (Richard, 1987). Obviously, such phylogenic inertia does not necessarily imply the presence of genetic constraints on social evolution: related taxa tend to have similar life histories and ecologies, and the similarities may therefore be adaptive (Pagel and Harvey, 1991). In support of the adaptive interpretation, the same traits that are phylogenetically conserved in one taxon can be highly variable within another equivalent higher-order taxon. For example, homi- noids have extremely diverse social organizations (McGrew et al., 1996), and dispersal among ceboids is not limited to one particular sex (Pope, 2000a). What Have We Learned So Far? The first four decades of primate socioecology have greatly advanced our understanding of variation in primate social systems and the underly- ing determinants and mechanisms. Following the first phase, characterized mainly by descriptive inventories, and a second phase dominated by corre- lational approaches, the ongoing hypothetico-deductive approach towards analyzing primate social systems has been by far the most productive and successful. Thus, there is no need to refrain from formulating and testing spe- cific predictions, simply because not all the basic natural history data from all species are available yet (Rodman, 1993; Sussman, 1999). After all, hy- potheses are erected to be tested with empirical data; if the relevant data do not exist yet, hypotheses provide focal points for field researchers to collect the relevant data (e.g., van Schaik and Kappeler, 1996; Wright, 1999). The result is always at least more pertinent information about the natural history of focal species. 726 Kappeler and van Schaik More specifically, much current research and many insights have been dominated by a cercopithecentric perspective, portraying semiterrestrial cer- copithecines as typical primates (Strier, 1994). Yet, much of the interesting variation in social systems is among New World primates and prosimians. They exhibit interesting similarities in social organization to one another, and idiosyncracies in social structure and mating systems that differ from the more familiar pattern of the few better-known Old World primates (Garber, 1997; Jolly, 1966, 1998; Kappeler, 1997a, 2000b; Kinzey and Cunningham, 1994; Pereira, 1995; Pereira and Kappeler, 1997; Strier, 1996; Wright, 1997, 1999). Accordingly, intensified study of platyrrhine and prosimian species and integration of the resulting insights into existing models is an important task for the next generation of field primatologists. PRIMATES AND EVOLUTIONARY BIOLOGY One goal of this special issue, and this paper in particular, is to link pri- matological research with the important developments in behavioral ecology and evolutionary biology. Hauser (1993) and Harcourt (1998), noted that primate and nonprimate socioecologists have largely ignored each other, so it is not surprising that most behavioral ecologists and evolutionary biolo- gists are familiar with only a few, mostly out-dated, primate studies (Krebs and Davies, 1992). Apart from the common mutual taxonomic ignorance of both primatologists and non-primatologists and the fact that researchers studying other taxa, especially other mammals, have fewer taxonomically specialized journals at their disposal, there may also be biological reasons for this isolation. Mainstream behavioral ecology has traditionally been dominated by students of birds and invertebrates. Many new questions and trends have originated via studies on them (Lack, 1947; Wilson, 1975). During the last decade, topics dealing with various aspects of sexual selection have domi- nated the main journals of behavioral ecology. Research addressing causes and mechanisms of noncontest aspects of sexual selection, such as female choice, reproductive skew and sperm competition, has reached new levels of detail and theoretical sophistication (Gibson and Langen, 1996; Johnstone, 1995, 2000). Questions dealing with the relevant costs of signalling, such as the genetic quality of mates and their phenotypic indicators, particularly fluc- tuating asymmetry, have received more attention from behavioral ecologists than any other topic (Møller, 1997; Watson and Thornhill, 1994; Widemo and Saether, 1999). Interestingly, relevant work on primates, with the possible exception of humans, is conspicuously absent from these areas of research (Manning, 1995; Manning and Chamberlain, 1993; Thornhill et al., 1995). Conservation biology and studies of the evolution of cooperation are the Evolution of Primate Social Systems 727 only examples of modern integrative disciplines in which information about social systems and other aspects of behavioral ecology are relevant and in which primates are not underrepresented (Jernvall and Wright, 1998; Noe ̈ and Hammerstein, 1995). Why, then, is primatology not in the mainstream of evolutionary bi- ology? We suggest the following 3 main reasons. First, the main topics of sexual selection dealing with aspects of female choice are by-and-large not applicable to primates. Striking ornaments of males with a potential function in mate choice are widespread among primates (Dixson, 1998), but the few existing studies have only suggested a function in intrasexual competition (Gerald, 2001; Setchell and Dixson, 2001a,b; Wickings and Dixson, 1992). This does not imply that primate females do not choose their mates (Manson, 1995; Paul, 2002; Small, 1989). We propose instead that mate selection based on arbitrary phenotypic traits is usually not as important to female primates because they know their potential mates from regular interactions as a result of and sometimes years of association. Second, the required experimental approach for conclusive studies of problems in sexual selection or other central topics in behavioral ecology is rarely, if ever, possible with primates. Both ethical and practical problems, having to do with slow reproductive rates or small sample size, explain this limitation (Janson, 2000). In addition, it is possible that the important ul- timate factors under investigation are hidden behind counterstrategies, so that theoretically major perturbations of the system are required to elicit meaningful responses experimentally. For example, adaptations against in- fanticide may be so effective that it rarely occurs under normal conditions, or female primates may opt to mate polyandrously to confuse paternity, rather than to chose a particular best male. Finally, the natural history of an organism typically guides the kinds of questions that can be asked meaningfully. Because big differences in natural histories are typically related to qualitative differences in key life-history traits, e.g., internal vs. external fertilization, ovipary vs. vivipary (Clutton- Brock, 1991), asking all of the same questions about all organisms does not make sense. Questions dealing with paternal care and quality are much more pertinent to understand female birds than female mammals. Simi- larly, primates may not always be the best or even an appropriate taxon to address some specific hypotheses in behavioral ecology (Harcourt, 1998). Thus, aspects of primate natural and life-histories may underly and explain our limitations. The Future of Primate Socioecology Now that primatologists have at least begun to study the rarest primates in the most remote places, often for several years, the amount of detailed 728 Kappeler and van Schaik information about primate natural history, social systems, and ecology is un- paralleled among vertebrate orders, with possible exceptions of some avian taxa. However, this does not signal the near end of socioecological research. Instead, the situation provides a unique opportunity to launch research of- fensives that explore several new frontiers which could have impact on future developments in behavioral ecology (Janson, 2000). First, the availability of large comparative data bases provide opportu- nities to test predictions of the socioecological model via comparative tests to demonstrate correlated evolution between pairs of traits, such as predation risk and group size, or group size and home range size (Hill and Lee, 1998; Nunn and Barton, 2000). The development of ever more refined comparative tests and more detailed phylogenies in recent years alone (Gittleman and Luh, 1992; Harvey and Pagel, 1991; Nunn and Barton, 2001; Purvis, 1995) provided an underused tool box for primatologists. Using a broad compar- ative approach holds the potential to identify and to measure the actual selective forces, and not merely their outcomes (Nunn and Barton, 2001). Second, the link between behavior and life-history is a key pillar in the evolution of social systems. Exploring their many causal and integrative interactions could provide a new frontier for primate socioecology. The slow life-histories of primates may have consequences for all 3 components of their social systems (Kappeler et al., 2003). The slow rates of growth and reproduction, in particular, make more conservative behavioral strategies, including a safer life in permanent groups, adaptive (Janson and van Schaik, 1993). Residence in stable groups over many years results in increased social complexity and perhaps cognitive abilities. Moreover, slow life-histories may also affect mating strategies and systems because they affect the operational sexratioandtheriskofinfanticide(KvarnemoandAhnesjo ̈,1996;Mitani et al., 1996a; Willson and Pianka, 1963). Third, extended periods of juvenility are defining features of slow life- histories, but the various developmental trajectories of juvenile primates and their integration into adult social systems remain virtually unstudied (Altmann, 1998; Pereira and Fairbanks, 1993; Pereira and Leigh, 2002). Exploration of these and many other potential links has only begun and provides great potential for future research. Finally, potential main determinants of some aspects of social systems, as well as some important consequences, remain virtually unexplored. There are few studies on diseases of natural primate populations, and their effects on behavior (Freeland, 1976; Davies et al., 1991; Heymann, 1999; Nunn et al., 2000). Similarly, the consequences of variation in social systems for conserva- tion and captive management need to be understood in much greater detail to implement effective strategies quickly (Dobson and Lyles, 1989; Ganzhorn, 1987; Gursky, 1998). Fortunately, both primates and primatologists already Evolution of Primate Social Systems 729 have prominent roles in international conservation programs (Jernvall and Wright, 1998; Myers et al., 2000).

Sexual Coercion

For example, males may try to force mating (coercion) that females do not want

Environmental Influence on Traits

Girls and boys are on average taller today then they were in the past due to increased nutrition

STDs and Circumcision

HIV, HPV, & Herpes-All reduced

Chromosomes and Replication

Homologous Chromosomes - each cell contains 2 similar copies of each chromosome (except for the sex chromosome). One of these copies is from mom and the other is from dad. Together, these are called a homologous pair and each alone is called a homologue. Chromatid - exact replicas of each chromosome

facial attractiveness

IF sxyhgsyx Theoriesofsexualselectionsuggestthat`goodgenes'ina potential partner may be advertised through some kind of indicator mechanism. Preferences for partners with traits that reliably indicate heritable bene¢ts in o¡spring should befavoured.Such`goodgenes'theorieshavebecome more popular with empirical demonstrations of herit- ability of ¢tness e.g. Petrie 1994) and theoretical models demonstrating that indicator mechanisms are possible under conditions characterized by rapidly changing selec- tion pressures e.g. host^parasite coevolution) Andersson 1994; Kirkpatrick 1996). One possible indicator that has received widespread attention is bilateral symmetry, which is thought to be an indicator of developmental stability. Fluctuating asymme- tries asymmetry in traits that are on average symmetric in a population) Van Valen 1962) result from an organ- ism's failure to cope with various inclement environ- mental e.g. malnutrition and parasitization) and genetic e.g.mutation)factorsMÖller1997).Better`quality'indi- viduals may resist environmental hazards better than poorer quality individuals and some of this quality may be heritable MÖller & Thornhill 1997). As such, a prefer- ence for symmetry in mate choice may have been favoured by selection. Males with low levels of £uctuating asymmetry have more mating success across multiple species and taxa MÖller & Thornhill 1998). Researchers have examined symmetry with respect to human behaviour and physical attractiveness see Gangestad & Simpson 2000) for a review). For example, men with low levels of body £uctuating asymmetry report more sexual partners and are involved in more extra-pair sexual encounters than men with greater levels of asymmetry Gangestad & Thornhill 1994; Thornhill & Gangestad 1994). Given the central role of the face in human social life and the importance of the face in * physical attractiveness judgements of potential partners, it is unsurprising that several studies have addressed the roleoffacialsymmetryinattractiveness.Studiesofasym- metry in natural faces e.g. Grammer & Thornhill 1994; Mealey et al. 1999) and digitally manipulated stimuli e.g. Perrett et al. 1999; Rhodes et al. 1998) indicate that facial symmetryispositivelycorrelatedwithattractivenessin both male and female faces. Recently, Scheib et al. 1999) presented data suggesting that symmetry is not an important cue to attractiveness in male faces, although it is a correlate. Raters preferred faces that originally possessed high levels of symmetry when presented with only left or right half faces faces split down a vertical midline bisecting the nose and mouth). As these faces had no or at most minor cues to symmetry, Scheib et al. 1999) concluded that a correlate of symmetry that is visible in half faces must drive attrac- tiveness judgements. They reported that facial `mascu- linity' as quanti¢ed by an index of two measured facial proportions, i.e. cheekbone prominence and relative lowerfacesize)correlateswithbothattractivenessand symmetry and is visible in half faces. Hence, masculinity seems to be a likely cue to attractiveness in male faces. Correlations between trait size and symmetry have been reported in some avian species and ¢t well with indicator models of sexual selection e.g. MÖller & Hoglund 1991; see Balmford et al. 1993) for another interpretation). Masculine facial traits large jaws and prominent brows) in males are thought to be testosterone dependent and, therefore, may represent an honest immunocompetence handicap Folstad & Karter 1992). As immunocompe- tence should be linked to developmental stability, a corre- lation between both symmetry and masculinity is predicted. However, the role of facial masculinity in attractiveness judgements is disputed. Cunningham et al. 1990) and Grammer & Thornhill 1994) used facial measurements and found a female preference for large jaws in males. Masculine features, such as a large jaw and a prominent brow ridge, are also reliably associated with ratings of Author and address for correspondence: Department of Psychology, University of Stirling, Stirling FK9 4LA, UK [email protected]). Proc. R. Soc. Lond. B 2001) 268, 1^7Received 18 December 2000 Accepted 9 April 2001 00pb1081.1 8 2001 The Royal Society doi 10.1098/rspb.2001.1703 Symmetry, sexual dimorphism in facial proportions and male facial attractiveness I. S. Penton-Voak*, B. C. Jones, A. C. Little, S. Baker, B. Tiddeman, D. M. Burt and D. I. Perrett School of Psychology, University of St Andrews, St Andrews, Fife KY16 9JU, UK Facial symmetry has been proposed as a marker of developmental stability that may be important in human mate choice. Several studies have demonstrated positive relationships between facial symmetry and attractiveness. It was recently proposed that symmetry is not a primary cue to facial attractiveness, as symmetrical faces remain attractive even when presented as half faces with no cues to symmetry). Facial sexual dimorphisms `masculinity') have been suggested as a possible cue that may covary with symmetry in men following data on trait size/symmetry relationships in other species. Here, we use real and computer graphic male faces in order to demonstrate that i) symmetric faces are more attractive, but not reliably more masculine than less symmetric faces and ii) that symmetric faces possess character- istics that are attractive independent of symmetry, but that these characteristics remain at present unde¢ned. Keywords: £uctuating asymmetry; facial attractiveness; facial sexual dimorphism 00pb1081.2 I. S. Penton-Voak and others Facial symmetry and attractiveness Author: Running head `Is Facial symmetry and attractive- ness' OK? Author: Perrett et al. 1994 not cited in refer- ences. Author: Burt & Perrett 1995. Change of date to1997 correct ? As cited in the refer- ences. Author: Re- dominance in photographic, identikit and composite stimuli by male and female raters McArthur & Apatow 1983^1984; McArthur & Berry 1987; Berry & Brownlow 1989; Berry & Wero 1993; Perrett et al. 1998). Facial domi- nance appears to correlate with status in some human hierarchies Mueller & Mazur 1997) and facial domi- nance in adolescent males is associated with an earlier age at ¢rst copulation Mazur et al. 1994). Nonetheless, the relationship between facial dominance and attractive- ness is unclearösome studies ¢nd a positive relationship Keating 1985) while others ¢nd the opposite McArthur & Apatow 1983^1984; Berry & McArthur 1985; Perrett et al. 1998). Other studies propose that a mixture of masculine and feminine traits are found attractive Cunningham et al. 1990) or that preferences for mascu- linity or femininity vary across the menstrual cycle as a function of the probability of conception Penton-Voak et al. 1999; Penton-Voak & Perrett 2000) The aim of the current study was to investigate the relationship between symmetry, masculinity and attract- iveness in male faces further. Study 1 investigated the rela- tionship between facial symmetry and attractiveness in a sample of male faces. Study 2 aimed to quantify sex di¡erences in the facial proportions visible in frontal facial photographs by measuring characteristics from a sample of male and female faces. This allowed the relationship between masculine features, symmetry and attractiveness to be studied and, secondarily, provided a composite measure of facial masculinity. Finally, in study 3 we investigated whether the corre- lates of symmetry in faces can be extracted using an alter- native, computer graphics technique. Composite images can be generated from samples of facial photographs using warping and blending techniques. Such composite images tend to retain properties that the sample faces have in common, e.g. sex Perrett et al. 1994, 1998) or age Burt & Perrett 1997), but lose the identities of individuals who make up the composite. This technique can be usefully applied to studies of facial symmetry, as there are no reported directional asymmetries in human faces at rest. Hence, a facial composite made from low-symmetry individuals should have a very low level of asymmetry, as should a composite constructed from individuals with high facial symmetry. Correlates of symmetry such as large jaws or other markers of masculinity should still di¡er between the two composite images. If, as Scheib et al. 1999) suggested, these cues are important to attractiveness judgements, raters should ¢nd the`high-symmetry'compositebothmoremasculineand moreattractivethanthe`low-symmetry'composite.As these techniques generate full facial stimuli rather than Figure 1. Points used in the calculation of facial metric measurements. Filled circles, points used in symmetry calculations; open circles, additional points used in measure- ments of scaled sexual dimorphism. Lines specify the dimensions used in masculinity calculations see table 1) occluding hair from the face as much as was possible. Each image was then normalized on the interpupillary distance. Some additional biographical information was collected height, weight and age). i) Attractiveness judgementsThe facial images were presented in a random order on computer and rated for attractiveness by 21 subjects 11 female) on a seven-point Likert scale 1 very unattractive and 7 very attractive). The subjects showed high consistency in their attrac- tiveness judgements Cronbach's 0.84). ii) Symmetry assessmentTwo techniques were used for assessing the symmetry of the individual male faces. A facial-metric technique estimated hori- zontalandverticalasymmetryfromx^ycoordinatesofseven bilateralpointsusingtechniquesdescribedinScheibetal.1999) and elsewhere see ¢gure 1). Thesecond`perceptual'measureofsymmetryinvolved creating two chimaeric stimuli from each original face Mealey et al. 1999). Each individual image was rotated so that the centre of the pupils lay on the same y-coordinate and then the face was split vertically along a line bisecting the distance between the pupils, thereby creating left and right half faces. These half faces weremirrored,creatingdoubleleftL^L)anddoublerightR^R) chimaeric images ¢gure 2). L^L and R^R image pairs were presented to ten subjects ¢ve female) who rated the faces for `similarity'onaseven-pointLikertscale1verydissimilarand 7 very similar). High ratings of similarity indicate small arrange-halffaces,thetaskmayhavesomewhatmoreecological ment with sub- sections a) Study 1 for of text into Methods and Results validity than the test employed by Scheib et al. 1999). sections PF wiryh Sixty-six young mean age 21.3 years), Caucasian, adult male More in undergraduates were photographed under standard conditions keeping studies 1^3 OK? with with di¡use £ash lighting from two lateral £ashguns. Images usual format were captured on a digital camera at a resolution of 12001000 ofpapers. pixels. Sitters assumed a neutral expression and removed Proc. R. Soc. Lond. B 2001) 00pb1081.3 Figure 2. Double left L^L) and double right R^R) chimaeric stimuli from a low-symmetry face top) and a high-symmetry face bottom). Author: Running head `Is Facial symmetry and attractive- ness' OK? Facial symmetry and attractiveness I. S. Penton-Voak and others perceptual di¡erences between L^L and R^R chimaeric faces, which may re£ect high facial symmetry. The subjects showed very high consistency in this task Cronbach's 0.93). Measured asymmetry and rated perceptual symmetry were signi¢cantly correlated Spearman's rs U0.477, p S 0.001 and n66). b) Study 2It has been proposed that the size of sexually dimorphic features correlates with both facial symmetry and attractiveness. In order to test this hypothesis, facial-metric measures were takenfrom27featurepointsmarkedontofacialfeaturesonthe 66 male faces used in study 1 and a further 49 female faces taken under the same photographic conditions see ¢gure 1). The identi¢cation of these features has been found to be reliable in several earlier studies e.g. Grammar & Thornhill 1994; Scheib et al. 1999). The female and male faces were standardized to the same interpupillary distance 100 pixels) in order to allow comparison between male and female facial proportions. This standardization technique prevents measurement of absolute trait size all traits are scaled relative to the distance between the pupils), but does eliminate the possibility that small varia- tions in head distance from the camera will lead to erroneous measurements. The measurements taken are reported in table 1. c) Study 3An alternative method of assessing the face shape correlates of facial symmetry is to use digital averages composites) of multiple individual faces. As human faces have no reported directional asymmetry at rest in frontal photographs, composite images tend towards symmetry regardless of the £uctuating asymmetries in each individual image. In order to construct composites, 172 feature points are marked on facial landmarks on each face for details of the choice of these landmark points see Rowland & Perrett 1995)). The mean XY position of each delineated feature point is then calculated in order to generate shapeinformation.An`average'colourisgeneratedbyrendering colour information from each individual into this average shape and calculating mean RGB colour values across the face set for each pixel location. A high-symmetry composite face was constructed from the shape and colour of the 15 faces with the highest rated symmetry and a low-symmetry composite face was generated from the 15 faces with the lowest rated facial symmetry ¢gure 3). Independent t-tests showed that the 15 faces in each group did not di¡er in any of the ¢ve sexually dimorphic measures from study 2 all values t S 1.25 and all values of p R 0.22 with d.f. 28) The high- and low-symmetry groups did not di¡er signi¢cantly in any of the other physical traits measured height, weight and body mass index). Author: Please give RGB in full. Proc. R. Soc. Lond. B 2001) Author: Figure 2 caption Changes to caption OK? 00pb1081.4 I. S. Penton-Voak and others Facial symmetry and attractiveness Table 1. Proportional dimorphic measurements from a sample of 66 male faces and 49 female facesn mean s.e. td.f.113) p eye size D1^D2)/2)female 49 male 66 lower face/face height D8/D7)female 49 male 66 cheekbone prominence D3/D6)female 49 male 66 face width/lower face height D3/D8)female 49 male 66 mean eyebrow heightfemale 49 male 66 39.44 0.25 37.14 0.16 0.60 0.00 0.62 0.00 1.17 0.01 1.14 0.01 1.20 0.01 1.16 0.01 23.77 0.4320.54 0.41 öö 7.986 0.000 ö ö U4.456 0.000 ö ö 3.356 0.000 ö ö 2.864 0.005 ö ö 5.385 0.000 Figure 3. Composites made from 15 faces with low rated symmetry left) and from 15 faces with high rated symmetry right). These two composites were presented to 77 university student raters mean age ca. 22 years) who selected which was the `most attractive', `most masculine', `most physically ¢t'andwhichhadthe`bestgeneralmedicalhealth'using a forced-choice paradigm. Apart from 15 subjects who rated both composites for both attractiveness and mascu- linity, all other samples were independent. In order to assess whether subjects were using cues from the face shape or from the hairstyles of the compo- sites, a second set of composites were prepared that were cropped along a horizontal line bisecting the forehead. Eighty participants mean age ca. 22 years) made forced- choice judgements about these faces on the same dimen- sions as the full-face stimuli. QF iv a) Study 1Given the consistency of the attractiveness ratings, the mean score across raters was taken as the attractiveness for each of the 66 faces. Attractiveness was not correlated signi¢cantly with measured asymmetry rs U0.226, p 0.068 and n 66) or perceptual asymmetry rs 0.217, p 0.08 and n 66). However, female raters appeared more sensitive to symmetry in male faces than male raters. Whentheattractivenessofeachfacewascalculatedsepa- rately from male and female ratings, both measures of symmetry were related to attractiveness for female subjects perceptual symmetry, rs 0.285 and p 0.02 and measured asymmetry, rs U0.284, p 0.02 and n 66), but neither measure was related to attractiveness for male subjects perceptual symmetry, rs 0.125 and n.s. and measured asymmetry, rs U0.136, n.s. and n 66). b) Study 2In order to calculate which features are proportionally sexually dimorphic in frontal photographs t-tests were performed on the mean distances between points for each sex or the mean value of the ratio for each sex as appro- priate see ¢gure 1 and table 1). In order to assess whether the signi¢cantly sexually dimorphic characteristics found in this sample see table 2) eye size, lower face height/ face height, cheekbone prominence, face width/lower face height and mean eyebrow height) are related to either Proc. R. Soc. Lond. B 2001) Facial symmetry and attractiveness I. S. Penton-Voak and others 00pb1081.5 Table 2. Responses to forced-choice comparisons of high- and low-symmetry composites in cropped and full-face conditions The table shows the number of subjects in each condition that selected the high-symmetry composite as being most representative of the question asked. Figures in parentheses indicate the number of female subjects. Fifteen female subjects completed full-face judgements of both attractiveness and masculinity. All other samples are independent.) Author: Table 1 Please add an expla- nation for the data in parenth- eses D1, D2, etc.) in a table note. n full-face compositespicked high symmetry cropped compositesn picked high symmetry attractivemasculine 3020) athletic ¢tness 22 11) medical health 20 10) 20p0.00001) 2010) 18p0.0004) 19p0.10) 2010) 13p0.13) 19p0.0007) 2010) 16p0.007) 19p0.00007) 2010) 17p0.002) 20 15) symmetry or attractiveness in male correlations were performed on eachteristic and an index of standardized measures. None of the individual measured traits or ratios corre- lated signi¢cantly with rated symmetry all values of rs S 0.23) or measured asymmetry all values of r S 0.18). None of the ¢ve dimorphic measurements were related to female-rated attractiveness all values of rs S 0.21). Intercorrelation between the ¢ve traits was very low. Only the closely related measures of lower face height/ total face height and face width/lower face height were signi¢cantly correlated r U0.32, p 0.008 and n 66). An index of standardized dimorphic measures was calculated in order to provide an overall masculinity score Zlower face height/face height) UZface width/ lower face height) UZeye size) UZmean eyebrow height) UZcheekbone prominence)). Despite the lack of correlation between each of the individual scores, this composite index correlated signi¢cantly with male judge- ments of male facial attractiveness rs 0.262, p 0.034 and n 66). Female judgements of attractiveness also showed a trend in the samep 0.091 and n 66). However,between measured asymmetry or rated symmetry and this composite masculinity measure r U0.013 rs 0.008, respectively, both non-signi¢cant). and c) Study 3Despite the lack of cues to symmetry in the composite image, the high-symmetry composite was rated as more attractive, more physically ¢t and healthier than the low- symmetry composite in both the full-face and cropped changesAuthor: conditions. There were also trends indicating that the to high-symmetry composite was judged to be more masculine sentence OK? than the low-symmetry composite in both conditions table 2). As a w2-test indicated that judgements of masculinity were not in£uenced by cropping the stimuli w21 0.014 and p R 0.05), a further binomial analysis was performed with the data collapsed across the cropped and full-face stimuli conditions and this indicated that symmetric composites were more likely to be selected as the `more masculine'ofthepair32outof50subjects)p0.016). RF hsgsyx The series of studies reported above supports several previous ¢ndings and con£icts with others and raises more questions about the role of male facial attractiveness Proc. R. Soc. Lond. B 2001) faces, a series of individual charac- in signalling biological properties. Study 1 agrees with previous literature that has demonstrated preferences for symmetry in both natural and digitally manipulated facial images Rhodes et al. 1998; Perrett et al. 1999). Humans are sensitive to asymmetries in complex natu- rally occurring stimuli. This sensitivity is increased when the axis of symmetry is vertical Evans et al. 2000), as is the case in human faces. However, symmetry covaries with at least one other cue in men's faces that is still visible when cues to symmetry are removed, as demonstrated by the compo- sites generated for study 3. Our study using composites from high- and low-symmetry individuals is analogous to the half-face study reported by Scheib et al. 1999): the two very di¡erent techniques generate convergent ¢nd- ings. These results are also consistent with earlier research demonstrating that humans attend to the left side of the face more than the right side when making judgements about attractiveness, age, sex or expression Burt & Perrett 1997). However, the cue that covaries with symmetry is not simply related to sexual dimorphism facial masculinity) as calculated from measurements of proportional male^ female di¡erences in two-dimensional frontal photo- graphs. This ¢nding con£icts with Scheib et al.'s 1999) ¢nding of trait size/symmetry covariation in male faces. This discrepancy can perhaps be explained by the assumptions made by Scheib et al. 1999) and others) who proposed that cheekbone prominence is a `masculine' characteristic when, in this sample at least, it is greater in females than in males. As this measure contributes 50% of the data in Scheib et al.'s 1999) `masculinity index', the conclusions of this earlier paper should be treated with caution. Any estimates of symmetry or masculinity may be prone to measurement errors in studies of two-dimensional facial photographs. The choice of facial landmarks for both symmetry and dimorphism measurements is some- what arbitrary, leading to the possibility that some trait size/symmetry relationships exist but remain unmeasured e.g. inspection of ¢gure 3 suggests that the high- symmetrycompositedoesindeedhaveamoremasculine hairline than the low-symmetry image). Although marking of feature points has been shown to be reliable Grammer & Thornhill 1994; Scheib et al. 1999), barely perceptible lateral or vertical rotations of the head could lead to inaccurate measurements of symmetry and/or masculinity. Furthermore, both this study and the Schieb direction rs 0.209, there was no link Author: Is 2001 the correct date for Jones et al.? Author: Is 2001 the correct date for Hume & Montgo- merie ? et al. 1999) study used proportional measures of di- morphism rather than absolute trait size. Although proportional measures lose valuable information, they do have the advantage of controlling for absolute trait size when estimating asymmetry/trait size relationships. The current study has employed perceptual measures of both symmetry and masculinity in an attempt to provide alter- native estimates, but such ratings may not re£ect the biological properties of the subject cf. Meyer & Quong 1999; but see also Perrett & Penton-Voak 1999; Evans et al. 2000). Clearly, however, symmetry does not appear to be the only cue to facial attractiveness in human males. Eluci- dating the characteristics that covary with symmetry is proving problematic. If facial symmetry is an indicator of developmental stability then one possibility is that symmetric faces also posses other cues to good health, such as a clear, unblemished complexion. This intuitive hypothesis receives some support from the data in study 3: although both composite images have unblemished skin see Alley & Cunningham 1991) for the e¡ects of averaging on skin texture), the high-symmetry composite isstilljudgedasmoreattractive,healthierandmore physically ¢t than the low-symmetry composite. Whilst these results could be interpreted as a psychological halo e¡ect of attractiveness the association of desirable char- acteristics with attractive faces), judgements of good health could equally be interpreted as the driving attractiveness responses e.g. Jones et al. 2001a). A recent study has demonstrated that socio-economic status is the best predictor of male facial attractiveness, even when the e¡ects of symmetry are partialled out Hume & Montgomerie 2001). At present the nature of male attractiveness remains somewhat mysterious. Despite the nearly complete lack of intercorrelation between individual dimorphic features found in study 2 and the lack of a relationship between these measures and symmetry, an overall masculinity index was related to men's facial attractiveness as rated by men with trends in the same direction present for females. In addition, when the two samples that had seen both cropped and full-face versions were collapsed together the high-symmetry composite was rated as more masculine than the low- symmetry composite. Although the picture is far from clear, these ¢ndings suggest that masculinity and symmetry are somewhat independent and perhaps signal di¡erent characteristics in male faces. In this sample, men consider male faces high in masculinity attractive, but symmetry has a smaller relationship with attractiveness. Women show the opposite pattern: there is a positive rela- tionship between symmetry and attractiveness, but the composite masculinity measure does not signi¢cantly predict attractiveness. In the current study, with relatively small samples and non-parametric correlation co- e¤cients, it is di¤cult to comment on the reliability of the di¡erences between male and female raters in response to masculinity and femininity. Two other recent studies have investigated the role of sex in symmetry judgements. Little et al. 2001) demonstrated that women's preference for symmetry was stronger in male than in female faces. In a further study, Jones et al. 2001b) found a positive relationship between perceived health and symmetry in opposite-sex face stimuli for both men and women that Proc. R. Soc. Lond. B 2001) was weaker in same-sex judgements. In conjunction with the current study, these two ¢ndings provide a basis for speculating that preferences for symmetry may be a mate choice adaptation rather than a by-product of strategies used to process visual stimuli e.g. Enquist & Arak 1994). It is possible that facial masculinity is more important in intrasexual competition than sexual display, with very masculine males succeeding through dominance of other males rather than female choice. However, given the high correlations between male and female ratings of attrac- tiveness found in other studies and the relatively low sample size employed in the current study, further work is necessary in order to investigate this hypothesis. Further complexities in the study of male facial attrac- tiveness arise from recent ¢ndings of di¡erences in women's preferences across time and between individuals that may represent strategic pluralism in response to social or environmental contingencies Gangestad & Simpson 2000). Preferences for masculine traits in male faces have been demonstrated to vary across the menstrual cycle Frost 1994; Penton-Voak & Perrett 2000) in interaction with the speci¢c context of the attractivenessjudgement`short-'or`long-term'relation- ship) and life-history factors the presence or absence of a partner) Penton-Voak et al. 1999). Women prefer rela- tively masculine faces in the follicular phase of their menstrual cycle, particularly when they have a long-term partner and are judging attractiveness for a short-term relationship, a ¢nding that has implications for the role of extra-pair copulations in the evolution of human sexu- ality. Little et al. 2001) demonstrated that female self- rated attractiveness also in£uences preferences: women who judge themselves as more attractive prefer relatively masculine and symmetric faces, a ¢nding that is analo- gous to condition-dependent mate choice in other species e.g. Bakker et al. 1999). Perhaps consideration of the interactions between the properties of both the raters and multiple cues within the faces to be judged will be neces- sary in order to provide a fuller understanding of attrac- tiveness judgements in our own species.

Genetics-The Study of the Inheritance of Traits

Molecular Genetics: Basic level (DNA)Mendelian Genetics: Inheritance of various traitsPopulation Genetics: Variation within and between populations Phylogenetics: Discerning evolutionary relationships Behavioral Genetics: Influence of genetics on behavior

Adaptations

Physical, morphological Behavioral as well Includes psychology

sperm competition

Sperm Competition Risk and Sexual Coercion Predict Copulatory Duration in Humans Nicole Barbaro1, Michael N. Pham1, and Todd K. Shackelford1 Abstract Evolutionary Psychology 2015: 1-8a The Author(s) 2015Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1474704915618411 evp.sagepub.com A man whose romantic partner is sexually unfaithful is at risk of sperm competition and cuckoldry—unwitting investment in offspring to whom he is genetically unrelated. Men, therefore, may have evolved mechanisms to solve the adaptive problems of sperm competition and cuckoldry. The current research investigates another potential anti-cuckoldry tactic: reducing in-pair copulation (IPC) duration, thereby more quickly placing his sperm into competition. We hypothesize that IPC duration will be negatively correlated with female infidelity (Hypothesis 1). We further hypothesize that IPC duration will be negatively correlated with sexual coercion (Hypothesis 2). Results of Study 1 (men's reports, n 1⁄4 410) indicate that both men's perceptions of female infidelity and men's sexual coercion predict shorter IPC duration. Results of Study 2 (women's reports, n 1⁄4 455) did not provide statistical support for the study hypotheses. The current research provides an initial investigation of men's adjustment of copulatory duration and suggests that men reduce IPC duration and ejaculate more quickly at the couple's most recent copu- lation, in response to greater risk of sperm competition and in the context of sexual coercion. Keywords sperm competition, sexual coercion, infidelity, in-pair copulation, cuckoldry Date received: August 31, 2015; Accepted: October 28, 2015 Sperm competition occurs when the sperm of two or more males simultaneously occupy a female's reproductive tract and compete to fertilize the ova (Parker, 1970). Sperm competition has been demonstrated or inferred to exist in many species, including humans (Baker & Bellis, 1993a, 1993b; Birkhead & Moller, 1992). A recurrent context for sperm competition in humans is female infidelity or extra-pair copulation (Baker & Bellis, 1993b; Gallup et al., 2003; Shackelford et al., 2002; Smith, 1984). A man whose long-term partner pursues extra- pair copulations is at risk of sperm competition and subsequent cuckoldry—unwitting investment in offspring to whom he is genetically unrelated. Cuckoldry has likely been a recurrent adaptive problem for humans (Baker & Bellis, 1995; Buss & Shackelford, 1997; Voracek, Haubner, & Fisher, 2008). Because human males often invest in their putative offspring (Chrastil, Getz, Euler, & Starks, 2006; Trivers, 1972), the costs of cuckoldry can be substantial (Leivers & Simmons, 2014). Men, therefore, may have evolved anti-cuckoldry tactics to prevent or to participate in sperm competition to guard their paternity (Platek & Shackelford, 2006; Shackelford et al., 2002, 2005). Anti-cuckoldry tactics include mate guarding behaviors (Buss, 1988; Buss & Shackelford, 1997), frequent in-pair copulations (IPCs; Pham, Shackelford, Holden, Zeigler- Hill, Hummel, & Memering, 2014; Shackelford, Goetz, Guta, & Schmitt, 2006), sexually coercing IPCs (Goetz & Shackel- ford, 2006; reviewed in Goetz, Shackelford, & Camilleri, 2008), semen-displacing copulatory behaviors (Gallup et al., 2003; Goetz et al., 2005), ejaculate adjustments (Baker & Bel- lis, 1993a, 1995), partner-abuse during pregnancy (Burch & Gallup, 2004), and adjustments in parental investment (Platek, Burch, Panyavin, Wesserman, & Gallup, 2002). The current research investigates another potential anti- cuckoldry tactic: reducing IPC duration. Men who perceive greater sperm competition risk report increased urgency to copulate with their in-pair partner, which, in part, reflects an urgency to submit sperm into competition (Shackelford et al., 2002; Shackelford, Goetz, McKibbin, & Starratt, 2007; see also 1 Department of Psychology, Oakland University, Rochester, MI, USA Corresponding Author: Nicole Barbaro, 108 Pryale Hall, Department of Psychology, Oakland University, Rochester, MI 48309, USA.Email: [email protected] Creative Commons CC-BY-NC: This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 3.0 License (http://www.creativecommons.org/licenses/by-nc/3.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access page (https://us.sagepub.com/en-us/nam/open-access-at-sage). 2 Evolutionary Psychology Spiess, Geer, & O'Donohue, 1984). Because sexual behavior is caused proximately by sexual arousal, reducing ejaculatory latency and, therefore, copulatory duration, affords quicker entry into sperm competition. We hypothesize that, in humans, IPC duration will be negatively correlated with female infide- lity (Hypothesis 1). Women attempt to delay IPC following copulation with another man (Gallup, Burch, & Mitchell, 2006), suggesting that women may manipulate sperm competition in favor of an extra- pair partner. In contrast, men report greater sexual interest in their partner and greater urgency to copulate with their partner following suspicions of her infidelity (Pham & Shackelford, 2013; Shackelford et al., 2002). As a consequence of this sexual conflict, men may attempt to counter women's resistance to IPC by deploying tactics of sexual coercion to gain sexual access to their partner and quickly enter into sperm competition. Men at greater risk of sperm competition report more fre- quent use of sexual coercion (McKibbin, Starratt, Shackelford, & Goetz, 2011), and women who self-report infidelity also report that their partner is more sexually coercive (Goetz & Shackelford, 2006, 2009). Men at greater sperm competition risk, including men who perceive greater risk of partner infi- delity, report greater upset in response to a partner's resistance to IPC (Pham & Shackelford, 2013; Shackelford et al., 2002, 2007), and this upset is positively associated with men's self- reports and women's partner reports of sexual coercion (Shack- elford & Goetz, 2004). Research in other animals corroborates the hypothesis that sexual coercion may function as an anti- cuckoldry tactic (Barash, 1997; McKinney, Cheng, & Brug- gers, 1984). In many socially monogamous birds, for example, forced IPC reliably occurs immediately following female extra-pair copulation (Bailey, Seymour, & Stewart, 1978; Barash, 1997; Birkhead, Hunter, & Pellatt, 1989). Research with other animals also indicates that forced copu- lations are shorter in duration than nonforced copulations. For example, males of several species of waterfowl (family Anati- dae) often perform precopulatory displays prior to nonforced IPCs (McKinney, Derrickson, & Mineau, 1983). During forced IPCs, however, males do not perform precopulatory displays (McKinney et al., 1983). This suggests that males may be attempting to inseminate the female quickly by eliminating precopulatory displays, affording males the opportunity to more quickly enter their sperm into competition. Similarly, sneak copulations in guppies (Poecilia reticulata) are shorter in duration than courtship copulations (Pilastro, Mandelli, Gasparini, Dadda, & Bisazza, 2007), and males who specialize in sneak copulations experience greater sperm competition—as indexed by their larger relative testes size (Taborsky, 1998). In other species, reduced copulatory duration may also function to minimize detection by another male. However, men in intimate relationships already have regular access to their partner and reduced copulatory duration is less likely to be an attempt to reduce detection from a rival male and more likely to be implemented as an anti-cuckoldry tactic. On the basis of this comparative research, we anticipate that sexually coercive men may also reduce IPC duration. Therefore, we hypothesize that IPC duration will be negatively correlated with sexual coercion (Hypothesis 2). However, sex- ual coercion in human intimate relationships can take more subtle forms than physically forced copulation (Goetz & Shackelford, 2006; Shackelford & Goetz, 2004). Men who use physical force to obtain sexual access to their partner risk incurring severe costs, including their partner's defection from the relationship (Goetz & Shackelford, 2006). Men, therefore, use more subtle forms of sexual coercion before resorting to physical force. Shackelford and Goetz (2004) identified three components of sexual coercion that men use in intimate rela- tionships: (1) Commitment Manipulation (e.g., ''I told my partner that if she loved me she would have sex with me''), (2) Defection Threat (e.g., ''I threatened to have sex with another woman if my partner did not have sex with me''), and (3) Resource Manipulation/Violence (e.g., ''I withheld bene- fits that my partner depends on to get her to have sex with me''; ''I threatened to physically force my partner to have sex with me''). In summary, we propose that, in response to suspicions of female infidelity, men deploy anti-cuckoldry tactics to enter their sperm into competition to guard their paternity. Specifi- cally, following perceptions of female infidelity, men may use sexual coercion to counter women's resistance to IPC and quickly inseminate their partner to increase their chances of success in sperm competition. We conducted two independent studies to test two hypotheses derived from sperm competition theory. In Study 1, we secured men's reports on the target variables, and in Study 2, we secured women's reports on these same variables. Study 1: Men's Reports Study 1 secured men's reports to test the hypothesized relation- ships between perceptions of female infidelity, sexual coer- cion, and IPC duration. Following Goetz and Shackelford (2006), we operationalized sperm competition risk as men's perceptions of their partner's past infidelity and likelihood of future infidelity. Method Participants Participants were 410 men in a committed, heterosexual, sexual relationship for at least 1 month. The mean age of participants was 24.6 years (SD 1⁄4 8.0), the mean age of the men's partners was 23.3 years (SD 1⁄4 7.4), and the mean relationship length was 39.0 months (SD 1⁄4 61.3). Approximately half of the participants were university students, and the other half of participants were from communities surrounding the university where the research was conducted. Materials Participants completed an anonymous survey that requested demographic information, including the participant's age, their Barbaro et al. 3 Table 1. Study 1: Men's Reports. Zero-Order Correlations and Descriptive Statistics.Variable 1234567 1. Relationship length — 2. IPC duration3. Female infidelity4. Sexual coercionTotal5. Sexual coercionCommitment6. Sexual coercionDefection7. Sexual coercionResource/Violence MeanStandard deviation Note. IPC 1⁄4 in-pair copulation. *p 1⁄4 .05. **p < .01. ***p < .001. .01 —.09 .15** — .04 .07 .02 .10* .02 .07 .06 .01 39.01 5.39 61.30 1.80 .16** —.14** .93***.09 .90***.22*** .74***1.25 3.71 1.85 0.86 1.01 1.58 13.88 5.73 6.38 3.83 3-5 times; 4 1⁄4 act occurred 6-10 times; 5 1⁄4 act occurred 11 or more times). Responses to each statement were recoded as the midpoint of the response category the participant reported. For example, if the participant reported an act occurring ''3-5 times in the past month,'' the response was recoded as occurring 4 times in the past month. Responses indicating ''act occurred 11 or more times'' were recoded as occurring 15 times in the past month. The SCIRS assesses three components: Commitment Manipulation, Defection Threat, and Resource Manipulation/ Violence (see above for sample items). Following Shackelford and Goetz (2004), composite scores were calculated by sum- ming the recoded response category midpoints for the appropri- ate items, yielding a composite score for each participant for overall sexual coercion and each sexual coercion component. Procedure The current research was approved by the institutional review board at the university at which the research was conducted. Participants who met the following criteria were eligible for the current study: (1) male, (2) at least 18 years old, and (3) cur- rently in a committed, heterosexual, sexual relationship for at least 1 month. Prospective participants arrived at a specified location and read a consent form. Participants who met the criteria and agreed to participate completed a survey and returned the completed survey to the researcher in an unmarked, sealed envelope. Results Zero-order correlations and descriptive statistics for the target variables are reported in Table 1. Relationship length was not correlated with the predictor variables (i.e., female infidelity and sexual coercion) and is therefore not included as a con- founding variable in subsequent analyses. We calculated a zero-order correlation between perceived female infidelity and IPC duration to test whether female infidelity—as a measure of sperm competition risk—is associated with shorter IPC dura- tion. The result supports Hypothesis 1 in that greater perceived sperm competition risk is correlated with shorter IPC duration (see Table 1). — .59*** .76*** — .47*** — partner's age, and the length of their relationship. Participants then answered questions about their partner's infidelities, their own use of sexually coercive tactics, and their most recent IPC. Partner's infidelity. Men were asked four questions about their partner's infidelity: ''As far as you know, has your current partner had sexual intercourse [fallen in love] with someone other than you since you have been involved in a relationship together?''; How likely do you think it is that your current partner will in the future have sexual intercourse [fall in love] with someone other than you, while in a relationship with you?'' Men responded to each question on a 10-point Likert- type scale ranging from 0 (definitely no/not at all likely) to 9 (definitely yes/extremely likely). We calculated the mean of the responses to the four questions to create a composite variable female infidelity (a 1⁄4 .71) for each participant as an indicator of sperm competition risk. IPC duration. Men reported on the relative duration of their last copulation with their partner in which the man ejaculated as a result of penetrative, penile-vaginal sex, by responding to the question ''In comparison to what is typical, how long did sex- ual intercourse with your partner last?'' on a 10-point Likert- type scale ranging from 0 (much less time than is typical) to 9 (much more time than is typical). We asked participants to indicate the relative length of time that copulation lasted, rather than absolute length of time (e.g., in minutes) that copulation lasted. On average, IPC duration in humans is approximately five to six minutes; however, IPC duration can, in rare instances, last upward of 45 minutes (Waldinger et al., 2005). By assessing relative copulation duration, we are able to con- trol for individual differences in participant's average length of IPC that, for example, may be influenced by motivations to sexually please a romantic partner. Sexual coercion. The Sexual Coercion in Intimate Relationships Scale (SCIRS; Shackelford & Goetz, 2004) was used to assess men's use of sexually coercive acts in their current relationship. The SCIRS asks how often men performed 34 sexually coer- cive acts in the past one month. Participants respond to each item on a 6-point scale (0 1⁄4 act never occurred; 1 1⁄4 act occurred 1 time; 2 1⁄4 act occurred 2 times; 3 1⁄4 act occurred 4 Evolutionary Psychology Table 2. Study 1: Men's Reports. Multiple Regression Analysis With IPC Duration as the Dependent Variable. other than your current partner since you have been involved in a relationship with your current partner?'' Dichotomous responses to the past sexual infidelity question were recoded such that a ''no'' response was recoded ''0,'' and a ''yes'' response was recoded ''9.'' Women also responded to the ques- tions: ''Have you fallen in love with someone other than your current partner since you have been involved in a relationship with your current partner?'' and ''How likely do you think it is that you will in the future have sexual intercourse [fall in love] with someone other than your current partner, while in a rela- tionship with your current partner?'' Women responded to these questions on a 10-point Likert-type scale ranging from 0 (definitely no/not at all likely) to 9 (definitely yes/extremely likely). We then averaged responses to the four questions to create a composite variable female infidelity (a 1⁄4 .60) for each participant as an indicator of sperm competition risk. IPC duration. In parallel with Study 1, women reported on the relative duration of their last copulation with their partner in which their partner ejaculated as a result of penetrative, penile- vaginal sex, by responding to the question ''In comparison to what is typical, how long did sexual intercourse with your partner last?'' on a 10-point Likert-type scale ranging from 0 (much less time than is typical) to 9 (much more time than is typical). Partner's use of sexual coercion. In parallel with Study 1, women completed the SCIRS (Shackelford & Goetz, 2004) to report their partner's use of sexual coercion. Following Shackelford and Goetz (2004), we calculated participant's composite scores for overall sexual coercion and each sexual coercion compo- nent (i.e., Commitment Manipulation, Defection Threat, and Resource Manipulation/Violence). Procedure The current research was approved by the institutional review board at the university at which the research was conducted. Participants who met the following criteria were eligible for the current study: (1) female, (2), at least 18 years old, and (3) currently in a committed, heterosexual, sexual relationship for at least 1 month. Prospective participants arrived at a specified location and read a consent form. Participants who met the criteria and agreed to participate completed a survey and returned the completed survey to the researcher in an unmarked, sealed envelope. Results Zero-order correlations and descriptive statistics for the target variables are reported in Table 3. Relationship length was not correlated with the predictor variables (i.e., female infidelity and sexual coercion) and is therefore not included as a con- founding variable in subsequent analyses. We calculated a zero-order correlation between self-reported infidelity and IPC duration to investigate whether female infidelity is associated with shorter IPC duration. The correlation between women's Predictor Variable Sexual coercionCommitment Sexual coercionDefection Sexual coercionResource/Violence Note. IPC 1⁄4 in-pair copulation. B b t-statistic .05 .16 0.18 .00 .01 0.17 .04 .08 1.28 p value .067 .866 .201 We calculated a zero-order correlation to test whether sex- ual coercion is associated with shorter IPC duration. The result provides support for Hypothesis 2 in that that men who report more frequent use of sexually coercive Commitment Manipu- lation acts also report shorter IPC duration (see Table 1). A multiple linear regression was conducted as a more stringent test of Hypothesis 2. The three sexual coercion components (Commitment Manipulation, Defection Threat, and Resource Manipulation/Violence) were simultaneously entered into the regression analysis, with IPC duration as the dependent vari- able (see Table 2). Men's use of sexually coercive Commit- ment Manipulation acts remained the best predictor of IPC duration, approaching statistical significance (p 1⁄4 .07). Study 2: Women's Reports Men's reports of their partner's infidelities and their use of sexual coercion may be inaccurate (Dobash, Dobash, Cava- nagh, & Lewis, 1998; Edleson & Brygger, 1986). Study 2 secured reports from an independent sample of women to examine the hypothesized relationships between female infide- lity, men's use of sexual coercion, and IPC duration. Method Participants Participants were 455 women in a committed, heterosexual, sexual relationship for at least 1 month. Women in this study were not necessarily partnered to the men in Study 1. The mean age of the participant was 22.1 years (SD 1⁄4 5.7), the mean age of the women's partner was 24.4 years (SD 1⁄4 6.9), and the mean relationship length was 31.4 months (SD 1⁄4 41.8 months). Approximately half of the participants were university stu- dents, and the other half of participants were from communities surrounding the university where the research was conducted. Materials The survey used in Study 2 was parallel to the survey used in Study 1, with appropriate replacement of gender-relevant terms. Self-reported infidelity. Women were asked four questions to assess their past infidelities and the likelihood of committing future infidelities. Women responded ''yes'' or ''no'' to the question: ''Have you had sexual intercourse with someone Barbaro et al. 5 Table 3. Study 2: Women's Reports. Zero-Order Correlations and Descriptive Statistics.Variable 1234567 1. Relationship length —2. IPC duration .13* — 3. Female infidelity .06 4. Sexual coercionTotal .00 5. Sexual coercionCommitment .00 6. Sexual coercionDefection .01 7. Sexual coercionResource/Violence .01 Mean 31.39 Standard deviation 41.83 Note. IPC 1⁄4 in-pair copulation. *p < .05. **p < .01. ***p < .001. Table 4. Study 2: Women's Reports. Multiple with IPC Duration as the Dependent Variable. .06 .04 .02 .08 .02 5.62 1.95 Regression Analysis — .20*** —.21*** .94*** —.11* .77*** .61*** —.18*** .92*** .80*** .57*** — 1.31 3.41 1.95 .60 0.86 1.73 22.48 10.22 5.95 9.07 as female infidelity. Men's reports indicate that as perception of sperm competition risk increases, the duration of IPC decreases. We suggest that circumstances indicative of sperm competition—specifically, a man's partner's extra-pair copula- tion—motivate men to enter their sperm into competition as quickly as possible, resulting in relatively shorter IPC duration. Women's reports do not replicate this relationship, although women's reports of their partner's use of sexually coercive Defection Threat acts approached statistical significance. Men may face a trade-off in copulatory strategies following female infidelity: Quickly enter sperm into competition or per- form semen-displacement behaviors to extract rival sperm from the vagina. Consequently, IPC duration may be relatively shorter or longer, respectively, depending on the copulatory strategy a man pursues. Men's copulatory strategy might be determined by whether they are attempting to prevent or cor- rect female infidelity (Shackelford, 2003). We operationalized sperm competition risk in Study 1 as men's perception of female infidelity. The results indicate that men engage in shorter IPC when they perceive greater risk or likelihood of female infidelity, which we suggest is a corrective sperm com- petition strategy. Goetz and colleagues (2005) suggested that men might pursue longer duration of IPC as a corrective strat- egy in response to sperm competition risk—operationalized in their research as female partner attractiveness. However, a partner's attractiveness can also motivate men to prevent infi- delity because her attractiveness portends that she is more likely to have extra-pair suitors. Thus, the relationship between female attractiveness and longer IPC duration might also be indicative of men's attempts to prevent partner infidelity (see Buss, 1988; Buss & Shackelford, 1997). The IPC duration and, consequently, the copulatory strategy men deploy, might be affected by a variable not measured in the current research or in Goetz et al.'s (2005) research: the time since a woman's suspected or known extra-pair copula- tion. The time since a women's extra-pair copulation might affect men's copulatory strategy at the next IPC. For example, longer copulation duration—which includes more semen- displacing behaviors—might be a more effective sperm com- petition strategy in circumstances in which the time between a woman's copulation with an extra-pair partner and an in-pair Predictor variable Sexual coercionCommitment Sexual coercionDefection Sexual coercionResource/Violence Note. IPC 1⁄4 in-pair copulation. B b .04 .19 .05 .17 .02 .08 t-statistic 1.21 1.81 0.68 p value .226 .071 .500 reports of their infidelity and IPC duration was not statistically significant (see Table 3). We calculated a zero-order correlation to investigate whether women's reports of their partner's sexual coercion is correlated with shorter IPC duration. The correlations between the three sexual coercion components and IPC duration were not statistically significant (see Table 3). A multiple linear regression was conducted as a more stringent test of Hypothesis 2. The three sexual coercion components (Commitment Manip- ulation, Defection Threat, and Resource Manipulation/Vio- lence) were simultaneously entered into the regression analysis, with IPC duration as the dependent variable (see Table 4). Women's reports of their partner's use of sexually coercive Defection Threat acts were the best predictor of IPC duration, approaching statistical significance (p 1⁄4 .07). General Discussion The results of Study 1 (men's reports) support Hypothesis 1 in that perceptions of female infidelity are correlated with shorter relative IPC duration. The results of Study 1 also support Hypothesis 2 in that men's use of sexually coercive Commit- ment Manipulation is correlated with shorter relative IPC dura- tion. The results of Study 2 (women's reports), however, do not provide statistical support for Hypotheses 1 and 2: The correla- tion between women's self-reported infidelities and IPC dura- tion, and between women's reports of their partner's use of sexually coercive acts and IPC duration, is not statistically significant. Tests of Hypothesis 1 investigated whether IPC duration is shorter in response to sperm competition risk—operationalized 6 Evolutionary Psychology partner is brief. Under these circumstances, a rival man's sperm and spermicidal substances have not yet been ejected from the vagina (e.g., less than 1 hour; Baker & Bellis, 1993b, 1995), and reduced ejaculatory latency may adversely affect semen displacement (Gallup & Burch, 2004). Shorter copulation dura- tion—quickly entering sperm into competition—might be more likely in circumstances in which the time between a woman's extra-pair copulation and IPC is longer (e.g., more than 1 hour; Baker & Bellis, 1993b), and it is unlikely that rival sperm are still present in the vagina. After sperm are ejected from the vagina, adjusting IPC duration to be shorter—ejacu- lating quickly—may be a more successful sperm competition strategy (but see, Gallup & Burch, 2004). This hypothesis is contingent on two factors: (1) the duration of how long sperm remain in the vagina (i.e., the ''corrective window''; Baker & Bellis, 1995; Johnson & Everitt, 1995; Morris, 1977; Smith, 1984), and (2) a woman's behavior after sexual intercourse (Gallup & Burch, 2006; but see, Baker & Bellis, 1995). Our understanding of men's strategies to correct or prevent female infidelity (Shackelford, 2003), such as extending IPC duration to facilitate semen displacement, would be informed by asses- sing the time between a woman's perceived or actual extra-pair copulation and the next IPC. Men's reports indicate a relationship between sexually coer- cive Commitment Manipulation and shorter IPC, supporting Hypothesis 2. This suggests that the duration of coercive sex is shorter than consensual sex, consistent with the results of research on nonhumans (e.g., waterfowl; McKinney et al., 1983). The current research is the first to investigate copulatory behavior—in particular, copulation duration—associated with reports of sexual coercion in humans. Men who employ sexual coercion to secure IPCs may be pursuing a copulatory strategy that affords quick entry into sperm competition, rather than a strategy to displace rival sperm that may be present. The nonsignificant relationship between women's reports of their infidelity and shorter IPC duration may be a result of a perceptual bias unique to the context of sexually coercive copu- lations. Because sexual coercion is often psychologically trau- matizing for women (Campbell, 1989; Thornhill & Thornhill, 1990, 1991), women's reports of copulation duration and/or the frequency of sexual coercion may differ from men's reports. Recent research (Be ́langer, Mathieu, Dugal, & Courchesne, 2015) conducted with romantic couples indicates that, com- pared to their male partner, women significantly underreport instances of sexually coercive behavior (measured by the Conflict Tactics Scale; Straus, Hamby, Boney-McCoy, & Sugarman, 1996). Discrepancies between male and female reports may account for the sex differences observed in the current research. Nevertheless, men's reports suggest that future research might profitably investigate sexually coercive copulatory behaviors and strategies. Future research could secure data from both romantic partners to obtain more accu- rate corroborating reports of sexual behavior in intimate relationships. The results of the current research are not conclusive regard- ing the relationships between sperm competition risk, sexual coercion, and copulation duration. Men's reports (Study 1) provide support for the study hypotheses, but women's reports (Study 2) do not corroborate these findings. The current research offers an initial investigation into adjustments of copulatory duration in humans—a domain that has not yet been explored. Future research that addresses the limitations of the current studies (see below) could continue to profitably inves- tigate copulation duration in humans. Limitations and Future Directions The current research secured data from independent samples of men and women. Although securing men's reports and women's reports of the same behaviors might afford triangu- lating on reliable assessments of the relevant phenomena, the current data are limited in that we secured data from men and women who were not partnered to one another. Future research might secure daily reports from both members of couples to afford more reliable assessments of sexual coercion and related copulatory behaviors in intimate relationships. Another limitation is that we assessed the duration of the most recent IPC. In contrast, we assessed the frequency with which men performed sexually coercive acts during the previ- ous one month. Because of the difference in assessment time frame for the two variables, we cannot be certain that sexual coercion was used to secure the copulation about which parti- cipants reported. Future research addressing sexual coercion in intimate relationships and copulatory behaviors could secure reports about sexually coercive acts used to achieve a specific IPC. Additionally, research could investigate the average time of IPCs or average time in the previous one month (e.g., in minutes) to afford congruence between measures of sexual coercion and copulatory duration. This would afford stronger claims about whether and how the use of sexual coercion is related to copulation duration in response to female infidelity. Moreover, we secured a relative measure of copulatory duration (how long copulation lasted compared to ''what is normal for you''), rather than an absolute measure of copula- tion duration (how long copulation lasted, in minutes). Although our measure of copulation duration afforded control of individual differences in copulation duration, we suggest that future research investigates absolute length of copulation duration. Because the upper limits of copulation duration in humans (e.g., 45 min; Waldinger et al., 2005) may indicate an increased motivation to sexually please one's partner, the absolute duration of copulation could provide useful informa- tion regarding whether—under circumstances of increased sperm competition risk—men are less inclined to sexually please their partner for the benefit of entering their sperm into competition more quickly. Finally, the results of the current research are correlational, and thus strong statements of causality are not defensible. Based on the available literature, however, we argue that men's copulatory behaviors and use of sexual coercion are motivated by increased sperm competition risk—in particular, by per- ceived or actual female infidelity. It is possible, however, that Barbaro et al. 7 men's sexual coercion motivates women to pursue extra-pair copulations or eventually to terminate the relationship. Conclusion Female infidelity is a primary context in which sperm compe- tition might have occurred over human evolutionary history and may have occasionally resulted in cuckoldry. Because cuckoldry can inflict substantial costs on paternally investing males, men may have evolved strategies to solve this class of adaptive problems. We tested the hypotheses that female infi- delity and men's use of sexual coercion are associated with shorter IPC duration. Men's reports indicate that perceptions of their partner's infidelity are correlated with shorter IPC duration, which we argue reflects a copulatory strategy to quickly enter sperm into competition. Men's reports also indi- cate that sexually coercive men reduce IPC duration. These findings contribute to our understanding of men's use of sexual coercion in intimate relationships and, more generally, to our understanding of human male adaptations to sperm competition.

