Evolution and Human Behavior 31 (2010) 48–58

Genetic dissimilarity, genetic diversity, and mate preferences in humansHanne. C. Liea,b,⁎, Leigh W. Simmonsb, Gillian Rhodesa

aSchool of Psychology, University of Western Australia, Crawley, WA, AustraliabCentre for Evolutionary Biology, School of Animal Biology, University of Western Australia, Crawley, WA, Australia

Initial receipt 2 April 2009; final revision received 5 July 2009

Abstract

It is clear that genes at the major histocompatibility complex (MHC) are involved in mate preferences in a range of species, includinghumans. However, many questions remain regarding the MHC's exact influence on mate preference in humans. Some research suggests thatgenetic dissimilarity and individual genetic diversity (heterozygosity) at the MHC influence mate preferences, but the evidence is ofteninconsistent across studies. In addition, it is not known whether apparent preferences for MHC dissimilarity are specific to the MHC or reflecta more general preference for genome-wide dissimilarity, and whether MHC-related preferences are dependent on the context of mate choice(e.g., when choosing a short-term and long-term partner). Here, we investigated whether preferences for genetic dissimilarity are specific tothe MHC and also whether preferences for genetic dissimilarity and diversity are context dependent. Genetic dissimilarity (number of allelesshared) influenced male, but not female, partner preferences, with males showing a preference for the faces of MHC-dissimilar females inboth mating contexts. Genetic diversity [heterozygosity (H) and standardized mean (d2)] influenced both male and female preferences,regardless of mating context. Females preferred males with greater diversity at MHC loci (H) and males preferred females with greaterdiversity at non-MHC loci (d2) in both contexts. Importantly, these findings provide further support for a special role of the MHC in humansexual selection and suggest that male and female mate preferences may work together to potentially enhance both male and femalereproductive success by increasing genetic diversity in offspring.© 2010 Elsevier Inc. All rights reserved.

Keywords: MHC; Mate preferences; Genetic diversity; Genetic similarity; Disassortative mate preferences; Facial attractiveness

1. Introduction

The major histocompatibility complex (MHC, or humanleukocyte antigen, HLA, in humans) is found in all jawedvertebrates and contains genes that are implicated in manyimportant biological functions, including immune function-ing (Doherty & Zinkernagel, 1975; Lechler & Warrens,2000) and mate preferences (reviewed in Havlicek &Roberts, 2009; Milinski, 2006). Because MHC genes areimportant for several aspects of individual fitness, they areparticularly good candidates for studying the genetic benefitsof mate choice (Apanius, Penn, Slev, Ruff, & Potts, 1997;Schwensow, Fietz, Dausmann, & Sommer, 2008; Tregenza& Wedell, 2000). MHC genes are directly involved inimmune functioning, where each allele codes for peptides

⁎ Corresponding author. School of Psychology (M304), University ofWestern Australia, Crawley, WA 6009, Australia. Tel.: +61 8 6488 3573;fax: +61 8 6488 1006.

E-mail address: hanneclie@graduate.uwa.edu.au (H.C. Lie).

1090-5138/$ – see front matter © 2010 Elsevier Inc. All rights reserved.doi:10.1016/j.evolhumbehav.2009.07.001

that can detect a restricted range of antigens derived frompathogens and parasites. MHC alleles are expressedcodominantly. Therefore, increased allelic diversity at theMHC should be beneficial as it broadens the range ofantigens an individual can detect and present to T cells fordestruction (Doherty & Zinkernagel, 1975). There isevidence that greater MHC diversity (heterozygosity)enhances immunocompetence in some cases (e.g., Carring-ton et al., 1999; Duggal et al., 2004; Froeschke & Sommer,2005; Hraber, Kuiken, & Yusim, 2007; McClelland, Penn, &Potts, 2003; Oliver, Telfer, & Piertney, 2009; Schwensow,Fietz, Dausmann, & Sommer, 2007; Thursz, Thomas,Greenwood, & Hill, 1997), although not in all (e.g., Hillet al., 1991; Meyer-Lucht & Sommer, 2005).