Allele

type of gene

Why are Primates Social?

Disadvantages competition for food visibility to predators competition for mates risk of social tension and violence Advantages predator defense access to food mate access assistance in care of young defense of food *Most mammals don't live in permanent social groups

Problems with Sexual Strategies Theory

Does not make predictions about when individuals should pursue short-term matings vs. long term matings Focuses only on variation between sexes, and not on variation within sexes Doesn't consider how women's control of resources should affect strategies Minimizes role of "good genes"

Travesti

Don Kulick studied Brazilian transgendered prostitutes These males inject themselves with silicone to appear more female

the evolution of human homosexual behavior

The Evolution of Human Homosexual Behavior1 And if there were only some way of contriving that a state or an army should be made up of lovers and their loves, they would be the very best governors of their own city . . . and when fighting at one an- other's side, although merely a handful, they would overcome all men. plato, The Symposium The brother-in-law is ally, collaborator, and friend.. . . In the same band, the potential brother-in-law, i.e., the cross-cousin, is the one with whom, as an adolescent, one indulges in homosexual activities which will always leave their mark in the mutually affectionate behaviour of the adults. claude le ́vi-strauss, The Elementary Structures of Kinship Homosexual behavior has existed throughout human history and in most, perhaps all, human cultures (e.g., Blackwood 1986a, Greenberg 1988, Duberman, Vicinus, and Chauncey 1989). Cross-cultural and historical stud- ies qualify the breadth of homosexual experience, while medical studies, primarily from the contemporary West, quantify its depth. The Melanesian example of homo- sexual behavior is well known, with 10 to 20% of Mel- anesian societies requiring all men to participate in ho- mosexual as well as heterosexual sex (reviewed in Herdt 1984a). In southern China at the turn of the last century, 100,000 women joined a marriage resistance movement that included, for many, lifelong homosexual partner- ships (Sankar 1986). And Mpondo miners of South Africa in the first half of the 20th century commonly entered into "mine marriages," forgoing sexual liaisons with nearby township women (Moodie, Ndatshe, and Sibuyi 1989). Homosexual behavior is much more systemic than the above isolated examples might suggest (table 1). In Pa- cific Island societies other than in Melanesia, such as Tahiti and Hawaii, homosexual behavior was common prior to Western influence (Gunson 1964, Morris 1990). In native North America, at least 137 societies had in- stitutional roles for transgenders (berdache) commonly associated with homosexual behavior (Callender and Ko- chems 1983, Williams 1986, Roscoe 1987, Lang 1998). Homosexual behavior has also been common in societies of native South America (reviewed in Greenberg 1988; Murray 1995a, b), Asia (e.g., Hinsch 1990, Leupp 1995), precolonial Africa (reviewed in Sweet 1996), and pre- modern Europe (e.g., Dover 1989, Rocke 1996). In Ford and Beach's (1951) world sample, homosexual behavior is normative in 64% of the societies with available data (n = 76), at least for certain classes of individuals. Studies over the past 70 years in contemporary soci- by R. C. Kirkpatrick Homosexuality presents a paradox for evolutionists who explore the adaptedness of human behavior. If adaptedness is measured by reproductive success and if homosexual behavior is nonrepro- ductive, how has it come about? Three adaptationist hypotheses are reviewed here and compared with the anthropological litera- ture. There is little evidence that lineages gain reproductive ad- vantage through offspring care provided by homosexual members. Therefore, there is little support for the hypothesis that homo- sexuality evolved by kin selection. Parents at times control chil- dren's reproductive decisions and at times encourage children in homosexual behavior. There is therefore more support for the hy- pothesis of parental manipulation. Support is strongest, however, for the hypothesis that homosexual behavior comes from individ- ual selection for reciprocal altruism. Same-sex alliances have re- productive advantages, and sexual behavior at times maintains these alliances. Nonhuman primates, including the apes, use ho- mosexual behavior in same-sex alliances, and such alliances ap- pear to have been key in the expanded distribution of human an- cestors during the Pleistocene. Homosexual emotion and behavior are, in part, emergent qualities of the human propensity for same-sex affiliation. Adaptationist explanations do not fully explain sexual behavior in humans, however; social and histori- cal factors also play strong roles. r. c. kirkpatrick is a biodiversity specialist with the Na- ture Conservancy's Yunnan Great Rivers Project (Kunming, Yun- nan Province, People's Republic of China [rckirkpatrick @mother.com]) and a research associate in the Anthropology De- partment of the University of California, Davis. Born in 1959, he was educated at Pomona College (B.A., 1981) and the University of California, Davis (Ph.D., 1996). His research interests include the feeding ecology of leaf-eating monkeys, ecological influences on primate social organization and behavior, the conservation of biodiversity in Asia, and the interaction of culture and biology in human behavior. He has published "Ecology and Behavior of the Snub-nosed and Douc Langurs," in The Natural History of the Doucs and Snub-nosed Monkeys, edited by N. G. Jablonski (Sin- gapore: World Scientific Press, 1998), "Colobine Diet and Social Organization," in The Nonhuman Primates, edited by P. Dolhi- now and A. Fuentes (Palo Alto: Mayfield, 1999), and (with Y. C. Long, T. Zhong, and L. Xiao) "Social Organization and Range Use in the Yunnan Snub-nosed Langur, Rhinopithecus bieti" (Interna- tional Journal of Primatology 17:13-51). The present paper was submitted 1 iv 99 and accepted 30 viii 99; the final version reached the Editor's office 29 xi 99. had their origin in a seminar on human mating led by S. Hrdy many years ago. I am grateful for her continuing encouragement and the encouragement of P. S. Rodman to develop these ideas. 1. Previous versions of this manuscript benefited from critical re- view by M. Borgerhoff Mulder, J. M. Dickemann, A. Fuentes, J. Gilardi, A. H. Harcourt, S. Hrdy, D. Judge, M. Small, K. Stewart, P. Vasey, and three anonymous referees, as well as comments from S. Connor, P. Coppolillo, and J. Scheib. The ideas expressed here 385 386 F current anthropology Volume 41, Number 3, June 2000 table 1 Selective Overview of Geographical Distribution and Forms of Homosexual Behavior Concurrent with Hetero- Female Juvenile sexual Continent/Society Form or Male? or Adult? Behavior? Frequency in Total Population Reference AfricaSiwah (Libya) Azande (Sudan) Dahomey (Nigeria) Mpondo (South Africa) EuropeClassical Athens Early Roman empire Dinaric (Serbia) Florence, 15th century AmericasLakota (U.S.A.) Mohave (U.S.A.) Nambikuara (Brazil) Yanomamo (Venezuela) OceaniaPrecolonial Tahiti Arunda (Australia) Big Nambas (Melanesia) Marind-anim (Melanesia) Patron/client, includ- ing "lending" of sons among promi- nent men M: patron/client among warriors; F: within polygynous marriages Among adolescent peers, predicts rit- ual best friend of adults Patron/client; client receives aid in household formation M: patron/client; F: reported but little described M: state-sanctified marriages; F: appar- ently much the same as male Within sworn friendships Patron/client across class, also among peers within guilds Men in conventional marriages sponsor transgender "berdache" Same-sex marriage be- tween transgenders and those of con- ventional gender Among classificatory brothers-in-law Among peers, prior to heterosexual marriage Within sworn friendships Among cross-cousins, following exogamy rules of heterosex- ual behavior M: patron/client; F: common but little described (appar- ently among peers) Adult male serves as mentor, ideally to sister's adolescent son M F, M J, A Yes J, A Yes ∼95%[?] 1 Common 2 Common 3 Common 4 Common 5 Common [?] 6 Unknown 7 1 50%[?] 8 Unknown 9 Limited 10 Common [?] 11 1 50% 12 Common 13 Common 14 ∼100% 15 ∼100% 16 F,M J,?A ? MA Yes F,M F,M M, ?F M M F, M M M M F, M F, M M J,A Yes J,A Yes ?A ? J,A Yes A Yes A Yes J No? J No J,A Yes J No J,A Yes J,A Yes kirkpatrick Evolution of Human Homosexual Behavior F 387 table 1 (Continued) Concurrent with Hetero- Female Juvenile sexual Continent/Society Form or Male? or Adult? Behavior? Frequency in Total Population Reference Sambia (Melanesia) AsiaChina, 700-400 b.c. China, a.d. 1865-1965 Japan, 16th, 17th century Pukhtun (Pakistan) Adult male serves as mentor, ideally to wife's younger brother Patron/client; client receives social prestige and politi- cal power Sworn alliances in marriage resistance of 100,000 women M: patron/client, common in both rural and urban ar- eas; F: little described More pleasurable than heterosexual behavior because of male-female antagonism M M F F, M M J, A Yes A Yes A No J, A Yes J, A Yes ∼100% 17 Unknown 18 Limited 19 1 50%[?] 20 Unknown 21 sources: 1, Cline (1936); 2, Evans-Pritchard (1970); 3, Herskovits (1938); 4, Moodie, Ndatshe, and Sibuyi (1989); 5, Dover (1989); 6, Boswell (1980); 7, Tomasic (1948); 8, Rocke (1996); 9, Williams (1986); 10, Lang (1998); 11, Le ́vi-Strauss (1943); 12, Chagnon (1966); 13, Gunson (1964); 14, Roheim (1933, 1950); 15, Deacon (1934); 16, van Baal (1966); 17, Herdt (1984b); 18, Hinsch (1990); 19, Sankar (1986); 20, Schalow (1989), Leupp (1995); 21, Lindholm (1982). eties quantify homosexual behavior over time (table 2). In a selected sample of U.S. women in the 1920s (n = 2,200), approximately 23% had homosexual sex (Davis 1929). Similarly, in Kinsey et al.'s (1953) sample of U.S. women in the late 1940s (n = 2,601), 20% had homosex- ual sex. In Kinsey, Pomeroy, and Martin's (1948) sample of U.S. men over age 20 (n = 2,830), 37% had orgasmic sex with other men, 10% were primarily homosexual for three years between ages 16 and 55, and 4% of Cauca- sians were exclusively homosexual after adolescence. In a representative sample of U.S. men in 1970 (n = 1,450), 20% had orgasmic sex with other men, 7% after age 19 (Fay et al. 1989). In the Darwinian view of natural selection, individuals should seek to maximize reproductive success. Humans are a sexually reproducing species, and children result only through mating with members of the other sex; homosexual acts do not appear to aid reproduction. Ho- mosexual behavior is too widespread to be a fluke or an aberration, but evolutionists in particular should be puz- zled by its ubiquity. One could look at homosexual be- havior as a value-free activity, such as grooming, but few societies do. In fact, much significance is attached to homosexual relations. Why is homosexual behavior so widespread? How could it have evolved? Attempts by evolutionary biolo- gists to reconcile homosexual behavior with natural se- lection have assumed that homosexual behavior is neg- ative for individual fitness and therefore must be maintained by indirect selection. The theories of kin se- lection and of parent-offspring conflict, in particular, have been the focus of adaptationist explanations. In the kin-selection hypothesis, homosexuals altruistically forgo reproduction to assist the offspring of relatives (Wilson 1975, Weinrich 1987a). In the parental-manip- ulation hypothesis, children are manipulated by parents to forgo reproduction, become homosexual, and assist the offspring of relatives (see Trivers 1974, Ruse 1988). A third hypothesis, presented by Hutchinson (1959) and called "balanced polymorphism," suggests that homo- sexual behavior is retained because it co-occurs with a second trait under positive selection. Hutchinson does not describe this second trait, although he suggests that it might involve rates of development. I propose a fourth possibility: that direct, positive se- lection for homosexual behavior occurs because of its nonconceptive benefits (see fig. 1). Heterosexual behav- ior serves nonconceptive functions such as the mainte- nance of long-term bonds (Morris 1967, Wilson 1975, Small 1995). If homosexual behavior also serves non- conceptive functions, such as the maintenance of same- sex alliances (long-term supportive relationships) that 388 F current anthropology Volume 41, Number 3, June 2000 table 2 Proportions for Homosexual Behavior Reported in Contemporary Societies, Primarily from the West Men n n Exclusive Study Population (Men) (Women) Homosexual Bisexual Sum Women ExclusiveHomosexual Bisexual Sum Reference U.S. nonrandom samples United States United Statesb United States - 2,200 111 - 2,830d 2,601 1,450 - ---n.r. n.r 23.0a 1 -- 2 n.r. 20.0 3 -- 4 3.0 4.0 5 3.3e 3.6e 6 - - 7 - - 8 2.2 3.0 9 2.0 2.6 10 3.2e 3.3e 6 2.5 2.6 11 1.6e 2.1e 6 -- 12 n.r. 12.0h 13 -- 14 0.3 1.2 15 -- 16 2.4 3.3 2.3 7.0 U.S. random samplesUnited StatesLosAngeles - 298 20.3 - 5.4e 6.2e 0.3e 1.2 2.3 - 4.9 5.6 - 2.9 3.5 0.8 2.4 3.1 0.6 10.0e 10.7e 0.1e 5.0 5.2 0.1 3.4e 4.5e 0.5e n.r. 38.0 - 15.0h 18.0h n.r. 16.3 16.9 - 3.1 3.3 0.9 n.r. 25.6 - United States United States United States European random samples 1,288 3,324 1,401 674 - - ∼3,150f 10,127 788 10,492 696 - n.r. - 1,285 - 0.8e 1.1 0.7 0.6 0.7 0.7e 0.2 1.1e n.r. 3.0h 0.6 0.2 n.r. Norway ∼3,150f France 9,928 France 1,506 United Kingdom United Kingdom Other societies, non- random samples Arab Near Eastb,gColombiabThailandb 173 Thailand 983 Thailandi Median Mean 8,384 1,137 n.r. n.r. 113 n.r. n.r. n.r. n.r. 27.0c - 37.0c n.r. n.r. n.r.- - - 1.0 0.7 4.9 8.5 0.6 0.9 6.3 14.2 0.5 sources: 1, Davis (1929); 2, Finger (1947); 3, Kinsey, Pomeroy, and Martin (1948), Kinsey et al. (1953); 4, Fay et al. (1989); 5, Essock- Vitale and McGuire (1988); 6, Sell, Wells, and Wypij (1995); 7, Billy et al. (1993); 8, Smith in Diamond (1993); 9, Sundet et al. (1988); 10, ACSF (1992); 11, Johnson et al. (1994); 12, Melikian and Prothro (1954); 13, Ardila (1986); 14, Whitam and Mathy (1986); 15, Sitti- trai et al. in Diamond (1993); 16, Nopkesorn et al. in Sittitrai and Brown (1994). note: Both median and mean numbers are unweighted and therefore bias totals toward U.S. and European samples. aMean of married and unmarried samples.bCollege students.cOrgasmic sex. dAge 21 years and older.eBehavior in past five years (i.e., not lifetime behavior).f63% return of 10,000 postal questionnaires; see also Veierod et al. (1997). gApproximately 90% from Lebanon, Syria, Jordan, and Iraq.hAverage of four cities; selected sample estimates homosexual:bisexual ratio. iMilitary conscripts, age 21. aid in resource competition or in cooperative defense, then homosexual behavior will be under positive selec- tion. In terms of the biological distinction between sex- ual and somatic reproduction, homosexual behavior is a survival strategy, not a reproductive strategy. If this is the case, homosexual behavior will be best explained by reference to the costs and benefits of reciprocal altruism (sensu Trivers 1971). After a brief review of correlates of human homosexual behavior, predictions of the kin selection, parental-ma- nipulation, and alliance-formation hypotheses are com- pared with the ethnographic literature. (Although my intent is to incorporate the homosexual behavior of women and men equally, examples in the literature are biased toward men.) Limited reference to nonhuman pri- mates is made in the course of a consideration of the evolution of homosexual behavior in hominids. In this review, sexual behavior is defined (after Vasey 1995) as genital contact or genital manipulation by individuals who have begun development of secondary sexual char- acteristics. By this definition, ephebophilic interactions (those involving pubescent adolescents) can be sexual, kirkpatrick Evolution of Human Homosexual Behavior F 389 Fig. 1. Curves for reproductive success both from sex- ual activities and from somatic (survival and mainte- nance) activities. The curve for sexual reproduction (thin solid line) assumes a monotonic decrease in fe- cundity with reduced proportion of heterosexual activ- ities. The curve for somatic reproduction (dotted line) assumes a sigmoidal increase in survival and mainte- nance with increased same-sex alliances to the mid- point, after which increases in same-sex alliances lead to a decrease in reproductive success. (The decrease results from the inability to sustain heterosexual alli- ances and thereby take advantage of the sexual divi- sion of labor and of the long-term bonds typically necessary for successful child rearing.) Averaging the two curves (heavy solid line) shows that the best strat- egy for net reproductive success is alliances with both sexes. If, within particular cultural systems, homosex- ual behavior is useful for maintaining same-sex alli- ances, it is then adaptive to include a moderate level of homosexual behavior in one's sexual repertoire. while pedophilic interactions (those involving prepubes- cent children) cannot. A "homosexual" has sex only with the same biological sex, while a "heterosexual" has sex only with the other biological sex. A person with a sexual history including both sexes is "bisexual." Bisexuality is as of much interest as homosexuality per se; the term "nonheterosexuals" refers to bisexuals and homosexuals together, and the term "homosexual behavior" refers to individual acts of both bisexuals and homosexuals. These behavior-based definitions are more restrictive than those typically found in the literature. Behavior- based categories are, however, more logically defensible and no less verifiable than assigning bisexuals to the homosexual or heterosexual categories. Behavior is also a better classifier than self-identification. In a sample of contemporary U.S. women (n = 1,384), 17% had homo- sexual sex, yet only 5% defined themselves as gay or bisexual; for men (n = 1,335) the numbers were 22% versus 9% (Janus and Janus 1993). Although intention- ality may differ between individuals who identify them- selves as gay and as straight, the a priori assumption should be that their behaviors arise from the same source. Behavior-based categories allow for comparisons across cultures—and species—while categories based on self-identification do not. Behavior-based categories fur- ther allow for objective comparison with the perform- ance measure of reproductive success. The use of restric- tive definitions in this review at times leads to the testing of predictions about homosexuals with data for nonheterosexuals (i.e., homosexuals and bisexuals to- gether). This is, unfortunately, unavoidable because of inconsistency among researchers in dividing the contin- uum of human sexual behavior. Correlates of Homosexual Behavior Any behavior to be explained by evolution must have inherited components. This is not to argue for direct genetic determinants of sexuality. Culture can be a pow- erful, consistent way of transferring traits between gen- erations (Boyd and Richerson 1985); behavioral poly- morphisms need not be genetically based and may result from mixed strategies within an individual's lifetime (Maynard Smith 1982). Recent research addresses the possibilities of biological and cultural influences on ho- mosexual behavior. Clearly, homosexual behavior has statistically significant correlates, at least in contem- porary Western societies. genes Individuals who identify themselves as homosexual or bisexual are found in particular family lines (reviewed in Bailey and Bell 1993, Pattatucci and Hamer 1995, Pil- lard 1996). Monozygotic twins have concordance for non- heterosexuality at about twice the rate of dizygotic twins, suggesting some genetic influence on the findings on family aggregation (e.g., Bailey et al. 1993, Whitam, Diamond, and Martin 1993; cf. Hershberger 1997). Fur- ther, there may be chromosomal concordances in non- heterosexuals who are brothers (Hamer et al. 1993). The chromosomal studies have been particularly contentious (see Risch, Squires-Wheeler, and Keats 1993, Byne 1995); initial findings have been replicated by Hamer and col- leagues (Hu et al. 1995), but replication has eluded in- dependent researchers (Rice et al. 1999). Twin studies also suggest the importance of environ- mental factors. Bailey and Pillard (1991; see also Buhrich, Bailey, and Martin 1991), for example, found that 52% of the monozygotic co-twins (n = 56) of male nonhet- erosexuals were also nonheterosexual, while only 22% of dizygotic co-twins (n = 54) were also nonheterosexual. Eleven percent of adopted brothers (n = 57) were non- heterosexual, compared with 9% of nontwin related brothers (n = 142). A large portion of homosexual be- 390 F current anthropology Volume 41, Number 3, June 2000 havior must therefore be free from the influence of genes and of the uterine environment, or monozygotic twins would have greater concordance and adopted brothers and nonadopted nontwin brothers—who share no genetic material—would have greater differences in homosexual behavior. hormones Data are contradictory concerning hormonal and devel- opmental differences between those who engage in ho- mosexual behavior and those who do not (reviewed in Small 1995, Byne 1995, LeVay 1996). Differences in pre- natal hormones between self-identified homosexual and heterosexual men are implicated by small statistically significant differences in fingerprints (Hall and Kimura 1994). (Fingerprints complete development by the fourth month of pregnancy and are quite heritable.) However, there are no apparent differences in the androgen recep- tors of self-identified homosexual and heterosexual men, and this weakens the case for differences in hormone exposure (Macke et al. 1993). There may be differences in brain morphology and associated differences in visuo- spatial ability between homosexuals and heterosexuals (reviewed in LeVay 1996, Sanders and Wright 1997), but any such differences are quite small and have almost complete overlap in range (Byne 1995, 1996; Wegesin 1998). The inconclusiveness of the available evidence for hormonal correlates to homosexual behavior is given a balanced treatment by Small (1995). Several longitudinal studies suggest that gender-atyp- ical children grow up to be homosexual (e.g., Green 1987, Zuger 1988). In some Western societies, children iden- tified by parents as gender-atypical tend to identify them- selves as homosexual as they become adolescents; many self-identified homosexuals recall a gender-atypical childhood (Whitam and Mathy 1986, 1991; Zucker 1990). However, twin studies suggest that sexual orientation does not covary with gender nonconformity (Bailey and Pillard 1991). Further, many populations known for male homosexual behavior do not have a reputation of effem- inacy (e.g., the samurai of feudal Japan [Leupp 1995], the knights of feudal Europe [Greenberg 1988; cf. Gilmour- Bryson 1996], the Yanomamo of South America [Chag- non 1966]). It remains possible that sexual orientation has a com- plex, dependent relation with the development of gender identity. Prenatal hormones contribute to differences in gender characters, and certain gender characters may tend to co-occur with homosexual behavior (reviewed in LeVay 1996). If a link exists between sexual orientation and gender identity, however, we know remarkably little of its details. Studies of gender identity and sexual ori- entation typically lack strict controls for study groups. Determining cause and effect is confounded by societal preconceptions of gender identity (Corbett 1998). While many self-identified homosexuals recall a gender-atypi- cal childhood, for example, so do many self-identified heterosexuals (Byne in Small 1993). Gender nonconform- ity is neither necessary nor sufficient for homosexual behavior. environment Homosexual behavior is correlated with social and dem- ographic variables. Military service increases the like- lihood of homosexual behavior by 50% (19% of 570 males versus 12% of ∼880 males, Yates-corrected x2 = 12.6, d.f. = 1, p ! 0.001 [Fay et al. 1989:348 n. 30]). Public school attendance in England doubles the likelihood of adolescent homosexual behavior (Johnson et al. 1994). Homosexual behavior also correlates with birth order, at least in U.S. and Canadian men: the likelihood that a man will engage in homosexual sex increases with the number of older brothers he has (reviewed in Blanchard and Bogaert 1996a, b). Blanchard and Bogaert (1996a, b) hypothesize that prenatal hormones mediate this effect, with immune response in mothers becoming stronger with each male child. The postuterine environment might also influence the correlation between birth order and homosexual behavior, however. The different op- portunities presented to boys in different positions within family hierarchies is an alternative, psychological explanation for this correlation (Sulloway 1996:488 n. 36). Human homosexual behavior aggregates in particular lineages, relates to birth order, and depends in part on one's social network, and some cultures encourage it while others do not. What is most striking about cor- relates of homosexual behavior is the small amount of variation explained by any single factor. Genes, hor- mones, childhood experience, and adult experience are multivariate and interact to produce multivariate life histories. Present evidence suggests that homosexual be- havior is weakly, if significantly, correlated with nu- merous traits, some genetic, some developmental, and some experiential. Subtle increases in propensity for ho- mosexual behavior on a broad range of fronts create an array of human experience. There may be certain con- stellations of traits that lead to a greater or lesser pro- pensity for an individual to participate in a greater or lesser proportion of homosexual acts. Sexual behavior and emotion are continuous variables, not dichotomous ones (Ellis, Burke, and Ames 1987, McConaghy, Buhrich, and Silove 1994). Focus on slight differences in modal tendencies due to single, isolated factors obscures the fundamental continuity between homosexual, bisexual, and heterosexual experience. This implicitly suggests that it is a mistake to divide individuals into "constitutional" versus "facultative" homosexuals (e.g., Bell, Weinberg, and Hammersmith 1981). Such a division is, in essence, an extreme position in the general "nature" versus "nurture" debate. For most mammals, social behaviors are based on a range of constitutional propensities interacting with a range of facultative opportunities; behaviors are condition-depen- dent and not typically fixed in individuals within pop- ulations but fluid as individuals' life histories accumu- late. Sexual orientation is not bimodal (contra LeVay kirkpatrick Evolution of Human Homosexual Behavior F 391 and (3) that homosexual behavior is typically seen in individuals of low reproductive potential. reduced individual reproductive success Present information is equivocal on the prediction of reduced reproductive success for individuals. Homosex- uals do not, by definition, have offspring, but most people who engage in homosexual behavior are bisexual (see, e.g., table 2). There is as yet no compelling evidence that the number of surviving offspring or even fecundity is limited by homosexual behavior as seen in the majority of individuals practicing it. Many individuals who engage in homosexual behavior have children. In a sample of contemporary British women (n = 3,180), bisexuals have significantly higher fecundity to age 25 and no significant difference in life- time fecundity when compared with heterosexuals (Baker and Bellis 1995). Of approximately 265 homosex- ual and bisexual men over 30 years old in contemporary Japan, 83% have offspring (Isomura and Mizogami 1992). Quinn (1996) provides data for lifetime fecundity (within the confines of marriage) for 13 men, all 19th-century Mormons known to have participated in consensual ho- mosexual sex. On average, 2.1 children were born to these men, with a range of 0 to 9. (Comparison of these numbers with a general sample of 19th-century Mor- mons could test for statistical differences.) In some so- cieties, many individuals—probably most individu- als—who produce children also engage in homosexual sex (e.g., 15th-century Florence [Rocke 1996], 17th-cen- tury Japan [Schalow 1989, Leupp 1995], Melanesia [Herdt 1984a]). There are limited data in support of this prediction, however. In Davis's (1929) sample of married U.S. women (n = 1,000), those having "intense emotional re- lationships" with other women tend to be childless more often than women without such relationships (39% of 18 women versus 21% of 982 women, Yates-corrected x2 = 2.40, d.f. = 1, p = 0.121). (Only 8 of the 18 women reporting "intense emotional relationships" are known to have had homosexual sex [see Davis 1929: chap. 11, table 4].) In a random sample of contemporary U.S. women (n = 298), the mean number of children born to women with homosexual experience is 1.2 versus a mean of 2.2 for women without homosexual experience (sig- nificance not reported [Essock-Vitale and McGuire 1988]). In some societies homosexual behavior is toler- ated only so long as procreative duties are fulfilled (e.g., 17th-century Fujian, China [Ng 1989], contemporary Thailand [Jackson 1989]), suggesting that parents may oppose homosexual behavior when it reduces a lineage's reproductive success. increased reproductive success for lineages If homosexuality occurs because of lineage benefits at the expense of homosexuals themselves, there will be greater offspring survival in lineages with homosexuals. Although there are no studies showing such ultimate 1996; e.g., see Johnson et al. 1994). Admittedly, the data are not clear on whether there is one type of human sexuality that ranges from homosexual to heterosexual or two types of sexuality—homosexual and heterosex- ual—that blur only when facultative opportunities do not match constitutional desires. However, while it is unclear whether sexual orientation is a dichotomous var- iable, it is clear that sexual behavior is a continuous one. And, as Small (1995:149) notes, "[A] strategy isn't a strat- egy if it's rarely implemented; it's just an unfulfilled wish. And evolution doesn't recognize unfulfilled wishes. In other words, what people say they want is not important, it is what they get that is of real significance [to natural selection]." Kin Selection and Parental Manipulation: Predictions and Evidence Humans have few children over a lifetime and invest heavily in each child. Reproductive success and lineage survival therefore depend more upon the quality and quantity of child care than on gross number of children. If a homosexual sufficiently increases the reproductive success of kin, then the genetic material transmitted to the next generation via kin and held in common with the homosexual by virtue of relatedness may offset the homosexual's lack of offspring. This may come from di- rect support of offspring, direct support of the lineage, or indirect support by not producing competitors to the lineage's designated heirs. Wilson (1975) suggests that if homosexuals in non- technological societies function as helpers, kin selection can maintain genes for homosexuality. Alternatively, nonreproductive individuals may increase the reproduc- tive success of parents, but the inclusive-fitness gains may not be enough to offset the losses to nonreproduc- tive individuals themselves (Alexander 1974, Trivers 1974). If so, parents may manipulate selected offspring to forsake sexual reproduction, become homosexual, and assist in raising siblings or the offspring of siblings. An implicit assumption of the proponents of both hypoth- eses has been that most individuals engaging in homo- sexual behavior are homosexuals and therefore homo- sexuality is the behavior to be explained. Kin selection and parental manipulation are con- founded with regard to homosexuality because the two hypotheses have a number of predictions in common. Dickemann (1995) joins the two under the more general hypothesis of "inclusive fitness" and suggests that since one need not be homosexual to assist kin, "the inclusive fitness hypothesis is a hypothesis about celibacy and nonreproduction. Whether and when celibates are ho- mosexual . . . are subsidiary questions" (Dickemann 1995:174). Three central predictions of both hypotheses are (1) that homosexual behavior reduces individual re- productive success, (2) that lineages with homosexuals have greater reproductive success than lineages without, 392 F current anthropology Volume 41, Number 3, June 2000 differences, the proximate mechanism presumably would be more or better aid to kin's offspring from ho- mosexuals than from heterosexuals or celibates. The data here are weak and not directly to the point. The most that can be said is that individuals engaging in homosexual behavior at times help their families. In the marriage resistance movement of southern China, two or three individuals would sometimes swear friend- ship, and this often led to homosexual behavior (Sankar 1986). Women in the movement were relatively wealthy from work in the silk industry and regularly sent wages back to their natal families; some parents encouraged daughters to join the sisterhood as a means of increasing lineage income. A different source of lineage support might be betrothal payments in homosexual marriages (e.g., 17th-century Fujian [Ng 1989], the Azande of Sudan [Evans-Pritchard 1970]). In the Siwah of Libya, boys com- manded a higher betrothal price than girls (Maugham in Murray 1997). In native North America, transgenders could provide lineage support through their roles as matchmakers and as mediators of disputes; they also had access to restricted technologies such as basket making and pottery, and this provided them material gain (e.g., the Navaho of North America [Hill 1935]). Data are unavailable, however, on the amount of care given by homosexuals to the offspring of kin. Arguments in support of this prediction rely on the potential of ho- mosexuals to aid their lineages. Even if it can be shown, for example, that North American transgenders received benefits in their roles, this is not the same as benefit to kin (Dickemann 1995). typical appearance in individuals of low reproductive potential If individuals forgo reproduction for the benefit of kin (altruistically with kin selection, under duress with pa- rental manipulation), these individuals should be those with a low probability of producing offspring directly. Factors associated with reproductive potential include birth order, biological sex, and demographic sex ratios. The data provide some support for this prediction. Birth order clearly has implications for the role one takes (or is assigned) in one's lineage. Also in keeping with this prediction, current information suggests more male than female homosexual behavior. There are as yet no sys- tematic studies to show that demographic skews lead to homosexual behavior, however. Birth order. Particularly in societies with unigeniture, sanctioned reproduction is reserved for those of high birth order (e.g., 15th- and 16th-century Portugal [Boone 1988]). Individuals of low birth order presumably have less opportunity for successful reproduction (e.g., in the priesthood, the military). Self-identified male homosex- uals in the United States and Canada tend to have more older brothers than do self-identified male heterosexuals (Blanchard and Bogaert 1996a, b), and this implicitly sup- ports the prediction that homosexuality correlates with low reproductive potential. Gender-role "reversal" of some individuals engaging in homosexual acts is com- mon in some societies—although this role reversal does not apply to their heterosexually married partners (e.g., North and South America [Williams 1986; Murray 1995a, b; Lang 1998], precontact Hawaii [Morris 1990]). Therefore, it is of interest that among the Lache of Co- lombia and in societies of the West Indies in the 1800s, sons of very low birth order were at times raised as daughters (Me ́ traux and Kirchhoff 1948, Waitz in Green- berg 1988). Further, later-born females apparently are more likely to be raised as sons among the Kaska of North America (Honigmann 1954). Such reports are po- tentially in keeping with the prediction that low birth order is correlated with homosexual behavior, particu- larly for males. Biological sex. Males have greater variance in repro- ductive success than do females (Trivers 1972), and if homosexuality results from low reproductive potential there should be more male than female homosexuals. This may be the case. Roughly twice as many males report homosexual behavior as do females in the con- temporary United States, the United Kingdom, and France (Sell, Wells, and Wypij 1995; see also table 2). Cross-culturally, male homosexual behavior is reported more frequently than female homosexual behavior (e.g., classical Athens [Dover 1989], 16th- and 17th-century Japan [Leupp 1995], native North America [Callender and Kochems 1983, Lang 1998]). This may be due in part to reporting bias (Herdt 1984a, Blackwood 1986b). Reduced homosexual behavior in females may also result from greater social control of female sexuality (Small 1995). In contemporary Britain, women and men report similar levels of homosexual attraction, but women report only half the level of homosexual behavior (Johnson et al. 1994). The control of female sexuality is variable across cultures, however, and there are as yet no data indicating a negative correlation between female homosexual be- havior and the control of female sexuality. Demography. Several writers suggest that less-pow- erful males engage in homosexual behavior primarily be- cause more-powerful males have multiple wives and thus reduce the availability of other-sex mates (e.g., Le ́ vi- Strauss 1943, Wilson 1959, Evans-Pritchard 1970). A study of 70 cultures randomly drawn from the Human Relations Area Files (HRAF) does not, however, find po- lygyny correlated with male homosexual behavior (Bar- ber 1998). It has also been suggested that organized ho- mosexual behavior in Melanesia may relate to imbalanced sex ratios or the spatial segregation of the sexes, but no systematic comparisons have been made (Davenport 1977, Herdt 1984a; see also Knauft 1987, Spain 1992). The implicit assumption is that homosexual acts occur only when heterosexual opportunities are ab- sent, but the data do not support such an assumption. Polygynous males in the Big Nambas of Melanesia have many male lovers, at times preferring them sexually to their wives (Deacon 1934). Homosexual behavior is also reported as more compelling than heterosexual behavior in numerous societies, including the Pukhtun of Paki- stan, migrant miners of several societies in South Africa, kirkpatrick Evolution of Human Homosexual Behavior F 393 and the classical Athenians (Lindholm 1982, Moodie, Ndatshe, and Sibuyi 1989, Dover 1989). Differentiating the Kin Selection and Parental Manipulation Hypotheses The hypotheses of kin selection and of parental manip- ulation differ in the mechanisms they envision (Trivers 1974). With kin selection, individuals should become ho- mosexual of their own accord. With parent-offspring con- flict, homosexuality is induced when parents channel resources or socialize offspring in ways that make them less competitive in reproductive roles. There is only anecdotal support for the prediction that parents manipulate selected offspring to take nonrepro- ductive roles and to become homosexual. Clearly, par- ents judge their children's reproductive value (e.g., co- lonial India [Dickemann 1979], 17th-19th-century Germany [Voland 1989]) and at times encourage children in homosexual behavior. Parents in 15th-century Flor- ence encouraged homosexual relations between their sons and the leaders of influential lineages, apparently to increase their political power (Rocke 1996). This is similar to paternal involvement in the selection of male patrons for sons in classical Athens and in Melanesia. Parents also manipulate selected offspring to take non- reproductive roles—castrating sons to qualify them for positions as eunuchs in the Byzantine court, for example (Ringrose 1996). In native North America, parents played significant roles in the selection of children for appren- ticeship as transgender shamans, although typically claiming such selection as a response to the child's in- nate disposition (Williams 1986; cf. Callender and Ko- chems 1983). Mohave transgenders were foretold by mothers' dreams, for example (Devereux 1937). Among the Araucanians of South America, male transgender shamans are "generally chosen for this office when they are children, and a preference is always [shown] to those, who at that early time of life discover an effeminate disposition" (Falkner 1935:117 [1774]). Data are absent, however, on whether nonreproductive, homosexual chil- dren significantly help the other children of their lineage. Alliance Formation: Predictions and Evidence The hypothesis of alliance formation through reciprocal altruism provides predictions that clearly distinguish it from both kin selection and parental manipulation. As pointed out by Trivers (1985:198), "the sexual and ro- mantic side of homosexual relations would seem to in- terfere with kin-directed altruism: insofar as one is sex- ually attracted to another individual, one will naturally be inclined to invest some resources in intrasexual com- petition to gain this individual's favors. Should the re- lationship blossom into a love relationship, it will be natural to devote some of the same resources and energy that would go into a loving heterosexual relationship." Contests for the resources at stake in intrasexual com- petition—such as material goods or social sup- port—allow direct selection to act upon the propensity for homosexual behavior. Predictions of this hypothesis include (1) that same-sex alliances aid survival and sub- sequent sexual reproduction, (2) that homosexual behav- ior aids in alliance formation, and (3) that bisexuality is more common than homosexuality. (An implicit as- sumption of this hypothesis is that bisexuality is the behavior to be explained.) importance for individual survival If homosexual behavior has fitness benefits due to same- sex alliances, those alliances must help individuals es- tablish and maintain households and, consequently, help offspring reach reproductive age. Data are limited on how same-sex alliances influence reproductive success. In- dividuals who engage in homosexual acts receive bene- fits applicable to their reproductive lives, however. This prediction is strongly supported in that same-sex alli- ances are important in networks that aid in survival and, potentially, marriage exchange. Among the K'ekchi' Mayans of Belize, men with same- sex alliances in agricultural labor have significantly more children surviving past six months because of increased productivity (Berte 1988). (Berte does not specify the mechanisms by which K'ekchi' Mayan males maintain their alliances, although precolonial Mayan culture con- doned homosexual behavior among unmarried men and may still today [Greenberg 1988; see also Reina 1959, Gossen 1974, Sigal 1997].) For the Azande and other Af- rican societies with polygynous marriages, Blackwood (1986b) suggests that female homosexual behavior helps negotiate alliances that extend trade networks and en- hance economic security. Alliances in the marriage re- sistance movement of southern China were fundamen- tally about mutual economic support (Sankar 1986); same-sex unions of the early Christian era were designed to create socioeconomic units, solemnized in religious vows and civil contracts (Boswell 1994). (Sexual behavior was clearly a part of some of these alliances, although population proportions are unknown.) Pair-bonds be- tween (non-transgender) males in many societies of na- tive North America reduced variance in food intake and provided for cooperative defense. As described by the 18th-century missionary Lafitau (in Katz 1976:289), "they become companions in hunting, in war, and in fortune [and] they have a right to food and lodging in each other's cabin." Within age-graded ("patron/client") same-sex alliances in Melanesia, the younger client provides labor in the fields (and sexual services) while the older patron pro- vides food and education (Marind-anim [van Baal 1966]; Big Nambas [Deacon 1934, Guiart 1953], Kiman [Ser- penti 1984]). In precolonial Tahiti and Hawaii, the clients of powerful patrons gained prestige (Gunson 1964, Mor- ris 1990), as did clients in classical Athens and 15th- century Florence (Dover 1989, Rocke 1996). In ancient Crete, men without same-sex sexual partners were at a 394 F current anthropology Volume 41, Number 3, June 2000 social disadvantage (Boswell 1980). Same-sex sexual part- ners of the Japanese samurai gained both martial training and land (Leupp 1995). In classical Athens and in To- kugawa Japan, same-sex alliances were thought to be more dangerous to entrenched interests than female- male alliances (Boswell 1980, Leupp 1995). Same-sex alliances with homosexual behavior often have exogamy rules and rituals similar to heterosexual marriages (e.g., Roheim 1933, Gunson 1964, Evans-Prit- chard 1970, Ernst 1991). This suggests that same-sex al- liances and heterosexual marriages are complementary. Le ́vi-Strauss (1943) treats as equivalent alliances based on the "reciprocal sexual services" of sister-exchange marriage and alliances based on homosexual behavior by classifactory brothers-in-law. Both female-exchange mar- riage and homosexual behavior in Melanesia link "in- dividuals and groups in complex chains of mutual de- pendency and obligation" (Lindenbaum 1984:345). Among the Etoro, for example, a patron's ideal client is his wife's younger brother, effectively linking wife and client in a sororal/fraternal "polygynous" marriage in which only one co-spouse is reproductively active (Kelly 1977). Among both the Jaqaj and the Onabasulu of Mel- anesia, homosexual alliances increase marriage exchange rights to the women controlled by one's partner (Boelaars 1981, Ernst 1991; see also Herdt 1984a). aid in the formation of same-sex alliances The prediction of aid in the formation of same-sex al- liances receives support, with the strong caveat that ho- mosexual behavior is not required for alliance formation. In some cultures, same-sex relationships that entail rights and responsibilities and are solemnized by public ceremonies or wealth exchange appear to be celibate (e.g., the Lovedu of southern Africa [Krige 1974], the ma- jority of 19th-century Mormons [Quinn 1996]). Same-sex alliances often do involve sexual behavior, however. Ho- mosexual behavior by adolescents appears to predict al- liances as adults among the Dahomey of Nigeria, the Nambikuara of Brazil, and the Barasana of Colombia (Herskovits 1938, Le ́vi-Strauss 1943, Hugh-Jones 1979; see also Sorensen 1984). Homosexual behavior also oc- curs as part of sworn friendships among the Khoisan of South Africa and the Dinaric of Serbia (Schapera 1930, Tomasic 1948). Sexual emotion can reinforce long-term supportive re- lationships (e.g., the Azande [Evans-Pritchard 1970]). Among the Kiman of Melanesia, homosexual behavior stops after adolescence but "nevertheless, a lifelong emo- tional relationship often results from homosexual rela- tionships" (Serpenti 1984:305); this also appears to be the case for the Onabasulu (Ernst 1991). (Homosexual be- havior does not appear to influence emotional ties at East Bay in Melanesia, however [Davenport 1977].) Among the Sambia, homosexual behavior occurs among initiates in a regional cohort of loosely joined militias. The Sam- bia are headhunters, often at war with neighboring groups; Herdt (1984b) argues that their homosexual be- havior solidifies bonds that are vital for mutual defense. bisexuality more common than homosexuality While the alliance-formation hypothesis predicts that self-motivated homosexual behavior typically enhances survival, it also predicts that homosexuality per se will be maladaptive in Darwinian terms. The alliance-for- mation hypothesis therefore predicts that bisexuals should outnumber homosexuals. The data strongly support this prediction. Most indi- viduals who engage in homosexual behavior are, in prac- tice, bisexual (e.g., Melanesia [Herdt 1984a], 17th-cen- tury Fujian [Ng 1989], classical Athens [Dover 1989], contemporary United States [Kinsey, Pomeroy, and Mar- tin 1948, Kinsey et al. 1953). Literary and biographical sources from 17th-century Japan and court records on sodomy accusations in 15th-century Florence suggest that the majority of the men in these two societies may have been bisexual (Schalow 1989, Leupp 1995, Rocke 1996). Across two millennia of records from China, in- dividuals known for homosexual activity are typically known for heterosexual activity as well (Hinsch 1990). In native North America, the majority of those engaging in homosexual sex were not transgenders (whose sexu- ality remains unclear) but gender-typical individuals who clearly were bisexual (Callender and Kochems 1983; Williams 1986, 1992a); the same is true with regard to Tahiti (Herdt 1997). Also in keeping with this prediction, homosexual emo- tion exists apart from homosexual behavior. Between 8 and 12% of Western men and women have homosexual attraction, though not behavior, after age 15 (Sell, Wells,and Wypij 1995; see also Davis 1929; Laumann et al. 1994: fig. 8.2; Pattatucci and Hamer 1995). Among both Australian and Malaysian medical students, most homosexual attraction is felt by those whose behavior is predominantly or exclusively heterosexual (Mc- Conaghy, Buhrich, and Silove 1994). This continuum of experience is implicit in the "sentimental effusions" re- ported by Le ́ vi-Strauss (1943) and others (e.g., Hugh-Jones 1979) when discussing same-sex sexual behavior that seems, to these investigators, to be unrelated to sexual gratification. Discussion homosexual behavior in humans In a Popperian sense, hypotheses cannot be proven, only soundly disproven. Further, even if it is shown that a particular behavior is adaptive, this falls short of showing that it is the product of natural selection (Caro and Bor- gerhoff Mulder 1987). In the final analysis, the available evidence does not allow rejection of any of the three hypotheses reviewed here (table 3). Kin selection and parental manipulation remain least satisfying, however, if only because the data are inadequate to test their cen- tral predictions. There is no strong evidence that ho- mosexual behavior, as seen in the majority of individuals kirkpatrick Evolution of Human Homosexual Behavior F 395 table 3 Predictions and Support for Three Evolutionary Hypotheses of Human Homosexual Behavior Prediction Kin selection and parental manipulation1. Homosexual behavior reduces individual reproductive success.2. Homosexuals provide more or better aid in raising offspring than do heterosexuals or celibates. 3. Homosexual behavior is seen predominantly in individuals of low reproductive potential. Parental manipulation1. Parents manipulate selected offspring to forgo reproduction, become homosexual, and aid in the rearing of siblings or the offspring of siblings. Alliance formation (reciprocal altruism)1. Same-sex alliances aid individual survival.2. Homosexual behavior aids the formation of same-sex alliances. 3. Bisexuality is more common than homosexuality. Support Equivocal Insufficient data Moderate Limited Strong Moderate Strong practicing it, reduces average reproductive success. There is no evidence for increased reproductive success in lineages with homosexual members. Although present information does not allow refutation, these two hy- potheses clearly do not have strong explanatory power. In particular, there is no evidence that homosexuals pro- vide greater service than celibates. A fundamental ques- tion remains why natural selection would select ho- mosexual nonreproductives when asexual nonre- productives would seem more efficient (Trivers 1985, Dickemann 1995). (Developmental processes may make a switch from "homosexual" to "heterosexual" easier than one from "asexual" to "sexual.") Nonreproductive offspring that are altruistic to sib- lings will always benefit their parents' reproductive suc- cess before reaching the threshold of benefit to their own (Trivers 1974). Parental manipulation is therefore more likely than kin selection if only because the threshold is lower. Parents manipulate children to accept nonre- productive roles; individuals who accept nonreproduc- tive roles potentially aid kin. With the possible exception of Native American transgenders, however, there is little evidence that (presumed) nonreproductive individuals are predominantly homosexual. The clergy in the early Christian era probably engaged in no more homosexual behavior than the laity, for example (Boswell 1980, Greenberg 1988). Present information rather suggests that parental evaluation of offspring reproductive poten- tial results in other behaviors, such as sex-biased infan- ticide, shunting of selected offspring to high-risk strat- egies (e.g., military service), or encouragement to form polyandrous households (Dickemann 1979, 1993; Voland 1989; Durham 1991). Alliance formation is the hypothesis that best explains the observations in the historical and ethnographic lit- erature. It does not depend on homosexual behavior's being detrimental to reproductive success or even require that lineages benefit. Same-sex alliances help individuals survive and subsequently reproduce, same-sex alliances are expressed sexually at times, and bisexuality is more common than homosexuality. Further, alliance forma- tion best explains the curious disjunction between the facultative expression of homosexual behavior and its genetic and developmental correlates. In many societies, people engage in homosexual behavior primarily at cer- tain stages of life, such as during adolescence or prior to heterosexual marriage, and then switch to relatively ex- clusive heterosexual behavior as adults or soon after household formation. This may be because same-sex al- liances aimed at somatic reproduction are more impor- tant early in life and female-male alliances aimed at sex- ual reproduction more important later. Also, the alliance-formation hypothesis is relatively parsimonious in that it does not posit a separation between the ho- mosexual behavior of humans and that of nonhuman primates, in which homosexual behavior is seen as one component of adaptive life histories (e.g., de Waal 1995). Reciprocal altruism often develops in long-lived, mu- tually dependent animals with repeated contact (Trivers 1971). When there is competition for social resources, it is important to prevent rivals from forming alliances with one's partners (Harcourt 1989). Bonds involving re- ciprocal exchange of lineage members, ritual ceremo- nies, and sexual emotion may help negotiate exclusive rights to alliance partners—which may explain the in- tense jealousy found in same-sex sexual partnerships (e.g., the Khoisan [Schapera 1930], the Big Nambas [Dea- con 1934], 17th-century Japan [Schalow 1989]). Layton (1989:435) suggests that "reciprocal altruism [explains] . . . much human action conducted in the idiom of kin" and that social kinship terms identify partners for recip- rocal exchange. This corresponds well to social kinship terms that regulate homosexual partnerships in many societies (e.g., the Nambikuara [Le ́vi-Strauss 1943], the Marind-anim [van Baal 1966]). The currencies and the costs and benefits of reciprocal altruism are complex. There are age-specific, sex-spe- cific, and culture-specific constraints to reproductive value, and these will vary with demographic and eco- logical conditions. The costs and benefits of altruistic acts are also influenced by asymmetrical limitations to reproductive success. Alliances within and between age/ 396 F current anthropology Volume 41, Number 3, June 2000 sex classes may therefore take different forms. In alli- ances among peers, the currencies of reciprocal altruism appear the same, and payoffs probably occur within short periods (i.e., over a number of years). In alliances between older patrons and younger clients, the currencies appear asymmetrical, and payoffs may span generations. The patron receives labor, the prestige of maintaining a social institution, and sexual gratification; the patron has, pre- sumably, already established a household and is within a reproductive system designed to ensure potential for fertilizations. The client is limited by access to both the knowledge and the resources that will allow establish- ment of a household. Nonreproductive sex is low-cost because the client is unlikely to be accepted socially as a member of the reproductive population. That patron/ client alliances are more common in males than in fe- males (Herdt 1997) may reflect greater intrasexual com- petition in males, with alliances holding reciprocal obligations such as access to mates (e.g., the Yanomamo [Chagnon 1982]). Selective pressure for alliances might occur when as- sisted labor increases individual productivity or when there is strong competition for defensible resources. In- creased fecundity can be detrimental to reproductive suc- cess when producing additional children now provides less benefit than investing in children already produced or amassing resources to support children one has a high probability of producing in the future (e.g., Hill and Kap- lan 1988). This leads to the prediction that same-sex al- liances will be seen in societies where investment in offspring is more strongly correlated with reproductive success than is raw number of offspring. Such societies include those with heavy sanctions on illegitimacy, those in predictable environments, and those in envi- ronments at carrying capacity. Culture is probably a more immediate influence on homosexual behavior than ecology, but the expression of homosexual behavior may be influenced, in part, by ecological variables (cf. Dickemann 1993). In societies where labor increases productivity, same-sex alliances may be based on considerations of labor (e.g., the Big Nambas [Deacon 1934]). In societies where status is rel- atively salient, alliances may be structured by status dif- ferentials (e.g., 15th-century Florence [Rocke 1996]). Lin- denbaum (1987) suggests that homosexual behavior in Melanesia is positively correlated with low population density, a marker of low-productivity habitats. In 70 cul- tures drawn randomly from the HRAF, homosexual be- havior occurs significantly more often in agricultural than in hunter-gatherer societies and more often in larger social groups (Barber 1998). This implicitly supports Dickemann's (1993) prediction that homosexual behav- ior will be seen more often when political networks rather than independent individuals are the prime social force. If patterns of social organization and alliance for- mation are predicted by ecological variables (Dickemann 1979, Flinn and Low 1986), then homosexual behavior should vary predictably with ecology. If, for example, certain forms of cousin marriage are predicted by eco- logical conditions that regulate male needs for coalitions (Flinn and Low 1986), male homosexual behavior should be predicted by such conditions (cf. Layard 1959, Lin- denbaum 1987). Again, if polygyny is predicted by eco- logical conditions and if female homosexual behavior strengthens economic security within polygynous mar- riages, female homosexual behavior should be predicted by such conditions. In some instances, alliances may be not between the patron and the client but between the patron and the client's parent, particularly if the patron is high in status (see Harcourt 1989). Parents in 15th-century Florence and in Siwah, Libya, in the early 20th century may have used their adolescent sons' sexuality to establish bonds with more powerful households (Rocke 1996, Murray 1997; see also Cline 1936). Among the Azande, the same- sex lover fulfilled kin obligations to the lover's parents (Evans-Pritchard 1970). When alliances result in the transfer of wealth to parents (e.g., betrothal payments in 17th-century Fujian [Ng 1989]), parents have an incen- tive to manipulate offspring alliances; they use children for their own purposes. As Hart and Pilling (1960:15) note, "in Tiwi culture daughters were an asset to their father, and he invested these assets in his own welfare. He therefore bestowed his newly born daughter on a friend or an ally, or on somebody he wanted as a friend or an ally." Parents in Tokugawa Japan sold both sons and daughters into indentured prostitution (Leupp 1995). In some Melanesian societies, initiation begins before puberty and is more properly considered pedophilic than ephebophilic, at least in its initial stages (e.g., the Sambia [Herdt 1984b]). In many societies, alliances are maintained without sexual behavior. Sexualization of alliances might be an- ticipated when commitment of social partners is partic- ularly important or competition for partners particularly severe. Sex can be a way of conferring pleasure— presumably a currency of long-term supportive relation- ships—and may signify to one's partner and to others a high level of affiliation and exclusivity. Insofar as main- taining an alliance has significant costs in terms of time, displays of commitment are adaptive because of the cost (cf. Zahavi 1975). Homosexual acts become powerful symbols of loyalty and affiliation. It remains unclear why sex is used to negotiate alliances that do not hold direct conceptive benefits. However, for some pri- mates—including humans—sex is rarely for procreation (Wrangham 1993, Small 1995, Manson, Perry, and Parish 1997). Wrangham (1993) estimates for bonobos that only 1% of copulations are conceptive. Sexual behavior and emotion can therefore be appropriated to establish and maintain alliances. Sexual behavior is not indispensable to same-sex alliances. Homosexual behavior is not at a selective disadvantage, however, as long as enough het- erosexual matings ultimately occur, on average, to pro- vide the individual with the average number of surviving offspring. Particularly in societies requiring heterosexual marriage, it is unlikely that homosexual behavior results in a net decrease to an individual's reproductive success (Weinrich 1987b). kirkpatrick Evolution of Human Homosexual Behavior F 397 homosexual behavior in the human ancestor Human homosexual behavior shows similarity across broad geographic regions. Organized homosexual behav- ior in Melanesia is thought to have originated in a culture that reached Melanesia perhaps 10,000 years ago (Herdt 1997). Widely divergent cultures across the Americas have a broadly similar although by no means unitary way of incorporating transgender homosexual behavior (Wil- liams 1986, Lang 1998), and this suggests that the orig- inal human migrants into the Americas 12,000 years ago had similar cultures. Homosexual behavior appears to have been part of the human behavioral repertoire at least since the beginning of the Holocene. In some species of nonhuman primates, homosexual behavior occurs at least as frequently as heterosexual behavior (reviewed in Vasey 1995). Abnormal captive conditions can induce homosexual behavior but do not explain the majority of observations. The strongest adap- tive explanation of homosexual behavior in nonhuman primates is the maintenance of social relationships (Va- sey 1995). In olive baboons at Gilgil, males use alliances in aggressive disputes; a primary form of expression of these alliances is holding the testes of alliance partners (Smuts and Watanabe 1990). In the Japanese macaques at Arashiyama, imbalances in demographic sex ratio ex- plain some female-female homosexual behavior, but this behavior also predicts alliances (Wolfe 1986, Vasey 1996). (These alliances appear primarily to maintain exclusive sexual access to same-sex partners [Vasey, personal communication].) In bonobos, homosexual behavior is integral to exchange networks that link sex with food (White 1989, Parish 1994). Immigrant bonobo females also use sex to solidify alliances with high-ranking resident females (Idani 1991). In gorillas at Virunga, homosexual behavior (including anal intromission) occurs between age-graded pairs in all-male units; silverbacks defend blackback sex- ual partners from the approaches of other silverbacks (Yamagiwa 1987). This intrasexual competition for same- sex mates is precisely the type envisioned by Trivers (1985). Interpretation of current evidence concerning early hominids suggests that their success was based on alli- ances with individuals of both sexes (Foley 1989). The genus Homo has extensively expanded its geographic dis- tribution over evolutionary time, presumably, in part, because of behaviors such as cooperative hunting and gathering, the separation of male and female spheres, and greatly enhanced transfer of knowledge between gener- ations. If such speculation is reasonably correct, then same-sex alliances almost certainly were key in hominid evolution. This should result, over evolutionary time, in a psychological predisposition for same-sex affiliation (cf. Trivers 1971). In nonhuman primates, homosexual be- havior is part of a complex network of reciprocal exchange. The great apes and humans all show homo- sexual behavior (Vasey 1995). We can never know the specific social behaviors of extinct hominids, but it is safe to speculate that homosexual behavior was part of their social repertoire. If sexual behavior indeed strength- ens a

the evolution of human physicl attractiveness

The Evolution of Human Physical Attractiveness Steven W. Gangestad and Glenn J. Scheyd Department of Psychology, University of New Mexico, Albuquerque, New Mexico 87111; email: [email protected], [email protected] Annu. Rev. Anthropol. 2005. 34:523-48 The Annual Review of Anthropology is online at anthro.annualreviews.org doi: 10.1146/ annurev.anthro.33.070203.143733 Copyright ⃝c 2005 by Annual Reviews. All rights reserved 0084-6570/05/1021- 0523$20.00 Key Words beauty, sexual selection, evolutionary psychology, evolutionary anthropology, costly signaling Abstract Everywhere the issue has been examined, people make discrimi- nations about others' physical attractiveness. Can human standards of physical attractiveness be understood through the lens of evolu- tionary biology? In the past decade, this question has guided much theoretical and empirical work. In this paper, we (a) outline the ba- sic adaptationist approach that has guided the bulk of this work, (b) describe evolutionary models of signaling that have been ap- plied to understand human physical attractiveness, and (c) discuss and evaluate specific lines of empirical research attempting to ad- dress the selective history of human standards of physical attractive- ness. We also discuss ways evolutionary scientists have attempted to understand variability in standards of attractiveness across cultures as well as the ways current literature speaks to body modification in modern Western cultures. Though much work has been done, many fundamental questions remain unanswered. 523 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. Contents INTRODUCTION................. 524 METATHEORY AND THEORY . . . 525 The Adaptationist Framework:Basic Concepts................ 525 Sexual Selection and Signaling Theory....................... 527 Are Honest Signaling Systemsof Quality Ubiquitous? ........ 529 Genetic versus Direct Benefits . . . . 530 Multiple Signals.................. 530 Mutual Mate Choice ............. 530 FEATURES ASSOCIATED WITH ATTRACTIVENESS: WHATDO THEY SIGNAL?............ 531 FacialSexualDimorphism........ 531 Facial Averageness................ 534 Facial Symmetry ................. 535 Female Body Form............... 536 Male Physique ................... 537 BODY MODIFICATION........... 538 CURRENT AND FUTURE DIRECTIONS .................. 539 The Precise Benefits Generating Preferences ................... 540 The Overlapping and Independent Contributions of Preferred Features ...................... 540 The Integration of Different Signals........................ 540 The Conditional Nature of Preferences ................... 540 The Moderating Effect of "Non-Physical" (or Nonstatic) Features ...................... 541 SUMMARY......................... 542 INTRODUCTION Intensive study of human physical attractive- ness in the social sciences was prompted by a serendipitous finding by Hatfield and col- leagues in 1966. These researchers randomly paired college men and women for a blind date. In return, participants filled out per- sonality and interest measures, background questionnaires, and a survey following the date. A key question was whether they were interested in a subsequent date with their partner. None of the researchers' own hy- potheses were supported. Rather, only one variable—added as an afterthought—reliably predicted interest: their partners' physical attractiveness. Research over the past four decades has demonstrated profound and broad-reaching implications of physical attractiveness for peo- ple's lives. As might be expected, attractive people have greater choice in mating mar- kets and hence are able to secure consen- sually more desired partners. But attractive people are treated differently from others more generally, leading them to have better jobs, higher incomes, and more friends than others—indeed, achieve more desirable out- comes in most spheres of life people consider important (e.g., Langlois et al. 2000). Physical attractiveness matters to mate choice not only among U.S. college students. Buss (1989) surveyed people from 37 cul- tures around the world about mate prefer- ences. The U.S. samples rated the importance of physical attractiveness only slightly above average for the cultures. In traditional groups as well, e.g., the Ache of Paraguay ( Jones & Hill 1993), the Shiwiar of Equador (Sugiyama 2004), the Machigenka of Peru (Yu & Shepard 1998), and the Hadza of Tanzania (Wetsman & Marlowe 1999), individuals have reliable standards of attractiveness. Indeed, a recent meta-analysis concluded that people in dif- ferent cultures generally agree on who is attractive (Langlois et al. 2000). Physical at- tractiveness appears to have important im- plications in traditional cultures too; at- tractive Ache women, for instance, have greater reproductive success (Hill & Hurtado 1996). Symons's (1979) Evolution of Human Sex- uality offered the first intensive analysis of physical attraction as an outcome of natu- ral selection. It inspired Buss's (1989) survey and spawned research on features considered 524 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. attractive cross-culturally (e.g., Jones & Hill 1993, Cunningham et al. 1995). Research guided by evolutionary thinking steadily in- tensified over the past 15 years. Since 1990, nearly 300 articles with the keywords "physi- cal (or facial) attractiveness" as well as "selec- tion" or "evolution" in their titles or abstracts have appeared in scientific journals. Those published in the five years since 2000 out- number those appearing in all of the 1990s (and, indeed, will probably soon outnum- ber those published in all of the twentieth century!). We first discuss the adaptationist frame- work that has guided evolutionary analyses of attractiveness, with emphasis on biolog- ical signaling theory. We then review re- cent research in light of this framework and conclude with key issues for future research. METATHEORY AND THEORY The Adaptationist Framework: Basic Concepts As a research strategy, adaptationism seeks to identify adaptations and elucidate the selec- tion pressures that forged them in an organ- ism's evolutionary past. In its current form, it crystallized in the 1960s and 1970s [heavily in- fluenced by the writings of George Williams (1966)] and now dominates the study of ani- mal behavior in biology (e.g., Krebs & Davies 1993, 1997). It also guides most evolution- ary work on human attractiveness. We review briefly some basic concepts and tools. (For a more extensive discussion, see also Andrews et al. 2002.) Traits. Biologists use the term trait to refer to aspects of organisms' phenotypes. A liberal definition allows a trait to be any aspect of the phenotype that can be discriminated on the basis of any criterion—its causes, effects, appearance, etc.—and includes dispositional traits (e.g., the disposition to develop calluses with friction; the disposition to choose mates with particular qualities). Adaptation. The word adaptation has two meanings in evolutionary biology (Gould & Vrba 1982). It is the process by which natural selection modifies the phenotype and gener- ates traits whose effects facilitate the propa- gation of certain genes over others, thereby causing evolution. It also refers to the end- products of that process—i.e., the traits that have been constructed by a process of phe- notypic modification by natural selection for a particular gene-propagating effect. An ef- fect refers to the way, or ways, in which an aspect of the phenotype interacts with the en- vironment (e.g., Williams 1992). At least one effect of an adaptation is beneficial in an evo- lutionary sense; this effect is referred to as the adaptation's function. A function is an ef- fect that caused the trait to evolve—that is, an effect that enhanced the reproductive suc- cess of the trait's bearers over others lacking the trait. Crudely put, the primary function of eyes is seeing; the primary function of wings is flight. Byproducts. A small minority of traits qualifies as adaptations. Byproducts (or span- drels; Gould & Lewontin 1979) evolved merely because they were linked with other traits that had beneficial effects and hence were carried along with selection for adap- tations. For example, vertebrate animals evolved skeletal systems of calcium phosphate (perhaps ancestrally for the function of stor- ing calcium, but modified for the function of structural rigidity necessary for locomotion; e.g., Ruben & Bennett 1987). Calcium phos- phate is white and hence so too are bones. The whiteness of bones, however, is not an adap- tation. It is a byproduct of selection for other properties of calcium phosphate. Energetic trade-offs. Organisms can be thought of as entities that capture energy from the environment (e.g., through foraging, hunting, or cultivating) and allocate it to re- production and survival-enhancing activities. Energy does not come for free. Were indi- viduals able to expend unlimited energy at no www.annualreviews.org • Evolution and Physical Attraction 525 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. cost, in principle they could evolve to grow and develop so rapidly they could begin re- producing immediately after birth, massively produce offspring, and preserve themselves such that they never age. In biological reality, however, individuals are constrained by finite energy "budgets" (themselves earned through energy and time expenditures). Allocation of finite energy and time budgets entails trade- offs and hence forces decisions about the rel- ative value of possible ways to spend. Individ- uals should be expected to evolve adaptations that lead them to allocate their budgets in ways that enhance their fitness. Hence, for exam- ple, selection does not favor ever-increasing clutch sizes in birds. Rather, selection presum- ably favors strategies of producing and raising offspring that maximize the overall reproduc- tive success of offspring through an optimal balance of offspring number and quality, as constrained by energy and time budgets (see, e.g., Seger & Stubblefield 1996, Kaplan & Gangestad 2005). As we emphasize later, a key notion in the study of attractiveness is that the amount of effort that individuals can dedicate to a biological signal found attractive by the other sex is constrained by their energy budgets. Cost-benefit modeling. Adaptationists often use quantitative optimization models to try to understand how selection operates on phenotypes, given trade-offs (e.g., Seger & Stubblefield 1996, Winterhalder & Smith 2000). A model has one or more actors (e.g., two or more individuals engaging in social interaction, etc.) expressing the phenotypes that the theoretician is trying to understand. The payoffs of the model are expressed in a currency, such as actual fitness units or some correlate of fitness (e.g., units of energy). The decision set is the suite of phenotypic or behavioral options available for pursuing payoffs, and selective constraints delineate how these options translate into costs and benefits (typically dictated by constraints of energy and time budgets). The optimal phenotypic or behavioral option is the one that maximizes net benefits under the constraints. Complexities can be taken into account. For instance, optimal strategies may be contingent on the condition or phenotype of the individual (conditional or phenotype- limited optima). Hence, as we'll see, an important theme of biological signaling theory is that the optimal intensity of a signal may vary as a function of individuals' own condition. (See Parker & Maynard Smith 1990). The special design criterion for identify- ing adaptations. Evolutionary biologists are interested in understanding adaptation. Because adaptations don't come with labels, however, their identification requires infer- ence. The major criterion used is special de- sign (e.g., Williams 1966; Thornhill 1990, 1997). A feature exhibits special design if it proficiently performs a specific function; and moreover, it is difficult to imagine any alterna- tive evolutionary process that would have led to the feature or its details other than selec- tion for that function. For instance, eyes are good for seeing and it's difficult to imagine an evolutionary process that would have led to many of their details other than selection for their optical properties, that is, for see- ing. Evidence for design is not only evidence that adaptation has been at work; it's evidence that the organism has been shaped by par- ticular selection pressures. Just as wings and eyes are telltale signs of historical selection for flight and seeing, respectively, certain psycho- logical features of modern humans (e.g., pat- terns of attraction to mates and their features) may be telltale signs that particular selection pressures shaped the minds of ancestral hu- mans (e.g., selection for a signaling system due to particular benefits associated with the signal). Attractiveness and beauty as an outcome of adaptation. Individuals find other in- dividuals attractive as a result of the lat- ter possessing specific favored traits. Obvi- ously, however, attractiveness reflects traits in 526 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. perceivers as well: Perceivers have traits lead- ing them to find some features more attractive than others. Adaptationist researchers have typically adopted the working hypothesis that perceiver traits are adaptations—evolved as a result of their having benefits for perceivers— or, as the title of an article by Symons pro- claims, "Beauty is in the adaptations of the beholder" (1995). Generally speaking, tendencies to be at- tracted to particular traits in individuals of the opposite sex are presumed to have ben- efited their bearers because individuals' own reproductive success (or that of their off- spring) is affected (or, more precisely, was affected ancestrally) by qualities of the in- dividuals with whom they mate. In broad terms, evolutionary biologists delineate two types of benefits that mates provide: first, ge- netic benefits to offspring—i.e., endow off- spring with superior ability to survive; sec- ond, material benefits to the perceiver such as food, care for offspring, physical protec- tion for self or offspring, or avoidance of dis- ease. Individuals disposed to mate with others who enhanced their own fitness would have, ceteris paribus, out-reproduced those not so disposed. Hence, selection may have favored dispositions to be attracted to mates who possessed qualities that signal (or ancestrally signaled) delivery of benefits. Adaptationist researchers typically adopt a working hypoth- esis that many of people's tendencies to find specific features attractive are outcomes of this kind of historical selection (though see our discussion of alternative, sensory bias models below). Once individuals of one sex prefer partic- ular qualities, the preferences exert selection pressures on the preferred traits. Individuals who possess preferred traits, by virtue of their enhanced ability to exercise choice in a mat- ing market, have greater reproductive success. Selection hence leads individuals to expend energy and time to display favored traits. As Charles Darwin himself recognized (1871), such selection can in theory lead to the evo- lution of extravagant ornaments, even those detrimental to an animal's viability, provided they gave a sufficient advantage in the realm of mating. In turn, selection may come to ex- ert pressures on perceivers' greater ability to discern truly favored traits. Preferences and the traits they prefer coevolve. Sexual Selection and Signaling Theory Physical traits individuals are selected to find attractive may be thought of as signals of un- derlying qualities. The coevolution of pref- erences and preferred traits, then, can be thought of as the evolution of a signaling system, which entails that one sex (signalers) possess signals and the other sex (receivers) possess psychological (cognitive and motiva- tional) capacities to perceive and act upon (e.g., be attracted to) those signals. (In a given species, both sexes may evolve preferences and hence be receivers in signaling systems, of course.) Signals of underlying quality or condition have received the greatest attention from biological signaling theorists. Quality or condition refers to an individual's ability to successfully interact with the environment to acquire and effectively expend energetic re- sources (e.g., Rowe & Houle 1996). Superior condition has been associated with the con- cept of health (e.g., Grammer et al. 2003). The concept of health it implies, however, is much broader than simply the absence of dis- ease; it implies greater phenotypic fitness or resourcefulness. Indeed, as discussed below, in particular circumstances individuals of su- perior condition may even be more prone to disease than others. The term health, then, is generally a poor stand-in for biologists' no- tion of condition. Individuals in superior condition may make better mates for a variety of reasons: fit- ter genes to pass on to offspring (e.g., a relative absence of mildly harmful mutations; Houle 1992); greater ability to provide material ben- efits such as protection or food; greater fertil- ity and ability to reproduce (e.g., more viable sperm in the case of males or greater ability to www.annualreviews.org • Evolution and Physical Attraction 527 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. conceive and carry offspring through gesta- tion and lactation in the case of females); and a relative absence of disease. General mod- els of signals of quality do not speak to the precise nature of the benefits associated with choosing a mate of superior quality, though researchers are often interested in determin- ing these benefits. A signaling system may be said to be at equilibrium when neither the signaling sex nor the receiving sex benefits from a change (i.e., in signal sent or preference exercised) given that the other retains its strategy. For a signal to be a valid indicator of one's quality at equilibrium, a reliable relation between the signaler's quality and the signal strength must persist. Zahavi (1975) introduced the idea that it is the very costliness of a trait that ensures its honesty. He specifically proposed that ani- mals may signal that they are of superior qual- ity with a "handicap"—a feature that imposes a cost on the individual. Zahavi did not pro- vide an optimization model of this process; his argument was a verbal one. The basic intu- itive notion is that individuals who can afford a large handicap must be more viable than indi- viduals who have smaller handicapping traits. Big signalers can afford to "waste" some of their viability and still have residual viabil- ity greater than that of small signalers, and this fact presumably renders the handicapping trait an "honest" signal of viability. (In this context, a "bigger" signal need not be larger. Rather, the term indicates greater cost for in- dividuals, on average, to produce. The cost itself may be due to the signal's size, its com- plexity, or any other characteristic requiring effort to produce. Costs can also be mediated socially, for some signals let competitors know that their bearers should be tested through competition.) Grafen (1990) was the first to quantita- tively model handicapping. He assumed that all individuals, regardless of quality, obtain the same fitness benefits from a particular level of a signal (however, see Getty 1998). The signal can evolve to display quality, according to this model, when the fitness costs (in the currency of mortality) associated with developing and maintaining a particular level of the signal are less for individuals of higher quality than for individuals of lower quality. In that instance, the size of the handicap that maximizes net fitness (benefits minus costs) is larger for indi- viduals of higher quality than for individuals of lower quality. The signal "honestly" con- veys fitness, then, simply because it is not in the interest of individuals of lower quality to "cheat" and develop a larger signal; the viabil- ity costs they would suffer exceed the fertility benefits they would derive from the increased signal size. Recent developments in honest signaling theory have furthered and revised our under- standing of it. Despite the intuitive appeal, systems of mate choice via signals of condi- tion need not imply that the biggest signals are sent by the most viable individuals. For signals to be reliable, higher quality individuals need only higher efficiency, i.e., a greater fitness return from a marginal increase in signal in- vestment (Getty 2002). Modeling has shown that individuals of highest quality may have the same, higher, or lower viability than small signalers at equilibrium, depending on spe- cific parameters of the system. A key parame- ter is the expense paid by receivers for prefer- ring those with big signals over others, most notably, the cost of search time. Individuals with strong preferences for a signal of a certain size may delay mating and hence lose valuable time reproducing. A preference may be very cheap in some species—e.g., lekking species in which males collectively gather and display to females, who can assess relative quality with- out sacrificing search time. In such species, a few males who "win" the display contest may garner nearly all the matings, which in turn boosts the intensity with which males capa- ble of sending strong signals will do so, some- times beyond the point at which high quality males have reduced their viability to that of their lower quality rivals. Oddly, then, qual- ity and mortality can actually become pos- itively correlated in a population, with the highest quality individuals dying, on average, 528 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. at younger ages than lower quality individ- uals (Kokko et al. 2002). When females pay relatively high costs for their preference in relation to its benefits, the system will not drive signal intensity to be as great (i.e., sig- nals themselves will be less costly at equilib- rium) and, furthermore, individuals of high- est quality are unlikely to invest in signals to the extent that they actually have lower viabil- ity than individuals of lower quality. Similarly, the association between quality and parasite load (disease-level) can be positive, negative, or negligible, depending on the system (Getty 2002). Are Honest Signaling Systems of Quality Ubiquitous? A model that does not assume an associa- tion between signal intensity and quality is the sensory bias model (e.g., Kirkpatrick & Ryan 1991). In this model, one sex has a bias to prefer individuals of particular qualities be- cause that bias has advantages in realms other than mating. For instance, redness may be preferred in features of mates because redness signals ripeness of fruit and a sensory bias to be attracted to redness spills over into domains other than food selection. (Here, the bias ap- plied to mates is a byproduct of an adaptation for food choice.) Fisher (1930) famously de- scribed a process whereby a small initial pref- erence ultimately leads to extreme traits and preferences through "runaway" selection. If a particular trait in one sex is preferred in mates due to sensory bias in the other sex, then genes disposing stronger preference for the trait could spread because they become linked with genes predisposing the preferred trait. In essence, genes for strong preference "piggyback" on the success of the preferred trait, leading to stronger and stronger preferences and more and more extreme traits. A distinct variant of a sensory bias model is the "chase-away" model (Holland & Rice 1998). It assumes that individuals of the choosing sex with a sensory bias nonadaptively applied to mate choice pay a cost for it (e.g., increased search time) and, hence, have lower reproductive success than those who are "re- sistant" to the bias. This purportedly leads to selection for more intense signals in the other sex, ones than overcome the resistance of a greater proportion of choosers, which in turn leads to selection for a higher threshold of re- sistance. Over time, the chase-away process results in extreme manifestations of the trait. Kokko et al. (2003) recently argued that runaway or chase-away processes due to small initial sensory biases are unreasonable mod- els of signaling equilibria. In each case, the signal that evolves is presumed to become in- creasingly costly. As costs increase, individuals of highest quality will be able to produce the signal more efficiently than others—precisely the condition in which a signal comes to honestly convey quality. Inevitably, then, sig- nals initially preferred merely due to sen- sory bias should become correlated with qual- ity through runaway or chase-away processes. Rather than conceptualizing sensory bias or chase-away models as competitors of honest signaling models, then, Kokko et al. (2003) argues that sensory bias may be the starting point of a process that leads to honest sig- naling. Such biases may explain how pref- erences for seemingly arbitrary traits—ones prior to the evolution of the signaling sys- tem were probably not associated with quality (e.g., a tail brighter or larger than average)— can get off the ground, such that these traits ultimately become correlated with quality. Honest signals of quality need not initially be uncorrelated with quality, however. Fea- tures may covary with quality prior to being signals because individuals of higher quality pay lower marginal costs for them—e.g., in some species, larger individuals or those bet- ter able to intrasexually compete may possess higher quality. Preferences for these traits (or natural correlates of them) may then evolve, which further intensifies them by increasing the benefit of investing in the development of these traits. As we discuss later, a number of features preferred by humans may well have www.annualreviews.org • Evolution and Physical Attraction 529 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. correlated with quality prior to their evolution as signals. Genetic versus Direct Benefits Again, honest signaling of quality can evolve through either benefits that directly enhance reproductive success (e.g., food, protection, lack of contagious disease) or genetic benefits passed on to offspring. In some instances, both may account for the preference. For instance, males in a multi-male primate group bet- ter able to protect offspring than others and hence providing direct benefits to choosers may well possess genes associated with quality as well. In mating systems in which males and fe- males form socially monogamous (or mildly polygynous) pairs and males invest heavily in offspring (e.g., many bird species), direct ben- efits and indirect benefits may covary nega- tively. This should particularly be true when females are not sexually monogamous and offspring are frequently sired by "extrapair" males (also true of many bird species; e.g., Petrie & Kempanears 1998). In such cases, males of superior condition may optimize their allocation of effort by investing less in offspring and more in efforts to secure extra- pair matings (as they may succeed more of- ten than males of lower quality). Where di- rect and indirect benefits correlate negatively, either can be stronger; females may have a preference for males offering less direct ben- efit or less indirect benefit. Moreover, female preferences may vary depending on whether they are selecting a social partner or an ex- trapair mate. In collared flycatchers on the is- land of Gotland, for instance, females have no clear preference for males having a relatively large forehead patch, an honest signal of qual- ity(Qvarnstro ̈m1999).Whenchoosingextra- pair partners, however, females clearly prefer large-patched males. In theory, direct and ge- netic benefits of social partners negatively co- vary and cancel out. (Males with larger patches feed offspring less.) Extra-pair partners, how- ever, provide no direct benefits and, hence, choice is based only on males' ability to pro- vide genetic benefits. Multiple Signals In many species, mate choosers attend to mul- tiple signals, sometimes transmitted through different sensory channels (e.g., both olfac- tory and visual cues). One possible explana- tion is error in signal transmission; multi- ple signals of the same quality evolve because each adds incremental information that oth- ers do not (e.g., each signal or its perception is not perfectly correlated with quality; Grafen & Johnstone 1993). Alternatively, such mix- ing might be explained through a combina- tion of constrained transmission time and a need to convey abundant information (Endler 1993). As a parallel, one might imagine the use of both hand gestures and oral instructions in giving directions to a driver stopped at a red light. A third explanation is that signals carry information about different qualities altogether. Mutual Mate Choice In many species, sexual selection on signals is much greater in one sex than the others. In mammalian species, females are typically a limiting reproductive resource and, hence, males compete through intrasexual competi- tion and through signaling for females. Se- lection on female signaling is much weaker, as males may seldom turn down sexual op- portunities with females. In species in which both males and females invest substantially in offspring, however, mutual mate choice may evolve (e.g., Kokko & Johnstone 2001), whereby both sexes are selected to display de- sired mate qualities. Although the extent to whichhumanmaleassistancetooffspringhas historically functioned to increase offspring quality or, alternatively, to increase access to mates is hotly debated (e.g., Kaplan et al. 2000, Hawkes et al. 2001), in any case mutual mate choice appears to characterize human cultures. Both men and women discriminate 530 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. the desirability of potential mates, partly on the basis of physical qualities. FEATURES ASSOCIATED WITH ATTRACTIVENESS: WHAT DO THEY SIGNAL? Researchers have focused attention on three facial features found attractive across many cultures: sexual dimorphism, averageness, and symmetry. In addition, researchers have ex- amined the effect of several nonfacial bod- ily features on attractiveness, most notably, female waist-to-hip ratio and male body type. Facial Sexual Dimorphism Men's and women's faces differ in a num- ber of ways. On average, men's chins are longer and broader than women's. Devel- opment of the brow ridge renders men's eyes smaller (relative to total face size) and narrower than women's (e.g., Gangestad & Thornhill 2003a). Women's cheekbones are more gracile and their lips fuller. During ado- lescence, testosterone promotes growth of the lower face (see Swaddle & Reierson 2002). Estrogens may cap growth of bones during puberty, contributing to sexual dimorphism as well. Despite variation in facial propor- tions across human groups, sex differences ex- ist wherever they have been examined (e.g., Jones & Hill 1993). We refer to the aggregate differences between men's and women's faces as facial masculinity versus femininity. The attractiveness of female facial fem- ininity. Facial femininity is attractive in women; highly attractive women's faces are more feminine than average (e.g., Johnston & Franklin 1993, Perrett et al. 1994). This find- ing has been replicated in a wide variety of human groups (e.g., United Kingdom, Japan, Russia), including traditional South American groups with little to no exposure to Western standards of beauty (e.g., Ache; Jones & Hill 1993). Men prefer female faces with relatively small chins, large eyes, high cheekbones, and full lips (e.g., Cunningham 1986). Several theories explaining men's attrac- tion to femininity have been offered. Facial femininity reflects babyness, preferred due to sensory bias. An early theory is that fem- inine features such as large eyes and small chin reflect "babyness" (Cunningham 1986, Johnston & Franklin 1993, Jones 1995). Peo- ple may be disposed to respond to babies with care and, hence, be attracted to baby-like fea- tures (Jones 1995). Ancestral females who ex- ploited this preference may have been advan- taged over those who didn't, driving the sexual dimorphism itself. This theory encounters two major prob- lems. First, some attractive feminine features are not baby-like. Babies have puffy, protrud- ing cheeks (e.g., McArthur & Apatow 1984), whereas men find gracile, high cheekbones at- tractive in women's faces (e.g., Grammer & Atzwanger 1994). Second, as noted earlier, even a sensory bias model should expect that, as a trait is driven to be more extreme, some individuals should be better able to display it and, hence, it should become correlated with quality (Kokko et al. 2003). Female facial femininity marks sex appropriate- ness. The view that attractive displays exag- gerate species- or sex-typical traits to foster choice of a species- or sex-appropriate partner was prominent in the first half of the twenti- eth century (see, for a review, Cronin 1991). It might also expect women to particularly prefer highly masculine faces, which as we discuss below is not the case. Furthermore, once again, selection due to a bias favoring a trait uncorrelated with quality should ulti- mately result in covariation between the trait and quality. Female facial femininity is a marker of repro- ductive value. Reproductive value (RV) is the expected residual reproductive success of an individual, generally based on age (Fisher www.annualreviews.org • Evolution and Physical Attraction 531 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. RV: reproductive value 1930). RV is maximal when women reach re- productive age and diminishes thereafter. If men ancestrally mated serially with the same partners throughout their reproductive ca- reer (e.g., Kaplan et al. 2000), men may have maximized their fertility by selecting mates with maximal RV (Symons 1979). As women age, their facial proportions become less fem- inized, probably due to accumulated exposure to androgens (see Thornhill & Gangestad 1993). Men's preference for facial femininity, then, could reflect selection for age-based RV. RV could account for selection and main- tenance of a preference favoring facial femi- ninity. Once again, however, women should have been selected to exaggerate and pro- long the period of facial femininity, with some women better able to do so than others, ulti- mately leading facial femininity to be a cue of quality. Facial femininity is a marker of female quality or condition. Female femininity may signal reproductive condition and ability to dedi- cate energy to offspring production. Women's production of estrogen and fertility status are designed to be sensitive to conditions that affect their ability to carry and lactate for offspring—their energy condition (total stored energy in fat), energy balance (resid- ual energy consumed minus expended avail- able for reproductive effort), and energy flux (total energetic output). Women have dimin- ished ovarian function when they do not have fat stored for reproduction, do not reliably take in calories that surpass energy expen- diture, or incur extreme (either very low or very high) energetic demands (e.g., Ellison 2001, 2003). Extreme instances result in amenorrhea, but normal variation affects the production of ovarian hormones, including estrodiol, and thereby ovarian function, such that female fecundity varies along a contin- uum. Energy balance and flux appear to have larger effects than energy status, at least in the United States (Lipson & Ellison 1996). Facial femininity does not change drastically with immediate circumstances, but it could index women's accumulated history of energy bal- ance and flux appropriate for reproduction. The advantages of attraction to women with such a history (at least ancestrally) could be twofold. First, it could be a cue of direct benefits, as it could reflect lower disease in- cidence, better development due to favorable levels of parental investment, greater ability to forage effectively, etc., all affecting cur- rent reproductive capabilities. Second, such a history could be associated with genetic benefits to offspring. Women with more ef- ficient metabolisms due to fewer mutations (and hence more favorable energy balance, controlling for energy intake) or genotypes more resistant to extant pathogens would pro- duce genetically more fit offspring. Evidence pertaining to whether women's facial femininity or attractiveness are as- sociated with health is mixed. Hume & Montgomerie (2001) and Henderson & Anglin (2003) found associations between fa- cial attractiveness and relative absence of past health problems and longevity, respec- tively (see also Langlois et al. 2000; com- pare Kalick et al. 1998). Rhodes et al. (2003) reported no association between rated fe- male facial femininity (in adolescence) and actual health records, whereas Thornhill & Gangestad (2005) found facial feminin- ity (measured from digitized photographs) to predict recent history of respiratory in- fections and antibiotic use. Body symmetry (aggregated across fingers, wrists, ears, el- bows, etc.) measures developmental stability, a lack of perturbations during development due to infection, mutations, and other stresses. Koehler et al. (2004) reported that women with greater body symmetry are perceived to have more feminine faces; Gangestad & Thornhill (2003a) found no association. No study has examined facial femininity in rela- tion to health in traditional societies exposed to ecological circumstances reasonably simi- lar to ones encountered by ancestral societies, in which the childhood mortality rates due to disease may typically have been 30-50% (e.g., Hill & Hurtado 1996) and energy budgets 532 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. were constrained. Moreover, no one to our knowledge has explored associations between ovarian function and facial femininity. Al- though we suspect that preference for female facial femininity has been maintained because of this trait's historical association with both RV and fecundity, the latter link requires ad- ditional evidence. Attractiveness in relation to male facial masculinity. One might suspect that, just as men find feminine faces attractive, women might be attracted to masculine faces. In fact, no clear preference exists; studies have found a female preference for somewhat masculine faces (Keating 1985, Johnston et al. 2001), preference for feminine faces (e.g., Perrett et al. 1998, Penton-Voak et al. 1999), and no systematic preference either way (e.g., Cun- ningham et al. 1990, Jones & Hill 1993, Swaddle & Reierson 2002). Male facial masculinity nonetheless co- varies with certain desired male traits. Across a variety of cultures, men with masculine faces are perceived to be and probably are (Mueller & Mazur 1997) socially dominant (e.g., Keating et al. 1981, Mazur et al. 1984). Kung San bushmen with broad chins and robust bodies have relatively high reproductive suc- cess (Winkler & Kirchengast 1994). One ar- gument is that testosterone promotes male mating effort, which partly involves male- male competition, and that men's condition affects the extent to which they can effectively dedicate effort to mating (in a way analogous to how women's condition affects their ability to dedicate effort to reproduction). Variation in condition therefore gives rise to variation in testosterone metabolism, which, during ado- lescence, causes variation in facial masculin- ity. Men with more masculine faces may be tested in male-male competition, a cost that may ensure its honesty as a signal of con- dition. Consistent with this interpretation, two studies have found associations between male facial masculinity and reported health (Rhodes et al. 2003, lower half of the distri- bution only; Thornhill & Gangestad (2005), Zebrowitz & Rhodes 2004). Gangestad & Thornhill (2003a) found that men with mas- culine faces have greater body and facial sym- metry, though Koehler et al. (2004) did not. Among childless men in a rural village in Dominica, body symmetry positively predicts testosterone levels (controlling for age; S.W. Gangestad, R. Thornhill, M.V. Flinn, L.K. Dane, R.G. Falcon, C.E. Garver-Apgar, M. Franklin & B.G. England, unpublished data). (Male facial masculinity may be weakly associ- ated with current testosterone level; Penton- Voak & Chen 2004.) As male testosterone varies as a function of a variety of factors, including mating status and paternity (e.g., Burnham et al. 2003), however, facial mas- culinity may better index levels during ado- lescence and early adulthood. If adolescence is characterized by relatively intense male- male competition, facial masculinity may nonetheless reflect condition. If male facial masculinity covaries with condition, as conjectured, why do women not consistently prefer masculine faces? Penton- Voak et al. (1999) speculated that women face a trade-off: Although more dominant and pos- sibly more fit, masculine men may be less willing to invest exclusively in partners and help care for offspring. Hence, just as female collared flycatchers do not particularly pre- fer males with large forehead patches as so- cial partners, women do not prefer more mas- culine men as mates. As this view expects, men with feminine faces are perceived to be warmer, more agreeable, and more honest than men with masculine faces (see Fink & Penton-Voak 2002). According to this view, attraction to male masculinity should have been shaped by selec- tion to be conditional—depend on conditions that affect (or ancestrally would have affected) the relative value of (possibly heritable) con- dition and paternal investment. Evidence suggests that it is. Preference varies with phase of the ovula- tory cycle. Women's mate preferences vary across their ovulatory cycles. When normally www.annualreviews.org • Evolution and Physical Attraction 533 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. ovulating (e.g., nonpill-using) women are close to ovulation (and hence fecund), they are particularly attracted to the scent of sym- metrical men (Gangestad & Thornhill 1998, Thornhill & Gangestad 1999, Rikowski & Grammer 1999, Thornhill et al. 2003), deep, masculine male voices (Putz 2005), and more confident, intrasexually competitive male be- havioral displays (Gangestad et al. 2004), par- ticularly (where it has been examined; e.g., Gangestad et al. 2004) when they evaluate men as sexual partners (their "sexiness") rather than as long-term mates. Changes in prefer- ences across the cycle may reflect female de- sign to weigh signals of heritable condition more heavily when they are fertile, particu- larly when selecting a sex partner. Interest- ingly, then, the face that women most pre- fer when close to ovulation is more mascu- line than the face most preferred when they are in the luteal phase [Penton-Voak et al. 1999 (United Kingdom, Japan), Penton-Voak & Perrett 2000 (United Kingdom), Johnston et al. 2001 (United States, Austria)]. Further- more, the effect is specific to female attraction to men as sex partners, not long-term social mates (Penton-Voak et al. 1999). Preference varies as a function of relation- ship context. The face women find most attractive in short-term mates is more mas- culine than the face they find most attrac- tive in long-term mates (Penton-Voak et al. 2003). More attractive women have a stronger prefer- ence for masculine faces. Little et al. (2001) reasoned that, attractive women need not trade-off male condition and investment as markedly as must unattractive women; mas- culine men should be more likely to invest in relationships with attractive women. Hence, attractive women should more strongly pre- fer facial masculinity. Studies by Little et al. (2001; using self-reported attractive- ness) and Penton-Voak et al. [2003; using others' ratings of facial attractiveness and waist-to-hip ratios (see below)] support this hypothesis. Preference for masculinity varies with cul- ture. Penton-Voak et al. (2004) proposed that women's preference for masculinity should have been selected to be sensitive to cues of the relative value of condition (and genetic bene- fits) and investment of male mates in their lo- cal ecologies. In Jamaica, infectious disease is more prevalent (due to higher parasite loads and poorer medical care) and male parental investment less pronounced than in the United Kingdom. They predicted and found that Jamaican women show greater prefer- ence for facial masculinity than do British women. Findings that women's preferences for fa- cial masculinity are conditional constitute provisional design evidence that women's at- traction to male faces have been shaped by how male condition and parental investment have traded-off. Facial Averageness In 1990, Langlois & Roggman published a highly influential study. They digitized pho- tographs of faces, then created composites of sex-specific sets of them. Raters of both sexes found the "averaged" faces to be at- tractive, consistent with earlier speculation by Symons (1979) and suggestive findings by Galton (1878). The finding has been repli- cated many times (e.g., Wehr et al. 2001), including in China and Japan (e.g., Rhodes et al. 2001a) and in the Ache (Jones & Hill 1993). This preference is not merely due to the fact that composite faces are symmetrical and have unblemished skin; people find av- erage face shape and morphology attractive (e.g., O'Toole et al. 1999, Rhodes et al. 1999, Valentine et al. 2004). As discussed above, some faces are more attractive than average faces. Nonetheless, extreme departures from average on even sexually dimorphic traits are not attractive. 534 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. Two main hypotheses may explain prefer- ence for averageness. A generalized sensory bias favors prototypes. Organisms may show preference for stimuli that are readily processed (e.g., Enquist & Arak 1994). People may build up cognitive prototypes of distinct categories of stimuli, representations consisting of average features, which are useful for dis- criminating new instances of the category. Stimuli that match a prototype well may then be preferred. Halberstadt & Rhodes (2000, 2003) found that people indeed find averaged instances not only of faces, but also of dogs, fish, birds, wristwatches, and automobiles, attractive. Although familiarity of the stimuli could account for preferences for average wristwatches and automobiles, it could not explain preferences for dogs, fish, or birds. They proposed that people have a preference for averageness per se, independent of familiarity, when judging the attractiveness of living organisms, which may reflect a preference for signals of quality. Averageness reflects quality. Departures from average may reflect the effects of ge- netic mutation, chromosomal abnormality, nongenetic congenital deformation, disease, or other factors affecting quality. Zebrowitz & Rhodes (2004) found that facial averageness predicted pubertal intelligence and adolescent health (the latter for women), but only in the lower half of the distribution. They therefore proposed that averageness (and other pur- ported indicators of condition, such as facial masculinity) discriminates average from "bad" genes (or poor condition) but not from "good" genes. In fact, however, this effect may not be robust, as the regression slopes for high and low averageness groups did not significantly differ. Moreover, a plausible alternative is that prediction at the high end of averageness is compromised because some nonaverage fea- tures indicative of good condition (e.g., sex- ually dimorphic features) are preferred. Ag- gregates of different signals may discriminate condition along a continuum, a possibility to be explored in future research. Facial Symmetry Bilateral asymmetry on features that, on aver- age within a population, are symmetrical may reflect perturbations occurring during devel- opment due to mutations, pathogens, toxins, and other stresses (e.g., Møller 1999). Ma- nipulations of symmetry of signals in some (but not all) other species affect attractiveness (see Møller & Thornhill 1998). Grammer & Thornhill (1994) found facial asym- metry to negatively predict attractiveness. Manipulations of symmetry generally en- hance attractiveness [Rhodes et al. 1998, 1999, 2001a (including samples from China and Japan); Perrett et al. 1999; Koehler et al. 2002; compare Noor & Evans 2003]. [Studies ma- nipulating symmetry by pairing left- or right- face mirror images, which look strange, have not found this effect. See Møller & Thornhill (1998).] Two empirical questions arise. The first is the size of the effect of symmetry on nor- mal variation of attractiveness. Whereas fe- male femininity and facial averageness reliably account for moderate amounts of variation in attractiveness, symmetry does not: Grammer & Thornhill (1994), Baudouin & Tiberghein (2004), Jones et al. (2004), Hume & Montgomerie (2001), and Scheib et al. (1999) report positive findings, whereas Rikowski & Grammer (1999), R. Thornhill & S.W. Gangestad (unpublished data), Rhodes et al. (2001b), Shackelford & Larsen (1997), Simmons et al. (2004), and Koehler et al. (2004) found null or mixed results. The cor- relation in most populations is probably in a predicted direction but weak (r < 0.2). A second question concerns the extent to which facial symmetry per se generates its association with attractiveness. Scheib et al. (1999) found that both facial and body asym- metry predict male facial attractiveness. But facial symmetry predicted just as well the attractiveness of half-faces, which possess www.annualreviews.org • Evolution and Physical Attraction 535 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. minimal cues of symmetry (see also Penton- Voak et al. 2001). Jaw size and prominent cheekbones (the first a masculine trait, the second not; see also Gangestad & Thornhill 2003a; compare Koehler et al. 2004) covaried with symmetry and was able to account for its association with attractiveness. Men with symmetrical faces may have healthier looking skin as well (Jones et al. 2004). Perhaps ironically, it is not clear that facial symmetry should be used as a cue of condi- tion. Single trait asymmetries are very weak indicators of underlying variation in develop- mental instability, with <10% of their vari- ation owing to it (Gangestad & Thornhill 1999, 2003b). Only by aggregating asymme- tries of multiple traits can researchers de- velop reasonably valid measures. Facial sym- metry reflects symmetry on multiple traits. Because the traits are developmentally linked, however, it is unclear how much nonre- dundant information about developmental stability individual facial features contain. Researchers have found near-zero to mod- est correlations between facial asymmetry and broad composites of body asymmetry (Rikowski & Grammer 1999, Scheib et al. 1999, Gangestad & Thornhill 2003a, Koehler et al. 2004). Quite possibly, facial symme- try weakly taps developmental stability. [In- deed, facial masculinity/femininity could pos- sibly be a better measure of it; Gangestad & Thornhill (2003a).] Enquist & Johnstone (1997) proposed that symmetry preferences may be a byproduct of generalization effects when individuals are exposed to asymmetrical variants of an ob- ject (e.g., faces) that are symmetrical on aver- age. Jansson et al. (2002) demonstrated that chickens repeatedly exposed to asymmetri- cal novel stimuli (around a symmetric mean) came to prefer symmetrical stimuli to which they had not previously been exposed. Little & Jones (2003) demonstrated a greater sym- metry preference for faces in normal orien- tation than for inverted faces, which can be explained by generalization (as people rarely see inverted faces). They also showed a pref- erence for symmetry even in familiar faces, which a prototype (though not a mere famil- iarity) account may explain. Simmons et al. (2004) found that, although faces are char- acterized by directional asymmetries (mean L > R or R > L differences in the population), these asymmetries do not affect attractive- ness; rather, deviations around them (atypi- cal asymmetries) do, a finding also consistent with either the perceptual bias account or the symmetry-marks-quality explanation. One finding favoring the symmetry- marks-quality explanation is that, just as attractive women have relatively strong mas- culinity preferences, they have strong sym- metry preferences (Little et al. 2001). At the same time, Koehler et al. (2002) found no ev- idence that normally ovulating women par- ticularly prefer symmetrical faces—as they prefer masculine faces—when near ovula- tion. Zebrowitz & Rhodes (2004) found an association of facial symmetry with child- hood intelligence but not health. In sum, we do not yet know the extent to which it drives attractiveness judgments and, to the extent that it does, what explains the preference. Female Body Form In 1993, Singh proposed that, although pref- erences with respect to female body weights vary cross-culturally, men universally prefer women with a low waist-to-hip ratio (WHR). Women have lower WHRs than do men, largely due to the fact that women tend to store fat in the hips as well as breasts, which is selectively available for gestation and lac- tation. It now appears that, across a wide va- riety of cultures (though see below), men do prefer a lower-than-average WHR (most pre- ferred typically being about 0.7, compared to a mean in most populations of about 0.75- .80; e.g., Singh 1993, 1994a,b; Singh & Luis 1995; compare Tassinary & Hansen 1998, but see Streeter & McBurney 2003). The primary benefit of this preference may ancestrally have been the same as a benefit of 536 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. a preference for feminine faces: Low WHRs reflect a history of energy balance and flux that promotes allocation of energy into re- productive effort. The proximate mechanism may involve estrogens. Indeed, Jasienska et al. (2004) found that, within a Polish population, women with lower WHRs and larger breasts have greater fecundity than other women, as assessed through precise measurements of es- trodial (E2) and progesterone. Quite possi- bly, reproductive women's tendency to store fat on the hips and breasts evolved as adapta- tions for creating a low center of gravity ap- propriate for carrying fetuses and babies and putting fat where it can readily be converted for lactation, respectively (e.g., Pawlowski & Grabarczyk 2003). As a correlate of fecundity, however, it may have coevolved as a signal of quality, with mating benefits possibly leading to some exaggeration and display. If facial femininity and attractive body shapes signal overlapping qualities, one might expect them to covary across women. Penton- Voak et al. (2003) reported a significant, mod- est correlation between facial attractiveness and WHR in a U.K. sample (r = −.28). By contrast, Thornhill & Grammer (1999) found a near-zero correlation (r = 0.013), despite correlations between attractiveness of distinct bodily features (e.g., face and back at- tractiveness). Similarly, R. Thornhill & S.W. Gangestad (unpublished data) found little cor- relation (or curvilinear relationship) of fe- male WHR with measured facial masculin- ity (r = −.07) or facial attractiveness (r = −.11). Body shape and facial femininity con- tain largely nonredundant information; future research should address their distinct sources of influence. The universality of a preference for low WHRs has been challenged. Men in two foraging or traditional societies have been claimed to largely disregard WHR and gener- ally prefer women viewed as relatively heavy: the Matsigenka of Peru (Yu & Shepard 1998) and the Hadza of Tanzania (Wetsman & Marlowe 1999). Sugiyama (2004) found a sim- ilar preference for large body size in the Shi- wiar of Ecuador. There, as in other forag- ing societies studied, however, women have higher WHRs (close to 0.9 on average) than do women in Western populations. Using a sample of figures with variation typical of the Shiwiar and controlling for weight, Sugiyama found that Shiwiar men do pre- fer smaller WHRs. In foraging societies, fe- male body weight may be a positive predic- tor of fecundity (e.g., Hill & Hurtado 1996) and obesity may rarely be a health problem (e.g., Brown & Konner 1987). Hence, men may use energy status (stored body fat) as a cue of fecundity and ability to lactate effec- tively. In Western cultures, energy status ap- pears to be weakly associated with fecundity (Ellison 2003) and hence men weight more heavily indicators of energy balance and flux (e.g., WHR). At the same time, they may prefer women of moderate body mass in- dex (BMI), argued to be a powerful predic- tor of female body attractiveness in Western samples (e.g., Tovee et al. 1999, 2002). A key unanswered question is whether and, if so, how selection has shaped adaptations for male preference for female body shapes or sizes to be conditional and depend on local ecological factors that affect predictors of fe- male fecundity. Sugiyama (2004) has proposed that men do have specialized adaptations for preferring women of high fecundity (see also Marlowe & Wetsman 2001), which have been shaped to be sensitive to local conditions, but the nature of such conditional inputs (e.g., food scarcity, distribution of body types, etc.) remains poorly understood. Cultural trans- mission processes (e.g., Boyd & Richerson 1985) may also play important roles, but these are not well understood either. Male Physique Scant research has addressed female prefer- ences for male body features. Dixson et al. (2003) found that women in both Britain and Sri Lanka prefer lean, muscular (meso- morphic) body types most, followed by av- erage and then skinny body types; heavy WHR: waist-to-hip ratio BMI: body mass index www.annualreviews.org • Evolution and Physical Attraction 537 Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. (endomorphic) body types are least preferred. Women also prefer men with broad shoulders, relative to waist or hip size (i.e., a "V-shaped" torso; see also Horvath 1981, Franzoi & Herzog 1987, Lindner et al. 1995, Mehrabian & Blum 1997, Hughes & Gallup 2003), aver- age WHRs (Singh 1995), and possibly chest hair (Dixson et al. 2003). To date, these pref- erences have been studied in few cultures. A number of potential benefits could ex- plain these preferences: physical protection, nutritional resources, and positive external- ities of male status, as well as indirect ge- netic benefits to offspring. Women particu- larly prefer muscularity in men as short-term (as opposed to long-term) partners (e.g., Buss & Schmitt 1993), and normally ovulating women may be particularly attracted to mus- cular men as short-term partners when near ovulation (Gangestad 2004), effects consis- tent with muscularity functioning partly as an indicator of genetic benefits (Frederick & Haselton 2004). Naturally, however, direct benefits cannot be ruled out. One question left begging to be asked is why women prefer masculine body traits (fos- tered by testosterone, which promotes muscle growth; see Ellison 2003) but do not system- atically prefer masculine facial features. Pos- sibly, though sharing some influences, body and facial masculinity do not signal precisely the same traits. Future research should exam- ine the independent influences on these traits to address why female preferences for them apparently differ. BODY MODIFICATION To the extent that one can produce the sig- nals that have been recurrently associated with high mate quality throughout the history of our species, he or she will activate, in po- tential mates, their psychological adaptations for choosing suitable mates. These adapta- tions are vulnerable to deception, and hu- mans have found innumerable ways of mod- ifying their bodies to just this effect. A wide range of methods, from the quotidian (diet and exercise) to the absurd (e.g., buttock im- plants), may be understood as techniques of enhancing the strength of mate value sig- nals. Men and women alike report using var- ious techniques of appearance enhancement for attracting and/or retaining mates (Buss & Shackelford 1997). Practically every inch of the body surface provides some indication of age, health, strength, and fertility. Although age, per se, may not be an important aspect of quality, age cues are also cues to health and fertility status. Substantial cross-cultural variation in the size of the sex difference notwithstanding, men throughout the world are more concerned with women's physical at- tractiveness than vice versa, and women who look relatively young are consistently rated as attractive (Buss 1989, Jones 1995). Among the ways people's faces reliably change with age are decreases in the visi- ble area of the eyes and the red area of the lips and an increase in the size of the nose ( Jones 1995). Duncan & Collison (2003) re- port that extracts from the toxic Atropa bel- ladonna (deadly night shade) were first used for pupillary dilation to enhance eye appear- ance over 2000 years ago. Today botulinum toxin is used for a number of cosmetic pur- poses, including lifting the eyelids. Blepharo- plasty (eyelid surgery) is the third most pop- ular cosmetic surgery in the United States (Meisler et al. 2000). Slightly less popular, but nonetheless routine, procedures include rhinoplasty and collagen lip injections. To some extent, a more youthful facial appear- ance can be achieved through makeup, but over a million cosmetic surgeries were per- formed in the United States alone in 1998, and the number is expected to rise in the coming decades (Meisler et al. 2000). The color, luster, and volume of hair all indicate age and health. The hair of younger women is judged to be of higher quality than that of older women, and it is primarily younger women who choose to wear their hair long (Hinsz et al. 2001). Hair growth tonics of varying efficacy have for hundreds of years been successfully marketed to men concerned 538 Gangestad ·Scheyd Annu. Rev. Anthropol. 2005.34:523-548. Downloaded from arjournals.annualreviews.org by University of Oregon on 03/06/08. For personal use only. with hair loss. Laser follicular stimulation and countless dietary supplements are purported to reverse or slow hair loss. Hair implants (plugs), toupees, strategic combing, and in a sense, shaving, are also popular ways to cover the problem. For those men and women who have no shortage of hair, but are unhappy with its appearance, the number of products to clean, condition, or color the hair defies a thorough cataloguing. Tiggemann & Kenyon (1998) describe a youthful feminine look as including a slim body, high taut breasts, and smooth hair- less skin. Dieting, exercise programs, weight loss surgeries and medications, breast aug- mentation and reduction surgeries, liposuc- tion, electrolysis, and shaving can be variously mixed and matched to achieve such an appear- ance. Removal of body hair by women in many societies is so widespread as to go unnoticed, except by its omission. A survey of female Aus- tralian students (Tiggemann & Kenyon 1998) showed that over 99% of female college stu- dents in the sample had shaved their leg or underarm hair at some point, and about 92% of the full sample, including high school stu- dents, regularly shave their leg and/or under- arm hair. The number one reaso