The exact effects of the MHC on mate preferences inhumans are still debated, but are likely to involve preferencesfor MHC diversity (heterozygosity) and/or compatibility ofparental MHC genotypes (e.g., Havlicek & Roberts, 2009;Roberts, Hale, & Petrie, 2006; Tregenza & Wedell, 2000).For example, the disassortative mating hypothesis proposes

49H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

that individuals may prefer mates with dissimilar MHCalleles to avoid inbreeding and to gain indirect benefits interms of increased genetic diversity and disease resistance inoffspring (Brown, 1997, 1999; Freeman-Gallant, Meguerdi-chian, Wheelwright, & Sollecito, 2003; Penn & Potts, 1998;Potts, Manning, & Wakeland, 1991; Tregenza & Wedell,2000; Yamazaki et al., 1976; Ziegler, Kentenich, &Uchanska-Ziegler, 2005). Additionally, the good-genes-as-heterozygosity hypothesis (Brown, 1997, 1999) proposesthat MHC diversity should be preferred in a mate if it isassociated with individual quality. Preferences for MHC-diverse mates could therefore be adaptive because high-quality mates should be able to provide direct benefits,including parental care, resources, and reduced risk ofcontagion to their partner and offspring (Kirkpatrick & Ryan,1991; Roberts, Little, Gosling, Perrett, et al., 2005;Sauermann et al., 2001). Moreover, choosing a geneticallydiverse, or heterozygous, mate could also potentiallyincrease genetic variability in the offspring since hetero-zygosity is on average heritable (Hoffman, Forcada, Trathan,& Amos, 2007; Mitton, Schuster, Cothran, & De Fries,1993). Mate preferences for genetic diversity and dissim-ilarity are not mutually exclusive, and both may contribute tothe maintenance of high levels of polymorphism observed atMHC loci (e.g., Apanius et al., 1997; Piertney & Oliver,2006).

Preferences for MHC-dissimilar mates have been foundin a range of nonhuman species, including lizards (Olsson etal., 2003), fish (Landry, Garant, Duchesne, & Bernatchez,2001), birds (Freeman-Gallant et al., 2003; but seeBonneaud, Chastel, Federici, Westerdahl, & Sorci, 2006),and mice (Penn & Potts, 1998; Potts et al., 1991; Yamazaki etal., 1976), while female sticklebacks choose mates to achievean intermediate level of MHC diversity in their offspring(Aeschlimann, Haberli, Reusch, Boehm, & Milinski, 2003;Milinski, 2003, 2006). In humans, studies of preferences forMHC dissimilarity in mates yield inconsistent results(recently reviewed in Havlicek & Roberts, 2009). Coupleswere found to be more dissimilar at the MHC than by chancein two reproductively isolated populations, in the Hutterites(Ober et al., 1997) and Mormons (Chaix, Cao, & Donnelly,2008), but not across the genome (Mormon sample),suggesting that the dissimilarity preference is specific tothe MHC (Chaix et al., 2008). However, no evidence ofMHC-based mate choice in couples was found in studiesacross a range of ethnic populations (Chaix et al., 2008;Hedrick & Black, 1997; Ihara, Aoki, Tokunaga, Takahashi,& Juji, 2000; Jin, Speed, & Thompson, 1995; Nordlander etal., 1983; but see Rosenberg, Cooperman, & Payne, 1983).Interestingly, increased MHC similarity in romantic coupleshas been associated with relationship dissatisfaction and atendency to seek extra-pair partners for females (Garver-Apgar et al., 2006).

In the laboratory, a female preference for the odor ofMHC-dissimilar males was found in two studies (Wedekind& Furi, 1997; Wedekind, Seebeck, Bettens, & Paepke, 1995)

but not in two larger studies (Roberts, Gosling, Carter, &Petrie, 2008; Thornhill et al., 2003). Additionally, Jacob,McClintock, Zelano, & Ober (2002) found a femalepreference for an intermediate level of MHC dissimilarityin males. Of three studies testing male preferences for femaleodor, two found a preference for MHC dissimilarity(Thornhill et al., 2003; Wedekind & Furi, 1997; see alsoSantos, Schinemann, Gabardo, & da Graca Bicalho, 2005).