sex selection in humans

Beauty and the beast: mechanisms of sexual selection in humans David A. Puts Department of Anthropology, Pennsylvania State University, University Park, PA 16802, USA Initial receipt 21 April 2009; final revision received 23 February 2010 Literature in evolutionary psychology suggests that mate choice has been the primary mechanism of sexual selection in humans, but this conclusion conforms neither to theoretical predictions nor available evidence. Contests override other mechanisms of sexual selection; that is, when individuals can exclude their competitors by force or threat of force, mate choice, sperm competition, and other mechanisms are impossible. Mates are easier to monopolize in two dimensional mating environments, such as land, than in three-dimensional environments, such as air, water, and trees. Thus, two-dimensional mating environments may tend to favor the evolution of contests. The two- dimensionality of the human mating environment, along with phylogeny, the spatial and temporal clustering of mates and competitors, and anatomical considerations, predict that contest competition should have been the primary mechanism of sexual selection in men. A functional analysis supports this prediction. Men's traits are better designed for contest competition than for other sexual selection mechanisms; size, muscularity, strength, aggression, and the manufacture and use of weapons probably helped ancestral males win contests directly, and deep voices and facial hair signal dominance more effectively than they increase attractiveness. However, male monopolization of females was imperfect, and female mate choice, sperm competition, and sexual coercion also likely shaped men's traits. In contrast, male mate choice was probably central in women's mating competition because ancestral females could not constrain the choices of larger and more aggressive males through force, and attractive women could obtain greater male investment. Neotenous female features and body fat deposition on the breasts and hips appear to have been shaped by male mate choice. © 2010 Elsevier Inc. All rights reserved.Keywords: Evolutionary psychology; Contest competition; Mate choice; Mating; Sexual selection 1. Introduction Viewing human mating in a developed nation, one surmises that success in heterosexual competition for mates entails attracting members of the opposite sex. Beauty, fashion, and physical fitness are so important in places like the United States that they have become multi-billion dollar industries. Men and women have virtual autonomy to choose their mates. These conditions are so pervasive that it is tempting to think that they have characterized our evolution —that humans evolved in a context where, in the mating arena, the preferences of the opposite sex were the primary forces shaping our phenotypes. With notable exceptions (e.g., Apostolou, 2007; Archer, 2009; Buss & Dedden, 1990; Buss & Duntley, 2006; Buss & Shackelford 1997; Daly & Wilson, 1988; Lassek & Gaulin, 2009; Sell et al., 2009), the recent literature in E-mail address: [email protected]. 1090-5138/$ - see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.evolhumbehav.2010.02.005 evolutionary psychology reinforces this impression. The vast majority of research on sexual selection in Homo sapiens focuses on mate choice. Of papers on human sexual selection, more than 75% (55 of 73) published from 1997 to 2007 in the journals Evolution and Human Behavior and Human Nature mainly concern mate choice (categorized by the present author and a trained research assistant into "mate choice," "dominance and status competition," and "other," according to the hypotheses tested in the paper). According to an influential researcher, in sexual species, "all genes must propagate through the gateway of sex, and mate choice is the guardian of that gateway. For this reason, sexual courtship was probably central in human evolution and remains central in modern human life" (Miller, 1998, p. 119). According to another leading researcher, "the desires of one sex establish the critical dimensions along which members of the opposite sex compete" (Buss, 1996, p. 307). The extensive evidence leaves little doubt that the preferences of each sex have been important selection pressures on the other. 158 D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 But has mate choice been the primary mechanism of human sexual selection, as the literature might suggest? I argue here that it has not. Rather, contest competition—in which force or threats of force are used to exclude same-sex rivals from mating opportunities—has been the main form of mating competition in men, whereas male mate choice has predominated as a mechanism of sexual selection operating on women. This argument will be built on theory developed from cross-species comparison and subsequently tested by examining evidence of apparent design in humans. 2. Sexual selection Darwin (1859; 1871) proposed sexual selection to explain traits that seemed harmful to survival—the hooked jaw of the male salmon, the stag's antlers, the cock's spurs, and the "gorgeous plumage" and "strange antics" (1859, p. 137) of the male rock-thrush and bird of paradise, for example. Although these traits might impair survival, Darwin postulated that they could nevertheless promote their own passage into the next generation by helping their bearers win mating opportunities: They could be favored under sexual selection. In the intervening years, a massive literature has validated this postulate and elaborated on the modes of sexual selection and the traits that each mode favors. Same- sex contests favor size, strength, weapons, and aggression, for example, whereas mate choice favors sexual ornaments and displays. (The term ornament is used here to refer to a trait that contributes to fitness primarily through a preference for this trait in the opposite sex.) Sperm competition favors the production of large ejaculates, motile sperm, and frequent copulation, scramble competition favors sensory and locomotory organs for swift mate location (Andersson, 1994, Table 1.1.1), and sexual coercion favors size, strength, and other traits that facilitate harassment, punishment, and forced copulation (Savalli, 2001). We have also made considerable headway in understand- ing why one sex, usually males, often exhibits more elaborate ornamentation, mating displays, aggression, weap- onry, and so forth. Clearly, the sex whose reproduction is more reliant on access to mates will experience stronger sexual selection. This depends on the operational sex ratio, the ratio of fertilizable females to sexually active males at a given time (Emlen, 1976). A scarcity of available mates forces the commoner sex into competition for them. The operational sex ratio, in turn, depends on the relative reproductive rates of males and females (Clutton-Brock & Vincent, 1991), and these rates typically depend on relative parental investment (Bateman, 1948; Trivers, 1972; Wil- liams, 1966). In general, the sex that invests less in producing and rearing offspring finishes each reproductive venture sooner and re-enters the mating pool to find a shortage of the more-investing, slower-reproducing sex. Because males usually invest less than females do, males are usually more competitive for mates. 2.1. Mechanisms of sexual selection Despite such advances toward our understanding of sexual selection, one question has attracted surprisingly little attention: What mechanisms of sexual selection will operate in a given species? That is, can we predict whether mating competition will take the form of ornaments and displays for attracting mates, weapons and aggression for excluding competitors by force, traits for winning fertilizations in sperm competition, another mechanism of sexual selection, or some combination of these (Fig. 1)? 2.1.1. Constraints on contest competition We can begin with the premise that same-sex contests can override other mechanisms of sexual selection. If contest outcomes are decisive, with winners excluding losers from proximity to potential mates, there will be no opportunity for choice (thus no need for coercion) and no sperm competition. Other mechanisms can occur only to the extent that the competing sex (for brevity, males, unless noted otherwise) cannot exclude competitors from potential mates by force or threat. Logically, the next step is to determine what might limit contest competition across species, affecting males' ability to monopolize females by force. One limiting factor is the spatial and temporal clustering of females (Emlen & Oring, 1977). Resource distribution Fig. 1. Different mechanisms of sexual selection are theoretically distinct avenues by which mates can be obtained. Mechanisms vary in importance to fitness across species. Distance along an axis represents the correlation between success in that form of mating competition and mating success. (Examples are approximations.) and predator risk determine the distribution of females, which determines the distribution of males (Leutenegger & Kelly, 1977; Lindenfors, Froberg, & Nunn, 2004). If females form groups or are solitary but closely dispersed, they may be collectively defensible by a single male. This can lead to intense male contests for control of multiple females (Emlen & Oring, 1977). However, if groups of females are large, a single male cannot exclude competitors, leading to multi- male, multi-female groups, thus reducing the importance of contests (Leutenegger & Kelly, 1977). For example, males have reduced canines and are smaller overall in relation to females among multi-male group-living primates, compared to those that exhibit single-male polygyny (Clutton-Brock & Harvey, 1984). However, testes are larger relative to body weight in multi-male groups (Harcourt, Harvey, Larson, & Short, 1981), indicating more sperm competition. This suggests that multi-male and single-male primates differ in the mechanism, but not necessarily the strength, of sexual selection. Temporal clustering due to breeding synchrony can also make multiple fertile females difficult for a single male to defend, especially if the females are widely distributed spatially (Emlen & Oring, 1977). A second constraint on contests may be ecological costs. Flight (Caizergues & Lambrechts, 1999; Tobalske & Dial, 2000) and arboreality (Crook, 1972) should limit the evolution of the large bodies that would otherwise be useful for winning contests. Contests are relatively rare among birds (Emlen & Oring, 1977) and most intense among largely terrestrial birds in which males compete on land for display territories (Payne, 1984). Arboreal primates also exhibit less body size dimorphism than terrestrial ones (Clutton-Brock, Harvey, & Rudder, 1977; Plavcan & van Schaik, 1997). Third, anatomical characteristics determine the benefits of contests. Gaulin and Sailer (1984) pointed out that the force generated by a blow increases with mass (a cubic function of length), whereas the ability to resist the blow increases at most as a square function (e.g., cross-sectional surface area of bone). Consequently, larger animals should be able to inflict more damaging blows. Males of larger species might therefore be more successful in excluding competitors by force or threat, and contests might be likelier to evolve. Although the above variables may affect males' ability to monopolize mates, none is satisfactory as a general explanation. Intense contests occur across all social struc- tures, including multi-male groups (e.g., yellow baboons), single-male groups (e.g., gorillas and red deer), and solitary species (e.g., beetles and orangutans). Moreover, ecological constraints on body size explain neither the rarity of contests in aquatic environments (Andersson, 1994), which impose few size limitations (consider blue whales), nor the frequency of contests in small-bodied animals, such as arthropods. Finally, contests require only the physical monopolization of mates, not the ability to injure competi- tors. Thus, contests may be intense in small animals, such as beetles (Eberhard, 1979). 2.1.2. The dimensionality hypothesis An ecological variable that has apparently escaped notice in this context is the dimensionality of the environment in which mating competition takes place. To succeed in contests, males must exclude same-sex competitors from mates or the resources necessary to attract them (Emlen & Oring, 1977). This may be feasible in the essentially one- dimensional environments of burrows and tunnels, and in two-dimensional environments such as dry land, but impossible in three dimensions (air, water, or trees), where there are too many in-routes for competitors. For example, the three-dimensional aquatic environment of bottlenose dolphins hinders individual males from monopolizing females, resulting in a promiscuous mating system (Connor, Richards, Smolker, & Mann, 1996), and presumably sperm competition. Fig. 2 illustrates differences in the relative size and dimensionality of the region that must be defended when mating competition takes place in (a) a one-dimensional environment (e.g., dung beetles competing for mates in a tunnel), (b) a two-dimensional environment (e.g., fur seals competing for mates on land), (c) a three-dimensional environment bounded by a territory (e.g., blue-headed wrasses guarding coral reefs), and (d) an open three- dimensional environment (e.g., bottlenose dolphins compet- ing for mates in the open ocean). Whereas the difficulty of defense increases linearly with the radius of the defense region in two dimensions, it increases with the square of this radius in three-dimensional environments. 2.1.3. Evidence for the dimensionality hypothesis Qualitative comparative evidence suggests that the dimensionality of the mating environment explains substan- tial interspecific variation in contest competition, and Fig. 2. Small discs, hemispheres and spheres represent females' movement in relation to each other. Cylinder and large disc, hemisphere and sphere represent the regions that a male must defend in order to monopolize the females depicted. Defense regions increase in size and decrease in defensibility from left to right and top to bottom and represent (A) one- dimensional, (B) two-dimensional, (C) three-dimensional hemispheric, and (D) three-dimensional spherical regions. D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 159 160 D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 consequently, other mechanisms of sexual selection. For example, contests are more extreme in arthropods that compete for mates on land (e.g., beetles, pseudoscorpions, and spiders: Eberhard, 1979; Watson, 1990; Zeh, 1987) and the floors of bodies of water (e.g., crabs and crayfish: Christy & Salmon, 1991; Crane, 1975; Snedden, 1990; Sneddon, Huntingford, & Taylor, 1997), than in those that fly, such as butterflies, katydids, and locusts (Andersson, 1994). Con- tests are especially intense among arthropods, such as some dung beetles, that compete for mates in the essentially one- dimensional environments of burrows or tunnels (Emlen, 2008; Emlen & Philips, 2006). Contests are rare in the three-dimensional aquatic environments of fish (Andersson, 1994), whereas mate choice (Noble, 1938) and sperm competition (e.g., Fuller, 1998) are more common. However, contests occur in species such as bluegills (Gross & MacMillan, 1981), wrasses (Robertson & Hoffman, 1977), and pufferfish (Gladstone, 1987), where males guard territories on lake, stream, or sea floors (Turner, 1993) and mating environments are thus reduced in volume. In sockeye salmon, mating competition occurs in the shallow, effectively two-dimensional head- waters of streams (Quinn, Hendry, & Buck, 2001), and males use body size and specialized hooked jaws to fight for females (Quinn, Adkison, & Ward, 1996). Mate choice is the predominant mechanism of sexual selection across bird species (Andersson, 1994; Emlen & Oring, 1977), which because of flight tend to compete for mates in three dimensions. Terrestriality is a derived trait in birds, evolving independently at least four times in Galliformes, Struthioniformes, Casuariiformes, and Rhei- formes. Each of these incidences of evolution toward a more two-dimensional mating environment is associated with a movement away from social monogamy and an increase in contest competition (Andersson, 1994; Emlen & Oring, 1977). Scrambles may be the primary sexual selection mecha- nism in aquatically-mating pinnipeds. Female Weddell seals are larger than males, and males obtain copulations underwater via speed and agility (Andersson, 1994). However, in terrestrially-breeding pinnipeds, such as sea lions, elephant seals, and fur seals, males are several times the size of females, and male contests are intense and bloody (Boyd, 1989; Le Boeuf, 1974; Lindenfors, Tullberg, & Biuw, 2002). Thus, among pinnipeds, male contests and large body size have evolved in two-dimensional mating environments despite the ecological costs of carrying a large body on land. Contests are also common among terrestrial primates, which exhibit greater body and canine size sexual dimor- phism than arboreal and arboreal/terrestrial species (Clutton- Brock et al., 1977; Leutenegger & Cheverud, 1982; Plavcan & van Schaik, 1997). Differences in body size dimorphism might reflect mass limitations of arboreality (Clutton-Brock et al., 1977). However, arboreality imposes no clear constraints on canine size, and selection in terrestrial species favoring larger canines for predator defense (Leutenegger & Kelly, 1977) should affect both sexes. Reduced contest competition in three-dimensional arboreal environments may have decreased body and canine size dimorphism and increased the importance of mate choice. Female preferences for brightly-colored males have been demonstrated in arboreal primates (Cooper & Hosey, 2003), which appar- ently exhibit greater sexual dichromatism than terrestrial primates (Crockett, 1987). The dimensionality of the mating environment thus helps explain why contests occur in some taxa and not others. Having developed a theoretical framework for predicting the mechanisms and relative intensity of sexual selection, we return our attention to humans. 3. Sexual selection in humans Women invest more in offspring than men do through gestation and nursing for up to several years in foraging societies (Eibl-Eibesfeldt, 1989) and through providing more parental care on average in all known societies (Hewlett, 1992). Greater investment slows women's reproductive rates, skewing the operational sex ratio so that there are more males than females available for mating. Across species, these conditions foster mating competition in the more rapidly-reproducing sex. Evidence suggests that men have indeed experienced stronger sexual selection. Com- pared to women, men have higher reproductive variance (e.g., Brown, Laland, & Borgerhoff Mulder, 2009; Howell, 1979), are larger, more muscular, mature later, and senesce and die sooner—all correlates of an effectively polygynous mating system (Daly & Wilson, 1983). However, men invest heavily in offspring compared with males of most animal species and virtually all mammals. This investment con- stitutes a basis for female competition. Evidence (below) suggests that sexual selection has influenced women's phenotypes as well. The variables proposed to predict sexual selection mechanisms are nearly unanimous: contests should be the dominant mode of sexual selection in men (Table 1). More precisely, ancestral men's mating success should have correlated more strongly with success in male contests than with success in other modes of sexual selection. [See Schwagmeyer and Woontner (1986) for a similar compar- ison of contests and scrambles in ground squirrels.] The two- dimensionality of the human mating environment, combined with female sociality and breeding asynchrony, should facilitate female defense. Men's large size enables them to inflict damaging blows, and phylogenetic relationships also predict male contests in humans, along with some sperm competition, mate choice, and coercion. All three genera of non-human great apes show intense male contests with some female choice, and significant sperm competition and sexual coercion in Pan and Pongo (Knott, Emery Thompson, Stumpf, & McIntyre, 2009; Muller, Kahlenberg, Emery Table 1Variables influencing contest competition and predictions regarding humans can tentatively conclude that it evolved for this function. This can be demonstrated by cross-species comparison. For example, the hypothesis that horns in male beetles function in combat is supported if hornlike structures are present in species with male combat and absent in species without it. In parallel, within species, functional hypotheses are bolstered by correlations between a trait (e.g., horn size) and its proposed function (fighting efficiency). The strongest within-species evidence involves experimental manipulation that affects the trait's utility in the predicted direction. The more efficiently a trait performs its hypothesized function compared with alternative functions, the stronger support for the adaptive hypothesis. 3.2. Male contests 3.2.1. Evidence of design for male contests All of these types of evidence support the prediction that male contests have been important in human evolution. Men are larger, stronger, faster, and more physically aggressive than women, and the degree of sexual dimorphism in these traits rivals that of species with intense male contests. The relatively modest 8% stature dimorphism in humans (Gaulin & Boster, 1985) and a difference of about 15-20% in body mass (Mayhew & Salm, 1990) might suggest that male contests are reduced compared with our closest relatives. However, human sex differences in size underestimate sex differences in the traits most relevant to contests. This is partly because women are unique among primates in having copious fat stores (Pond & Mattacks, 1987), perhaps for building the large, fatty brains of human offspring (Lassek & Gaulin, 2008), and as sexual ornamentation (see below). When fat-free mass is considered, men are 40% heavier (Lassek & Gaulin, 2009; Mayhew & Salm, 1990) and have 60% more total lean muscle mass than women. Men have 80% greater arm muscle mass and 50% more lower body muscle mass (Abe, Kearns, & Fukunaga, 2003). Lassek and Gaulin (2009) note that the sex difference in upper-body muscle mass in humans is similar to the sex difference in fat- free mass in gorillas (Zihlman & MacFarland, 2000), the most sexually dimorphic of all living primates. These differences in muscularity translate into large differences in strength and speed. Men have about 90% greater upper-body strength, a difference of approximately three standard deviations (Abe et al., 2003; Lassek & Gaulin, 2009). The average man is stronger than 99.9% of women (Lassek & Gaulin, 2009). Men also have about 65% greater lower body strength (Lassek & Gaulin, 2009; Mayhew & Salm, 1990), over 45% higher vertical leap, and over 22% faster sprint times (Mayhew & Salm, 1990). Contrary to earlier claims, sex differences in anaerobic sprint speeds are not narrowing (Cheuvront, Carter, Deruisseau, & Moffatt, 2005; Seiler, De Koning, & Foster, 2007). Men and boys are more physically aggressive than women and girls (Archer, 2004, 2009). Boys engage in more play-fighting, and the amount of play-fighting a boy engages Males2D (Contests) Large (Contests) Yes (Contests) No (Contests) Multi-male/ multi-female groups (Contests reduced) No (Contests) Yes (Contests) Females2D (Contests) Large (Contests) No (No Contests) No (Contests) Multi-male/ multi-female groups (Contests reduced) No (Contests) No (No contests) D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 161 Variable influencing contest competition Dimensionality of mating environment Body sizeCapability of physically constraining opposite sex Temporal clumping of available matesSpatial clumping of mates and competitors Body-size constraints of flight or arboreality Contests in close phylogenetic relatives Value of variable (prediction regarding contests) Thompson, & Wrangham, 2007; Nishida & Hiraiwa- Hasegawa, 1987; Rodman & Mitani, 1987; Smuts, 1987). Only multi-male groups are predicted to reduce individual males' abilities to monopolize females, elevating the importance of other forms of sexual selection, but the influence of this variable may be mitigated in humans (see below). Female contests are absent in humans' close phylogenetic relatives, and monopolization of mates is likely to be unfeasible in women, as well (Table 1). The multi-female structure of human groups should have hindered ancestral females from excluding their competitors from mates. Indeed, in the presence of multiple same-sex competitors, successful mate defense probably depends partly on mates' cooperation. However, evolutionary models suggest that such cooperation will tend to evolve in one sex when the other is physically dominant and thus capable of sexual coercion (Clutton-Brock & Parker, 1995). Yet, men are greater in size and physical prowess than women are, and thus, men should be prohibitively difficult to constrain in their choices. We can therefore predict that female mating competition would favor traits to attract men, rather than physically monopolize them. Before these predictions can be evaluated, it is necessary to clarify what constitutes evidence of a trait's adaptive function in winning mates. For example, is a particular trait a weapon or an ornament, a dominance signal or a mate attraction display? 3.1. Testing evolutionary predictions One can infer ancestral selection pressures by studying the adaptations that they produced. Natural selection is the only evolutionary process to systematically produce traits that appear engineered for specific functions (Williams, 1966). If, under scrutiny, a trait looks well-suited to a purpose that would have benefited ancestral bearers, then we 162 D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 in correlates positively with peers' rankings of his domi- nance (Pellegrini, 1995; Pellegrini & Smith, 1998). Men report engaging in, and inclinations to engage in, nearly one standard deviation more physical aggression than women (Buss & Perry, 1992). Men perpetrate more offensive physical aggression, defined as non-defensive "attacking, hitting, and/or restraining another individual" in all societies studied (Ellis et al., 2008). The vast majority of same-sex homicides (about 95%), from every society and time period for which data are available, are committed by men (Daly & Wilson, 1988; M. Wilson & Daly, 1985). Importantly, these data do not include war killings, which occur almost entirely at the hands of men (Adams, 1983). Traumatic injuries in ancient skeletal remains indicate that interpersonal violence was especially prevalent among men throughout human history and prehistory (Walker, 2001). Such sex differences in traumatic skeletal injuries may help explain why some aspects of men's skeletons, particularly in the face, are more robust. For example, in modern populations, the incidence of mandibular fractures is approximately five times higher in men than in women, young men are disproportionately represented, and the primary cause is typically found to be violent assault with a fist or blunt object (Adi, Ogden, & Chisholm, 1990; Haug, Prather, & Indresano, 1990; Scherer, Sullivan, Smith, Phillips, & Robson, 1989; Sojat, Meisami, Sandor, & Clokie, 2001). A similar pattern in the evolutionary past could have selected for more robust mandibles in men than in women. Certainly, size, strength, speed, and aggression in men correlate with physical competitive ability, and manipula- tions that increase these variables lead to greater physical prowess. This is why many athletes abuse anabolic steroids. Relatively greater male upper-body (compared with lower- body) muscle mass and strength in particular suggest an evolutionary history of fighting (Sell et al., 2009). These traits also characterize male contests across species; males are larger, stronger and more aggressive in diverse species with male contests across the animal kingdom (Andersson, 1994). Close relatives of humans with minimal male contests, such as gibbons, lack substantial sex differences in size, strength, and aggression. Men possess several traits that appear to function primarily in threatening rivals. For example, beards and eyebrow hair grow at puberty in males and may signal dominance through association with testosterone levels and by increasing the apparent size of the jaw and brow (Guthrie, 1970; Muscarella & Cunningham, 1996; Neave & Shields, 2008). Male faces with beards are rated as more dominant than the same faces clean-shaven (Muscarella & Cunning- ham, 1996; Neave & Shields, 2008). Likewise, deep, low- pitched voices increase men's apparent size (Feinberg, Jones, Little, Burt, & Perrett, 2005) and dominance (Puts, Gaulin, & Verdolini, 2006; Puts, Hodges, Cardenas, & Gaulin, 2007). Perhaps deep voices signal dominance in men partly because they correlate with high testosterone levels (Bruckert, Lienard, Lacroix, Kreutzer, & Leboucher, 2006; Dabbs & Mallinger, 1999; Evans, Neave, Wakelin, & Hamilton, 2008). Across animals, low-pitch vocalizations generally signal dominance, and high-pitch vocalizations signal submissiveness (Morton, 1977; Morton & Page, 1992). In the laboratory, men who reported superior fighting ability tended to lower their voice pitch during mating competition, whereas those who perceived themselves to be poor fighters raised their pitch (Puts et al., 2006). It is often claimed (e.g., Lorenz, 1966) that men lack antlers, long canines, or other weapons typical of intense contest competition. However, these traits are called "weapons" by analogy with human weapons, and for as long as there is any record, weapons have been manufactured and wielded almost entirely by men (Christensen, 2004; D'Andrade, 1974; Darwin, 1871; Gat, 2006). The first known combat weapons were also used in hunting (Christensen, 2004), which is performed nearly exclusively by men in foraging societies (Murdock, 1967). It is therefore appropriate to consider weapons a part of men's phenotype. The use of projectile weapons must have selected for targeting abilities, which show the largest known "cognitive" human sex difference, with an effect size of about 1.5 standard deviations favoring men (Hines et al., 2003; Kimura, 1999). Thus, the already large sex differences in muscle mass, strength, speed, and aggression probably underestimate the intensity of contest competition in men. Men's anatomy and behavior seem well designed for contest competition. The alternative hypothesis that these traits evolved in the service of hunting (e.g., Lancaster & Kaplan, 2009) is unsatisfying partly because it is unparsi- monious: across the animal kingdom, where one finds large, strong, aggressive males with weapons, it is almost always because males employ these traits in fights for females (Andersson, 1994). The hunting hypothesis also has difficulty explaining traits such as beards, deep voices, robust skulls, and male-male aggression that are easily explicable by male contest competition. Even targeting ability may have emerged in the context of male contests: male chimpanzees throw rocks and branches with far greater frequency than do females, and the targets are other males, not prey (Goodall, 1968; Van Lawick-Goodall, 1971). Although superior targeting ability in men was likely shaped for both male contests and efficient hunting, men also exhibit superior intercepting abilities (Watson & Kimura, 1991), which are difficult to comprehend as adaptations for hunting (Mark Flinn, personal communication). Finally, orangutans, gorillas, chimpanzees, and early hominins, such as Austra- lopithecus afarensis and A. anamenis (Leakey, Feibel, McDougall, Ward, & Walker, 1998; Plavcan, Lockwood, Kimbel, Lague, & Harmon, 2005; Reno, Meindl, McCollum, & Lovejoy, 2003, 2005; Ward, Leakey, & Walker, 1999), all exhibit moderate to high degrees of sexual dimorphism, and intense male contests occur in all genera of extant great apes. Substantial sexual dimorphism and contests thus probably existed in the earliest hominins, yet hunting became a major source of food only as early as perhaps 2.5 million years ago (Dominguez-Rodrigo, Pickering, Semaw, & Rogers, 2005). Greater male size and strength probably preceded hunting in our lineage, and thus could not have resulted from it. 3.2.2. Human mating systems and the role of male contests Although it seems clear that contest competition shaped many of men's traits, it is less clear how success in male contests increased mating opportunities over human evolu- tion. Several possibilities exist. First, coalitional aggression could have facilitated acquisition and defense of mates against other groups of males. Second, males could have used force or force threat within their groups to acquire and defend one or more long-term mates, or to obtain disproportionate short-term mating opportunities. Finally, contests could have contributed indirectly to mating success if dominant males could acquire resources, territory, or status needed to attract females. To what extent did these possibilities apply to ancestral humans? 3.2.2.1. Male coalitions and between-group competition. The tendency of males to form alliances may have evolved in the common ancestor of humans and our closest living relatives, Pan, as a means of cooperative female capture and defense (Fig. 3), although coalitions may also have evolved independently in these lineages for this purpose (Geary & Flinn, 2001; Wrangham, 1999). Male coalitions are rare among primates but common in humans and Pan, especially common chimpanzees (P. troglodytes), and are strengthened by kinship (Nishida & Hiraiwa- Hasegawa, 1987). The capture of women was a primary objective of early warfare (Darwin, 1871; Hrdy, 1997; Lerner, 1986; Spencer, 1885), and among foragers, groups of men commonly raid other villages and abscond with women Fig. 3. From a great ape progenitor with single-male polygyny and single- male social groups (1) (Harrison & Chivers, 2007), two trajectories for African apes are depicted: a continuation of this pattern in Gorilla, and the evolution of multi-male groups for female defense, as in African lions, in the common ancestor of Pan and Homo (2). In Pan, within-group monopoli- zation of females was difficult, and greater sperm competition thus predominated (3), resulting in lower sexual dimorphism (Jungers & Susman, 1984) but larger investments in testicular tissue (Short, 1979). In Homo and their immediate ancestors, individual males more effectively monopolized females, perhaps due female cooperation, between-group competition, and more intensive use of weapons. (e.g., Chagnon, 1988). Such raids may also function in mate defense by deterring future attacks. These behaviors would tend to favor not only aggression and physical prowess, but also social intelligence for negotiating alliances (e.g., Alexander, 1989; Geary & Flinn, 2002; Mueller & Mazur, 1996; Wrangham, 1999). 3.2.2.2. Within-group competition. Despite their relevance in Pan and Homo, alliances are open to subversion; if one member can gain a reproductive advantage, he may defect. Male chimpanzees not only cooperate to defend their community range against outside males but also fight within groups over estrous females (Nishida & Hiraiwa-Hasegawa, 1987). As noted, multi-male groups should make females more difficult to defend from other group members. The cross-cultural ubiquity of within-group aggression between men (Daly & Wilson, 1988), male violence against women over suspected infidelity (Daly & Wilson, 1988; Smuts, 1996), high levels of male sexual jealousy (Buss, Larsen, Westen, & Semelroth, 1997; Daly, Wilson, & Weghorst, 1982; Schützwohl & Koch, 2004), and evidence of moderate sperm competition (Harcourt, Purvis, & Liles, 1995; Wyck- off, Wang, & Wu, 2000) indicate that men have had to defend their mates within their groups, as well. Yet, within multi-male groups, men form enduring mateships with one or more females and usually have near-exclusive sexual access (Daly & Wilson, 1983; Flinn, Quinlan, Ward, & Coe, 2007). Among Yanomamo hunter- gatherers, approximately 10% of offspring are sired by a man other than the mother's social partner (Chagnon, 1979). This number has been estimated to be 2% (Simmons, Firman, Rhodes, & Peters, 2004), 4% (Bellis, Hughes, Hughes, & Ashton, 2005) and 9% (Baker & Bellis, 1995) across modern societies. The cross-cultural frequency of these conditions— enduring mateships and relatively low rates of extrapair paternity—suggests that they characterized early Homo sapiens. Women's inconspicuous estrus also suggests a single-male mating system (Clutton-Brock & Harvey, 1976; Hrdy, 1997; Nunn, 1999) and, thus, that ancestral males could monopolize females. Several factors may have contributed to ancestral males' ability to monopolize females. First, the importance of male alliances might have reduced within-group conflict, enabling individual males—especially those distinguished in inter- group conflict—to monopolize one or more females (Hrdy, 1997; Smuts, 1995). Among the horticultural Yanomamo of Venezuela, men who have killed enemies in intergroup conflict have more wives (Chagnon, 1988). However, male chimpanzees also form alliances but are unable to similarly monopolize females for prolonged periods. Second, acquisition and long-term defense of females by individual males may have been promoted by enhanced cooperation from females as a result of male investment (Geary & Flinn, 2001). Serial long-term mating in particular may have been an important means by which dominant males could monopolize the reproductive careers of multiple D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 163 164 D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 females while having to defend (and invest in) only one female at a time (Hill & Hurtado, 1996; Johanna, Forsberg, & Tullberg, 1995; Kaar, Jokela, Merila, Helle, & Kojola, 1998; Marlowe, 2004). However, in most societies, some men are simultaneous polygynists (Murdock, 1967), and this was probably true of all human societies until a few hundred years ago (Betzig, 1986, 1995). Finally, ancestral males likely acquired and defended females within groups through the use of force or threat of force. This may have been facilitated by the use of lethal weapons that can be wielded at a distance. Among the Hadza hunter-gatherers of Tanzania, courtship of a female by more than one male can lead to violent, possibly fatal conflict between the males (Marlowe, 2004). Bloody and sometimes fatal club fights erupt between Yanomamo men when one suspects the other of trysting with his wife (Chagnon, 1992). In these cases, the ability to inflict serious damage to a competitor apparently functions in mate acquisition and mate retention, respectively. Certainly, the ability to inflict physical harm on competitors would have helped ancestral males win such skirmishes over mates. A reputation for—or advertisement of—fighting ability would also have enabled ancestral males to win and defend mates while avoiding many costly fights. Indeed, some male traits, such as such as deep voices and facial hair, seem more explicable as within- group dominance signals than as functioning in between- group competition. Human males have probably not competed over short- term access to fertile females to the extent that such competition occurs in other multi-male primates, such as chimpanzees. Unlike chimpanzees, humans do not exhibit conspicuous estrus. Without a reliable indicator of female ovulatory status, males' expected reproductive return from copulations over a short duration is low (Bongaarts & Potter, 1986; Leridon, 1977). Consequently, short-term sexual access was probably seldom worth incurring the potentially large costs of physical competition. However, contest competition leading to ascension in a male dominance hierarchy would likely have ramifying reproductive benefits, including greater access to short-term mates and less retaliation after trysting with already-mated females. 3.2.2.3. Competition for mate-acquisition resources, or as sexual display. Perhaps dominant males were also more likely to be chosen by females because dominant males had access to better resources or territory, or because the victors of male contests tended to provide high quality genes. Female preferences must have affected males' ability to acquire mates and ensure their fidelity over human evolution. Thus, we now direct our attention to female choice. 3.3. Female choice Women's preferences for men with resources and willingness to invest appear culturally ubiquitous (Buss, 1989; Cashdan, 1996; Hill & Hurtado, 1996). Men likely provide resources partly because this attracted more mates (Buss, 1989; Hawkes, 1990, 1991) and increased mates' fidelity (Geary & Flinn, 2001), although resource provision- ing can also function as parental investment (Marlowe, 2003; Trivers, 1972). However, foragers typically cannot accumu- late great wealth or resources (Marlowe, 2005), and all humans were foragers prior to about 10,000 years ago. Nevertheless, women can choose mates based on less tangible benefits, such as foraging ability (Marlowe, 2004) and high status for offspring (Hill & Hurtado, 1996). Women can also obtain protection from rape and harm to offspring (Smuts, 1996). Male infanticide of unrelated offspring is prevalent among primates (Hrdy, 1979), including humans (Smuts, 1996), and is mitigated by the presence of the biological father (Daly & Wilson, 1988). Because most human reproduction occurs within long-term mateships such as marriage (Apostolou, 2007), preferences for such benefits may have evolved primarily in this context (see, e.g., Marlowe, 2004). However, women may also extract male investment from short-term mating (Buss & Schmitt, 1993; Greiling & Buss, 2000; Hawkes, 1990, 1991). 