Only one study has directly tested preferences for MHCdissimilarity in faces. Roberts, Little, Gosling, Jones, et al.(2005) conducted a study where each female rater was“prematched” with three MHC-similar and three MHC-dissimilar males. Contrary to past studies where a femalepreference for the odor of MHC-dissimilar males has beenreported, Roberts, Little, Gosling, Jones, et al. (2005) founda female preference for the faces of MHC-similar men (seealso Roberts, Little, Gosling, Perrett, et al., 2005). Roberts etal. suggested that the MHC-similarity preference is con-sistent with familial imprinting studies showing assortativepreferences for facial appearance among actual couples (e.g.,Bereczkei, Gyuris, Koves, & Bernath, 2002; Bereczkei,Gyuris, & Weisfeld, 2004). Thus, in humans, the currentevidence for the MHC-disassortative mating hypothesisis mixed.

Preferences for genetic diversity (heterozygosity) at theMHC have been found in some nonhuman animals(reviewed in Kempenaers, 2007). For example, female fat-tailed dwarf lemurs (Cheirogaleus medius) prefer their socialand extra-pair mates to be genetically diverse at both MHCand non-MHC loci (Schwensow, Fietz, et al., 2008). MHCdiversity also predicted male mating success in rhesusmacaques (Macaca mulatta, Sauermann et al., 2001). Inhumans, laboratory studies suggest that females prefer boththe odor and the faces of men who have greater geneticdiversity at MHC loci compared to less diverse males (Lie,Rhodes, & Simmons, 2008; Roberts, Little, Gosling, Perrett,et al., 2005; Thornhill et al., 2003), even when controlling fordiversity at non-MHC loci (Lie et al., 2008). However, malesdo not show a preference for MHC diversity in female faces(Coetzee et al., 2007; Lie et al., 2008; Thornhill et al., 2003),but may prefer genetic diversity at loci outside the MHC (Lieet al., 2008). Thus, at least female preferences for male MHCdiversity appear to be consistent with the good genes asheterozygosity hypothesis in humans.

Although there is evidence for MHC-related influenceson mate preferences in many species (Havlicek & Roberts,2009; Milinski, 2006; Penn, 2002; Penn & Potts, 1999), littleis known about the relative importance of the MHC versusgenetic background since most studies do not control for thepotential influence of non-MHC loci (but see Chaix et al.,2008; Lie et al., 2008; Reusch, Häberli, Aeschlimann, &Milinski, 2001; Schwensow, Fietz, et al., 2008). Moreover,in humans, little is known about whether preferences forMHC dissimilarity or diversity are dependent on the matingcontext. That is, are they equally preferred in short-term andlong-term mates? Roberts, Little, Gosling, Jones, et al.

50 H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

(2005), the only study to compare MHC effects in the twomating contexts, found that the preference for MHCsimilarity in faces was strongest in the long-term condition.Similarly, although they did not include a short-term context,Roberts, Gosling, Carter, & Petrie (2008) found that ratingodors specifically for desirability as a long-term partnerelicited stronger MHC-related effects than when rating odorsfor “pleasantness.” Thus, MHC-related preferences may beinfluenced by the mating context. Here, we explored whethergenetic dissimilarity and genetic diversity, within and outsidethe MHC, influence partner preferences, and whether suchpreferences are context dependent.