3.3.1. Good genes mate choice A growing body of evidence suggests that women also choose mates partly on genetic quality (Roberts & Little, 2008). Because selection culls alleles associated with inferior phenotypes, the traits most relevant to fitness should lose heritability (Fisher, 1930), degrading the basis for good genes mate choice. However, mutation and fluctuating selection can maintain genetic variation. Hamilton and Zuk (1982) suggested that parasites generate important tempo- rally varying selection on hosts. As parasites' short generation times facilitate rapid evolution, resistance to infection is an essential fitness component that might remain heritable in hosts. A preference for indicators of heritable parasite resistance could spread and be maintained in a population (Hamilton & Zuk, 1982). Some animal research indicates that females choose mates based on parasite resistance, and this resistance is heritable (e.g., Hillgarth, 1990). An important genomic region in mediating disease resistance is the major histocompatibility complex (MHC). Greater MHC allelic diversity theoretically enables recognition of more invaders and stronger immune function. Hence, preferences for MHC-dissimilar mates are predicted. Olfactory preferences for MHC-dissimilar mates have been demonstrated in several vertebrate taxa, including fish, reptiles, birds, and rodents, and in most human studies (reviewed in Roberts & Little, 2008). Preferences for MHC dissimilarity may produce greater MHC-specific genetic dissimilarity within human couples than occurs between random pairs of individuals (Chaix, Cao, & Donnelly, 2008) and may produce attractive, healthy-looking offspring; heterozygosity at MHC loci (Lie, Rhodes, & Simmons, 2008), has been associated with facial attractiveness and healthy-looking skin (Lie et al., 2008; Roberts et al., 2005). Women are thus expected to prefer mates with genes that confer disease resistance, and with few harmful mutations. Two putative good-genes indicators are androgen-dependent traits and fluctuating asymmetry. Androgen-dependent, masculine traits may indicate heritable disease resistance because androgens suppress immune function (Grossman, 1985) and may be produced in proportion to inherited immunocompetence (Folstad & Karter, 1992). High andro- gen levels may increase competitive ability but attenuate inherited immunocompetence, so that good-genes males end up little healthier than average. Males with few harmful mutations may also be able to produce and maintain more elaborate androgen-dependent traits (Zahavi & Zahavi, 1997). Regulation of androgen levels and patterns of response to androgens may thus have evolved as a means of producing sexually selected traits in proportion to a male's ability to safely bear them (Folstad & Karter, 1992). Fluctuating asymmetry (FA) refers to asymmetry in anatom- ical traits that are normally bilaterally symmetric. FA may negatively indicate genetic quality because it results from developmental stresses such as mutation and parasitic infection (Moller & Pomiankowski, 1993; Parsons, 1990, 1992; van Valen, 1962) and is moderately heritable in several species (Moller & Thornhill, 1997). As expected, women have been found to exhibit preferences for the odors (Gangestad & Thornhill, 1998; Rikowski & Grammer, 1999; Thornhill & Gangestad, 1999), faces (Gangestad, Thornhill, & Yeo, 1994; Scheib, Gangestad, & Thornhill, 1999), and voices (Hughes, Harrison, & Gallup, 2002) of men whose external features are symmetrical. Women also prefer men who are somewhat more masculine than average in height (Paw- lowski & Jasienska, 2005), body build (Frederick & Haselton, 2007; Horvath, 1981), voice (Feinberg et al., 2005; Puts, 2005), and perhaps face (e.g., Johnston, Hagel, Franklin, Fink, & Grammer, 2001; but see Perrett et al., 1998). Some studies have found that men's symmetry and masculinity correlate, as might be predicted if both features index underlying genetic quality (Gangestad & Thornhill, 2003; Scheib et al., 1999). Presumably because they are more sexually attractive to women, men of putatively high genetic quality tend to expend more effort acquiring additional mates and less effort investing in mates (Gangestad & Simpson, 2000). For example, men with high testosterone levels are more likely to have extramarital sex (Booth & Dabbs, 1993) and less likely ever to have been married (Booth & Dabbs, 1993), and unmarried men have higher testosterone levels than do married men (Gray, Kahlenberg, Barrett, Lipson, & Ellison, 2002). Similarly, symmetrical men have more extra-pair sex partners (Gangestad & Thornhill, 1997, 1999) and invest less in their current mates (Gangestad, 1993; Simpson, Gang- estad, Christensen, & Leck, 1999). Consequently, masculine and symmetrical men should tend to make better sires than long-term mates. Predictably, women's preferences for these men are greater for short-term, sexual (vs. long-term) relationships and strongest during the fertile phase of the ovulatory cycle (reviewed in Gangestad & Thornhill, 2008). Symmetrical, physically attractive men are also more often the extra-pair sexual partners of women (Gangestad & Thornhill, 1997). Women's extra-pair, but not intra-pair, sexual interest increases near ovulation (Gangestad, Thornhill, & Garver, 2002), and this cyclic shift occurs mainly in women whose primary partners are putatively of low genetic quality (Gangestad, Thornhill, & Garver-Apgar, 2005; Haselton & Gangestad, 2006; Pillsworth & Haselton, 2006). These findings suggest that women's sexual preferences are designed partly to recruit genetic benefits from men who may not be the women's long-term partners. Evidence of sperm competition and moderate rates of extra-pair paternity across human societies (Simmons et al., 2004) support this possibility. 3.3.2. Sexual ornaments in men? Women's preferences raise the question of whether men's traits are better viewed as sexual ornaments or armaments. Female mate choice has been observed in virtually every primate species studied (Smuts, 1987). Paradoxically, few male primates exhibit clear sexual ornaments (Andersson, 1994). Part of the reason may be that females choosing males on genetic quality might prefer dominance and the traits associated with it, and thus male secondary sex traits adopt dual functions (Berglund, Bisazza, & Pilastro, 1996). Berglund et al. (1996) argue that secondary sexual characters usually originate through contests rather than female choice partly because signals used in contests tend to be costly to produce, are constantly tested by competitors and, thus, should provide accurate information about male quality to both competitors and potential mates. In many species, females prefer traits that function in contests (Berglund et al., 1996; Kodrick-Brown & Brown, 1984). Low, closely-spaced vocal formant frequencies intimidate rivals among red deer stags (Reby et al., 2005) and are preferred by estrous females (Charlton, Reby, & McComb, 2007). Likewise, spur length among ring-neck pheasants is associated with both male dominance (Mateos & Carranza, 1996) and attractiveness to females (von Schantz, Wittzell, Göransson, Grahn, & Persson, 1996). Estrous females choose dominant males in a variety of nonhuman primates (Smuts, 1987); for example, dominant male orangutans are larger and possess cheek flanges, and females show greater willingness to mate with these males near ovulation (Knott et al., 2009). These ideas help explain the origin of women's preferences for traits such as size, muscularity and deep voices that have clear utility in contests. The idea that such traits should be especially good indicators of male quality also helps explain why women tend to prefer them more for sexual relationships and during the fertile phase of the ovulatory cycle. However, even if men's traits initially arose through male combat, it is possible that female choice could have become a stronger selection pressure, D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 165 166 D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 and men's traits subsequently evolved to assume a more ornamental role. This hypothesis that men's traits evolved to be ornamen- tal gains little support from the literature. Some male traits, such as beards, have been found to decrease attractiveness to women, yet have strong positive effects on men's appearance of dominance (Muscarella & Cunningham, 1996; Neave & Shields, 2008). In other cases, it is unclear whether women prefer more masculine males. For example, women have been found to prefer both slightly feminine-looking (Perrett et al., 1998; Rhodes, Hickford, & Jeffery, 2000) and slightly masculine-looking (DeBruine et al., 2006; Johnston et al., 2001) male faces. Despite these equivocal results, facial masculinity strongly increases the appearance of dominance, and has substantially larger effects on dominance than attractiveness when both are examined (DeBruine et al., 2006; Perrett et al., 1998). Finally, although some masculine traits, such as muscular builds (Frederick & Haselton, 2007) and deep, masculine voices (Feinberg et al., 2005; Puts, 2005, 2006) are sexually attractive, masculinity in these traits increases perceptions of dominance to a much greater degree than it increases attractiveness (Fig. 4). Even near ovulation and in the context of short-term mating (where masculine traits are maximally attractive), masculinity has been found to produce smaller positive effects on attractiveness to women than on dominance as judged by men (Puts et al., 2006) (see Fig. 4). One might argue that there is considerable cross-cultural variation in the importance of masculine traits for attractive- ness and dominance. This is certainly true, but cultural differences are unlikely to reverse the findings that masculine traits have greater positive effects on perceptions Fig. 4. Masculinity in facial hair, voice, facial structure and body build has larger positive effects on perceptions of dominance than on perceptions of attractiveness. Note: the important comparisons are the effects of masculinity on attractiveness versus dominance within each study. Between-studies comparisons are confounded by differences in the magnitude of manipulations and other methodological details. of dominance than they do on attractiveness. For example, in societies in which nearly all adult men have beards, women may prefer facial hair on men. Yet, in such societies, facial hair will probably be even more essential to perceptions of dominance, with a lack of facial hair being perceived as childlike in men. One might also argue that, although masculine traits appear better at increasing dominance, mating success is determined more by attractiveness. Thus, masculine traits may actually have been favored primarily through their utility as sexual ornaments. Existing evidence contradicts this possibility. For example, men's voice attractiveness predicts mating success (Hughes, Dispenza, & Gallup, 2004). Yet, when dominance and attractiveness ratings of men's voices were used simultaneously to predict mating success, dominance strongly predicted number of sex partners, whereas attractiveness did not (Puts et al., 2007). Finally, one might argue that the masculine traits reviewed above bias a functional analysis toward the conclusion that ancestral men competed for mates mainly via contests, and that choosing a different set of traits would lead to a different conclusion. However, traits that show a high degree of sexual dimorphism, especially those that develop at sexual maturity, are the best candidates for targets of sexual selection (Andersson, 1994; Darwin, 1871). Choosing such traits in men necessitates choosing masculine traits (i.e., traits that are present to a greater degree in males), which could be ornaments or weapons. In other words, choosing mascu- line traits for a functional analysis amounts to choosing the best candidates for sexually selected traits without implying anything about the traits' utility in contests, mate attraction, or any other mechanism. [In long-tailed widowbirds, long tail feathers are masculine, yet male widowbirds use their long tails to attract females (Andersson, 1982).] The function of masculine traits in men was an open question, but men's traits seem better designed for contests than for attracting mates. Are any of men's traits properly considered sexual ornaments (i.e., function primarily in mate attraction)? Men's penises are longer and thicker, both relatively and absolutely, than those of our closest relatives, chimpanzees and gorillas, and could have evolved to signal mate quality. Women report greater satisfaction with larger penises (Lever, Frederick, & Peplau, 2006), so penis size may affect a man's ability to stimulate orgasm in women (Miller, 2000). Female orgasm may boost sperm retention, facilitate sperm activation, and encourage additional copulations (reviewed in Puts & Dawood, 2006, see also Gallup et al., 2003). However, it has also been suggested that penis size may advertise vigor to other men (Diamond, 1997). Geoffrey Miller (1998; 2000) is the leading proponent of a theory that men's brains and creativity are designed to attract females. In this view, displays of neurophysiological efficiency advertize heritable fitness. Miller's "mating-mind" hypothesis offers explanations for the evolution of large human brains and behaviors such as humor, music, and poetry that do not have obvious survival value. It also accords with widespread preferences for intelligent mates and the moderate heritability of intelligence, which affords the opportunity for genes associated with intelligence to be favored by mate choice. Despite these advantages, Miller's hypothesis suffers shortcomings (Miller, 2001). Betzig (2002) notes that sexual selection tends to produce sexual dimorphism, but men's brains are only slightly larger relative to body size than women's (Ankney, 1992), and sex differences in overall intelligence, if present, are small (Irwing & Lynn, 2005). Although women may produce fewer creative displays than men do (Miller, 2000), it is unclear how ancestral women could offset the costs of producing and maintaining large brains under Miller's hypothesis. The mating-mind hypothesis better explains men's investment in creative displays, rather than human intelligence generally (Miller, 2001). Even this narrower version encounters difficulties. Men's mental displays lack features of an ornament advertising genetic quality. As we have seen, putative good-genes indicators are preferred more in short-term, sexual mating contexts, and near ovulation (Gangestad & Thornhill, 2008). Yet, women prefer intelligence, creativity, and humor more in long-term relationships (Bressler, Martin, & Balshine, 2006; Gang- estad, Garver-Apgar, Simpson, & Cousins, 2007; Kenrick, Sadalla, Groth, & Trost, 1990; Prokosch, Coss, Scheib, & Blozis, 2009). Women also do not seem to prefer intelligent men more strongly near ovulation (Gangestad et al., 2007; Prokosch et al., 2009, but see Haselton & Miller, 2006). An intriguing possibility is that creative displays represent an alternative mating tactic to male contests. Contest competition favors not only fighting ability but also "alternative mating tactics of inferior competitors, avoiding contests with superior rivals" (Andersson, 1994, Table 1.1.1). Alternative mating tactics probably occur in most nonhuman primates, and often involve mate attraction (Smuts, 1987). For example, alpha male chimpanzees monopolize estrous females through posses- sive behavior, while non-alpha males form sexual consortships away from other males (Smuts, 1987). Consortships often appear to occur through mate choice, although they may sometimes result from sexual coercion (Smuts, 1987; Wrangham & Peterson, 1996). Given the importance of male investment in human reproduction (Kaplan, Hill, Lancaster, & Hurtado, 2000), creative displays may represent "super-stimuli" (Miller, 2000) designed to trigger female preferences for investing mates. Thus, while large brains cannot presently be regarded as sexual ornaments, men's creative displays may represent courtship behaviors shaped by female choice, perhaps as part of an alternative mating tactic to contest competition. 3.3.3. Does female choice drive male dominance competition? It is possible that men engage in contest competition primarily because women prefer dominant men. Women have been shown to prefer dominant male behavior in short- term/sexual mating contexts and during the fertile phase of the menstrual cycle (Gangestad, Simpson, Cousins, Garver- Apgar, & Christensen, 2004; Snyder, Kirkpatrick, & Barrett, 2008), apparently because these preferences garnered genetic benefits ancestrally. As noted above, women also prefer the perquisites of dominance, including protection and access to resources. However, strong sexual selection through mate choice tends to produce sexual ornaments and displays—traits favored primarily through their utility in mate attraction. Even in lekking species, where females choose dominant males, males tend to evolve sexual ornaments and displays (Andersson, 1994). But we do not see these in humans, except possibly for creative displays and large penises. Furthermore, some evidence suggests that women had less freedom of mate choice ancestrally than they do in many modern industrial societies (Apostolou, 2007; Low, 2005; Smuts, 1995). Among foragers, reproduction occurs mainly within the context of marriage, and familial control over women's marriage decisions is prevalent (Apostolou, 2007). [Men's marriages are not as strongly influenced by family members (Apostolou, 2007).] In most societies, men may purchase women from women's families through bride-price or bride-service (Murdock, 1967), or steal women from other villages (e.g., Chagnon, 1992). Men also constrain women's choices through the threats of loss of investment and physical injury to mates and competitors. Among both traditional (Chagnon, 1992; Marlowe, 2004) and industrial societies (Daly & Wilson, 1988), men injure or kill wives whom they suspect of infidelity and injure or kill other men in competition over women. Men's greater size, strength, and physical aggressiveness contribute to an ability to sexually coerce females (Clutton-Brock & Parker, 1995). Women may have less influence over mating when they lack kin support (Smuts, 1996), and female exogamy was likely the ancestral human condition (Chapais, 2008; Fox, 1980; Wrangham, 1987). In general, men everywhere appear concerned with controlling women's reproductive capacity (e.g., Daly & Wilson, 1983; Dickemann, 1981). This is not to say that ancestral women had little choice. In some traditional societies women choose their spouses (Marlowe, 2004), and across traditional societies women frequently influence their family's choice of a spouse, are often consulted, and occasionally elope (Apostolou, 2007). Divorce is also commonly initiated by women (Betzig, 1989). Women have affairs, although men's proprietariness over women makes this behavior risky (Kaighobadi & Shackelford, 2009; Kaighobadi, Starratt, Shackelford, & Popp, 2008; Wilson & Daly, 1993; Wilson & Daly, 1996). Given the cross-cultural frequency of these conditions (Apostolou, 2007), women probably also utilized these avenues of mate choice over human evolution (Small, 1992). D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 167 168 D.A. Puts / Evolution and Human Behavior 31 (2010) 157-175 Indeed, preferences are costly and generally do not evolve without some compensatory benefit (but see, e.g., Ryan & Rand, 1995). In a species such as Homo sapiens, where mate choice and contests occur simultaneously, it can be difficult to disentangle their relative contributions. Several lines of evidence reviewed above are relevant. Men's traits look designed to make men appear threatening, or enable them to inflict real harm. Men's beards and deep voices seem designed specifically to increase apparent size and domi- nance. Size, musculature, aggression, and the manufacture and use of weapons directly increase fighting ability. Even masculine facial structure may be designed for fighting; heavy brow ridges protect eyes from blows, and robust mandibles lessen the risk of catastrophic jaw fractures, for example. Men's traits and a human tendency toward polygyny (Murdock, 1967) indicate an evolutionary history of male monopolization of females through force. Many of men's traits are probably not generally preferred by females, and those that are appear better designed for contests than mate attraction. Various constraints on female mating, including familial control of marriage, female exogamy, male sexual coercion and mate guarding, and male exclusion of competitors by force, also challenge the notion that ancestral men competed for dominance primarily to be chosen by women. Although more work is needed before firm conclusions can be drawn, cross-species comparison, empirical evidence of design, and multiple features of human mating support the theoretical prediction that contest competition was the predominate form of sexual selection in men. The contention that "adult male hominids must have been rather peripheral characters in human evolution, except as bearers of traits sexually selected by females for their amusement value or utility" (Miller, 1998, p. 109) is almost certainly false. 3.4. Male choice Although contest competition may have predominated in shaping men's traits, male mate choice is expected to have been more important in shaping women's. Why might women compete for mates? While men contribute far less parental care than women do across societies (Hewlett, 1992), men protect mates and offspring from predators and other men, and can provide high quality food through hunting. Men procure more food (in kcal) than women do in foraging societies, far more than they consume (Kaplan et al., 2000). In contemporary foragers, men use meat to obtain mating opportunities (Hawkes, 1991) and to invest in current mates and offspring (Kaplan et al., 2000; Marlowe, 2003), and meat probably served this dual function over human evolution. The proportion of a man's resources channeled toward parenting likely depended on his paternity confidence (Anderson, Kaplan, Lam, & Lancaster, 1999; Anderson, Kaplan, & Lancaster, 1999, 2007). To the extent that men vary in their ability to protect and provide, women are expected to compete for these benefits (Cant, 1981; Geary & Flinn, 2001; Symons, 1979). Compared to mechanisms of sexual selection in men, the mechanism of sexual selection in women is uncontroversial. Multi-female groups and strong selection for fighting ability in men militated against female monopolization of men through force. Although ancestral females could not make themselves the only option, they could compete to be the most attractive one. Men generally value mates' physical attractiveness highly a

Tinbergen's Four Questions

Function (adaptation)-see, find food, avoid danger Phylogeny (evolution)-blind spot Causation (proximate)-consider all Development (ontogeny)-genes/environment

Subsistence Strategy / Diet

More sophisticated hunting techniques Dietary diversity(aquaticresources&birds) Some prey specialization Larger bodied species

Culture of Homo erectus

"Committed to a cultural way of life" Complex tools Hunting-Gathering Behavioral flexibility Long learning period

Costs and Benefits

"The most profitable strategy for the typical man may have been to devote most of his reproductive effort to enhancing the phenotypic quality of his offspring."

Genetic Hypotheses

Evidence - For some (but not all) homosexuals, the tendency can be found in family lines along maternal lines There is an area of 100 genes or so on tip of X chromosome that is shared by many but not all gay relatives Monozygotic twins are twice as likely to both be homosexual than dizygotic twins If one twin homosexual/bi, then there is a 52% chance a monozygotic twin will be homosexual, versus 22% for dizygotic twins Adopted brothers 11%Genetic non-twin brothers 9%.So it is not just a gene, but also the environmental or developmental effects.

Female Reproductive Potential

"Wife of Feodor Vassilyev": 69 children from 27 pregnancies (Russia 1800s) Madalene Granata: 15 sets of triplets (Italy, 1839 to 1886) Fatma Saygi (at 28): 6 sets of triplets But... Males can have a lot of offspring, but are also more likely to have few Females have less variance in reproductive success

Opportunity for Sperm Competition

479 undergraduates in New York"Have you had sex with more than one male within a 24 hour period?" 13% of females said yes 8% had participated in a ménage à trois with two males 25% of females reported having had sex within someone else while in a relationship 15% of males

Runaway Sexual Selection (Fisher 1930)

A preference for a trait arises through a random mutation Since females prefer this random trait, the trait is selected for Thepreferenceforthetraitisalsopassedon along with the trait through sexual recombination Thepreferenceforthetraitandthetraitbecome genetically correlated Thisprocesscancontinueuntilthecoststo survival become too great to bear

Neanderthal Behavior

Adaptations to cold stress Behavioral Fire (true hearths) Cavewindbreaks Crude clothing High physical activity levels Coordinated big game hunting Complex stone tool technology Complex social structure Ritual? Probable cannibalism Interbreeding with modern humans

Self-Semen Displacement

After ejaculation Loss of erection (head may shrink first) Shallower thrusting after ejaculation Change in penile sensitivity Continuedthrustingcanbecomeunpleasantfollowingejaculation The refractory period Usually30minutesto24hoursdependingonage The Coolidge effect can shorten the refractory period

The Human Revolution

Anatomically modern humans evolve >100,000 years ago But... Major cultural changes only after 50,000 years ago Why? Cultural change? Biological change? Niche construction? Language? Richard Klein

Alliance Formation

And if there were only some way of contriving that a state or an army should be made up of lovers and their loves, they would be the very best governors of their own city . . . and when fighting at one another's side, although merely a handful, they would overcome all men. Plato, The Symposium The brother-in-law is ally, collaborator, and friend. . . . In the same band, the potential brother-in-law, i.e., the cross-cousin, is the one with whom, as an adolescent, one indulges in homosexual activities which will always leave their mark in the mutually affectionate behaviour of the adults. Claude Levi-strauss, The Elementary Structures of Kinship

Costly Signaling

Both signaler and receiver benefit from honest communication Honesty can be ensured when there is some cost associated with producing the signal The cost prevents the signal from being faked because it is: impossible for a faker to produce the signal too expensive to be worthwhile

Other Considerations and Ideas

Brains: research on differences in male and female brain also includes homosexuals, but results can be contradictory. Maybe homosexual behavior is just a byproduct of the fact that we enjoy sex If it were simply pleasure, then why do individuals prefer one mate over another or one sex over another?

Brains

Brains: research on differences in male and female brain also includes homosexuals, but results can be contradictory. Suggestions that male homosexual brain more like female in some regions. Is this just a correlate or a consequence or a cause?

The Comparative Method

By comparing similarities and differences between species we can get information about: When traits evolved (Did they evolve before or after a shared common ancestor?) Why they evolved (How do the niches of species with the trait differ from niches of species without the trait?) How likely they were to evolve (Analogous traits: fins in fish and mammals, horns in insects and deer, etc...)

Veblen: The Theory of the Leisure Class (1899)

Conspicuous consumption Conspicuous consumption and costly-signaling are really the same thing

Why does this matter for studying human sexuality?

Continual changes in subsistence may have resulted in changes in social structure and sexual systems Increased brain size led to an increase in need for male investment and increased consumption of high nutrient food (meat and fruit) PI Theory: Increased need for male investment should affect mating strategies According to Miller: The huge increase in brain size may have been the result of sexual selection. The brain is like a big, costly, peacock tail for producing art and other mating displays! Miller argues that stone axes were the first form of art and were a way to show off genetic quality Hunting is also a way to advertise genetic quality

In Humans

Continuous sexuality Hypotheses about concealed ovulation Evolved to facilitate monogamy Evolved to confuse paternity Evolved to keep women from choosing not to have children Didn't evolve at all (this is the ancestral state) Jared Diamond Birth-Control Hypothesis

Anatomy

Cranialcapacity:1350cm3 Small teeth & jaws Small & flat face Chin Rounded skull with high forehead Prominentmastoidprocess

Females Can Gain from Cuckoldry

Cuckoldry is when a female cheats on her pair bonded partner Female gains good genes, extra investment, etc... The pair bonded male ends up raising another male's child This is costly to the male, since he is helping someone else's genes rather than his own But females should be choosy in picking their extra pair partners as well

How Does DNA Work?

DNA is composed of four types of nucleotides Adenine (A) Thymine (T) Guanine (G) Cytosine (C) Nucleotides pair up in predictable base pairs (A-T, G-C) The base pairs are linked together in a chain with a sugar- phosphate backbone

Monogamous Primate Penises

Monogamous primates have relatively unadorned penises

Dowries and Bride Prices

Dowries = pay to take daughter Bride price = buy a wife Sudan Nuer: 40 Cows, Papa New Guinea: pigs Men compete for access to females so would expect bride price to be more common, and it is (66% of societies) Verycommon(90%)inpolygynous societies (where 20% of men have more than one wife): i.e., wives = rare resource therefore pay Dowries are 50 times more likely to occur in a socially stratified, monogamous society than in a polygynous or non-stratified one

Parental Investment and Survival

Due to the increased benefits to males of greater competition and also the greater costs of not competing, males put more energy into mating effort This increased energy into mating effort may be detrimental to survival, much as costly armaments and ornaments can decrease survival

Art & Symbolism

Elaborate cave paintings e.g.,Lascaux,France- 17,000 years ago Portableart Personal ornamentation

Sex Biased Investment

Elites favor sons in terms of care and inheritance Lower social scales favor daughters who can marry into wealthier families Often related to the marriage systems and the "value" of sons and daughters Canonlyhavepolygynywheresufficientresourcesthatmalecareisless needed Expect monogamy where male care needed Whereconditionsareveryharshandneedevenmoremalecare,thenwe expect polyandry

Hypotheses about the loss of the baculum

Encourage extended foreplay, which strengthens the pair bond (monogamy) Byproduct of penis elongation for sperm competition (promiscuity) Extended copulation produced an increased chance of phallus injury, which could be minimized by baculum loss

Hormones

Everyone starts out the same in uterus, then burst of pre-natal hormones that organize or prime the brain to be male or female, and set up to respond to hormones at puberty. But male homosexuals have same amounts of testosterone, females same estrogen. Testosterone injections just increase sex drive, not direction. Some evidence that male homosexuals respond to female hormones differently, but it isn't really relevant. Birth order effects (but flexibility here) Present evidence suggests that homosexual behavior is weakly, if significantly, correlated with numerous traits, some genetic, some developmental, and some experiential.

Life History Theory

Evolutionary life history theory identifies fundamental trade-offs faced in the: Life history of species in general, and Those within the socio-ecology of a given species in particular (e.g., Clutton-Brock 1991; Williams 1966). Life history theory examines how natural selection produces adaptations yielding age-related and context-dependent tradeoffs between allocation of available resources to:

Elites Favor Sons

Farmers in 18th century Germany Castes in 19th century India Genealogies in medieval Portugal Wills in modern Canada Pastoralists in modern Africa

Consequences of Sex Differences in Parental Investment

Female PI is usually greater than Male PI A mating act that is inexpensive for a male triggers a large, costly investment for a female (even w/ paternal care). What limits Male RS and Female RS? Female RS - ability to convert resources (e.g., food) into offspring Male RS - limited by access to females

Female Orgasm

Female orgasm may serve to "suck" sperm up The cervix dips into the sperm pool with vaginal contractions Female orgasm may be a way for women to exercise cryptic female choice

Sensory Bias

Females evolve a preference for a certain signal (like color) through natural selection Since females have this preexisting preference, males evolve signals to match

Implications of Differences in Parental Investment

Females evolved to be choosy; mate-choice error very costly Males evolved to be competitive; take every mating opportunity To summarize PI logic ... Females specialize in parenting effort Males specialize in mating effort So, male & female anatomy, psychology, & behavior differ in predictable ways

Infidelity

Females may cheat on monogamously mated partners, causing them to invest in offspring that are not theirs Males may cheat, depriving females of investment

Intra and Inter-sexual Selection

Females may choose males based on the same traits that males use to compete or fight This accelerates the rate of sexual selection

Peak Fertility and Fecundity

Fertility - number of offspring produced Fecundity - chance of getting pregnant Often these terms are used interchangeably by people in different disciplines Not the same thing as RRV Peaks later, around age 22-24

Flowback: Sperm Retention

Flowback (the wet spot) is sperm and vaginal fluids that leak back out of the vagina after intercourse, or are expelled by urination By measuring ejaculate size (with a condom) and flowback size (by having females catch it) researches can figure out how much sperm females retain

Strategic Pluralism

Focus on trade-offs Parenting effort versus mating effort Expands sexual strategies as conditional strategies All individuals possess evolved capacity for multiple strategies Cues in the environment are used to "decide" which strategy is best

Behavior Based Categories

For cross-cultural and evolutionary purposes, behavior based definitions are the most useful For example, Kirkpatrick uses these definitions: A "homosexual" has sex only with the same biological sex A "heterosexual" has sex only with the other biological sex A person with a sexual history including both sexes is "bisexual." The term "homosexual behavior" refers to individual acts of both bisexuals and homosexuals

What is a gene?

Gene: a sequence of DNA bases on a chromosome that specifies order of amino acids in a protein Structural genes-make proteins Regulatory genes-control expression of other genes Two different terms make up what we refer to as a gene: LOCUS ALLELE

Development of food preferences

General preferences are strongly genetic, e.g., desire for salt, sugar, fat Babies can acquire additional preferences based on what foods their mothers eat during gestation and lactation

Monogamy

Monogamy can be life long or serial, that is one monogamous relationship after another Even in monogamous societies, males are more likely to remarry and have a second family than females When males remarry, they often marry a younger female

Primate Trends

Grasping Hands & Feet Opposable Thumbs (& Big Toes) Nails Tactile Pads Reliance on Vision Binocular Stereoscopic Color Increased Encephalization Especially for Memory, Learning, Thinking Increased Sociality & Parental Care Increased Parental Investment Reduced Litter Size (Single Offspring*) Long Gestation Especially development of the mother-infant bond "K" vs. "r" selection Mammals more "K" selected

Australopithecine Reconstruction

Habitual bipeds Small-bodied Retention of arboreal adaptations Pronounced sexual dimorphism Primarily vegetarian Hunted not hunter Limited tools & culture Limited pop. size & distribution

Clark and Hatfield 1989

Had research assistants approach strangers on campus and say: "I've been noticing you around campus. I find you to be very attractive." "Would you go out with me tonight?" "Would you come over to my apartment tonight?" "Would you go to bed with me tonight?" Men said: "Why do we have to wait until tonight?" "I can't tonight but how about tomorrow?" Apologies: "I'm sorry, I'm married." Women said: "You've got to be kidding!" "What's wrong with you, leave me alone."

Mating System and Resources

Highly polygynous societies are found in areas of very high plant productivity Mildly polygynous and monogamous societies (which are equivalent) at about the same level Polyandrous in harsher conditions It is rare - and usually co-husbands are brothers Among northern India Lepcha, the land is poor and requires two men to work to raise one family

The Frequency of Self-reported Homosexuality

How common: depends on society and survey Early survey in 1951 found that 64% of 76 societies tolerated, accepted, or considered normal some forms of homosexual behavior or individuals Homosexuality is found in all parts of the world Some cultures have required periods of homosexuality for certain classes Young males entering a patron-client relationship with an older man Examples 10 to 20% of Melanesian societies require all men to participate in homosexual as well as heterosexual sex In southern China at the turn of the last century,100,000 women joined a marriage resistance movement that included, for many, lifelong homosexual partnerships Mpondo miners of South Africa in the first half of the 20th century commonly entered into "mine marriages," forgoing sexual liaisons with nearby township women In Pacific Island societies other than in Melanesia, such as Tahiti and Hawaii, homosexual behavior was common prior to Western influence In native North America, at least 137 societies had institutional roles for transgenders (berdache) commonly associated with homosexual behavior Homosexual behavior has also been common in societies of native South America, Asia, precolonial Africa, and premodern Europe

Loss of the Baculum

Humans have no baculum (os penis) Boner is a misnomer

The Evolution of Homosexual Behavior

Hypotheses Kin selection: Do homosexuals increase inclusive fitness by helping relatives raise offspring rather than having their own? Manipulation by parents: Do parents benefit (in terms of grandkids produced) from "helper of the nests" even though the offspring don't? Same sex alliances: Homosexual sex as a way to solidify alliances Dominance interactions: Homosexual sex as a way to be dominate over someone else Pleiotropic gene effects: Does a gene conveying benefits in one area create costs elsewhere? Balancing selection: Is bisexuality beneficial and exclusive homosexuality a byproduct of selection for bisexuality?

Genetic kin selection model

Idea: Homosexuality isn't adaptive but the siblings or mothers of homosexuals have higher fitness Some evidence that relatives of homosexuals have higher fertility For example- a gene that increases female fertility (maybe through hormone levels) could also increase the chance of homosexuality

Women and Long-term Mating Problems

Identification of men who can and will invest in self and offspring over the long term Identifying men with good parenting skills Identifying men who are willing to commit Identifying men who will protect them from other aggressive men

Trivers- Willard

If you can only produce a poor quality offspring: produce a female with a good chance of reproducing NOT a poor quality male who will not reproduce If you can produce a good quality offspring: Produce a male that has a high reproductive potential NOT a female whose reproductive potential is limited Subordinates produce females; dominants produce males

Importance of Learning

Importance of Mother-Infant Bond Monkeys raised in isolation Social & reproductive problems Males: couldn't obtain mate Females: couldn't raise own offspring Reproductive & Social Behavior Learned

Ribbed for her pleasure?

In fact, the coronal ridge appears to facilitate sperm displacement In an artificial simulation, penises with realistic ridges and deep thrusting displaced the most semen from a fake vagina Gallup, et al 2003

Cautions about Parental Investment Theory

In some cases, females are selected to be promiscuous particularly in primates like chimpanzees, to combat infanticide or to encourage sperm competition In some cases, males may be choosy (especially when males invest equally or more than females) Females can be competitive, especially when males invest in offspring Sex ratio in a population can affect which sex is the limited resource

Extensions of Parental Investment Theory

Individuals are expected to invest differentially in different juveniles as if in response to three questions: How likely is it that the juvenile is my own offspring? How likely is it that the juvenile will be able to translate investment into future reproductive success? What is the phenotypic state of the juvenile? How much does the juvenile "need" the investment? Can juvenile get investment elsewhere, or acquire the resources on his/her own? What are my alternate potential uses of this investment (e.g., mating effort, other offspring, future offspring)?

Sexual selection includes

Intra-sexual competition (between members of the same sex) Inter-sexual selection (mate choice) Inter-sexual conflict Both pre- and post-copulatory selection In most species males are competitive and females choosy because females invest more in offspring, while males are limited only by access to mates There are many exceptions to this general pattern

Homo neanderthalensis

Large brain 1300-1740 cc (avg. 1520cc) Distinctive cranial & post-cranial features Cold adaptation

Gibbons

Lesser Apes Genus Hylobates 13 species Southeast Asia Small Body Size

Other Ways to Divide Tradeoffs

Life history theory also divides effort into tradeoffs between Past and future reproduction Quantity vs. quality of offspring Mating vs. parenting effort Parental investment (PI) theory is a subset of life history theory PI examines how individuals allocate resources between: different juveniles current versus future offspring quantity versus quality of offspring

Human Mating Systems

Like primates in general, human cultures show variety in mating systems There are costs and benefits of each for men and women

Behavioral Changes after Threat of Infidelity

Males and females report greater depth and vigor of thrusting when there is some threat of infidelity Threats of infidelity increase male arousal, but decrease female arousal An analysis of Internet porn showed men downloaded images with more than one male and a single female more often than other images

Maximizing Reproductive Success

Males emphasize mating Rarely contribute to the cost of gestation so may father more offspring Females emphasize rearing of offspring Once fertilized, females will not increase

Ways Human Females Advertise Ovulation

Males find partner body odors (saliva, vagina, underarms, and loin) most pleasant when from ovulatory phase of cycle Male strangers rated body odor of women from T-shirts sexier when women in fertile phase The skin is lightest, smoothest, and most free of blemishes around ovulation Men are able to distinguish fertile post-pubescent females from infertile pre-pubescent females on the basis of their skin condition Breasts may swell and become more symmetrical

WHR and attractiveness

Males typically prefer a WHR of about 0.7 (Singh 1993 and many other studies)

Does Homosexual behavior reduce reproductive success?

Many individuals who engage in homosexual behavior have children. In a sample of contemporary British women (n = 3,180), bisexuals have significantly higher fecundity to age 25 and no significant difference in lifetime fecundity when compared with heterosexuals (Baker and Bellis 1995). Of approximately 265 homosexual and bisexual men over 30 years old in contemporary Japan, 83%have offspring (Isomura and Mizogami 1992). In some societies, many individuals—probably most individuals— who produce children also engage in homosexual sex (e.g., 15th- century Florence [Rocke 1996], 17th-century Japan [Schalow 1989, Leupp 1995], Melanesia [Herdt 1984a]).

Sexual Swellings

Many primate females advertise oestrus with bright perineal swellings A graded signal that encourages mating with all males during fertile period A defense against infanticide Gradual swelling makes mating with the most dominant male most likely to actually conceive (Dominant males mate at peak swelling, subordinate males, outside peak.)

Homo habilis

Mary and Louis Leakey 1964 TIME: 2.5 - 1.5 mya SITES: East & South Africa Expanded brain (363-600 cc) Moderate body size Smaller molars Thin enamel Stone tools Homo rudolphensis Koobi Fora, Lake Turkana, Kenya 700cc KNM-ER1470

Sex Differences and Strategic Pluralism

Mean sex differences do not imply universal sex differences Most males could want the same number of sexual partners as most females, but if a few males want more partners, this will make the mean higher "Generic claims that women desire certain mating arrangements whereas men desire others raise the questions "which men?" and "which women?"

Results for Men

Men found the women standing alone to be just as attractive (average rating of 2.7) as the woman interacting positively with the child (also 2.7) In fact, the varying contexts made no difference to men in their judgments of how attractive the woman was as a marriage partner; whether the woman ignored the child, stood next to the child, vacuumed a rug, or interacted with the child did not matter—the men's attractiveness ratings remained stable

How Concealed is Human Ovulation?

Meredith Small: undergraduate students in Anthropology courses: Females: 74% said they knew when they ovulated (related to regular cycles, and knowing process) Males with girlfriends: 49% said they knew when she ovulated

A few factors affecting frequency

Military service increases the likelihood of homosexual behavior by 50% (19% of 570 males versus 12% of »880 males, Yates-corrected x2 = 12.6, d.f. = 1, p < 0.001 [Fay et al. 1989:348 n. 30]). Public school attendance in England doubles the likelihood of adolescent homosexual behavior (Johnson et al. 1994).

Female Costs of Reproduction

Milk production & long- term energy costs Limit food access Increased risk of predation

Last Common Ancestor

Mixed arboreal and terrestrial behavior Omnivorous—probably with some meat eating Complex social system Highly social Learning Individual recognition Limited tool use Limited cultural behavior Locomotion?

Primate Social System Terms

Natal group: the group an individual was born into. Female resident groups: females remain in the groups they are born into, males leave (e.g. baboons, macaques). Non-female resident groups: females leave their natal groups and join unrelated females (e.g. spider monkeys, Howlers, chimpanzees, gorillas). Males in the group may or not be related. Matrilineal groups: social systems based on females that are related through maternal descent (e.g. baboons and macaques) Matrilines: within a matrilineal group, a set of closer related individuals, often mother, daughters, granddaughters. A group may contain more than one matriline. Matrilines can be related to each other Rank inheritance in females - daughters rank directly below mother, no matter how "good" the daughters are Age-related rank: female rank by age

Male-male competition

No evidence that human males actually use erect penises in competition Doesn't explain loss of baculum Males might have evolved the ability to use penis size to assess sperm competition threat

Giraffe Necks

Not used for reaching food like Lamarck thought Giraffes use their long necks for fighting

Summary

Offspring are expensive. In most species, the death of the mother equals the death of the offspring. With increased investment through grandmothers, sex determination, alloparenting, offspring and fitness (through kin selection) increase. The juvenile period is especially important. It is often the juveniles who die first during periods of hardship as they are largely on their own as mothers tend to focus on younger offspring.

Olive Baboons

Olive baboon males hold testes to show alliances in aggressive disputes Baboon males also mount each other when forming alliances Or have one way mounting to show dominance

Preferences for Ornaments can Evolve for a Variety of Reasons

Ornaments are honest indicators of: Genetic quality Ability to protect/provide Runaway selection Sexy son hypothesis Sensory bias

Is Ovulation Concealed?

Ovulation is relatively concealed in humans Many primates mate only at estrus

Pygmy Chimpanzee / Bonobo

Pan paniscus Congo River Basin Known as the "Hippy Apes"

Female Adaptations for Reducing Infanticide

Paternity confusion By mating promiscuously, all males have some chance of being the true father

The Human Penis

Physical adaptations provide evidence about the kinds of conditions they evolved in

Orangutan

Pongo pygmaeus Indonesian Islands: Sumatra & Borneo

physical attractiveness

Preferences evolve to motivate behavior in an adaptive direction Attractiveness preferences are like tastes for food Preferences for mates are similar to preferences for food Preferences signal: Health Genetic Quality Fertility (Fecundity) Reproductive Value Resources Status All else being equal, those who are attracted to healthy, fertile mates will have more surviving offspring

Bipedalism

Problem with bipedalism: standing up causes flowback Face to face copulation encourages a horizontal orientation of female reproductive track Mechanisms for maintaining a horizontal posture Post-copulatory cuddling Nocturnal copulation Sedative like effects of orgasm

Female Inheritance of Rank

Rank can be inherited in some social systems (humans, baboons, macaques) Offspring that stay home compete with the mother Competitionhaslesseffectifyouaredominant If you are subordinate, female offspring would inherit low rank and do poorly Malescouldleaveandpotentiallydobetterelsewhere Manybaboonsandmacaques: Subordinatesmakesonswholeaveandcanpotentiallygethigh rank elsewhere and fewer daughters to compete and inherit low rank Dominants make daughters who inherit high rank and can be partners

Residual Reproductive Value

Residual Reproductive Value The expected future contribution of an individual to the next generation at a given age Takes into account age specific chance of dying RRV peaks around menarche since at this point remaining reproduction is maximal and chance of dying before reproducing is minimal

Costs to Men from Short-term Mating

Risk of sexually transmitted disease Reputation as a womanizer could impair mate value when seeking a long-term mate Risk violence at the hands of jealous husbands Cost of repeated mating efforts Testosterone

Earliest Bipeds

Sahelanthropus tchadensis Found in Chad in 2002 Between 6-7 million years old 350 cc brain volume Orrorin tugenensis Kenya in 2001, 6 million years old Fragmentary fossils Ardipithecus ramidus, A. kadabba 4.4 & 5.6 mya, Afar region of Ethiopia 300-350 cc

Inter-sexual Selection

Selection by members of the opposite sex (mate choice) Usually females are choosy Leads to the evolution of ornaments

Semen Coagulation

Semen coagulates within seconds after ejaculation and then liquefies 15-30 minutes later Viscous semen is more difficult to displace Many species of primates form copulatory plugs

Primate Life History Trends

Single Offspring Extended Ontogeny (Prolonged Life Histories) Learning / Behavioral Flexibility

Different kinds of homosexuality?

Some studies group all types of same sex sexual interactions together Others separate them out based on age (i.e. male sex with young boys is excluded) or direction Are client/patron relationships different from partnerships What about homosexual behavior in countries where it is defined based on penetrator/penetrated

Byproducts of heterosexuality?

Some types of "homosexuality" like the behavior of the boyfriends might be byproducts of heterosexual preferences Travesti manipulate themselves to have cues of female reproductive value: Low WHR Butt, hip, and breast "fat" Hair removal Certain patron/client relationships might be similarly explained In study of heterosexual 19 year old males found the bodies of 14 year old males attractive when genitals were removed Boys at this age are long and skinny, low WHR, and without cues of masculinity (shoulders are still narrow, hair not grown, muscles still small)

The Environment of Evolutionary Adaptiveness (EEA)

The EEA is a concept that evolutionary psychologists use to describe the environment in which human adaptations arose Roughly, the EEA is the sum of the conditions experienced by humans in the Pleistocene (1.8 MYA - 10,000 YA) Hunter-gathering lifestyle Small, family based social groups Small scale warfare Human behavior should be adaptive for the EEA, not necessarily our present environment

Alliances

The alliance formation hypothesis is really about bisexuality Homosexual behavior cements male-male alliances which supports survival and (heterosexual) reproduction

Are Humans "Naturally" Monogamous?

The classic argument (Desmond Morris): Human children require investment from two parents The pair-bond evolved to facilitate parental investment Ovulation was concealed to keep males around Constant sex maintains pair-bond, depletes male sperm

What to take away

The evolution of bipedalism occurred along with changes in climate and food Social systems follow food - sexual systems are related to social systems Differences between human and chimp sexuality are related to differences in evolutionary environment As food resources became "patchier," male ability to control females may have decreased, leading away from a chimp like social system to a mildly polygynous system

Cryptic Female Choice

The female's body constitutes the field on which males compete, and her behavior and physiology set some of the rules by which they must abide...small changes in female morphology, behavior, or physiology can tilt the playing field and change the rules, thus biasing paternity. Certain males may be allowed to copulate longer Mate guarding Sperm storage Interactions between male and female genotypes

Promiscuity in Humans

There is a distinction between marriage and mating Similarly, mating systems are idealized versions and do not always reflect what people actually do

The Politics of Studying Homosexuality

There is opposition from multiple sides to the study of homosexuality as a biological phenomena Fear that if homosexuality is biological then it can't be changed (so moral claims about it from the right are unfounded) humans can find a way to change it, i.e. a "cure" (a large fear of the gay rights activists)

The Problem

These studies had participants rate females who were: Only within their reproductive lifetimes (with central tendency toward peak fertility) Either stated to be females and/or exhibiting additional secondary sex characteristics Nulliparous or low parity Unlikely to be pregnant at the time stimuli images were made For example, one stimuli set used in several of these studies had male and female undergraduates rate color photos of 50 real women with a mean age of 26 (range 18 - 42, s.d. 8 years) (Tovee and Cornelissen 2001, Tovee et. al 1999, 2006)

Inter-sexual Conflict

This is competition between males and females Males and females employ different strategies based on different biology

Humans

Through the 42nd president of the US, there have been 90 sons and only 61 daughters. Is this different from 50/50? In general, it has been found that there is a tendency for royalty, aristocrats, and well-fed American settlers to produce more sons than expected. Wives of test pilots, abalone divers, clergymen, and anesthetists tend to have sons. Hormones may influence sex ratio (testosterone and gonadotrophin).

Assessing Symmetry

To assess fluctuating asymmetry, eight bilaterally symmetrical features were measured using 6-inch digital calipers: wrist breadth elbow breadth, ear breadth ear length finger length for four fingers Left and right sides of these features were measured independently to the nearest millimeter

Principle of Independent Assortment

Traitsare inherited independently E.g.flowercolor doesn't affect flower height

TW vs. LRC

Trivers-Willard and LRC predict opposite effects TW -> high rank = males LRC -> high rank = females In baboons and macaques we find TW when less crowding and LRC when crowded (more local competition) Same population can switch with population pressure

Herto, Ethiopia

Two adults & 1 child 196,000 years old Stone tools & butchered hippo bones

Defining Homosexuality

Until relatively recently, Western culture saw homosexuality as something one did, not something one could be Other cultures may define sexuality based on roles taken during sex: penetrator / penetrated e.g. Latin cultures, some Middle-Eastern cultures Sexual identity categories are culture specific

To Test if Are Humans are "Naturally" Monogamous We Need to Know

What people do: Are they behaviorally monogamous? What selective pressures do our anatomy appears designed for: Are we shaped like monogamous creatures? Is ovulation actually "concealed"?

Mating Systems and Sex Biased PI

Where polygyny is practiced, males are a good investment as get better return, can get more wives so have male-biased inheritance Also true in Western society - get bias in wills, when poor, daughters get more, when rich sons get more Bottom line: differential production of sexes is adaptive but balanced by the advantage of the rarer sex at 50/50 ratio

Examples of Alliances

Within age-graded ("patron/client") same-sex alliances in Melanesia, the younger client provides labor in the fields (and sexual services) while the older patron provides food and education In precolonial Tahiti and Hawaii, the clients of powerful patrons gained prestige, as did clients in classical Athens and 15th century Florence In ancient Crete, men without same-sex sexual partners were at a social disadvantage Same-sex sexual partners of the Japanese samurai gained both martial training and land In classical Athens and in Tokugawa Japan, same-sex alliances were thought to be more dangerous to entrenched interests than female-male alliances

Monogamy vs. Polygyny

Women Costs Less ability to have the best available male as a mate No shared parenting with co- wives Benefits High male investment Less competition with co- wives Men Costs Limited reproductive success Benefits Offspring quality (rather than quantity) Paternity certainty Lower male-male competition Higher quality mate

Foreskin and the Pair Bond

Women felt more intimacy with unaltered partners Circumcised: partner cared little about me, glad it's over, discomfort, he's working hard for an orgasm, disinterested, pumping until it hurts me Uncircumcised: relaxed, "complete as a woman", afterglow, "gee that was great", "what a lover"

Results for Women

Women found the man interacting positively with the child to be more attractive as a marriage partner (average rating 2.75) than the man standing alone (2.0) or standing neutrally next to the child (2.0) Women found the man who ignored the child in distress to be the lowest in attractiveness as a marriage partner (1.25) and lowest in attractiveness overall Finally,theeffectofinteractingpositivelywiththechildprovednottobea result of the man showing domestic abilities in general, as the women found the man vacuuming to be less attractive (1.3) than the man standing alone and doing nothing (2.0) Fromthisstudy,LaCerraconcludedthatwomen'sratingsofattractiveness of men as potential mates are increased by cues that they would be a good father and decreased by behaviors indicating poor paternal skills

WHR and health

Women with low WHR have lower risk of cardiovascular disease and diabetes Fat on the abdomen is much worse for health than fat of the hips and thighs

hadza pair bonding

A critical period for provisioning by Hadza men Implications for pair bonding Frank W. Marlowe Peabody Museum, Department of Anthropology, Harvard University, 11 Divinity Avenue, Cambridge, MA 02138, USA Received 9 November 2002; received in revised form 7 March 2003 Abstract Human pair bonding is often attributed to the importance of male provisioning. However, this has been called into question in recent years. Among tropical hunter-gatherers like the Hadza of Tanzania, the foods that men acquire contribute less to the diet than women's foods, are acquired with less regularity, and are shared more widely outside the household. This forces us to ask what benefits forager women gain from being married. Here, I present data suggesting that Hadza women benefit from a husband's provisioning when they have young nurslings. During this critical period, women have lower foraging returns and return rates, while their husbands have higher returns. These higher returns are not due to more meat, but to less widely shared foods, like honey. Even if women are subsidizing husbands much of the time, provisioning by husbands during this critical period of lactation could be enough to favor pair bonding. D 2003 Elsevier Science Inc. All rights reserved. Keywords: Foraging; Hadza; Hunter-gatherers; Lactation; Male provisioning; Pair bonds; Parenting effort 1. Introduction Human pair bonding has long been considered an adaptation to the demands placed upon mothers by needy offspring and the benefits of provisioning by a husband (Darwin, 1871; Lancaster & Lancaster, 1983; Lovejoy, 1981; Westermarck, 1929). This paternal investment theory of human pair bonding has been challenged over the past two decades. Using data on Hadza hunter-gatherers, Kristen Hawkes et al. have argued that pair bonds may have little to do with male provisioning (Hawkes, 1991; Hawkes, O'Connell, & Blurton Jones, 2001a). E-mail address: [email protected] (F.W. Marlowe). 1090-5138/03/$ - see front matter D 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S1090-5138(03)00014-X 218 F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 They note that big game is acquired so sporadically that a woman may go long periods without receiving any meat from her husband. Big game is also shared so evenly across households that the wife of a good hunter may receive no more meat than the wife of a poor hunter, or a single woman. Good hunters give out much more meat than they ever get back, so in-kind reciprocity does not appear to favor their households (Hawkes, O'Connell, & Blurton Jones, 2001b; Woodburn, 1998). Finally, among hunter-gatherers in warm climates, women usually contribute more to the diet than men (Kelly, 1995; Lee, 1968; Marlowe, 2001), which is true of the Hadza. Nevertheless, the trait most frequently cited by Hadza women as important in a potential husband is ''good hunter'' (Marlowe, ND-a). I found that Hadza men with reputations as good hunters had more surviving offspring (Marlowe, 2000). Eric Smith reports similar results for other foragers. He argues (as does Hawkes) the link between successful hunting and reproductive success may be due to costly signaling (Smith, 2002). Because hunting is difficult, a woman might prefer to mate with a good hunter even if she gets no extra food, so long as hunting success is a reliable signal of male quality, for example, vigor (Hawkes & Bliege Bird, 2002; Smith & Bliege Bird, 2000). Hawkes et al. (2001a) found that Hadza men who were better hunters had wives and children with better nutrition. They argue that this is only because better hunters are married to better gatherers, because children's weight gains correlated with the mother's foraging returns, but not the father's. However, they did not distinguish stepfathers from fathers. This makes a difference in my data, with one third of all Hadza children V8 years old being stepchildren (Marlowe, 1999b). A cross-species analysis of birds and mammals revealed that species in which parental care interferes with a mother's foraging were more likely to form pair bonds (Ember & Ember, 1979). Here, I present data suggesting that this same factor could be important in maintaining pair bonds among the Hadza and, by extension, other foragers. Even if a woman subsidizes her husband most of the time, she might still benefit if he subsidizes her during the period when she is nursing and has reduced foraging efficiency. Hadza men may be doing just that. Women with infants had lower foraging returns than other women. Their husbands had higher returns and brought in more food than other men, so long as the children at home were their own offspring and not their stepchildren. The extra calories they brought in did not come from big game but from nonmeat foods, especially honey. Therefore, despite the problem with big game hunting, which Hawkes et al. have pointed out, Hadza women may prefer good hunters because they do receive direct benefits from a husband's overall provisioning. 2. The Hadza The Hadza are nomadic hunter-gatherers who number about 1000 and live in a savanna- woodland habitat in northern Tanzania. The average camp has 29 people and moves once every month or two, with people frequently moving in and out. Polygyny is rare; about 4% of men have two wives. Some couples stay married until death, but most people divorce at least once or twice. Couples shift their residence, but more often live with wife's kin than husband's kin. F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 219 Women dig wild tubers, gather berries, and collect baobab fruit. Men collect wild honey and baobab, and they hunt mammals and birds with bow and arrow. Hadza children forage for themselves from a very early age and are able to meet about half their needs by the time they are 10 years old (Blurton Jones, Hawkes, & O'Connell, 1989; Hawkes, Blurton Jones, & O'Connell, 1995). The Hadza are egalitarian (Woodburn, 1982). Although some men are much better hunters than others, there is no clear dominance or status hierarchy. When medium to large game is brought into camp, it is shared quite evenly across households. When really large animals are killed, people will go to the kill site and carry their own shares back to camp. Honey is also sometimes shared widely, but its distribution can be targeted to particular individuals more easily because it is easier to hide than large game. There is a dry season from June through November. During this time, camps grow larger because they must be near the few permanent waterholes. Normally, men hunt alone. However, during the late dry season, men hunt in pairs, waiting all night at a waterhole to ambush animals coming to drink. Rains last from December through May. During these months, honey, which the Hadza say is a good weaning food, becomes very important. Men collect honey alone or in pairs, but also often go collecting with their wives. Hadza women forage in groups of three to eight, and they take their infants in a sling. Infants nurse on demand, nursing about 20% of the time soon after birth and about 5% of the 1 time when they are over 1 year old. Children are usually completely weaned by age 2=2-3 years (Marlowe, 2002). Women usually do not take toddlers foraging because they are too small to keep up, but too big to carry. Therefore, someone must stay in camp to watch toddlers, and food must be taken back to provision them. 3. Methods The data presented here were collected over 9 months in 1995-1996 in five different camps (population mean=37, median=24). All food arriving in camp was weighed and edible portions multiplied by caloric value of food type (Marlowe, 1997). This caloric score was divided by the number of days sampled in order to compute daily kilocalories (Kcal), which makes data comparable across camps. These data include only food that is brought back to camp, not what is consumed while out foraging, which is a considerable amount. However, because men forage alone most of the time, and because I am interested in provisioning, food arriving in camp is what matters. These data are only proxies for the actual amount of food eaten in the household, something that is difficult to measure, given how much sharing occurs and how quickly food disappears. Instantaneous scan observations were conducted each hour to measure the time a person was in and out of camp. Time out of camp was used as a proxy for foraging time. Continuous observations of focal individuals were conducted for 30 min per bout on children ( V8 years old) within three time blocks during daylight hours. From these, and from scan observations, I calculated nursing frequency. 220 F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 Ages of all individuals were estimated and compared with previous data where birthyears and/or relative age rankings were known (Blurton Jones, personal communication). Paternity was assessed by asking men and several other people whether a child was the man's own offspring or his stepchild (for more details, see Marlowe, 1999a). I use the term ''offspring'' to refer to a man's own genetic offspring as opposed to his stepchild. Hunting reputation was calculated by asking adults to name the three best hunters they know and summing up the number of times a man was nominated. In the five camps in this sample, there were 183 individuals altogether, but four of these appeared in two different camps, so N = 187 in Table 1 and the top row of Table 2. Seven of these people were excluded from the statistical analyses of the critical period effect, that is, when the sample consists of people with and without young children (for a complete explanation, see Appendix A). Statistical analyses included Pearson correlations, t tests, and multiple linear regressions. 4. Results Table 1 shows the Hadza diet as measured by daily Kcal of food by type. The 6.6% of the diet consisting of maize and millet was given to the Hadza in the largest camp by a missionary, or acquired through trade with neighboring agro-pastoralists. Because much consumption occurs while foraging in the bush, the percent of domesticated food in the total diet was actually less than 6%. During the berry season, the bulk of the diet is berries, and because they are mostly eaten while foraging, they surely comprised a greater percent of the diet than Table 1 implies. Hadza men (z18 years old, n = 51) foraged 5.7 h/day, and women (z18, n = 59) 4.2 h/day, on average. There was a slight dip in foraging time during the reproductive years for women (age 19-48 years). Table 1 shows that the food brought into camp by women, was comprised mostly of baobab and several species of tubers and berries. Of the food men brought into camp, only 40% was meat of all kinds, from small birds and mammals like bush babies Table 1The Hadza diet Food (a) contributed by: Berries (%) (1) Both sexes of 17.2 all ages (2) Women (z18 24.8 years old) (3) Men (z18 6.3 years old) (b)Tubers (%) 23.5 38.8 5 (c)Honey (%) 14.2 0.7 30.2 (d)Baobab (%) 19.2 25.6 13.8 (e)Meat (%) 19.3 1.2 39.6 (f)Maize/millet (%) 6.6 8.9 5.1 (a-f) Total (%) 100 100 100 Percentages show the amount that each food type contributes to the total diet, as measured by daily Kcal of food brought into camp by: (1) people of all ages, (2) adult women, and (3) adult men. F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 221 (Kcal/h) Kcal/h 343S.D. = 555 (n=93) 597S.D. = 646 (n=51) 666S.D. = 706 (n=40) 642S.D. = 855 (n=18) 678S.D. = 867 (n=17) 690S.D. = 137 (n=6) Percentages inand females within the sample in each row. Note that while women's returns decline with younger children, men's returns increase. (Galago senegalensis), to the largest game, giraffe (Giraffa camelopardalis). Honey accounted for 30% and baobab 14% of the daily Kcal men brought into camp. This is important because men often sneak honey into their huts. It is only medium-sized and larger game that is shared so evenly across households. Big game is not only shared with others in camp, but even with people from other camps. For example, 27.2% of all meat arriving in camp in 1995/1996 was killed by a man from a different camp. Table 2 shows how many daily Kcal males and females of various categories brought into camp. Adult females brought into camp 3076 daily Kcal (57% of the adult total) and adult males 2792 daily Kcal (43% of the adult total). The relative contribution by sex was about equal for adults who were married. Among couples with young offspring, paternal contribution was greater than maternal contribution. Among those with an offspring under 3 years old, men accounted for 58% of daily Kcal, and 69% of daily Kcal among those with an offspring under 1 year old. These data alone suggest that men might be provisioning in a way that benefits their wives when they have young offspring. It is clear that women with nursing children could benefit from help with provisioning. Married women who were nursing brought in fewer daily Kcal than married women who were not (t = 2.41, P = .021, df = 35), and the more frequently a woman nursed, the lower were her returns (r=.329, P=.047, n=37). In addition, among women with a child under 3 Table 2Hadza mean foraging returns as measured by daily Kcal of food brought into camp, and return rates Males SampleAll ages (N = 187) Females Kcal/day 2226 (58%) S.D. = 1866 (n=94) 3076 (57%) S.D. = 1726 (n=59) 3016 (50%) S.D. = 1900 (n=41) 2697 (47%) S.D. = 2056 (n=19) 2346 (42%) S.D. = 1650 (n=17) 1713 (31%) S.D. = 1409 (n=6) Kcal/h 561S.D. = 463 (n=94) 795S.D. = 414 (n=59) 764S.D. = 435 (n=41) 768S.D. = 550 (n=19) 693S.D. = 476 (n=17) 451S.D. = 332 (n=6) Kcal/day 1633 (42%) S.D. = 1962 (n=93) 2792 (43%) S.D. = 1966 (n=51) 2990 (50%) S.D. = 2025 (n=40) 3049 (53%) S.D. = 2369 (n=18) 3227 (58%) S.D. = 2316 (n=17) 3851 (69%) S.D. = 1283 (n=6) Adults (18 and Married adults Married adults <8 years old Married adults <3 years old Married adults <1 year old up) with offspring with offspring with offspring parentheses next to daily Kcal are the proportion of the whole diet (100%) contributed by males 222 F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 years old, the younger a woman's youngest child was, the lower was her hourly return rate (r=.629, P=.009, n=16) (Fig. 1). To assess the potential benefit of a husband's provisioning more precisely, I analyzed the returns of couples. I subtracted from each woman's own returns the returns of her husband. The more negative the number, the more a husband compensates for his wife's returns. I limited the sample to people aged 18-65 to exclude the very old, but not exclude some fathers. Even though women are not caring for their own young children beyond their early 50's, some men do have very young children at 65. Therefore, I controlled for age in the following analyses. Among all married women without a child under 1 year old, women actually brought in slightly more daily Kcal than their husbands. However, among married women with a child under 1 year old that was the offspring of her husband, women brought in an average of 2138 daily Kcal less than their husbands. Having an offspring under one year old was a significant predictor of the gap between a man's returns and those of his wife (b=.387, P=.030, df = 30, controlled for the woman's age) (Fig. 2). Husbands appear to compensate for their wives' returns until about the time of weaning. Women who had children less than 3 years of age that were the offspring of their husband brought in less food than their husbands, whereas those who did not brought in more food than their husbands. Having an offspring under 3 years old was also a significant predictor of Fig. 1. Foraging return rate, as measured by daily Kcal of food taken back to camp by women per hour of foraging time, by the age of a woman's youngest offspring (r = .629, P = .009, n = 16). The data show a newborn lowers a woman's foraging efficiency. F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 223 Fig. 2. Foraging returns, as measured by daily Kcal that a woman takes back to camp, minus the daily Kcal that her husband takes back to camp, for couples (age 18-65) with, and without, a child under 1 year old who is the offspring of the husband. The presence of an offspring under 1 year old predicts a greater amount of food brought back by a husband relative to his wife (b = .387, P = .030, df = 30, controlled for the woman's age). Lines inside boxes show medians, box boundaries show quartiles, and T's show extremes. the gap between a couple's daily Kcal (b=.366, P=.048, df=30, controlled for the woman's age). This disparity between a couple's returns is not due merely to the lower returns of women with young children. Men who had an offspring V8 years old at home, young enough to still require considerable provisioning, brought in more daily Kcal than men who did not (b = .358, P = .029, df = 36, controlled for the man's age). Some men had both their own offspring and stepchildren at home, and the presence of a stepchild was associated with lower returns for men. After controlling for the presence of a stepchild V8 years old, married men brought in more daily Kcal if their youngest offspring was younger. The presence of an offspring V8 years old predicted higher returns for men (b = .379, P = .022, df = 35, controlled for the man's age). Without controlling for the presence of a stepchild, this relationship was obscured. The difference between the returns of men who did, and did not, have offspring V8 years old at home was greater when meat was excluded from the analysis. Men who had offspring V8 years old at home had higher nonmeat returns than men who did not (b = .464, P = .004, df = 36, controlled for the man's age). This was due largely to those men getting more honey 224 F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 (b = .387, P = .018, df = 36, controlled for the man's age). There was no difference between men who did, and did not, have young offspring in terms of daily Kcal of meat. Focusing only on meat, therefore, obscures the benefits of a husband's provisioning, just as does ignoring the presence of a stepchild. As long as men do not focus exclusively on big game, hunting may not be such a poor strategy for household provisioning. Hunting reputation was correlated with men's meat returns (r=.275, P=.051, n=51), but more strongly correlated with nonmeat returns (r=.341, P=.014, n=51). It was most strongly correlated with total returns of all foods (r=.375, P=.007, n=51). Returns of all foods were even more highly correlated with nonmeat (r=.862, P=.000, n=51) than meat returns (r=.486, P=.000, n=51). Men with reputations as good hunters, therefore, are not just good at getting big game, but also good at acquiring all types of food. 5. Discussion These data show that an infant lowers a Hadza mother's foraging efficiency. Similar findings have been reported for the Hiwi and Ache (Hurtado, Hill, Kaplan, & Hurtado, 1992), and even for the Hadza (Hawkes et al., 2001a). What I add here is that Hadza husbands compensate for their wives' lower returns if they have young offspring. My data are only proxies for household provisioning. However, because men with young offspring brought in the same amount of meat as other men, plus more of other foods like honey, their higher returns probably resulted in more food going to their own households. Hawkes et al. have argued that hunting is not an efficient strategy for provisioning one's household. They have described Hadza men as targeting big game, whereas I found them willing to shoot at almost any mammal or bird they felt they could hit (see Appendix B). Hadza men do express a preference for larger game because, as they say, ''there is more meat.'' Even though big game is shared more widely in camp than small game, in each encounter with a prey animal the hunter gets more absolute kilograms of meat (and so do others) when he kills a big animal than a small animal. For example, 1/20th of a 200-kg zebra shared by 20 people is greater than 1/2 of a 5-kg rock hyrax shared by two people, so Hadza men are happier when they encounter a zebra than a rock hyrax. Therefore, as long as men are not targeting big game exclusively, taking big game when possible is not a bad strategy for household provisioning. Foraging men may specialize in meat and honey because these foods have the highest trade value with women. A Hadza woman is undoubtedly more pleased if her husband comes home with meat or honey rather than extra piles of tubers and berries, which she can get herself. It is because women acquire foods that can be counted on everyday that men are free to pursue less predictable, but potentially more rewarding foods. Still, when men failed to kill game or find honey, I often saw them stop on the way home to get baobab fruit to take back. Hawkes et al. interpret much food sharing as tolerated scrounging, and I agree (Blurton Jones, 1987; Hawkes, 1992; Hawkes et al., 2001b; Marlowe, ND-b). When a Hadza man has food (especially meat) and others see it, he feels he simply must share it. However, what others do F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 225 not see, one does not have to share. In larger camps, people sometimes waited just outside camp until dark then signaled me to come and weigh their food, but to be discreet about it. When Hadza men sneak food into their huts, they tell me that they are doing so because they want to feed their family. Men do indeed lose much of their food to tolerated scrounging, as Hawkes et al. note, but this should not count as evidence that they are uninterested in provisioning their households. That would only be true if men were passing up foods that are less likely to be scrounged but still highly valued by their wives, which is something I did not see. There are several ways besides costly signaling that it could pay a woman to marry a good hunter even if his food is widely shared. (1) Variance in daily family consumption should be reduced if, in general, he targets different foods than she does (Winterhalder, 1986). (2) If his gifts of meat were repaid with nonmeat foods, variance would be reduced and diet breadth increased. (3) If many of those in camp receiving his meat were her close kin, she may gain inclusive fitness benefits. (4) She should always be living in a camp with at least one good hunter (her husband). Even if she receives only a little more food than other women do in the short term, this could add up to a significant benefit in the long run. Hawkes et al. (2001a) have argued that provisioning by a woman's mother is more important than provisioning by her husband. However, there is no reason not to get help from both, or whichever one is available, as Hadza women do with direct childcare (Marlowe, 2002). There are reasons to suspect that Hadza women gain some economic benefit from a husband. They cite ''good hunter'' as the single most important trait in a potential husband (Marlowe, ND-a). They get angry when they catch their husbands pursuing other women; if they did not risk losing some benefit, it is difficult to see why they should get jealous. Men say they feel pressure from their mothers-in-law to be productive foragers. Men with reputations as good hunters (who, as I show here, are good foragers in general), have higher reproductive success. Good hunters do not have more wives over their lifetimes, but they do have a better chance of marrying reproductive-aged women after their previous wives reach menopause (Marlowe, 2000). The most parsimonious explanation to account for all of these observations is that a wife can receive direct benefits from her husband's provisioning. He does not even need to bring home more food than her, just more than he consumes (of his food and hers). There is an alternative interpretation of these data that is more consistent with costly signaling. If the best hunters are desired by females for their inherent qualities, not for their household provisioning, then better hunters should have first choice of wives. If better hunters choose younger, more fecund women, they should be more likely to have young offspring. This could result in a correlation between having a young offspring and having higher foraging returns (as observed here). One strike against this interpretation, however, is that most Hadza men do get married and reproduce. It seems unlikely that the chance of having a young offspring would be so much greater for the best hunters, great enough to account for these results. It is more plausible that men increase the amount of food that they take back to camp, at least a bit, when they have young offspring. Women's foraging strategies surely reflect parenting effort more than men's do (Bird, 1999; Hawkes et al., 2001a; Marlowe, 1999b), yet male provisioning could still be an 226 F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 important factor in pair bonding. Men's foraging might be motivated by gains from increased mate access and retention, rather than increased offspring survivorship. Pair bonds could be a mate-guarding strategy for males, but a way to get help with rearing children for females. Still, in the data presented here, the presence of a stepchild did make a difference. This means that either (a) men are less motivated to provision unrelated children, or (b) better foragers are less likely to be living with a stepchild. If the latter is true, it would presumably be because good foragers have higher mate value, which affords them greater choice of mates, and they more often choose to marry women who do not already have other men's young children. Either way, paternity matters, as it should. Even if mate acquisition and retention is what motivates a man's foraging, he gets more fitness benefit if his food goes to feed his own offspring rather than his stepchildren. The paternal investment theory of pair bonding tends to put the cart before the horse by suggesting men provision because offspring are needy. Surely, ancestral females did not begin to bear offspring so needy that they could not rear them on their own before others began to acquire more food than they consumed. The surplus production that makes provisioning possible must have come first. Life history then coevolved with provisioning. Women in many societies like the Hadza can rear offspring without husbands. However, they may take longer to conceive the next child, especially if they do not have a mother around to help. If so, women should be able to rear more offspring in less time when they have a husband providing food (Marlowe, 2001). In conclusion, Hawkes and colleagues have greatly advanced our understanding of hunter- gatherer mating systems by exposing problems with the paternal investment theory. However, the pendulum may now be swinging too far in the opposite direction because they have focused too much on big game hunting and too little on the difference between fathers and stepfathers. Hadza men acquire foods other than big game. This means that despite widespread sharing, much of their food probably stays in the household. Hadza women are less efficient foragers when they are burdened with infants and toddlers. During this period, their husbands bring in more food, but only if the young children are not their stepchildren. Women stand to benefit most from provisioning when lactation places the greatest demands upon them. At that time, they are less likely to be ovulating, and should be less sexually desirable. Without pair bonds then, this is just when they should be least likely to get food from males. Offering men increased paternity confidence through pair bonding may be the best strategy for women to gain provisioning for themselves and their children during this critical period. Acknowledgments This research was supported by L.S.B. Leakey grant no. SL952012 and NSF grant no. 9529278. I thank Nicholas Blurton Jones for mentoring, Melissa Gerald, Peter Gray, Helen Fisher, Kristen Hawkes, Martin Daly, Margo Wilson, and an anonymous reviewer for comments, David Bygott and Jeannette Hanby for hospitality, COSTECH for research permission, and the Hadza for tolerance and companionship. F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 227 Appendix A. People excluded I worked in six camps but one whole camp was excluded because it was close to a village where both men and women, but especially men, went almost everyday in search of alcohol and where foraging was quite altered by normal bush standards. All 187 people in the other five camps were included in Table 1, and in the top row of Table 2 where N = 187 (actually 183 people, but 4 people appeared in two camps). In statistical analyses of the critical period, that is, when the sample is married couples with and without young children, or men with and without young children, the people described below were excluded from the total of 187. Two married men with young offspring were excluded because they were disabled and did not forage at all. One had recently broken his arm badly and could not use a bow. He stayed in camp with his young offspring, while his wife brought in more daily Kcal than any other person. The other man had severe asthma. He and his wife normally spent most of their time in a village. Both he and his wife stayed in camp all day long and brought in no food, being fed by relatives. A third man who was blind, not married, had no children, and brought in no food was also excluded. One woman appeared in two different camps with her infant. Her husband was the only man in the sample with two wives, and he frequently traveled between the two different camps where his wives lived. While the wife with an infant resided in the first camp, he appeared only sporadically. In the second camp, she moved in with him and his other wife. Data used on this couple come only from the second camp because only there was the couple actually frequently together. Another woman with a young child had a husband who was gone most everyday for a month pursuing a younger woman in another camp. When he returned briefly his wife yelled at him all day. This couple is also excluded. Both of these couples divorced soon afterwards. One might object to excluding these subjects, but their exclusion reveals that when a man is disabled, or when a marriage is unraveling, provisioning levels are affected, just as they are if men are unmotivated to provision a stepchild. The inclusion or exclusion of these cases actually sheds light on the potential for women to benefit. Finally, two other men were excluded. One is an outlier excluded simply because, if he is not, his extremely high returns for a brief period of extraordinary hunting success obscure the results presented here. Because this man had such high returns and happened to have no young offspring, the results presented here are not significant unless another man, who did have a young offspring, is included. However, I excluded this other man because scoring his returns is problematic. If I score all the meat from a buffalo as his returns, then he also has incredibly high returns, and because he does have a young offspring, these results are significant. This buffalo, however, was actually killed, along with another one, by neighbor- ing Datoga pastoralists. They gave one buffalo to him as repayment of meat that they had received from him in the past, according to him. Although he did not kill the buffalo, perhaps it should be counted as his. However, because several men and women went to carry the meat back, I gave credit to each one of these people for the amount each brought in (in Table 1, these meat Kcal are assigned to the total for males). Because I must exclude the one 228 F.W. Marlowe / Evolution and Human Behavior 24 (2003) 217-229 successful hunter without any young child to find these significant results, it seemed only fair to also exclude the other man as well, because he did not actually kill the buffalo. The important thing is that if I leave both of these men in, all the significant results reported here do hold. Appendix B Perhaps I observed less big game hunting in 1995-1996 than Hawkes et al. did in 1985- 1986. Although they provide no figures on the percent of the diet that was meat, I found it to be 19.3% of daily Kcal, which appears to be no less than what Woodburn (1968) observed in the early 1960s. He estimated 20% of the diet by weight was meat and honey combined. Meat and honey accounted for 22.2% of the diet by weight in my data.