We investigated for the first time whether MHCdissimilarity predicts face preferences while controllingfor genetic dissimilarity at loci outside the MHC. Ratherthan comparing attractiveness ratings of high and low MHCdissimilar individuals (as in Roberts, Little, Gosling, Jones,et al., 2005; Wedekind et al., 1995), we used geneticdissimilarity as a continuous variable (as in Jacob,McClintock, Zelano, & Ober, 2002; Wedekind & Furi,1997), and investigated its effect on attractiveness prefer-ences. This study is also the first to test whether malesprefer the faces of MHC-dissimilar females. Althoughfemales are expected to be the choosier sex according toparental investment theory (Trivers, 1972), it should also beadaptive for males to maximize the genetic quality ofoffspring, especially in species, such as humans, wheremales have low potential to reproductively monopolizemany females and tend to invest heavily in their partner andoffspring (Cunningham & Birkhead, 1998; Schwensow,Eberle, & Sommer, 2008; Sefcek, Brumsbach, Vasquez, &Miller, 2006). We therefore examine both female and malepreferences for genetic dissimilarity in potential partners.

Second, although previous research has linked MHCdiversity to male facial attractiveness, and found anonsignificant trend for general genetic diversity to predictfemale attractiveness (Lie et al., 2008; Roberts, Little,Gosling, Perrett, et al., 2005), we do not know whether thesepreferences are influenced by the mate-choice context.Human mate preferences are expected to be contextdependent (e.g., Buss & Schmitt, 1993). Past researchsuggests that females show greater preference for facialmasculinity and symmetry, which are used as putative facialsignals of genetic quality, in short-term partners compared tolong-term partners (Little, Jones, Penton-Voak, Burt, &Perrett, 2002). Thus, we also examine whether preferencesfor genetic dissimilarity and for genetic diversity arecontext dependent.

2. Methods

2.1. Participants

The participants were 79 (32 female), heterosexual,Caucasian students at the University of Western Australia(aged between 18 and 32). They received course credit in

return for their participation, and they had all beenpreviously genotyped and photographed (Lie et al., 2008).

2.2. Face photographs

Photographs and genotype details of the faces of 79 malesand 80 females were available from our database, whichincluded the current participants. Details of the photographsare provided in Lie et al. (2008). In brief, the photographswere taken under standardized conditions, with the partici-pants displaying a neutral expression, and with any makeupor facial hair removed. To minimize any effects of clothingand hairstyle, the color photographs were digitally fitted withan oval mask, leaving only the face and minimal hair visible(see also Roberts, Little, Gosling, Perrett, et al., 2005).

2.3. Genotyping and genetic analysis

Each individual was typed at 12 MHC and at 11 non-MHC microsatellite loci (for details, see Lie et al., 2008).The 12 MHC loci are known to be in linkage disequilibriumwith functional MHC genes, including the classical HLAloci A, B, C, DRB1 and DQB1 (Malkki, Single, Carrington,Thomson, & Petersdorf, 2005). Genetic dissimilarity wasmeasured as the number of shared alleles between twoindividuals. For each rater and face (person in thephotograph) pair, number of alleles shared was calculatedusing SHAREDST, available from www2.biology.ualberta.ca/jbrzusto/sharedst.php (Brzustowski, n.d.). This programuses an algorithm to estimate allelic distance between allindividuals in a sample and can be used to count how manyalleles each individual shares with all other individuals in thesample. Following past research (Roberts, Little, Gosling,Jones, et al., 2005; Thornhill et al., 2003; Wedekind & Furi,1997; Wedekind et al., 1995), allele sharing was calculatedas the total number of alleles shared between two individualsat MHC or at non-MHC loci. If two individuals werehomozygous for the same allele at a locus, they were thencounted as sharing two alleles. Observed ranges of allelesharing between all possible male–female pairs were 0–13for non-MHC loci (mean=5.36, S.D.=0.81; possible range,0–22) and 0–21 for MHC loci (mean=6.67, S.D.=0.91;possible range, 0-24).

Following Lie et al. (2008), genetic diversity wasmeasured separately for the MHC and non-MHC lociusing heterozygosity (referred to as H), the proportion ofheterozygous loci among all loci measured, and standardizedmean d2, the squared difference in number of repeat unitsbetween the two alleles at a given locus, standardized by themaximum observed value at that locus, and averaged acrossall measured loci (Amos et al., 2001). Heterozygosityreflects whether the two alleles at a locus are the same ornot, while standardized mean d2 (referred to as d2) reflectsthe genetic distance between two alleles at a locus, averagedacross measured loci (Coulson et al., 1998).

It should be mentioned that the observed effects of MHCgenotype on appearance could potentially be driven by

51H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

substructure within the sample, reflecting regional differ-ences in MHC genotypes and facial characteristics, ratherthan actual effects of MHC diversity or similarity. Althoughsuch a confounding effect cannot be completely ruled out, itis unlikely because no evidence of substructure was found inthis sample (see Lie et al., 2008).

2.4. Attractiveness ratings

Participants rated photographs of opposite-sex indivi-duals for attractiveness in two mate-choice contexts, as ashort-term partner and as a long-term partner. The tworatings were completed at two different rating sessions,conducted over the internet, with a minimum of 2 weeksbetween rating sessions. Of the 32 female raters, 22 femalesrated the males in both contexts (six rated short-termattractiveness only and four rated long-term attractivenessonly). Of the 47 male raters, 38 males rated the females inboth contexts (four rated short-term attractiveness only andfive rated long-term attractiveness only). The photographswere presented in one of two random orders for both short-term and long-term attractiveness ratings. The participantswere assigned randomly to order of presentation and viewedall the faces in the sample before rating them forattractiveness as a short- or long-term partner on a scalefrom 1 (not attractive at all) to 10 (extremely attractive).Description of short term and long term was based onStewart, Stinnett, and Rosenfeld (2000). A short-termrelationship was defined as a single date accepted on thespur of the moment, an affair within a long-term relationship,and the possibility of a one-night stand. A long-termrelationship was defined as a relationship with someoneyou may want to move in with, someone you may considerleaving your current partner to be with, and someone youmay, at some point, wish to marry (or enter into a relationshipon similar grounds as marriage). We standardized theattractiveness ratings from each rater to control for individualdifferences in scale use (Roberts, Little, Gosling, Jones, etal., 2005; Roberts, Little, Gosling, Perrett, et al., 2005).Ratings of each individual face were averaged across theraters, while excluding faces familiar to the rater (on average

Table 1Means, standard deviations (S.D.), ranges, and Spearman’s ρ correlation coefficienfor males (below the diagonal, n=79) and females (above the diagonal, n=80)

Non-MHC-H MHC-H Non-MHC-d2 MHC-d2 Mating

Short-

Non-MHC-H 0.12 0.33⁎⁎ 0.09 0.08MHC-H 0.17 0.00 0.55⁎⁎⁎ 0.07Non-MHC-d2 0.31⁎⁎ −0.12 0.10 0.30⁎

MHC-d2 0.13 0.23⁎ 0.00 −0.05Short-term −0.04 0.21 −0.14 0.04Long-term −0.02 0.29⁎⁎ −0.16 0.11 0.93⁎

⁎ pb.05.⁎⁎ pb.01.⁎⁎⁎ pb.001.

3.75% per rater). Interrater agreement was high across allratings, with mean Cronbach's α=.93 (range, .89–.95),indicating that there was high interrater consistency in theattractiveness rankings of the faces.

3. Results

3.1. The relationship between short-term and long-termpartner attractiveness ratings

We first investigated the degree of consistency betweenattractiveness ratings across the two contexts (short-term andlong-term partners). Table 1 shows the descriptive statisticsand Spearman ρ correlation coefficients for the geneticdiversity and attractiveness variables for male faces (n=79)and female faces (n=80). The mean attractiveness ratings inthe short-term and long-term partner context correlatedhighly for both males and females (both rS's N.92, Table 1),suggesting a strong relationship between which faces peopleprefer across contexts.

However, these high correlations may misrepresent such arelationship because they do not reflect differences at thelevel of the individual but rather at a group level. Toinvestigate differences in preference within the individual,we correlated each rater's attractiveness ratings of the samefaces in the short-term and in the long-term context. We usedSpearman's ρ correlation coefficients (rS) as some of theindividual attractiveness rating distributions were skewed.The mean correlation coefficients between short-term andlong-term attractiveness across individuals were substantialbut considerably lower than for the group-level coefficientsfor male raters (mean rS=0.54; S.D.=0.16; range, 0.13–0.78;n=37) and among female raters (mean rS=0.58; S.D.=0.14;range, 0.26–0.81; n=22). Thus, there was potential forgenetic dissimilarity and diversity to influence preferencesfor short-term and long-term partners in different ways.