Homosexuality and epigenetics

THE COMMON occurrence of homo- sexuality is perplexing from an evolu- tionary perspective. Simple logic suggests that a fitness-reducing phenotype should be selected against, but homosexuality is none- theless surprisingly common in human pop- ulations—e.g., a prevalence of about 8% in both sexes was reported in a large and sys- tematic sample in Australia (Bailey et al. 2000). Existing genetic models for the evo- lution of human homosexuality can be sep- arated into two major classes: one based on kin selection (Wilson 1975) and another based on sexually antagonistic alleles and/or overdominance (Camperio-Ciani et al. 2004, 2008; Gavrilets and Rice 2006; Bailey and Zuk 2009; Iemmola and Camperio-Ciani 2009). These models are all based on special cases of selection that directly, or indirectly, maintain genetic variation at loci contributing to the homosexual phenotype. However, despite nu- merous studies over the last decade searching for polymorphisms associated with homosexu- ality, no convincing molecular genetic evi- dence has been found despite the fact that pedigree and twin studies clearly show that ho- mosexuality is familial (reviewed in Ngun et al. 2011). Homosexuality has also been hy- pothesized to be caused by nongenetic fac- tors such as maternal antibodies against male-specific antigens (reviewed in Bogaert and Skorska 2011). This hypothesis may in- deed explain some cases of homosexuality, but cannot account for most cases in men and none in women (Cantor et al. 2002). The poor correspondence between current models and data calls for a new conceptual framework to understand the evolution of homosexuality. Here we integrate theory from evolution- ary genetics with recent developments in the regulation of gene expression and 50 years of research on androgen-dependent sexual development. We first find that the existing paradigm of mammalian sexual develop- ment is incomplete, with the missing compo- nent being a system to canalize androgen signaling during fetal development such that the response to circulating testosterone is boosted in XY fetuses and blunted in XX fetuses. We integrate these data with recent findings from the epigenetic control of gene expression, especially in embryonic stem cells, to develop and empirically support a mathe- matical model of epigenetic-based canalization of sexual development. The model predicts the evolution of homosexuality in both sexes when canalizing epi-marks carryover across generations with nonzero probability. We will use the term epi-marks to denote changes in chromatin structure that influ- ence the transcription rate of genes (coding and noncoding, such as miRNAs), including nucleosome repositioning, DNA methylation, and/or modification of histone tails, but not including changes in DNA sequence. It is now well established that a parent's epi-marks some- times carryover across generations and influ- ence the phenotypes of offspring (reviewed in Morgan and Whitelaw 2008). Epigenetics is a relatively new subdiscipline in genetics and its importance in evolution, especially as a major contributor to realized heritability, is currently being developed and debated (e.g., Slatkin 2009; Furrow et al. 2011). Nonetheless, there This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 345 is now clear evidence that environmentally induced epigenetic modifications of genes expressed in male mice (e.g., DNA methyl- ation; Franklin et al. 2010) that feminize their brains and behavior can be transgen- erationally inherited by their offspring (Mor- gan and Bale 2011). Our study examines the ramifications of transgenerational epigene- tic inheritance to the phenomenon of hu- man homosexuality. The first half of our analysis is general and applied to all sexually dimorphic traits in mammals that are strongly influenced by fe- tal/neonatal androgen exposure. The sec- ond half focuses on homosexuality and its similarity with other common gonad-trait dis- cordances that have important medical sig- nificance. By homosexuality we mean any same-sex partner preference, spanning all Kinsey scores 0 (e.g., including bisexual- ity). Our model of homosexuality may also apply to transsexualism, but we do not de- velop this application here. Classical View: Sex Hormone Differences Fully Determine Sexual Dimorphism Beginning with Phoenix et al. (1959), a long succession of studies have consistently and unambiguously demonstrated that sex- ual dimorphisms of the genitalia and brain of mammals are strongly influenced by an- drogen exposure during fetal development. The foundation for this conclusion is that XY fetuses experimentally exposed to androgen antagonists during gestation develop femi- nized genitalia, brains, and behavior, whereas XX fetuses exposed to elevated androgens de- velop masculinized phenotypes for these same traits. Studies of untreated fetuses demon- strated that circulating androgen levels differ between XX and XY genotypes, with signifi- cantly higher average androgen levels in XY fetuses at a time at or before the genitalia, brain, and behavior become sexually dimor- phic. The logic of the "prenatal androgen par- adigm" (also known as the Jost paradigm) be- gins with the observation that only XY embryos express the Y-linked gene SRY (Figure 1A). This gene product induces development of the testes in XY embryos, which in turn pro- duce androgens that influence later sexual development. The absence of elevated circu- lating androgens during fetal development leads to the female phenotype. Although many aspects of sexual dimorphism are a response to the "organizational" effects of sex-specific dif- ferences in circulating androgens during fetal and neonatal development, full manifestation of sexual dimorphism sometimes depends on the "activational" effects of androgens and estrogens at and after puberty. In hu- mans, the fact that XY individuals (with fully formed testicles and normal levels of circu- lating androgens) that are homozygous for a null allele at the androgen receptor locus (and therefore cannot respond to circulat- ing androgens) develop fully female-typical genitalia and reproductive behavior (re- viewed in Wisniewski et al. 2008) provides strong support for the prenatal androgen paradigm. Sex Hormone Differences are not Sufficient to Produce Sexual Dimorphism Although prenatal androgen levels play a fundamental role in sexual development, there is also evidence that the prenatal andro- gen paradigm is at least partially incomplete (reviewed in Davies and Wilkinson 2006). Stud- ies in the mouse "four core" model system (in which a the male-determining Sry gene has been translocated to an autosome, enabling gender and sex chromosome karyotype to be experimentally manipulated independently) clearly demonstrate that some aspects of sexu- ally dimorphic behavior and brain anatomy are strongly influenced by the sex chromosome karyotype rather than the level of fetal andro- gen exposure alone (reviewed in Arnold and Chen 2009). These studies are, however, con- sistent with the conclusion that androgen sig- naling is the predominant factor controlling sexual dimorphism in this model system. Here we provide evidence that the prena- tal androgen paradigm is missing a major component. This conclusion is based on our reanalysis of studies of circulating prenatal androgens in human and rat fetuses. In hu- mans, the testes begin to secrete testosterone (T) in XY male fetuses beginning around the eighth week of gestation (Wilson et al. 1981). This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms 346 THE QUARTERLY REVIEW OF BIOLOGY Volume 87 Figure 1. The Sexual Dimorphism Signaling Pathway The classical view of sexually dimorphic development (A) is that higher androgen levels in XY fetuses and adults masculinize sexually dimorphic traits and lower androgen levels in XX fetuses and high estrogen in adults feminizes development. Our analysis (B) indicates that androgen signaling includes an additional component: it is canalized by epi-marks that are produced during the embryonic stem cell stage of develop- ment. However, T is also present in XX female fetuses in substantial amounts, originating from the fetal adrenals and from placental/ maternal sources. Secretion of T by the testes increases its concentration in the blood of XY fetuses. The maximum average differ- ence in T concentration between male (XY) and female (XX) fetuses occurs between weeks 11-17 (Reyes et al. 1974). After this time, fetal secretion of T by the testes de- clines markedly, causing average T values in males to become indistinguishable from lev- els in females (Reyes et al. 1974). Although male fetuses have higher average T than fe- males starting around week 11, overlap in T levels between the sexes (i.e., some XX fetuses having higher T than some XY fe- tuses) was observed at all times except be- tween weeks 15-19 (Reyes et al. 1974). This transient lack of overlap (and hence an un- ambiguous signal of fetal gender) may be genuine or an artifact due to small sample size. The latter explanation is supported by a much larger study (166 female and 185 male fetuses) of amniotic fluid collected between weeks 15-19 (T diffuses from the fetal circu- lation into the amniotic fluid via the skin at this stage of development), in which there was about 5% overlap (i.e., 5% of XX fetuses had higher T than some XY fetuses) in T concentration between male and female fe- tuses (Perera et al. 1987). In a large study of rats, significantly higher circulating T in male compared to female fetuses occurred only between days 17-21 of gestation (Weisz and Ward 1980). Despite this window of sig- nificantly elevated T in male fetuses, T levels overlapped between the sexes throughout all time points during gestation (Weisz and Ward 1980). Collectively, these studies indi- cate that the level of circulating T alone is not an unambiguous indicator of gonadal sex at any time during fetal development because T levels overlap between the sexes at nontrivial frequencies at all developmental time points. This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 347 Overlap in T concentrations between the sexes (despite highly significant differences in average T between XX and XY fetuses) would make the prenatal androgen para- digm incomplete (i.e., missing an important component of androgen-induced sexual di- morphism) unless discordance between the gonad and sex-specific traits is observed to be correspondingly common. This is, however, not the case. To illustrate this point, we can focus on the ontogeny of the genitalia. The human male phallus and female vulva are formed during weeks 9-15 of gestation, al- though the phallus requires T to continue to grow during later fetal development (summa- rized in Wilson et al. 1981). During this time of genital differentiation, data on fetal T collected by Reyes et al. (1974) show high overlap in T concentrations between the sexes. The same relationship is found in the rat, in which T concentrations strongly overlap between the sexes during the time (and all points previous) when the phallus and vulva are differentiating (Weisz and Ward 1980). Yet discordance be- tween the gonad and the genitalia (including ambiguous genitalia) is rare in both humans (Sax 2002) and rats (Ostby et al. 1999; Hotch- kiss et al. 2007). Therefore, the available data do not fully support the prenatal androgen paradigm because there is too much overlap in circulating androgens to be consistent with the observed low discordance between the gonad and the genitalia observed in both humans and the rat model system. Differential Sensitivity of XY and XX Fetuses to Androgens One can fully rescue the prenatal andro- gen paradigm if XY fetuses have higher sen- sitivity to circulating androgens compared to XX fetuses. In this case, XX and XY fetuses would respond differently even when T lev- els overlap to a limited degree between the sexes. Many lines of evidence indicate that this is the case. First, in humans the expression of the 5-- reductase-2 gene, which converts T into the more potent androgen dihydrotestosterone (DHT), is three times higher in XY fetuses than XX fetuses within the urogenital swell- ings and tubercles (structures that develop into the phallus or vulva; Wilson et al. 1993). Boehmer et al. (2001) review evidence that strongly supports the conclusion that this sex-specific difference in gene expression is not androgen-induced via a feed-forward process (i.e., due to changes induced by higher T in XY fetuses during earlier devel- opment). Higher conversion of T to DHT would permit XY fetuses to develop male traits even when T levels overlap (to a limited degree) with XX female fetuses, thereby promoting phallus development despite low circulating T. Similarly, lower 5--reductase production in XX females would prevent or reduce masculinization of the vulva when T levels overlapped (to a limited degree) with those of XY males. Second, sex hormone binding globulin (SHBG) binds circulating T and makes it unavailable for uptake by cells. In human fetuses in which T levels overlap between genetic males and females, SHBG is mark- edly higher (approximately 50%) in female fetuses compared to male fetuses (Ham- mond et al. 1983). This elevated SHBG in XX fetuses would reduce sensitivity to circu- lating T when it overlaps with XY fetuses. Third, in rhesus monkeys (but not hu- mans), levels of circulating progesterone are markedly higher (three times) in female fe- tuses compared to male fetuses (Hagemenas and Kittinger 1972). Progesterone acts as an anti-androgen because it has a high binding affinity for the androgen receptor (AR), which it inactivates. Its higher concentration in female fetuses is expected to lower their sensitivity to androgen levels that overlap with males. Fourth, human XX female fetuses homo- zygous for loss of function alleles at the CYP21 locus cannot produce the steroid hor- mone cortisol due to a block in its synthetic pathway (a form of Congenital Adrenal Hy- perplasia, CAH). Buildup of intermediate products leads to their conversion to T, and consequently highly elevated circulating T in affected XX CAH fetuses. This elevated level of T begins in the seventh week of gestation (Speiser and White 2003; Trakakis et al. 2009) and "the developing fetus is exposed to the excessive adrenal androgens, equiva- lent to the male fetal level, secreted by the hyperplastic adrenal cortex" (New 2004), This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms 348 THE QUARTERLY REVIEW OF BIOLOGY Volume 87 including the period of maximal average T excess in XY fetuses (Forest 1985). Despite a male-typical level of T throughout fetal devel- opment, the genitalia of XX newborns with CAH are usually only partially masculinized— about halfway between a typical male and fe- male genital (Hall et al. 2004), as is childhood sexually dimorphic behavior (Hines 2011). Al- though rates of homosexuality and transsexu- ality are elevated in CAH patients, the vast majority have female-typical sexual behavior (reviewed in Hines 2011). These data provide strong evidence that androgen-induced mascu- linization is blunted in the XX fetuses. Fifth, human XY male fetuses with acute 17-HSD-3 deficiency are homozygous for loss-of-function alleles at the 17-HSD-3 lo- cus and cannot produce T in the testes due to a block in its synthetic pathway (reviewed in Rey and Grinspon 2011). Buildup of the precursor androstenedione occurs in affected individuals, but this steroid is a much weaker androgen than T (about a hundredfold lower binding affinity for the androgen receptor; Fang et al. 2003). These males experience highly reduced circulating T throughout fetal development, although T is somewhat elevated later in fetal development due to allozymes expressed outside the testes that convert circulating androstenedione to T. Androgen-induced Wolffian duct structures (epididymis, vas deferens, seminal vesicles, ejaculatory ducts) that are in close proximity to the testes developed normally despite the low level of T they experience during their ontog- eny. The more distant genitals of affected indi- viduals, however, are highly feminized and most affected newborns are reared as females. At puberty, these individuals experience a surge in T due to the nontesticular conversion of circulating androstenedione to T (via allo- zymes of 17-HSD-3) and about one-half of these individuals, reared as girls, change their sex to male (reviewed in Wisniewski et al. 2008). This is the same rate of sex change as XY individuals with normal levels of T through- out life that were raised as girls because they were born without a penis due to cloacal ex- strophy (Reiner and Gearhart 2004). Only a few reports on the sexual orientation of males with acute 17-HSD-3 deficiency are avail- able, but they suggest a predominance of male heterosexual orientation (Imperato- McGinley et al. 1979; Meyer-Bahlburg 1993). The high level of masculinization (of the Wolffian duct structures, gender identity, and sexual orientation) indicate that, despite low fetal levels of T, the XY genotype leads to increased sensitivity of the fetuses to the ac- tion of T. Sixth, as described in the previous section, the low prevalence of gonad/genital discor- dance in both XX and XY fetuses, despite substantial overlap in concentrations of cir- culating T when the genitals develop, indi- cates that the sex chromosome karyotype somehow modulates sensitivity to T prior to the onset of sex-specific androgen signaling. The examples discussed above strongly support the conclusion that XX and XY fe- tuses have different sensitivities to circulating androgens. Because most of the genes re- sponsible for this asymmetric response to an- drogens are autosomal (see next section), they must be transregulated in response to the XX versus XY sex chromosome karyo- type. Transregulation can occur in many ways, but recent studies demonstrate that the sex chromosome karyotype alone, indepen- dent of sex hormones, epigenetically regu- lates many autosomal genes (reviewed in Wijchers and Festenstein 2011). Epigenetic modification (i.e., methylation, histone tail modifications, and nucleosome reposition- ing) is emerging as a pivotal factor control- ling gene expression. For example, variation in the level of a single histone modification (trimethylation of lysine residue-4 on his- tone-3; the H3K4me3 epi-mark) of gene pro- moters can account for almost 50% of the variation in genome-wide gene expression lev- els in the early mouse embryo (Mikkelsen et al. 2007). From these studies and others (see be- low), we conclude that XX- and XY-specific epi-marks almost certainly contribute to the differential sensitivity to androgens of XX and XY fetuses (Figure 1B). The remainder of this article explores the potential for sex-specific epi-marks to contribute to the canalization of sexually dimorphic phenotypes and, as a side effect of pleiotropy and transgenerational in- heritance, contribute to the evolution of ho- mosexuality and other gonad-trait discordance This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 349 Figure 2. Androgen Signal Transduction Steps in the androgen signaling pathway that can boost or blunt signal transduction. T testosterone; AR androgen receptor; ARE androgen response element (DNA); CoFacts androgen receptor cofactors. such as hypospadias, cryptorchidism, and idio- pathic hirsutism. Mechanisms by which Sex-Specific Epi-Marks can Canalize Androgen Sensitivity Canalization occurs when a developmen- tal endpoint is reached despite environmen- tal interference that can potentially disrupt it (Waddington 1942). The androgen signal- ing pathway can be disrupted by natural vari- ation in androgen levels about their mean value as well as environmentally introduced androgen agonists and antagonists. Studies demonstrating high natural variation in fetal T (e.g., Reyes et al. 1974; Weisz and Ward 1980; Perera et al. 1987), as well as the com- mon occurrence of environmental androgen agonists and antagonists (Fang et al. 2003), demonstrate that there is strong selection to canalize the androgen signaling pathway. We can represent the androgen signaling pathway as flux (i.e., rate of flow) through a series of steps, each capable of being aug- mented or depressed, ultimately leading to androgen influence on a gene's expression (Figure 2). There is a surprisingly large num- ber of mechanisms by which the androgen signal can be strengthened or weakened. For example, varying the SHBG concentration in the blood, converting T to the more potent DHT, posttranslational modification of the AR (phosphorylation, ubiquitylation, and acetyla- tion) that change its activity, nucleosome place- ments that influence access to androgen response elements (AREs) in the DNA, and especially the concentration of the numerous and diverse array of AR cofactors. All of the steps in Figure 2 could also be influenced by sex-specific regulation of miRNA levels that are known to influence sexually dimorphism of mRNA concentrations in the brains of mice, and to be influenced by epigenetic control that is heritable across at least one generation (Mor- gan and Bale 2011). To canalize the impact of natural variation in T, the XY karyotype must lead to one or more epigenetic modifications that boost signal transduction through the pathway, and XX karyotypes must do the re- verse. To illustrate canalization of the androgen signaling pathway, suppose that the average fetal concentration of circulating T 10 in males and T 5 in females at a time when a sexually dimorphic trait develops. Further suppose that boosting epi-marks by the XY karyotype convert the actual T in the blood to a doubled endpoint signal (TEndPoint 20) affecting gene expression, and blunting epi- marks in XX fetuses lead to a halving of the endpoint value (TEndPoint 2.5). If natural variation in T causes overlap between the sexes (e.g., T in males varies between 4-16 and T in females between 2-8), then epige- netic modifications by the XX versus XY kar- yotype would cause the functional TEndPoint values to be nonoverlapping (male T 8 -32 and female T 1- 4). XY epi-marks that boost androgen signal- ing, and XX epi-marks that blunt androgen signaling, can also protect against androgen antagonists. For example, rats fed daily on a naturally occurring anti-androgen found in licorice root (that blocks the action of 17- This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms 350 THE QUARTERLY REVIEW OF BIOLOGY Volume 87 HSD in the T synthesis pathway) developed significantly reduced circulating T (Zaman- soltani et al. 2009). Epi-marks that boost an- drogen signaling in XY fetuses (such as higher conversion of T to DHT or lower SHBG concentrations) would protect them from the action of anti-androgens in the same way that they protect them from en- dogenous variation in T, as described in the above paragraph. In the same way, epi-marks that blunt androgen signaling in XX female fetuses would protect them from environ- mental agents that elevate the level of circu- lating androgens. An androgen mimic (e.g., as found in the Indian medicinal herb Tinospora cordifolia; Kapur et al. 2009) can be canalized in a different manner. It is well established that T is modified by the enzyme 5--reductase-2 to the more potent DHT in some tissues and this modification is necessary to achieve sufficient androgen signal to induce a male- specific phenotype, e.g., in the external geni- tals and the internal prostate. Work on the prostate indicates that T and DHT are inter- changeable (qualitatively identical) in promot- ing prostate growth, but DHT is two-and-a-half times more potent (Wright et al. 1996). The conversion of T to the more potent DHT will canalize androgen signaling in the presence of an androgen mimic whenever the 5-- reductase-2 enzyme does not catalyze the con- version of the androgen mimic to DHT. For example, again suppose that the normal con- centration of circulating T 10 in males and T 5 in females at a time when a sexually dimorphic trait differentiates. Next suppose that a T-mimic (TMimic) is present in the fetal blood at a concentration emulating T 5. In males, the androgen mimic poses no problem with respect to feminization, but in females the mimic would produce TSignal T TMimic 5 5 10 and a female fetus would be expected to incorrectly develop the male- specific trait. However, if the target tissue converted T to DHT, with a threshold of TEndPoint 25 (i.e., two-and-a-half times av- erage T in males) to induce the male trait, then the female would be "canalized" against the androgen mimic assuming the androgen mimic was not converted to the more potent DHT (because in females TEndPoint 2.5*5 1*5 17.5 25). Interestingly, this last mode of canaliza- tion may provide an explanation for the enigmatic within-cell conversion of T to es- tradiol (E) by the enzyme aromatase in the androgen signaling that occurs in the brain of rodents. Our review of many published studies of levels of circulating T and E indi- cates that, at its peak during the estrus cycle, unbound E is at least tenfold less common than peak unbound T in males (e.g., Bao et al. 2003; Travison et al. 2007). E is therefore a steroid hormone that is at least tenfold more potent than T (i.e., it functions at a concentration tenfold lower). By converting T to E, and assuming aromatase does not convert the androgen mimic to E, canaliza- tion will occur by the same logic as the con- version of T to the more potent DHT. A more formal model of canalized androgen signaling is provided in Appendix 1. Timing of XX/XY-Induced Epi-Marks that Canalize Sexual Development Epi-marks that boost androgen signaling in XY male fetuses and blunt it in XX female fetuses could, in principle, be produced any time prior to the onset of androgen signaling (when the fetal testes begins to secrete T). However, it is already established that epi- marks that are dimorphic between XX and XY embryos are produced during the nearly genome-wide episode of epigenetic repro- gramming that occurs at the embryonic stem cell stage of early development (reviewed in Bermejo-Alvarez et al. 2011). Epi-marks pro- duced during this early embryonic stage are known to strongly influence gene expression later in development (Mikkelsen et al. 2007). In addition, epi-marks produced so early in development would be transmitted to cell lineages leading to both the soma and the germline and would therefore have the po- tential to be heritable across generations. Be- low we develop these recent findings in more detail. During early development, there is a nearly global erasure of epi-marks (DNA methylation and histone tail modification) that originated in the sperm and egg stages. As reviewed in Hemberger et al. (2009), the erasures occur: This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 351 when protamines are replaced by histones on the paternal genome prior to fusion of pronu- clei; and during the first few cell divisions of embryonic development due to low availability of enzymes that methylate DNA (erasure on both the maternally and paternally inherited DNA). Although the majority of genes lose their gameticly inherited epi-marks at this time, some—such as imprinted genes and active transposons—somehow escape epi-mark era- sure (Hemberger et al. 2009). The expansive epigenetic erasure that occurs over the first few cell divisions is immediately followed by nearly genome-wide de novo epi-marking (gene pro- moters remodeled by histone modification and DNA methylation; reviewed in Hemberger et al. 2009). Genes essential to stem cell func- tioning (including housekeeping genes) are marked by an activating epi-mark on their pro- moters (trimethylation of the lysine-4 residue of histone H3, H3K4me3). There is a strong genome-wide correlation (0.67) between the level of this histone modification and the level of gene expression (Mikkelsen et al. 2007). Genes used only in later development are marked by a silencing epi-mark on their promoter's histones (trimethylation of the lysine-27 residue of histone H3, hereafter H3K27me3; Mikkelsen et al. 2007), or via methylation of their promoter's DNA (Fouse et al. 2008). Interestingly, thousands of re- pressed genes (that are expressed later in development) are bivalently marked with both a repressing (H3K27me3) and an acti- vating (H3K4me3) epi-mark (Mikkelsen et al. 2007). The nearly genome-wide epi-marking that occurs in early development provides a poten- tially simple and efficient way for epi-marks that influence androgen signaling to be manifest across all androgen-sensitive tissues. Consider the abundant bivalent epi-marks— containing both a silencing (H3K27me3) and an activating (H3K4me3) epi-mark—discov- ered by Mikkelsen et al. (2007). The overriding repressive effect of the silencing epi-mark turns off these genes early on in development while the activating epi-marks enable immediate strong gene expression when the silencing part of the epi-mark is removed later in develop- ment. By increasing the size of the activating epi-mark (e.g., more CpGs of the promoter histones methylated) in genes that boost an- drogen signaling (e.g., those whose gene prod- ucts acetylate the AR) and decreasing the size of these same epi-marks in genes that blunt androgen signaling (e.g., those coding for SHBG and/or its up-regulators), XY fetuses would be protected (canalized) from low cir- culating T in the fetus. The reverse pattern would protect (canalize) XX fetuses when T was atypically high. There is clear evidence that XX and XY em- bryos differ epigenetically at the earliest stages of mammalian development, i.e., in the preim- plantation embryo (blastocyst stage). At this time, XY embryos are physiologically distinct from XX embryos, having a higher metab- olic rate, faster growth rate, and increased resistance to some stress agents (reviewed in Gardner et al. 2010). Correspondingly, by the preimplantation blastula stage, the two sexes are reported to have widespread differences in gene expression levels at many hundreds of genes, most of which are autosomal (see Bermejo-Alvarez et al. 2010 and reference therein). Regulation of gene expression in complex eukaryotes is usually accomplished via epigenetic modifications (methylation of CpGs on the DNA or modification of histone tails; see Gordon et al. 2011). Recent studies have demonstrated that, at this early embry- onic stage, there are also sex-specific differ- ences in DNA methylation on the promoters of specific loci (reviewed in Bermejo-Alvarez et al. 2011). It has also been established that the Y-linked transcription-regulating genes SRY and ZFY are expressed in the preimplan- tation human embryo (Fiddler et al. 1995). There is also evidence for XX/XY-induced differences in gene expression later in devel- opment, but prior to secretion of T by the fetal testes in XY males. At this time in the mouse model system there is differential expression of 51 genes in the brains of XX versus XY embryos, most of which are autosomal (Dew- ing et al. 2003). These data indicate that the XX/XY karyotype somehow influences (in trans) the expression of many genes during later embryo development (but before the testes start secreting T) in a manner that is independent of androgen signaling. Such XX/XY karyotype-specific transregulation is known to occur in adult humans (Wijchers This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms 352 THE QUARTERLY REVIEW OF BIOLOGY Volume 87 and Festenstein 2011). For example, a gene on the human Y chromosome (TSPY) transregu- lates the level of expression of the X-linked androgen receptor in the adult germline (Akimoto et al. 2010). Collectively, these studies indicate that XX and XY embryos are epigenetically differenti- ated by the stem cell stage of the blastocyst—far in advance of androgen production by the tes- tes. Epi-marks produced at this time are there- fore strong candidates for the causative agents underlying the canalization of sexually dimor- phic development later in development in re- sponse to circulating androgens. The recent finding that environmentally induced epi- marks that reverse sexually dimorphic brain development (i.e., feminize male develop- ment) can carryover and produce the same reversal in the following generation (Frank- lin et al. 2010; Morgan and Bale 2011) dem- onstrates the potential for epi-marks laid down during very early development to in- fluence androgen signaling later in develop- ment. Heritable Epi-Marks A consequence of epi-marks being laid down at the stem cell stage of development, before the division between soma and germ- line, is that such epi-marks have the potential to be transmitted across generations, but only when the cycle of epi-mark erasure and re- newal, within and between generations, is somehow circumvented. Studies in both mice and humans clearly demonstrate that trans- generational inheritance of epi-marks occurs at nontrivial rates (reviewed in Morgan and Whitelaw 2008). When the epi-marks are sexually dimorphic, their transgenerational inheritance would be expected to influence the sexual development of opposite-sex off- spring, as described in the next section. A Model for Heritable Sexually Antagonistic Epi-Marks Any physiological mechanism that protects XX fetuses from atypically high T and/or en- vironmental androgen agonists during devel- opment would be favored by natural selection, assuming no counterbalancing harmful side effects. The same logic applies to mechanisms that protect XY fetus from atypically low T and or environmental androgen antagonists. As described above, sex-specific epi-marks (i.e., XX- or XY-specific) laid down during early em- bryonic development represent one mecha- nism to achieve such adaptations. However, such epi-marks would be sexually antagonistic if they sometimes carryover to the next gener- ation and redirect development in a gonad- discordant direction. We will refer to these sexually antagonistic epi-marks as SA-epi-ma- rks. SA-epi-marks can be favored by natural se- lection. In the autosomal case, an XX- and XY-dependent epi-mark always increase the fitness of the individual in which it is formed, and when there is carryover across genera- tions, it has only a 50% chance of decreasing fitness by being expressed in the opposite sex. The situation is somewhat more com- plex on the sex chromosomes but, as we show more formally below, sexually antago- nistic epi-marks can be favored across the entire genome under feasible selective pa- rameters. autosomal mutation We next more formally solve for the param- eter space that supports the evolution of muta- tions that produce SA-epi-marks. Throughout, we assume that the mutation has some expres- sion in the heterozygous state and our selec- tion coefficients apply to heterozygotes. First consider an autosomal mutation that produces an XX- or XY-dependent epi-mark (in cis, at its own location) that increases fitness of one sex (say, females) by an increment s, but with prob- ability q it carries over to the next generation and decreases the fitness of the opposite sex (sons) by a decrement . Because of the trans- generational effects, we need to consider the number of grandchildren of a mutant. The expected number of copies (w) of a mutant allele in the grandchildren's generation is (see Appendix 2): w 1⁄2 *1*1/2*1 1/2*1 s)) 1⁄2 *1 s)*[q/2*1 1 q)/2*11/2*1 s)], This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 353 where the first term on the right side repre- sents the number of grandchildren when the mutation originates in a male and the sec- ond term when it originates in a female. The allele invades if w 1, which for small s and q is equivalent to s q*/4, cost/benefit /s 4/q. Next assume that the X-linked mutation is expressed in XY embryos and canalizes de- velopment toward the male phenotype. The expected number of copies of a mutant al- lele in the grandchildren's generation is (see Appendix 2): w 1/3*1s)(q*1 1 q*1 2/3*1*1/2 *1 s) 1/2*1, where the first term on the right side repre- sents grandchildren when the mutation orig- inates in a male and the second term when it originates in a female. The allele invades if w 1, which for small s and q is equival- ent to: or Even when the rate of transgenerational car- ryover is 100% (q 1), the mutation invades when the costs are fourfold larger than ben- efits. When transgenerational carryover is much smaller, there is essentially no con- straint on the invasion. The above inequality is compatible with a two-generation general- ization of Hamilton's rule: the benefit b goes to the carrier (r 1) and its same sex off- spring (r 1/2), while the cost c goes to the opposite-sex offspring (r 1/2). x-linked mutation When the mutation is X-linked, it occurs in a female with probability 2/3 and in a male with probability 1/3. First assume that the mutation is dominant and canalizes de- velopment toward the female phenotype and is expressed in XX fetuses. The ex- pected number of copies of a mutant allele in the grandchildren's generation is (see Ap- pendix 2): w 1/3*1*1 s*1 2/3*1 s*q/2*1 1 q/2*11/2*1 s, where the first term on the right side repre- sents grandchildren when the mutation origi- nates in a male and the second term when it originates in a female. The allele invades if w 1, which for small s and q is equivalent to: s q*/4 cost/benefit /s 4/q. or s q*/2 cost/benefit /s 2/q. or In this case, the cost/benefit ratio must be half as large as the autosomal case for the mutation to invade. Nonetheless, the mutation can in- vade under a broad and feasible range of pa- rameter space. For example, if transmission across generations (q) is 0.25, the mutation will invade when costs are eight-times larger than benefits. Our X-linked analysis has assumed dominance of the epi-mark producing muta- tion. For partial dominance, the selection co- efficients s (female canalization) or (male canalization) must be multiplied by a domi- nance scaler h (0 h 1). x-linked trans-effect Our model has assumed that an X- or autosome-linked mutation producing an SA- epi-mark makes the epigenetic modification in cis at its own location (i.e., it epi-marks itself). When the mutation produces an SA- epi-mark in trans anywhere else in the ge- nome, the same equations as described above can be applied by replacing the param- eter q with q/2, since the mutation cosegre- gates with its SA-epi-mark with probability 1/2. When the SA-epi-mark is produced at another locus on the same chromosome, the parame- ter q must be replaced with q * (1 r), where r is the recombinational distance between the This is the same constraint that was found for autosomal linkage. This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms 354 THE QUARTERLY REVIEW OF BIOLOGY Volume 87 mutation and the SA-epi-mark it produces, and (1 r) is the probability that the mutation and its epi-mark cosegregate. y-linked mutation Lastly, when the mutation is Y-linked and canalizes development toward the male phe- notype, the allele invades when s 0 irre- spective of the value of and irrespective of where the epi-mark is produced. generalizations These calculations demonstrate that mu- tations causing sexually antagonistic epi- marks can invade even when the cost to the harmed sex far exceeds the benefit to the favored sex. This conclusion holds irrespec- tive of linkage to the sex chromosomes or autosomes. Such invasions are expected to lead to the eventual fixation of mutations producing SA-epi-marks, unless there were some additional factors such as frequency- dependent fitness. Although our model predicts that muta- tions causing SA-epi-marks will go to fixation, the androgen-induced phenotypes they af- fect may nonetheless be highly variable. This is expected because epi-marks can be highly variable despite genetic monomorphism. As described in the next section, monozygotic twins at birth show strong differences in methylation levels of individual promoters and large differences in gene expression lev- els at as many as hundreds of gene loci. These data indicate that epi-marks are intrin- sically variable and that the same fixed mu- tation can produce variable epi-marks. More extreme epi-marks (e.g., with denser or lon- ger tracts of histone modification and/or DNA methylation) would span more chro- matin locations and hence have a higher probability of serendipitously achieving at least partial transgenerational inheritance. Homosexuality and SA-Epi-Marks Although pedigree studies indicate a fa- milial association of homosexuality in both males (e.g., Hamer et al. 1993) and females (e.g., Pattatucci and Hamer 1995), more than a decade of molecular genetic studies have produced no consistent evidence for a major gene, or other genetic marker, con- tributing to male homosexuality (reviewed in Ngun et al. 2011). Moreover, the most recent genome-wide association study using exceptionally high marker density found no significant association between homosexual- ity in males and any SNPs (Ramagopalan et al. 2010). These negative/inconsistent re- sults may reflect insufficient statistical power, but they also support another agent causing the familial association of homosexuality: ep- igenetic inheritance. There is a consensus among studies com- paring homosexuality in monozygotic versus dizygotic twins that one or more coinherited elements (assumed to be genes, but which could just as well be heritable epi-marks) con- tribute substantially to this trait—accounting for an estimated 20-50% of the phenotypic variation in sexual orientation in both sexes (Kirk et al. 2000; Alanko et al. 2010; Lang- strom et al. 2010; Burri et al. 2011). How- ever, estimates of proband concordance among monozygotic twins (i.e., the probabil- ity that a twin is homosexual given that the other twin is homosexual) are surprisingly low in both sexes (about 20%) for a trait predominantly influenced by genetic factors (Bailey et al. 2000; Langstrom et al. 2010). Correspondingly, studies of twins consistently report a high "nonshared environment" con- tribution to homosexuality, typically account- ing for at least 50% of the phenotypic variation in both sexes (Kirk et al. 2000; Alanko et al. 2010; Langstrom et al. 2010; Burri et al. 2011). The substantial estimated heritability of homo- sexuality, low proband concordance between monozygotic twins, and negative results from numerous molecular genetic association stud- ies are collectively consistent with an epigenetic causation for homosexuality that contains two independent components: monozygotic twins share inherited (transgenerational) gonad- discordant SA-epi-marks influencing androgen signaling (contributing to the observed sub- stantial heritability estimates and negative re- sults from genetic association studies), but do not share one or more de novo gonad- concordant epi-marks (including erasure of a coinherited SA-epi-mark) that are laid down during fetal development (independently in each twin) that also influence androgen signal- This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 355 ing (contributing to the observed low concor- dance between monozygotic twins). To understand how the homosexual pat- tern of substantial estimated heritability and low concordance between identical twins can be feasibly caused by epigenetic inheritance, we summarize results from studies of twins. In the remainder of this section, we first focus on arbitrary phenotypic traits and how epige- netics can contribute to both phenotypic simi- larity and dissimilarity between monozygotic twins. Next we focus on twin studies of traits other than homosexuality that are strongly influenced by fetal androgen signaling and well characterized at birth by numerous stud- ies. Finally, we extend these studies to homo- sexuality. arbitrary traits Empirical studies indicate that epigenetics can contribute substantially to the similarity of identical twins. For example, Gartner and Baunack (1981) used an isogenic mouse line to created monozygotic and dizygotic "iden- tical" twins and compared them for a variety developmental traits. They consistently found higher phenotypic similarity between monozy- gotic compared to dizygotic identical twins, de- spite the fact that all mice were isogenic and developed (in utero and postnatally) in surro- gate mothers. This finding, and others sum- marized in Wong et al. (2005), supports the conclusion that there is a substantial contribu- tion of shared epi-marks to the phenotypic sim- ilarity of twins. Empirical studies also indicate that epige- netics can contribute substantially to the dis- similarity of identical twins. Bouchard et al. (1990) compared a large sample of human identical twins who were reared together or apart since birth. They found that pheno- typic dissimilarity for a wide diversity of traits (with a substantial "environmental" compo- nent of variation) was commonly no higher when twins were reared apart. This finding indicates that many phenotypic differences between monozygotic twins developed prena- tally, i.e., at a time of extensive de novo epige- netic programming. This finding also indicates that nonshared epi-marks laid down indepen- dently in each individual twin (or other types of relatives) during fetal development may contribute importantly to the "environmental variance" that causes heritability to be less than one for most traits. This conclusion is sup- ported by a study that compared methylation levels at the promoters of four genes in new- born monozygotic human twins (Ollikainen et al. 2010). Median differences in methylation levels were only about 3-4% (about half the value for dizygotic twins), but values as high as 54% were seen when looking at individual CpG units. At the level of genome-wide gene expression, Gordon et al. (2011) found that some newborn monozygotic twins had over 600 genes at which expression differed by at least twofold. They also found that newborn identical twins that separated earlier in devel- opment (1-3 days compared to 4 -9 days post- fertilization) had larger differences in their gene expression profiles, indicating that "this short period, very early in development, repre- sents an important window for epigenetic vari- ability" (Gordon et al. 2011). twin studies of androgen-influenced traits other than homosexuality Evidence that epigenetics contributes to high levels of phenotypic variation for traits influenced by fetal androgen exposure comes from studies on two phenotypes in humans: hypospadias (subterminal opening of the ure- thra on the phallus) and cryptorchidism (one or both testes fail to descend into the scrotum by birth). Like homosexuality, in which the genitals are concordant with the gonad but sexual preference (and brain anatomy; e.g., Savic and Lindstro ̈m 2008) is not, both traits represent a gonad-trait discordance in which one aspect of androgen signaling matches the gonad while the other does not. In hypospadias, the phallus is generally male- typical in size, shape, and internal com- position, but the length of the urethra is feminized (shortened). In cryptorchidism, the phallus is usually normal in all respects, but the position of the gonad is feminized (nondescended). The prevalence of cryptorchidism is sub- stantial and similar to that of human homo- sexuality (2-9%; Bay et al. 2011), while that of hypospadias is substantial but somewhat lower (prevalence of about .3- 4%; Ahmed et al. 2004; Boisen et al. 2005). Animal models This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms 356 THE QUARTERLY REVIEW OF BIOLOGY Volume 87 indicate that exposure to androgen antago- nists during a short period when the genitals differentiate (but not later in development) leads to highly elevated levels of both hypo- spadias and cryptorchidism, confirming that both traits are strongly influenced by fetal an- drogen signaling. In humans, however, the si- multaneous expression of naturally occurring hypospadias and cryptorchidism is rare (i.e., they usually occur in isolation; Weidner et al. 1999). This pattern indicates that the two traits are caused, in large part, by different disrup- tions to androgen signaling. Like homosexual- ity, both traits display a familial association: when one brother in a family is affected, the prevalence is elevated in other brothers by about tenfold for hypospadias and threefold for cryptorchidism (Weidner et al. 1999). Sim- ilar to homosexuality, both traits are usually not shared among monozygotic twins (approx- imately 25% concordance for each trait; Fre- dell et al. 2002; Jensen et al. 2010). Also like homosexuality being elevated in individuals with loss of function at the CYP21 gene (but this gene not being a major cause of female homosexuality), extensive genetic studies have found that while loss of function at some can- didate genes can lead to both hypospadias and cryptorchidism, the majority of cases are not associated with any known mutations (re- viewed in Bay et al. 2011; Kalfa et al. 2011). Further evidence for a substantial nongenetic contribution in the case of cryptorchidism is: higher concordance (twofold) between dizy- gotic twins than that between singleton brothers (Jensen et al. 2010); and the high incidence of cryptorchidism (up to 70%) ob- served in some isolated wildlife populations despite no genetic evidence for inbreeding or a founder effect—presumably due to an environmental hormone-signaling disruptor (Latch et al. 2008). The familial association observed in both androgen-influenced traits (cryptorchidism and hypospadias) indicates a substantial con- tribution of some coinherited factor (gene or epi-mark), while the low concordance for both traits observed in monozygotic twins indicates a substantial contribution of a "nonshared environment." Since the traits are measured in newborns, the nonshared environment must occur during gestation. But since monozygotic twins would be ex- pected to share nearly all environmental effects during gestation (like exposure to an- drogen antagonists), something other than traditional environmental variation is almost certainly responsible for their observed low concordance. Epi-marks influencing andro- gen signaling that are laid down indepen- dently between monozygotic twins are the most feasible candidate to account for the strong "nonshared environment" component of both androgen-influenced traits. If these epi-marks sometimes escaped transgenera- tional erasure, they could also account for the familial association of both traits. homosexuality As described above, there is compelling evidence that epi-marks contribute to both the similarity and dissimilarity of family members, and can therefore feasibly contrib- ute to the observed familial inheritance of homosexuality and its low concordance be- tween monozygotic twins. We also showed that two other androgen-sensitive pheno- types, cryptorchidism and hypospadias, show the same pattern of high prevalence, strong familial associations, low monozygotic twin concordance, and discordance between the gonad and the trait (i.e., testes paired with undescended gonad(s) or testes paired with short urethral length). Just as epigenetics is a probable etiological agent contributing to cryptorchidism and hypospadias, so too is it a probable agent contributing to homosexuality. In this case, epi-marks that sometimes carry- over across generations would contribute to the causation of homosexuality and its obser- ved heritability while de novo epi-marks pro- duced independently in each monozygotic twin would account for the low observed con- cordance for homosexuality between monozy- gotic twins. An inherited gonad-discordant epi-mark causing homosexuality must not be masked by any gonad-concordant epi-marks pro- duced during the recipient's ontogeny. This would occur most simply when an inherited epi-mark is stronger than average and is com- bined with a relatively weak de novo epi- mark(s) produced in the recipient. These two processes (i.e., transgenerational inheri- This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 357 tance of an epi-mark and its penetrance once inherited) are collectively subsumed in our model parameter q probability that an SA-epi-mark carries over to the next genera- tion and decreases the fitness of the opposite sex (see above modeling section). An epi-mark producing homosexuality must also have a highly restricted effect, i.e., cause discordance between the gonad and sexual preference, but no discordance for other traits such as the genitalia and sexual identity. As described above, most cases of cryptorchid- ism and hypospadias are not associated with gonad-trait discordance for other androgen- influenced traits. This observation demon- strates that gonad-trait discordances can occur independently for different traits in the same individual. The most feasible explanation for this independence is the exceptionally wide diversity of AR cofactors that are known to oc- cur and their high tissue specificity (Heemers and Tindall 2007)—each of which may be epi- genetically regulated independently. Although we cannot provide definitive evidence that homosexuality has a strong epigenetic underpinning, we do think that available evidence is fully consistent with this conclusion. For example, we now have clear evidence that epigenetic changes to gene pro- moters that influence their expression (e.g., levels of CpG methylation) can be transmitted across generations (Franklin et al. 2010) and that such heritable epigenetic changes can strongly influence, in the next generation, both sex-specific behavior and gene expression in the brain (Morgan and Bale 2011). As a consequence, we next apply our model of sex- ually antagonistic epi-marks to the human ho- mosexual phenotype (as described in Table 1 and Figure 3). Discussion Sexually antagonistic selection is now well appreciated as a powerful factor in biological evolution (Parker 1979; Haig 1993; Rice and Holland 1997; Partridge and Hurst 1998; Chapman et al. 2003; Arnqvist and Rowe 2005; Bonduriansky and Chenoweth 2009). It has been shown to significantly affect or drive a number of biological phenomena and pro- cesses, including survival and fertility (Rice 1996), mate choice (Gavrilets et al. 2001), ge- netic differentiation (Hayashi et al. 2007), reproductive isolation (Gavrilets 2000) and speciation (Parker and Partridge 1998; Gavri- lets and Waxman 2002; Gavrilets and Hayashi 2005), sex chromosome evolution (Rice et al. 2008), sib competition (Rice et al. 2009), ma- ternal selection (Miller et al. 2006), and grand- parental care (Rice et al. 2010). This paper argues that sexually antagonistic selection can also be involved in epigenetic effects and ex- plain the enigmatic high prevalence of several fitness-reducing human characters. As de- scribed below, our model and its predictions are consistent will the major empirical patterns associated with male and female homosexual- ity, and other common gonad-trait discor- dances. Homosexuality is frequently considered to be an unusual phenotype because it repre- sents an evolutionary enigma—a trait that is expected to reduce Darwinian fitness, yet it persists at substantial frequency across many different (possibly all) human populations. However, from the perspective of other traits influenced by fetal androgen signaling, and in which there is gonad-trait discordance, the high prevalence of homosexuality is not un- usual. For example, the prevalence of hypos- padias (gonad-trait discordance for urethral length) varies from 0.4% to 1% in newborns, and when including milder cases (ascertained in three years postpartum), its prevalence can be as high as 4% (Boisen et al. 2005). This phenotype is expected to interfere with sp- erm transfer during copulation, but despite this fitness cost, it persists at substantial fre- quency. Cryptorchidism (gonad-trait discor- dance for the position—abdominal versus descended—of the gonads) is associated with reduced fertility and increased rates of testicu- lar cancer. The prevalence of this androgen- influenced trait is 2-9% (Bay et al. 2011). Examples of other androgen-influenced phe- notypes with a high prevalence of gonad-trait discordance, but less obvious fitness effects, are male childhood cross-gender behavior (3.2%; van Beijsterveldt et al. 2006), female childhood cross-gender behavior (5.2%; van Beijsterveldt et al. 2006), and female idiopathic hirsutism (i.e., male-like pattern of body hair in the ab- sence of both atypical menstrual cycles and elevated circulating androgens, 6%; Carmina This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms 358 THE QUARTERLY REVIEW OF BIOLOGY Volume 87 TABLE 1 Sexually antagonistic epi-mark hypothesis of homosexuality We describe our hypothesis for an epigenetic cause of homosexuality as a series of statements (see Figure 3 for a graphical summary): a) Empirical studies demonstrate that XX fetuses are canalized to blunt androgen signaling (lower sensitivity to T) and XY fetuses are canalized to boost androgen signaling (higher sensitivity to T). b) Empirical studies demonstrate the production of XX- and XY-induced epi-marks in embryonic stem cells and extensive sex-specific differences in gene expression at this time. Epi-marks laid down during the embryonic stem cell stage are also established to influence gene expression later in development. This stem cell period is the most plausible candidate time point for the production of epi-marks influencing sensitivity to androgens later in development (canalization of fetal androgen signaling). c) Epi-marks produced in embryonic stem cells are mitotically transmitted to cell lineages leading to both the soma and the germline, and hence can contribute to pseudo-heritability when they escape erasure across generations (nonerasure in the primordial germ cells and in the zygote and first few cell divisions of the next generation). Animal models as well as human data unambiguously demonstrate that such a multistep escape from erasure does occur at nontrivial frequency. d) Epi-marks blunting (in XX fetuses) or boosting (in XY fetuses) androgen signaling will be sexually antagonistic (SA-epi-marks) when they have a nonzero probability of carryover across generations and are expressed in oppose sex descendants. Such carryover will contribute to discordance between the gonad and one or more sexually dimorphic traits. e) Our modeling work shows that SA-epi-marks are favored by natural selection over a broad span of parameter space because there is a net benefit to the carrier (due to canalization of sexually dimorphic development) that is not offset sufficiently by transmission (and fitness reduction) to opposite sex descendants. f) Genetic mutations causing SA-epi-marks are expected to fix in populations and are therefore not expected to be polymorphic except transiently during their initial spread within a population. Therefore, no association between genotype and homosexuality is predicted. g) Because the androgen signaling pathways differ among organs and tissues (e.g., use of different AR cofactors), the same inherited SA-epi-mark can affect only a subset of sexually dimorphic traits, e.g., no effect on the genitalia, but a large effect on a sexually dimorphic region of the brain. h) Shared, gonad-discordant SA-epi-marks that carryover across generations would contribute to the observed realized heritability of homosexuality, e.g., monozygotic twins share the same SA-epi-marks coinherited from a parent. i) Unshared, gonad-concordant SA-epi-marks, produced during fetal development, would contribute to the low proband concordance of homosexuality observed between monozygotic twins, i.e., they need not share SA-epi-marks generated during development that occurs after the twins have separated. j) Homosexuality occurs when an individual inherits one or more gonad-discordant SA-epi-marks that are not masked nor erased by the production of de novo gonad-concordant SA-epi-marks that accrue during ontogeny. The SA-epi-mark(s) influence androgen signaling in the part of the brain controlling sexual orientation, but not the genitalia nor the brain region(s) controlling gender identity. 1998). From these examples it is clear that the substantial prevalence of homosexuality (a gonad-trait discordance) is not unusual for a phenotype strongly influenced by fetal andro- gen exposure. Why should phenotypes associated with fe- tal androgen signaling commonly have high frequencies of gonad-trait discordance? We do not know. The simplest hypothesis is that environmental stress and androgen agonists and antagonists are sufficiently common that they generate constant selection for new, more effective epi-marks that protect (cana- lize) each sex from their effects. Some of these newly evolved epi-marks escape the normal generational cycle of erasure/re- programming and thereby carryover across generations (by happenstance and with mod- erate to low probability) and lead to gonad-trait discordance. Since it is now well established that environmentally induced epi-marks, like those from prenatal/perinatal stress, are com- mon and can be heritable with sex-specific ef- fects on the brain and behavior (Franklin et al. 2010; Morgan and Bale 2011), it seems inevi- table that some epi-marks produced by new mutations (coding for epi-marks that cana- lize sex-specific development) will also some- times carryover across generations and form SA-epi-marks. Our modeling analysis clearly This content downloaded from 140.211.24.51 on Tue, 24 May 2016 19:21:18 UTC All use subject to http://about.jstor.org/terms December 2012 EPIGENETICS, CANALIZATION, AND HOMOSEXUALITY 359 Figure 3. SA-Epi-Marks and Homosexuality Our SA-epi-mark model predicts that homosexuality is produced by transgenerational epigenetic inheri- tance. As in nature, epi-marks are assumed to sometimes carryover across generations (depicted by the "*" superscript) and epi-mark strengths are assumed to be variable, irrespective of genetic polymorphism (depicted by the intensity of their letter symbols). Masculinizing (superscript "M") epimarks are produced in response to the XY genotype in the early embryo stage (stem cells) and canalize male development by increasing sensitivity to fetal T. Feminizing (superscript "F") epi-marks similarly canalize female development by reducing sensitivity to fetal T. Homosexuality occurs when one or more stronger than average epi-marks—bold, that canalize sexual preference (Sp), but not the genitals (Ge) nor sexual identity (Si)—carryover across generations into an opposite-sex descendent and cause gonad-trait discordance when combined with weaker than average (light) de novo sexually concordant epi-mark(s). demonstrates that mutations that cause epi- marks that blunt androgen signaling in XX fetuses, or boost it in XY fetuses, can have a selective advantage even when they carryover across generations at nontrivial frequency and reduce fitness by feminizing or mascu- linizing opposite-sex descendants. Modifiers that restrict androgen blunting/boosting epi-marks to the appropriate sex would be expected to eventually evolve and produce nearly invariant canalization, but new muta- tions creating new epi-marks may continue to evolve (requiring new modifiers) if andro- gen agonists and antagonists are sufficiently variable across time and/or space. Another possibility leading to persistently high levels of gonad-trait discordance is an arms race between male- and female-benefit SA-epi-marks that blunt/boost androgen sig- naling during fetal development. The accumu- lation of such SA-epi-marks favoring one sex generates selection in the other sex to evolve new epi-marks that protect them from oppo- site-sex SA-epi-marks that sometimes carryover across

Levels of Biological Adaptation

Adaptation often used generally to reflect "change by which organisms surmount challenges to life" (Lasker 1969: 1481) Genetic/Darwinian Adaptation: Many generations; reflects natural selection Measured through differential reproductive success Functional AdaptationsAdjustments during the lifetime of an individual Acclimatization: Short-term, reversible, e.g., increased sweating ability in hot climates (within 10 days) Developmental Acclimatization/Plasticity (Adaptability): Long term, irreversible Habituation: Gradual reduction in response to stressor

Why bipedalism?

Advantages -long distance travel -carrying/manipulating -improved view -effective displays Disadvantages -birth more difficult -lower back stress -infant clinging

Darwin's Mechanism of Evolution Natural Selection

All species are capable of producing offspring faster than the food supply increases (Malthus). All living things show variations; no two individuals of a species are exactly alike. Because more individuals than can survive, there is a struggle for existence. Those with favorable variation possess advantage over others (Malthus). These favorable variations are inherited and tend to be more likely to be passed on to the next generation. Over long periods of geologic time (Lyell), these successful variations produce great differences that result in new species. In the process, animals acquire traits that appear functionally well designed for the tasks they carry out.

Adaptation

An adaptation is a characteristic of an organism that has been favored by natural selection and increases the fitness of its possessor. The concept is central to biology, particularly evolutionary biology.

Genghis Khan

As many as 8 percent of the men dwelling in the confines of the former Mongol empire bear Y chromosomes that seem characteristic of the Mongol ruling house. Some 16 million men, or half a percent of the world's male population, can probably claim descent from Genghis Khan.

Chromosomes

Chromosomes: long piece of DNA found only in the nucleus. Bacteria have only one chromosome. Higher organisms have multiple.

Males Adjust Sperm Number and Quality

Baker and Bellis

Sperm Competition

Behavioral and Morphological adaptations

Behavioral Ecology

Behavioral ecology is the study of the ecological and evolutionary basis for animal behavior, and the roles of behavior in enabling an animal to adapt to its environment.

Proximate and Ultimate Functions

In ethology, the study of animal behavior, causation can be considered in terms of these two mechanisms. n Proximate causation: Explanation of an animal's behavior based on trigger stimuli and internal mechanisms. n Ultimate causation: Explanation of an animal's behavior based on evolution. Requires that behavioral traits, like physical ones, are genetically heritable, and then explains behavior using an explanation of why this specific behavioral trait was favored by natural selection.

Predictions from Hypotheses of Human Homosexual Behavior

Kin selection and parental manipulation 1. Homosexual behavior reduces individual reproductive success. 2. Homosexuals provide more or better aid in raising offspring than do heterosexuals or celibates.3. Homosexual behavior is seen predominantly in individuals of low reproductive potential. Parental manipulation 1. Parents manipulate selected offspring to forgo reproduction, become homosexual, and aid in the rearing of siblingsor the offspring of siblings. Alliance formation (reciprocal altruism) 1. Same-sex alliances aid individual survival.2. Homosexual behavior aids the formation of same-sex alliances. 3. Bisexuality is more common than homosexuality.