3.2. Genetic dissimilarity and facial attractiveness

Next, we examined whether attractiveness ratings wereinfluenced by genetic dissimilarity, measured as the number

ts for the genetic diversity and attractiveness variables presented separately

context Males Females

term Long-term Mean±S.D. Range Mean±S.D. Range

0.11 0.83±0.12 0.55–1.00 0.84±0.13 0.37–1.000.02 0.84±0.14 0.42–1.00 0.83±0.13 0.30–1.00

⁎ 0.27⁎ 0.15±0.06 0.04–0.34 0.16±0.07 0.05–0.40−0.11 0.17±0.07 0.07–0.35 0.17±0.07 0.05–0.360.95⁎⁎⁎ 3.53±0.94 1.91–6.25 3.93±0.91 1.98–6.56

⁎⁎ 3.52±1.05 1.65–6.46 4.13±0.93 2.30–6.70

Table 2The mean regression slopes (β) of the number of non-MHC and MHC alleles shared with male raters predicting female facial attractiveness and their 95%confidence intervals averaged across 80 hierarchical multiple regression models

Contexta Mean β (95% CI) t(80) p Range of β's Negative/total β's

Short-termNon-MHC −0.03 (−0.07 to 0.01) −1.69 .096 −0.35 to 0.34 48/80MHC −0.10 (−0.13 to −0.06) −5.72 b.0001 −0.48 to 0.29 59/80⁎

Long-termNon-MHC −0.01 (−0.05 to 0.02) −0.75 .454 −0.43 to 0.31 47/80MHC −0.11 (−0.14 to −0.07) −5.54 b.0001 −0.48 to 0.29 59/80⁎

Number of non-MHC and MHC alleles shared was entered in Block 1 and Block 2, respectively. The t and p values are from single-sample t tests comparing themean β to zero. The range of β's and the number of negative β of total number of β are also provided.

a Context refers to whether the faces were rated for attractiveness as short-term or as long-term partners.⁎ Sign test pb.001, comparing the number of observed negative β's to chance (40/80).

52 H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

of alleles shared between the rater and the person being rated.We used analyses that treat the face (the individual beingrated) as the unit of analysis. This approach has the greatestpower to detect preferences for MHC dissimilarity because itcontrols for factors other than allele sharing that mayinfluence a person's attractiveness (Roberts, Little, Gosling,Jones, et al., 2005; Thornhill et al., 2003; Wedekind et al.,1995). Because we were interested in the effect of MHCdissimilarity on attractiveness ratings (both short-term andlong-term) after controlling for effects of non-MHC dissim-ilarity, we used hierarchical multiple regression modelswhere non-MHC dissimilarity was entered in the first blockand MHC dissimilarity in the second block. This procedureallowed us to assess whether MHC dissimilarity affectsattractiveness ratings over and above non-MHC similarity.Thus, for each face being rated (80 female faces and 79 malefaces), we used the number of non-MHC and MHC allelesshared with the rater to predict the attractiveness ratings inboth the short-term and long-term contexts. Normality ofresiduals was checked for each model. For female faces, thisresulted in 80 β's for each predictor variable (i.e., non-MHCand MHC dissimilarity) in each context, and for male faces,in 79 β's for each predictor in each context. We thencalculated the average regression slope, “mean β,” for eachpredictor in each context, separately for the female and themale faces. To test for non-MHC or MHC-dissimilarpreferences, we compared their respective mean β's to

Table 3The mean regression slopes (β) of number of non-MHC andMHC alleles shared witintervals averaged across 79 hierarchical multiple regression models

Contexta Mean β (95% CI) t(79)

Short-termNon-MHC 0.04 (−0.09 to 0.01) −1.75MHC 0.00 (−0.05 to 0.05) 0.00Long-termNon-MHC −0.02 (−0.05 to 0.01) −1.16MHC −0.01 (−0.06 to 0.04) −0.23

Number of non-MHC and MHC alleles shared was entered in Block 1 and Block 2,mean β to zero. The range of β's and the number of negative β of total number of

a Context refers to whether the faces were rated for attractiveness as short-te

zero using single-sample t tests (all reported p values aretwo-tailed).

3.2.1. Genetic dissimilarity and female facial attractivenessThere was a significant effect of MHC dissimilarity on

female attractiveness. After controlling for non-MHCdissimilarity, males preferred the faces of MHC-dissimilarfemales. That is, females who shared fewer MHC alleleswith the males received higher attractiveness ratings thanfemales who shared more MHC alleles, in both the short-term and the long-term attractiveness context (Table 2).There was one β value for MHC allele sharing that wasmore than three standard deviations below the mean inthe short-term context. Removing this outlier did notinfluence the effect of MHC dissimilarity on attractivenessratings [mean β (95% CI)=−0.09 (−0.12 to −0.06), t(78)=−5.72, pb.0001].

There was also a nonsignificant trend for female faces thatshared fewer non-MHC alleles with the males to be rated asmore attractive in the short-term attractiveness context(Table 2). This trend became significant when removingone β value for non-MHC allele sharing that was more thanthree standard deviations above the mean in the short-termcontext [mean β (95% CI)=−0.04 (−0.07 to −0.01), t(78)=-2.26, p=.026]. Thus, genetic dissimilarity at the MHC, andpossibly outside the MHC, appears to influence males'preferences for female facial attractiveness.

h female raters predicting male facial attractiveness and their 95% confidence

p Range of β's Negative/total β's

.085 −0.57 to 0.40 46/791.00 −0.41 to 0.46 40/79

.252 −0.34 to 0.32 45/79

.818 −0.50 to 0.59 37/79

respectively. The t and p values are from single-sample t tests comparing theβ are also provided.rm or as long-term partners.

Table 4Hierarchical multiple regression models of female attractiveness as short-term or long-term partner regressed on non-MHC- andMHC-heterozygosity, or on non-MHC- and MHC-standardized mean d2

Contexta B±95%CI S.E. β t p R2/ΔR2

HeterozygosityShort-term Non-MHC-H 0.85 (−0.80 to 2.51) 0.83 0.119 1.03 .309 0.03/0.01

MHC-H 0.59 (−1.03 to 2.22) 0.82 0.085 0.73 .470Long-term Non-MHC-H 1.17 (−0.45 to 2.78) 0.81 0.167 1.44 .154 0.03/0.00

MHC-H 0.12 (−1.46 to 1.71) 0.80 0.018 0.15 .878Standard mean d2

Short-term Non-MHC-d2 3.67 (0.56 to 6.78) 1.56 0.261 2.35 .021 0.07*/0.00MHC-d2 −0.55 (−3.48 to 2.36) 1.47 −0.042 −0.38 .706

Long-term Non-MHC-d2 3.37 (0.33 to 6.41) 1.53 0.245 2.21 .030 0.06*/0.01MHC-d2 −1.03 (−3.88 to 1.83) 1.44 −0.080 −0.72 .476

Only Block 2 results are presented, n=80 (full models are available online as Supplementary Data Appendix A, Table S1).a Context refers to whether the faces were rated for attractiveness as short-term or as long-term partners.

53H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

3.2.2. Genetic dissimilarity and male facial attractivenessWe found no evidence for an effect of genetic dissim-

ilarity, at either non-MHC or MHC loci, on attractivenessratings of male faces made by female raters in either context(Table 3).