Fitness

Not necessarily based on traits like speed or strength Fitness is actually defined by the number of genes an individual passes on to the next generation Fitness: Which of these women do you think has higher fitness? Fitness refers to the "fit" between an organism and its environment Roughly, fitness can be thought of as the number of surviving offspring one produces Technically this is direct fitness What is indirect fitness? Indirect + Direct = Inclusive Fitness

cycle effects on tips

Ovulatory cycle effects on tip earnings by lap dancers: economic evidence for human estrus?☆ Geoffrey Miller⁎, Joshua M. Tybur, Brent D. Jordan Department of Psychology, University of New Mexico, Albuquerque, NM 87131, USA Initial receipt 16 April 2007; final revision received 26 June 2007 To see whether estrus was really "lost" during human evolution (as researchers often claim), we examined ovulatory cycle effects on tip earnings by professional lap dancers working in gentlemen's clubs. Eighteen dancers recorded their menstrual periods, work shifts, and tip earnings for 60 days on a study web site. A mixed-model analysis of 296 work shifts (representing about 5300 lap dances) showed an interaction between cycle phase and hormonal contraception use. Normally cycling participants earned about US$335 per 5-h shift during estrus, US$260 per shift during the luteal phase, and US$185 per shift during menstruation. By contrast, participants using contraceptive pills showed no estrous earnings peak. These results constitute the first direct economic evidence for the existence and importance of estrus in contemporary human females, in a real-world work setting. These results have clear implications for human evolution, sexuality, and economics.© 2007 Elsevier Inc. All rights reserved. Keywords: Estrus; Female sexuality; Behavioral economics; Sexual service industries; Hormonal contraception 1. Introduction Estrus is a phase of increased female sexual receptivity, proceptivity, selectivity, and attractiveness. It is common across mammalian species (Lange, Hartel, & Meyer, 2002; Lombardi, 1998), including primates (Dixson, 1998; Nelson, 2000), and seems functionally designed to obtain sires of superior genetic quality (Gangestad, Thornhill, & Garver- Apgar, 2005; Thornhill, 2006). However, the conventional wisdom holds that human female estrus became uniquely "lost" or "hidden" over evolutionary time (e.g., Burt, 1992), perhaps to promote male provisioning and paternal care in long-term pair-bonded relationships (Strassmann, 1981; Turke, 1984). Contrary to this "hidden-estrus" view, recent laboratory-based studies show that women near the most fertile point of their cycle (just before ovulation) are more attractive to males, as manifest through more attractive body scent (Havlíček, Dvořáková, Bartoš, & Flegr, 2006; ☆ Brent D. Jordan's contribution to this project was supported by a McNair/ROP Scholars Program Fellowship. ⁎ Corresponding author.E-mail address: [email protected] (G. Miller). 1090-5138/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.evolhumbehav.2007.06.002 Kuukasiarvi et al., 2004; Singh & Bronstad, 2001), greater facial attractiveness (Roberts et al., 2004), increased soft- tissue body symmetry (Manning, Scutt, Whitehouse, Leinster, & Walton, 1996), decreased waist-to-hip ratio (Kirchengast & Gartner, 2002), and higher verbal creativity and fluency (Krug, Moelle, & Fehm, 1999; Symonds, Gallagher, Thompson, & Young, 2004). While such laboratory-based findings are theoretically important, only four studies have, to our knowledge, investigated the real-world attractiveness effects of human estrus outside the laboratory. Haselton, Mortezaie, Pillsworth, Bleske-Recheck, and Frederick (2007) photographed 30 young women—all in steady relationships and not using the pill—twice each, wearing their self-chosen clothing, once during estrus (as confirmed by hormonal assay) and once during a lower- fertility (luteal) cycle phase. Then, 42 mixed-sex raters made a forced-choice judgment ("In which photo is the person trying to look more attractive?") between the two photos of each woman (with faces obscured, leaving only body and clothing cues). They chose the woman when she was in estrus about 60% of the time—modestly but significantly above chance. This result confirmed that both male and 376 G. Miller et al. / Evolution and Human Behavior 28 (2007) 375-381 female observers are perceptually sensitive to women's choice of more conspicuous and fashionable clothes during estrus. (A related real-world study by Grammer, Renninger, & Fisher, 2004, found that mated women attending Vienna discotheques without their partners tended to dress more provocatively if they had higher estradiol levels; however, that study did not directly assess cycle phase.) Gangestad, Thornhill, and Garver (2002) found con- vergent evidence that men's real-world behavior is sensitive to their female partners' estrous cues. (Note that estrous is the adjectival form of the noun estrus.) Among 31 mated women not using the pill, their sexual partners were reported as using more mate-guarding behaviors, including higher proprietariness, attentiveness, and vigilance (e.g., calling the women's cell phones at random times to see what they were doing) when the women were in estrus. This effect was especially strong when the relationship was not yet steady or exclusive—consistent with the theory that increased mate guarding of estrous women by male partners is functionally designed to deter extra-pair copulation. Haselton and Gangestad (2006) replicated this mate- guarding result among 25 normally cycling women, using a stronger repeated-measures design based on daily reports. They further found that estrous mate guarding by male partners was mediated by both men's sexual attractiveness (less-attractive men mate guarded especially more during estrus) and women's own attractiveness (less-attractive women were especially mate guarded during estrus, whereas more-attractive women were mate guarded all the time). In a subsequent study, Pillsworth and Haselton (2006) found similar results on measures of male attentiveness and expressed love. These four studies provide some evidence that men are sensitive to estrous cues in real-world situations. However, the photo-choice results were rather modest in strength (Haselton et al., 2007), and the mate-guarding results were rather indirect measures of estrous female attractiveness (Gangestad et al., 2002; Haselton & Gangestad, 2006; Pillsworth & Haselton, 2006). In this article, we build upon such research by presenting the first real-world economic evidence of male sensitivity to cyclic changes in female attractiveness. Specifically, we measured the tips earned by professional lap dancers in gentlemen's clubs over a 2-month span. These dancers are highly motivated to maximize tip earnings during every shift they work, which they do by appearing more sexually attractive than the other 5 to 30 rival dancers working the same shift and by doing the "emotional labor" of "counterfeiting intimacy" with male club patrons (Barton, 2006; Beasley, 2003; Deshotels & Forsyth, 2006; Pasko, 2002). Hence, if the hidden-estrus view is right, there should be no reliable cues of fertility available or expressed to patrons and no ovulatory cycle effects on dancer tip earnings. On the other hand, if women retain perceivable cues of estrus—if they become more attractive to male club patrons at midcycle, just before ovulation—then they might be in higher demand and earn higher tips. 2. Background Because academics may be unfamiliar with the gentle- men's club subculture, some background may be helpful to understand why this is an ideal setting for investigating real-world attractiveness effects of human female estrus. The following information was gathered from interviews with local club managers and from the sociological and feminist literature on erotic dancing (Barton, 2006; Beasley, 2003; Brewster, 2003; Deshotels & Forsyth, 2006; Enck & Preston, 1988; Forsyth & Deshotels, 1997; Hall, 1993; Hochschild, 1983; Lewis, 2006; Linz et al., 2000; Pasko, 2002; Ronai & Ellis, 1989; Thompson, Harred, & Burks, 2003). All participants in this study worked as lap dancers in Albuquerque "gentlemen's clubs" circa November 2006 through January 2007. The clubs serve alcohol; they are fairly dark, smoky, and loud (with a DJ playing rock, rap, or pop music). Most club patrons are Anglo or Hispanic men aged 20 to 60, ranging from semiskilled laborers to professionals; they typically start the evening by getting a stack of US$20 bills from the club's on-site ATM and having a couple of drinks. Dancers in these clubs perform topless but not bottom- less; law requires them to wear underwear, bikinis, or similar garments to cover the pubis. Thus, menstruating dancers can wear tampons (with strings clipped short or tucked up) and change them often during heavy-flow days, without reveal- ing any visual signs of menstruation. Dancers typically wear very little perfume, but they often have breast implants, dye their head hair, trim their pubic hair, shave their legs and underarms, and adopt a "stage name" different from their real first name. They typically do regular aerobic and resistance exercise to maintain a fit, lean body shape. During work, each dancer performs one to three "stage dances" on an elevated central stage about every 90 min to advertise her presence, attractiveness, and availability for lap dances. These result in only modest tip earnings (typically US$1-5 tips from the men seated closest to the stage, totaling only about 10% of her earnings). The rest of the dancer's time is spent walking around the club asking men if they want a "lap dance." A lap dance typically costs US$10 per 3-min song in the main club area or US$20 in the more private VIP lounge. Dancers typically make about two thirds of their income from the main club area and one third from the VIP area; thus, average income from each lap dance is about US$14. Lap dances require informal "tips" rather than having explicit "prices" (to avoid police charges of illegal "solicitation"), but the economic norms of tipping are vigorously enforced by bouncers. Dancers thus maximize their earnings by providing as many lap dances as possible per shift. In each lap dance, the male patron sits on a chair or couch, fully clothed, with his hands at his sides; he is typically not allowed to touch the dancer. The topless female dancer sits on the man's lap, either facing away from him (to display her buttocks, back, and hair) or facing him (either leaning back to display her breasts, and to make conversation and eye contact, or learning forward to whisper in his ear). Lap dances typically entail intense rhythmic contact between the female pelvis and the clothed male penis (Barton, 2006; Beasley, 2003). Thus, lap dances are the most intimate form of sex work that is legal in most American cities—much more intimate than the stage dances and table dances that characterized such clubs until the late 1990s. However, most lap dance clubs strongly discourage more intimate patron-dancer contact, even in private VIP rooms, since clubs can be closed swiftly if undercover police discover that prostitution is occurring. Rather, the dancer's earnings are maximized by inducing the man to agree to further lap dances after the first 3-min song is over (Deshotels & Forsyth, 2006; Pasko, 2002). Thus, a dancer can make up to US$400 per hour performing for a rich, attentive patron in a VIP lounge. Club patrons will often "sample" several different dancers with one lap dance each before picking one for a more expensive multisong bout of dancing. Thus, patrons can assess the relative attractiveness of different women through intimate verbal, visual, tactile, and olfactory interaction, and those attractiveness judgments can directly influence women's tip earnings, through the number of 3-min dances that patrons request from each dancer. In these ways, estrous attractive- ness effects on lap-dancer earnings in gentlemen's clubs may be stronger than in other kinds of psychology research that use photo ratings (e.g., Haselton et al., 2007) or other kinds of sex work (e.g., visual pornography, phone sex) that give fewer fertility cues across fewer modalities. 3. Methods Participants were recruited through indirect e-mails (forwarded through local industry contacts), newspaper advertisements, and flyers posted near clubs. To minimize possible response biases through demand characteristics, we mentioned the ovulatory cycle only in recruitment and consent and we did not suggest that tip earnings would be examined specifically as a function of days since menstrual onset. We believe that this brief mention of possible cycle effects on tip earnings in the consent form did not bias participant responses in any particular direction. To enter the study, each participant collected an experi- ment packet (containing a consent form, 14-page ques- tionnaire, and instructions for using the online web site) from a public location on the University of New Mexico (UNM) campus. The questionnaire (with an attached subject ID number) asked about age, ethnicity, work experience, sexual experience and attitudes, menstrual cycle characteristics, contraception use, physical characteristics, education, intel- ligence, and personality. Signed consent forms and com- pleted questionnaires were returned to separate boxes at UNM to maintain anonymity. Participants were also asked to use their anonymous subject ID number to log in to the web site every day for 60 days. Each day, they were to report their mood, work hours, work location, and tip earnings in US dollars and whether they had begun or ceased menstruation. Participants were offered a payment of US$30 upon completion of the study. A total of 18 women gave analyzable data (signed consent forms, completed questionnaires, and reported menstrual cycle data over the 60-day period); 1 additional woman completed the study but reported highly irregular menstrual activity (e.g., three different menstrual periods beginning in the same 10-day interval), rendering her unsuitable for relevant analyses. Because recruitment was through for- warded e-mails, advertisements, and flyers, it is hard to estimate the proportion of women who responded. 4. Results 4.1. Participant demographics and traits The 14-page questionnaire completed by each participant gave extensive background information, only some of which is relevant to this study. All participants were exclusively or primarily heterosexual. All women reported regular cycles of 28-30 days; 7 were using the hormonal contraception pill, and 11 were not (and had not within the previous 3 months); none used any other form of hormonal contraception (e.g., the patch or the implant). On average, participants were 26.9 years old (S.D.=5) and had 6.4 years of experience as exotic dancers (S.D.=2.1). Their demographics seem representative of exotic dancers studied in previous sociological research (e.g., Barton, 2006; Deshotels & Forsyth, 2006; Forsyth & Deshotels, 1996; Pasko, 2002). 4.2. Daily online reports and fertility estimates We asked participants to log in to our study web site daily to report whether they had started or stopped menstruating on that day and to report their tip earnings and other details of any shift they worked that day. Each participant's menstrua- tion data were plotted on a calendar, and we recorded how many days into the cycle each participant was for each shift reported. The online data revealed that all 18 participants showed quite regular cycles ranging from 28 to 29 days in cycle length—a good match to their self-reported cycle lengths of 28-30 days. Actuarial data (e.g., Wilcox, Dunson, Weinberg, Trussell, & Baird, 2001) suggest that fertility is high around Days 9- 15 of the cycle and is low around Days 1-8 (early follicular days including menstruation) and Days 16-28 (days in the luteal phase). We divided nonestrous parts of the cycle into menstrual and luteal phases because we expected that menstrual side effects (e.g., fatigue, bloating, muscle pains, irritability) might reduce women's subjective well-being and tip earnings and we wanted to be able to distinguish an estrous increase in tips from a menstrual decrease, relative to the luteal phase. Also, because fertility estimation is imperfect and fertility may be high a few days before or G. Miller et al. / Evolution and Human Behavior 28 (2007) 375-381 377 378 G. Miller et al. / Evolution and Human Behavior 28 (2007) 375-381 Fig. 1. Effects of ovulatory cycle (Days 1-28) on average tip earnings per shift, for normally cycling women versus women using hormonal contra- ception (pill users); each data point represents a 3-day average of the indicated day, the previous day, and the following day. after the typical fertile window, we used methods similar to those of Haselton and Gangestad (2006) by conservatively estimating cycle phase. Thus, the cycle was broken up into three phases: menstrual [Days 1-5 of the cycle (Days 6-8 were dropped because participants could have been fertile and were likely not menstruating)], fertile (Days 9-15 of the cycle), and luteal [Days 18-28 of the cycle (Days 16 and 17 were dropped because participants could have been fertile)]. We then calculated an average tip-earning level for each participant in her menstrual, fertile, and luteal phases, based on all available online data in each category. Apart from the ambiguous-fertility days dropped from our analysis, participants reported tip earnings in a total of 296 online entries, averaging 16.4 entries each (S.D.=5.2, range=9-29), out of the 60 days requested. Participants only logged on to the web site on about half of the days when they worked (27% of the 60 requested days). The average work shift lasted 5.2 h (S.D.=1.7, range=0- 12), usually starting between 5 and 10 p.m. and ending between midnight and 4 a.m. Participants reported mean earnings of US$248.73 per shift (S.D.=US$125.30). With lap dances yielding an average of about US$14, this mean earnings level of about US$250 reflects about 18 dances per shift. Shift length was unrelated to shift earnings (r=−.03, p=.63); thus, "earnings per shift," rather than "earnings per hour," was used in subsequent tip analyses. Additionally, number of hours worked per shift did not differ across cycle phase [F(2, 249)=1.19, p=.306; menstrual mean=5.44, fertile mean=5.26, luteal mean=5.04]. Fig. 1 shows average tip earnings for Days 1-28 in the cycle for normally cycling women versus pill-using women. 4.3. Effects of ovulatory cycle and contraception on tip earnings Our design involved multiple observations (i.e., tips per shift) for dancers, who were nested within contraception use and crossed with cycle phase; hence, we analyzed effects of cycle phase and contraception use on tip earnings using multilevel modeling (hierarchical linear modeling). Multi- level modeling is most appropriate in this context because it allows interpretable tests of cycle phase and contraception use despite multiple observations and the nonindependence of tip earnings and despite the differential number of observations between participants. Analyses were thus performed using SPSS 14.0 Mixed Models. Restricted maximum likelihood criteria were employed. Shifts (nested within individuals) were the Level 1 units of analysis, and participants were Level 2 units of analysis. Accordingly, cycle phase was a Level 1 factor and contraception use was a Level 2 factor. We report significance levels using both traditional p values for null-hypothesis testing and prep values (Killeen, 2005). Main effects of cycle phase [F(2, 236)=27.46, pb.001, prep=1.00] and contraception use [F(1, 17)=6.76, pb.05, prep=.929] were moderated by an interaction between cycle phase and pill use [F(2, 236)=5.32, pb.01, prep=.964]. Figs. 1 and 2 show these key results by plotting average tip earnings within each running 3-day period (Fig. 1) or each cycle phase (menstrual, fertile, or luteal; Fig. 2) for normally cycling participants versus pill-using participants. We conducted two planned contrasts of the interaction to investigate how cycle phase moderated the difference in tip earnings between normally cycling participants and pill users. The first contrast investigated how pill use moderated differences in tip earnings between the menstrual and luteal phases. We predicted that pill-using and normally cycling participants would demonstrate a similar difference between these two phases. The second contrast investigated how pill use moderated differences in tip earnings between the fertile phase and the other two phases. We predicted that normally cycling participants would demonstrate a larger increase in the fertile phase relative to the other phases than pill-using participants. Consistent with our predictions, the first contrast indicated that pill use does not moderate the difference between tips earned in the menstrual and luteal phases [F(1, Fig. 2. Effects of ovulatory cycle phase (menstrual phase, fertile estrous phase, or luteal phase) on tip earnings per shift, for normally cycling women versus women using hormonal contraception (pill users). Error bars represent 95% confidence intervals. 234)=0.012, p=.911], with normally cycling participants demonstrating a similar increase in tip earnings from menstrual to luteal phases (+70.45; 95% confidence limits from +25.68 to +115.16) as compared to pill-using participants (+66.60; 95% confidence limits from +20.72 to +112.26). Also consistent with our predictions, the second contrast indicated that pill use does moderate the difference between tips earned in the fertile phase and the other two phases [F(1, 238)=10.52, pb.01, prep=.983], with normally cycling participants demonstrating a greater increase in tip earnings at the fertile phase relative to the other two phases (+135.63; 95% confidence limits from +111.20 to +160.07) than pill users (+47.62; 95% confidence limits from +11.50 to +83.74). Although the sample size of individuals was small (N=18), the sample size of work shifts was much larger (N=296), and these work shifts reflected a sample of about 5300 lap dances (about 18 per work shift). 5. Discussion We found strong ovulatory cycle effects on tip earnings, moderated by whether the participants were normally cycling. All women made less money during their menstrual periods, whether they were on the pill or not. However, the normally cycling women made much more money during estrus (about US$354 per shift)—about US$90 more than during the luteal phase and about US$170 more than during the menstrual phase. Estrous women made about US$70 per hour, luteal women made about US$50 per hour, and menstruating women made about US$35 per hour. By contrast, the pill users had no midcycle peak in tip earnings. As in other previous research, the pill eliminates peak- fertility effects on the female body and behavior by putting the body in a state of hormonal pseudopregnancy (e.g., Gangestad, Simpson, Cousins, Garver-Apgar, & Christen- sen, 2004; Gangestad et al., 2005; Macrae, Alnwick, Milne, & Schloerscheidt, 2002). This also results in pill users making only US$193 per shift compared to normally cycling women making US$276 per shift—a loss of more than US $80 per shift. This is the first direct economic evidence for the existence of estrus in contemporary human females. Under the "revealed preference" doctrine in behavioral decision theory (Camerer, 2003; Hensher, Louviere, & Swait, 1999), real consumer spending patterns reveal human preferences more reliably than verbally stated judgments do, especially for socially stigmatized products such as pornography or sex work (Salmon & Symons, 2001). When women and men interact intimately over the course of several minutes through conversation and body contact, women apparently either "signal" or "leak" cues of their fertility status, and these cues influence spending patterns by male consumers. These results argue against the view that human estrus evolved to be lost or hidden from males (e.g., Strassmann, 1981; Turke, 1984). Indeed, the standard argument for women's "extended sexuality" (sexual receptivity outside the fertile estrus phase) is that it evolved to help women extract material resources from males (Gangestad et al., 2005; Thornhill, 2006), and some evidence suggests that women outside estrus place a higher value on male wealth relative to other male traits (Haselton & Miller, 2006). By this reasoning, nonestrous women in their extended sexuality phase should be better adapted to maximize tip earnings through displays of (nonfertile) sexual receptivity such as lap dancing. The fact that tip earnings peak during estrus suggests that men can detect female fertility more accurately than the "concealed ovulation" model suggested—but not so accurately that tips during the luteal and menstrual phases drop to zero (as they might if men found women generally unattractive during low-fertility parts of the cycle). As in so many coevolutionary arms races between the sexes (Arnqvist & Rowe, 2005), this outcome is not a clear victory for either sex. These cycle effects are notable because in previous research on gentlemen's clubs (e.g., Barton, 2006; Deshotels & Forsyth, 2006; Pasko, 2002), summarizing thousands of hours of interviews, dancers are never reported as noticing cycle effects on tip earnings (thus making it less likely that participant expectations or demand characteristics can explain the results). Yet, dancers have rich opportunities to learn how to maximize tip earnings. Learning optimal performance in any economic game requires immediate cash feedback across many iterations of the game (Camerer, 2003), and this is exactly what lap dancers get, every few minutes, in every shift, throughout the average of 6.4 years of dance experience. For this reason, we suspect that cyclic shifts in women's attractiveness are driving our tip earnings results—rather than the well-documented shifts in sexual receptivity, proceptivity, or selectivity for good genes (e.g., Feinberg et al., 2006; Gangestad et al., 2002, 2004, 2005; Haselton & Miller, 2006; Thornhill, 2006). Although an estrous increase in sexual receptivity and proceptivity toward clients who exhibit good-gene cues may bias dancers to approach certain men, it is unclear how this bias would lead to greater tip earnings during estrus. Indeed, it seems that the optimal strategy for obtaining tips is to focus on men who are profligate, drunk, and gullible rather than those who are intelligent, handsome, and discerning. This study has several limitations. The sample size of participants is small (N=18), although we gathered many data points per participant, which allowed us to use a statistically powerful repeated-measures design (including 296 work shifts reflecting about 5300 lap dances). Although the modest number of participants does not increase type I errors (i.e., false positives) in our statistical tests, it may reduce the generalizability of the results across populations —although it is unclear why different populations of sexually mature, normally cycling, human females would show different ovulatory cycle effects on tip earnings, if they work in the same industry. Another limitation is that our key measures (tip earnings, menstrual cycle phases, hormonal contraception G. Miller et al. / Evolution and Human Behavior 28 (2007) 375-381 379 380 G. Miller et al. / Evolution and Human Behavior 28 (2007) 375-381 use) were self-reported, to maximize participant anonymity and confidentiality in this stigmatized and suspicious population (see Thompson et al., 2003). Future studies could use larger samples, could identify ovulation more precisely using the luteinizing hormone surge as measured in urine samples, and could investigate whether these effects hold in other less-stigmatized types of sex work and service and entertainment industries, in other settings and cultures. A final limitation is that our study did not identify the precise proximal mechanisms that influence tip earnings. These might include the previously documented shifts in body scent, facial attractiveness, soft-tissue body symmetry, waist-to-hip ratio, and verbal creativity and fluency—or they might include shifts in other phenotypic cues that have not yet been studied. We can, however, exclude some possible mediators based on previous exotic dancer research. Tip earnings are unlikely to be influenced by cycle shifts in stage-dance moves, clothing, or initial conversational content because these cues just do not vary much for professional dancers (Barton, 2006; Beasley, 2003). The tip earnings pattern in Fig. 1 is similar to the pattern of estradiol levels across the cycle (with a main estrous peak and a secondary midluteal peak); hence, it is plausible that estradiol levels might mediate the tip- earning effects. Perhaps, most importantly, from an evolutionary view- point, further research could clarify whether women have evolved special adaptations to signal estrus through such cues—or whether the cues are "leaking" to sexually discriminating men as unselected side effects of cycle physiology. Distinguishing between estrous "signals" and "leaked cues" may be difficult in practice because estrous females (seeking extra-pair copulations with good-gene males) and extra-pair males (offering good genes) may have shared interests in female fertility signals being "conspiratorial whispers" that are accurate but inconspicuous (Pagel, 1994). In serially monogamous species such as ours, women's estrous signals may have evolved an extra degree of plausible deniability and tactical flexibility to maximize women's ability to attract high-quality extra-pair partners just before ovulation, while minimizing the primary partner's mate guarding and sexual jealousy. For these reasons, we suspect that human estrous cues are likely to be very flexible and stealthy—subtle behavioral signals that fly below the radar of conscious intention or perception, adaptively hugging the cost-benefit contours of opportunistic infidelity.

The Scientific Method

Scientific method refers to bodies of techniques for investigating phenomena, acquiring new knowledge, or correcting and integrating previous knowledge. Hypothesis: an explanation of observed results. You can make predictions from hypotheses. Collect data. nData can support hypotheses but does not prove them. Scientific hypotheses must be testable. Hypotheses that cannot be tested are not scientific.

Anthropology

The study of human and non-human primates Cultural Anthropology Ethnographies qLinguistics How humans share meanings derived from symbols and signs Archaeology Recovery and analysis of material remains Physical Anthropology The study of human biology within the framework of evolution

post-coital time interval

This study investigated experiences with partners during the time interval immediately following sexual intercourse. There is a tremendous volume of research on human sexuality, and in recent decades, evolutionary researchers have generated a large body of literature on variance in human reproductive strategies (see Buss, 2005). Much of this literature has focused on differences between male and female reproductive strategies and how these differences are represented in psychology and behavior. In comparison to AUTHOR NOTE: Please direct correspondence to: Daniel J. Kruger, Institute for Social Research and School of Public Health, University of Michigan, Ann Arbor, 1415 Washington Heights, Ann Arbor, MI 48109. Email: [email protected] 2010 Journal of Social, Evolutionary, and Cultural Psychology 254 Variation in Reproductive Strategies topics such as mate selection preferences, courting behavior, and sexual activities prior to full sexual intercourse, there has been relatively little attention paid to psychology and behavior following acts of sex in the evolutionary literature. Others have noted that discussions of the time spent together after sex has been conspicuously absent in the mass market products on sexuality and also underrepresented in the initial empirical literature (Halpern & Sherman, 1979). We believe that the Post-Coital Time Interval (PCTI), the time in which couples spend together after sexual intercourse before one partner leaves or falls asleep, is an important component of sexual relationships. Specifically, we argue that sex differences in PCTI experiences reflect divergence in the evolved reproductive strategies of men and women. We also predict that individual variation in PCTI experiences within each sex is related to other psychological aspects of variation in life history strategy, particularly tendencies towards engaging in committed long-term monogamous relationships. We designed an exploratory investigation of PCTI experiences and tested predictions derived from evolutionary theory regarding the psychology of human sexuality. Sex Differences in Reproductive Strategies The differential costs and benefits for various aspects of reproduction result in partially divergent reproductive strategies for women and men (Buss & Schmitt, 1993; Gangestad & Simpson, 2000). Women's average reproductive investment considerably outweighs average male investment (Fisher, 1992). Because women are obligated to a greater minimal investment in offspring than men, and are much more limited in the number of offspring they can produce, they tend to be more selective of sexual partners, and tend to place more emphasis on pair-bonding and having more enduring romantic relationships (Trivers, 1985). On the other hand, men have larger returns on reproductive success from having a greater number of mating partners (Bateman, 1948), and thus men tend to show a greater desire for a higher number of sexual partners and more variety in these partners (Gangestad & Simpson, 2000). Men have less of an incentive to commit to long-term monogamous relationships than women, as this would foreclose on multiple partnerships that could enhance their reproductive success more so than for women. Although men's average investment in offspring is considerably outweighed by women's average investment, it is still relatively high amongst mammalian males, particularly primates (Fisher, 1992). Substantial paternal investment may be necessary because of the extended development of offspring, who are far more altricial than those of closely related species (Fisher, 1992). Paternal investment is generally thought to enhance offspring reproductive success (Geary, 2005) and children who are raised with a father present have lower mortality rates in foraging cultures (Hill & Hurtado, 1996). Women seek male partners who have the ability and willingness to sustain long-term relationships with substantial contributions of resources (Buss, 1989, 1994). A mate who abandons his partner would remove paternal resources and care that are important for offspring survival (Gallup & Suarez, 1983). Previous models of sex differences in human sexuality are consistent with the expected consequences of sexual selection. For example, Hatfield, Sprecher, Pillemer, and Wexler (1989) describe the main goal of women's sexual behavior as expressing affection to another person in a committed relationship. In contrast, the main goal of men's sexual behavior is described as physical gratification. Likewise, preferences for certain activities during sex were consistent with this model. Compared to women, men Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 255 Variation in Reproductive Strategies were less concerned about receiving indications of love from their partner during intercourse, and more concerned about variety of sexual experiences and partner initiative (Hatfield, et al., 1989). Pair Bonding in the Post-Coital Time Interval Halpern and Sherman (1979) believe that the potential for bonding and sharing may be at its peak in the post-coital period, and satisfaction with this experience is the most important aspect of a sexual relationship. Despite women's efforts in screening and selecting partners prior to first sexual intercourse, women's feelings of uncertainty in the future of the relationship are likely due to the differential costs and benefits for commitment described above. Women's desires for expressions or signals of relationship bonding and commitment by one's partner may be particularly salient in the PCTI. One woman in her 30s remarked, "I think women have always known how important afterplay is. Many marriages have died because men don't." (Halpern & Sherman, 1979, p. 3). Indeed, in possibly the first empirical investigation on the topic, women reported greater desires for more loving behavior and less physical separation after intercourse than men, whereas men reported not enjoying "excessive" expression or affection after intercourse (Halpern & Sherman, 1979). Denny, Field, and Quadagno (1984) reported that women wanted to spend more time engaging in both foreplay and afterplay activities than their partners, with this being especially the case for afterplay. Women also had higher valuations for both verbal and physical affectionate behavior, whereas men were more likely to report enjoying sexual intercourse itself (Denny et al., 1984). Furthermore, it appears that certain hormones contribute to sex differences in desires for pair bonding; sexual activity is associated with a rise in oxytocin levels (thought to be important for pair bonding), and women have more neuronal oxytocin receptors than men (Hiller, 2004). Kissing is often thought of as an activity that can increase sexual arousal and receptivity and may also serve as a mechanism by which pheromones and sebum are exchanged to induce bonding. It has been shown that men tend to use kissing so as to increase sexual arousal and the likelihood that intercourse will occur, whereas women tend to use kissing more for promoting bonding (Hughes, Harrison, & Gallup, 2007), particularly during the PCTI. For instance, women rated both kissing after sex (Hughes, et al. 2007) and sleeping next to their partner after sex (Hughes, Harrison, & Gallup, 2004) as being more important than did men. In addition, both sexes indicated that women are usually the ones to initiate sleeping together after sexual intercourse (Hughes et al., 2004). Sleeping next to a romantic partner following copulation may serve multiple functions, including reducing the likelihood of philandering or abandonment by each partner, promoting greater paternity confidence through mate guarding, increasing sperm retention by remaining in a horizontal position after sex, and perhaps increasing the likelihood of pair-bonding after sex (Hughes et al., 2004). Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 256 Variation in Reproductive Strategies Hypotheses for Sex Differences in Post-Coital Time Interval Experiences Hypothesis 1: As women have greater incentives for establishing long-term relationship commitments than do men because of sex differences in minimal obligatory parental investment, we predict that women will report a greater desire for signals of bonding and commitment from their partner in the PCTI than will men. Hypothesis 2: Men, on the other hand, will be more likely to report that their partner is more interested in talking about relationship issues than they are during the PCTI. Hypothesis 3: Because individuals' levels of satisfaction with their PCTI experiences should stem from the sexual divergence in commitment incentives, we predict that women's satisfaction with PCTI experiences will be inversely related to the extent to which they desire higher levels of bonding and commitment signals from their partner than what is currently being given. Hypothesis 4: On the other hand, we anticipate that men's satisfaction with PCTI experiences will be inversely related to the extent to which they are less interested in discussing relationship issues than their partners. Variation in Reproductive Strategies We also predict that PCTI experiences will be associated with several psychological constructs that have been proposed to underlie or mediate human sexual strategies. For instance, there is considerable empirical evidence documenting the relationship between attachment styles and reproductive strategies (see Del Giudice, 2009). Specifically, Bowlby (1969) noted the relationship between attachment insecurity and unstable romantic relationships. Belsky, Steinberg, and Draper (1991) view the attachment process as an evolved psychological mechanism to evaluate social conditions and choose an effective reproductive strategy for those conditions. This model proposes that insecure attachment may be a response to environmental cues where long-term monogamous relationships are not the most viable strategy. Attachment avoidance is characterized by having discomfort with being close to partners and hiding true feelings from them. Del Giudice (2009) recently proposed that avoidant attachment styles are a component of high mating effort reproductive strategies (i.e., those with relatively low parental investment, and an emphasis on short-term and uncommitted mating), and are particularly related to variation in male reproductive strategies. Women may adopt avoidant strategies when conditions are particularly harsh and paternal investment is very unlikely, and adopt anxious attachment styles when environmental risk is moderate. Attachment anxiety is characterized by worries about being abandoned by one's partner and worries that one's feelings for a partner are not reciprocated. Attachment anxiety may be a mechanism to elicit relationship commitments and additional investment from mates and/or alloparents such as close kin. On average, men tend to have higher avoidance scores and lower anxiety scores than women (for a review, see Del Giudice, 2009). Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 257 Variation in Reproductive Strategies Figueredo and colleagues (e.g., Figueredo, Vásquez, Brumbach, Schneider, Sefcek, Tal, Hill, Wenner, & Jacobs, 2006) believe that a common factor underlies human life history parameters and reproductive, familial, and social behaviors. They propose a continuum of strategies ranging from a focus on short-term gains at the expense of long-term costs, high mating effort, and low parenting effort to long-term strategies with selective mating and high parental effort. This continuum is related to impulsivity, social support, disregard for social rules, and risk taking behaviors. Figueredo et al. (2006) developed the Mini-K Short Form of the Arizona Life History Battery as a brief inventory to assess life history strategy based on a factor analysis of data from the Midlife in the United States (MIDUS) study. Those who score higher on the inventory have greater tendencies for long-term, committed relationships, and thus may be more attentive to their partners in the PCTI and behave in other ways that facilitate emotional bonding. Hypotheses for Variation in Reproductive Strategies and Post-Coital Time Interval Experiences Following the models of life history variance in reproductive strategies described above, we predict that such psychological constructs will be associated with PCTI experiences. Hypothesis 5: We predict that attachment avoidance will be inversely related to the degree of PCTI bonding. Hypothesis 6: We also predict that attachment avoidance will be directly related to the degree to which they are less interested in discussion relationship issues than their partners. Hypothesis 7: Similarly, we propose that attachment anxiety will also be inversely related to the degree of bonding during the PCTI. Hypothesis 8: Attachment anxiety will be directly related to the extent to which participants desire greater levels of signals of bonding and commitment from one's partner. Following Del Giudice's (2009) model, attachment anxiety may be particularly relevant to women, so these relationships may be more evident and pronounced in women than in men. Hypothesis 9: Those scoring higher on the Mini-K (indicating tendencies for long-term, committed relationships) will report a greater degree of bonding in the PCTI. Hypothesis 10: Those scoring higher on the Mini-K will also be less likely to report that their partner is more interested in talking about relationship issues following sex than they are. Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 258 Participants and Procedure Variation in Reproductive Strategies Method Ethnically diverse (52% indicated some non-Western European ancestry) undergraduates (analytical sample N = 160, 93 female) completed anonymous surveys at their convenience over the Internet to either fulfill a course requirement or simply on a voluntarily basis to partake in a research study. Additional respondents who reported not yet having full sexual intercourse (n = 97; 34% of total participants) as well as those who reported being equally or more attracted to others of their own sex (n = 19; 7% of total participants) were excluded from our sample and analyses. Participants included students in the introductory Psychology pool at a large public Midwestern research university (n = 103) and evening Psychology students at a small private Mid-Atlantic liberal arts college (n = 57). Age of respondents (M = 23, SD = 7) ranged from 18 to 67 years. There was no significant sex difference in age. Respondents described their ethnic descent as Western European (47%), Eastern European (25%), African American (7%), Latino/Latina (6%), South Asian (2%), East Asian or Pacific Islander (2%), Native American (1%), and Other (11%). Respondents identified themselves as Christian (51%; including Catholic, 6%, Protestant, 21%, and Orthodox, 3%), Jewish (15%), Hindu (2%), Muslim (1%), and "Other" religious affiliation (8%). Measures We generated 16 items that assessed experiences with one's partner in the post- coital time interval (PCTI). Some items were positively worded, (e.g., "It is easy for us to have a heart-to-heart talk after sex,"), whereas other items were negatively worded, (e.g., "The time we spend together after sex feels like a chore," others were framed as desires "I wish that my partner were more loving and caring with me after we have sex.") General themes for the items measured included bonding, affection, communication, focus of attention, satisfaction, and responsiveness. The survey presented these items in a randomized order. Additional questionnaire measures included five items from each of the avoidant attachment and attachment anxiety dimensions of the Experiences in Close Relationships (ECR) inventory (Brennan, Clark, & Shaver, 1998) and the 20 items of the Mini-K life history strategy inventory (Figueredo et al., 2006). The brief versions of the ECR scales were developed with data from a previous study (N = 807, 51% female, M age = 19, SD age = 1): attachment avoidance Cronbach's alpha = .820, r(807) = .923 with full scale score; attachment anxiety Cronbach's alpha = .855, r(807) = .883 with full scale score. Demographic items included age, sex, race/ethnicity, religion, sexual orientation, and whether respondents have had full sexual intercourse with their partners. Analyses We used Principal Axis Factoring to reduce the novel PCTI items into dimensional factors and rotated factors with the Varimax Method using Kaiser Normalization. Items that loaded above .40 on a factor and did not load above .30 on any other factor were retained to calculate scale scores. Independent samples t-tests evaluated Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 259 α Women Variation in Reproductive Strategies Hypotheses 1 and 2, and Pearson correlations assessed the relationships predicted in the remaining hypotheses. Results We extracted three factors from the 16 PCTI items (see Table 1 for scale descriptives). The first factor identified was "Satisfaction and Bonding," (SB, Cronbach's alpha =.870) and the four items unique to this factor included "I am satisfied with the amount of time that my partner and I spend together immediately after intercourse," "It is easy for us to have a heart-to-heart talk after sex," and "The time spent together after sex is an important bonding experience." The second factor identified was "Longing for Connection," (LC, Cronbach's alpha = .910), with four items unique to this factor including "I wish that my partner were more romantic with me after we have sex," "I wish that my partner would communicate with me more after we have sex," and "I wish that my partner were more loving and caring with me after we have sex." The third factor identified was "Partner Neediness," (PN, Cronbach's alpha = .615) and the two items unique to this factor included "After we have sex, my partner wants to talk about our future more than I would like to" and "After we have sex, my partner always wants to talk about our relationship." Table 1. Scale Descriptives Satisfaction and Bonding Longing for Connection Partner Neediness A voidance Anxiety Mini-K Men d Items M SD M SD .870 4.13 .910 2.28 .615 2.03 .856 2.96 .900 4.14 .751 9.01 0.73 3.89 0.78 1.05 2.24 0.79 0.79 2.44 0.82 1.45 2.87 1.37 1.70 3.79 1.43 1.03 8.07 0.97 .32 4 .04 4 .51 2 .06 5 .22 5 .94 20 Hypothesis 1 was not supported;which participants desired greater levels of signals of bonding and commitment from their partner (LC) in the PCTI, t(158) = 0.380, p = .71. However, Hypothesis 2 was supported; men were more likely to report that their partner is more interested in talking about relationship issues than they are (PN), t(158) = 3.10, p = .002. Desires for greater bonding and commitment signals from one's partner (LC) were inversely related to satisfaction with PCTI experiences for both sexes (SB, see Table 2), supporting Hypothesis 3. The degree to which respondents thought their partner was more interested in talking about relationship issues than they were (PN) was inversely related to PCTI satisfaction and bonding (SB) for male participants only, supporting Hypothesis 4. Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 260 there was no sex difference in the extent to Variation in Reproductive Strategies Attachment avoidance was inversely related to the degree of PCTI satisfaction and bonding (SB) for both sexes, supporting Hypothesis 5. Attachment avoidance was directly related to the degree to which respondents felt their partner was more interested in talking about relationship issues than they were after sex, supporting Hypothesis 6. Attachment anxiety was inversely related to the degree of PCTI satisfaction and bonding for female participants only, supporting Hypothesis 7. Attachment anxiety was directly related to the degree of desire for signals of bonding and commitment from one's partner for female participants only, supporting Hypothesis 8. There was also a substantial positive relationship between degree of desire for signals of bonding and commitment by one's partner (LC) and reports that they are less interested in discussion relationship issues than their partners (PN) for men only. Both men and women scoring higher on the Mini-K reported a greater degree of satisfaction and bonding (SB) in the PCTI, supporting Hypothesis 9. Results did not support Hypothesis 10, as there was no relationship between Mini-K scores and reports of whether respondents thought their partner was more interested in talking about relationship issues than they were after sex (PN). Post-hoc analyses on attachment avoidance indicated that the sex X attachment avoidance interaction was the strongest (and only unique) predictor of satisfaction and bonding, t(273) = 8.11, p < .001, and partner neediness, t(273) = 7.24, p < .001. In both cases, the association with attachment avoidance was stronger for men. The sex X attachment anxiety interaction was the strongest (and only unique) predictor of satisfaction and bonding, t(273) = 5.06, p < .001, this relationship was stronger for women (See Table 2). Table 2. Associations Among Constructs for Female (top) and Male Participants (bottom) Satisfaction and Bonding Longing for Connection Partner Neediness A voidance Anxiety Mini-K Satisfaction and Bonding - -.299* -.387*** -.533*** -.118 .434*** Longing for Partner Connection Neediness -.610*** .001 - .126 .487*** - .137 .240* .182 .152 -.067 -.118 A voidance -.380*** .301** .284** - .489*** -.448*** Anxiety Mini-K -.274** .397*** .389*** -.149 .037 .083 .376*** -.350*** - -.203 -.099 - Note: * indicates p < .05, ** indicates p < .01, *** indicates p < .001. Discussion This study provides an initial inquiry in an area of sexual relationships currently underrepresented in psychological research. We believe that the Post-Coital Time Interval (PCTI) is an important aspect of sexual relationships and can reveal variation in Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 261 Variation in Reproductive Strategies reproductive strategies, particularly differences in strategies between women and men. The results confirmed most of our predictions, which follow from an evolutionary framework for the psychology of human sexuality. The PCTI may be particularly important for pair bonding, and women have a greater incentive to use the PCTI for pair- bonding and as a time to gain commitment promises from their partner because of the differential costs and benefits of mating effort and parental investment. We have operationally defined the PCTI as the time partners are awake together following sexual intercourse. This definition prioritizes the importance of communication between partners when pair-bonding issues are especially salient. These interactions would exemplify both the convergent and divergent aspects of female and male reproductive strategies. We acknowledge, however, that one partner could have experiences promoting bonding even if the other partner is asleep. This may influence relationship dynamics despite the lack of reciprocity from the sleeping partner. There was no sex difference in the reported desire for signals of bonding and commitment in the PCTI. Some may argue that participants may feel that they already have a commitment from their partner because they are in a relationship and thus do not need or seek additional commitment. Yet, the degree of this desire had a stronger relationship to satisfaction with PCTI experiences for women than for men. This finding is consistent with studies showing that women exhibit positive emotional shifts towards their partner following sex so as to promote pair-bonding, whereas men show opposite, negative shifts in affect toward their partner following sex (Haselton & Buss, 2001; Townsend, 1995). The degree to which respondents thought that their partners wanted to talk about relationship issues following sex more than they did was inversely related to male PCTI satisfaction, but was not a predictor of satisfaction for women. Overall, men reported a lesser interest in talking about relationship issues than their partners following sex as compared to women. Our results also demonstrated several relationships between PCTI concerns and psychological constructs related to variation in reproductive strategies. Each of the three extracted dimensions in this study was significantly associated with attachment style. PCTI satisfaction and bonding was inversely related to both forms of insecure attachment, avoidance and anxiety. Consistent with Del Giudice's (2009) model of attachment styles and reproductive strategies, attachment avoidance was associated with indicators of lower proclivity to establish relationship commitment. We found that women (but not men) who scored higher on attachment anxiety had stronger desires for additional signals of relationship commitment from their partners. PCTI experiences were also related to an index of life history strategy. Those individuals who were more future-oriented and prone to long-term committed relationships with substantial investment reported a greater degree of PCTI satisfaction and bonding. Curiously, male responses suggested that they would feel a better connection with a partner who was more detached. Because men tend to have a greater proclivity for short-term mating than do women (Buss, 1994), and men are better at avoiding emotional attachment following sex (Townsend, 1995), perhaps men feel more connected to and understanding of women who demonstrate similar tendencies. We consider these results promising for an initial investigation of the post-coital time interval (PCTI). There are several ways in which future research may enhance understanding of PCTI experiences. For instance, the primary limitation of the current study may be the collection of data from one partner in a pair, and perhaps gathering data from both individuals in a couple may provide a better understanding of relationship Journal of Social, Evolutionary, and Cultural Psychology - ISSN 1933-5377 - Volume 4(4). 2010. 262 Variation in Reproductive Strategies patterns. It may also be the case that the temporal location of a woman's menstrual cycle could reflect variations in PCTI behavior. In addition, the first iteration of our scales of PCTI experiences proved fruitful, however they may be refined considerably. All three scales are quite brief and should be extended, although two of the scales demonstrated excellent inter-item reliability despite their brevity. We also acknowledge the possibility of item associations due to the ways in which items were worded. Whereas we can correct for this potential artifact of design, we maintain that scale items have adequate face validity. We note that our results are based on samples of college students in the USA. The issues studied in psychological research are influenced by our evolutionary heritage of adaptations from selection pressures across numerous generations, which interact in complex ways with an individuals' socio-developmental environment. There are likely reliable aggregate differences in romantic relationship dynamics between populations which are consistent with selection pressures from the ecologies of the ancestral populations. The patterns of PCTI experiences may be different with groups experiencing endemic father absence, for example. There are groups and cultures with arranged marriages, normative polygyny, polyandry, and even sequential polygynandry. These cultural features and influences may shape PCTI phenomenon in predictable ways and we welcome collaborators who could assist us in addressing cross-cultural issues. Our results further demonstrate the power and fecundity of an evolutionary framework for understanding human psychology and behavior. Despite the exploratory nature of this study, we were able to generate and support several interconnected a priori predictions. We hope that this paper brings additional attention to an important aspect of sexual relationships which is rarely acknowledged by researchers regardless of theoretical orientation. As our participants reveal to us, fulfillment of goals in sexual relationships does not end with sex.

Primatology

We will look in depth at the behavior and ecology of primates. Child care Paternal investment Communication Mating systems Social systems Environments

Choosy Females, Amorous Males

Why do we usually assume males will be competitive and amorous and females choosy?

Genotype

genetic makeup (PP, Pp, or pp) PP = purple =Homozygous dominant Pp = purple =Heterozygous dominant pp=white = Homozygous recessive

Parental investment (PI)

investment in an offspring that increases its chance of survival and reduces parent's ability to invest in other offspring. Examples: protecting young, feeding young, producing gametes - eggs or sperm. Conceptually - PI is costly; PI is a limited resource.

Reproductive potential

number of surviving offspring that could potentially be produced by an individual (number of eggs, number of ejaculations)

Phenotypes

what one can observe