3.3. Genetic diversity and facial attractiveness

We also investigated whether genetic diversity predictedattractiveness in both rating contexts. Again, we usedhierarchical multiple regression models and regressed eithershort-term or long-term attractiveness on non-MHC diversity(H or d2, entered in Block 1), and MHC diversity (H or d2,entered in Block 2) to assess whether MHC diversity affectsattractiveness ratings over and above non-MHC diversity.Analyses were conducted separately for males and females.Attractiveness ratings in both the short-term and long-termcontexts were normally distributed, but all of the geneticdiversity distributions were negatively skewed. The skeweddistributions were left untransformed, because the residualsof each model were normally distributed (and usingtransformed variables did not change the results reported,data not shown).

Table 5Hierarchical multiple regression models of male attractiveness as short-term or longMHC- and MHC-standardized mean d2

Contexta B±95%CI S.E.

HeterozygosityShort Non-MHC-H −1.04 (−2.99 to 0.91) 0.98

MHC-H 1.72 (0.08 to 3.38) 0.83Long Non-MHC-H −0.82 (−2.54 to 0.90) 0.86

MHC-H 2.07 (0.61 to 3.52) 0.73Standard mean d2

Short Non-MHC-d2 −1.83 (−5.53 to 1.87) 1.86MHC-d2 −0.02 (−3.52 to 3.48) 1.76

Long Non-MHC-d2 −2.27 (−5.57 to 1.03) 1.66MHC-d2 0.89 (−2.23 to 4.02) 1.57

Only Block 2 results are presented, n=79 (full models are available online as Supa Context refers to whether the faces were rated for attractiveness as short-te

3.3.1. Genetic diversity and female facial attractivenessWhen measuring genetic diversity as heterozygosity,

neither non-MHC-H nor MHC-H predicted female attrac-tiveness in either context (Table 4). However, whenmeasuring genetic diversity as genetic distance (d2), femalefacial attractiveness was significantly predicted by non-MHC diversity (non-MHC-d2) in both the short-term and thelong-term context. More genetically diverse females wererated as more attractive than less diverse females. Non-MHC-d2 explained 7% of the variance in short-termattractiveness ratings and 6% of variance in long-termattractiveness ratings. MHC-d2 did not, however, predictfemale facial attractiveness in either context. Thus, malesappear to prefer non-MHC, but not MHC, diversity in apartner regardless of the mate-choice context.

3.3.2. Genetic diversity and male facial attractivenessWhen measuring genetic diversity as heterozygosity,

male facial attractiveness was significantly predicted byMHC diversity (MHC-H), after controlling for non-MHC-H, in both the short-term and the long-term context(Table 5). That is, females preferred the faces of males withgreater MHC diversity compared to less diverse males.

-term partner regressed on non-MHC- and MHC-heterozygosity, or on non-

β t p R2/ΔR2

−0.121 −1.07 .289 0.06*/0.05*0.237 2.08 .041

−0.106 −0.95 .346 0.10**/0.10**0.315 2.83 .006

−0.112 −0.99 .328 0.01/0.00−0.001 −0.01 .992−0.155 −1.37 .175 0.03/0.000.064 0.57 .571

plementary Data Appendix A, Table S2).rm or as long-term partners.

54 H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

MHC-H explained 5% of the variance in short-termattractiveness ratings and 10% of the variance in thelong-term attractiveness ratings. Non-MHC-H did notpredict attractiveness of male faces in either context.MHC-H explained significantly more variance than non-MHC-H in both short-term [F change(1,76)=4.34, p=.041]and long-term attractiveness [F change(1,76)=7.99,p=.006]. Neither non-MHC-d2 nor MHC-d2 predictedmale facial attractiveness in either context (Table 5).Thus, females show a preference for MHC diversity (H)in both short-term and long-term partners.

4. Discussion

We investigated whether genetic dissimilarity and geneticdiversity at both MHC and non-MHC loci influence humanmate preferences in short-term and long-term contexts. Wefound for the first time that males show a specific preferencefor MHC-dissimilar females, after controlling for non-MHC-dissimilarity, in both short-term and long-term partners.Moreover, there may also be a weak effect of non-MHCdissimilarity on male preferences for female faces, at least inshort-term partners.

A second novel finding was that preferences for geneticdiversity in a mate do not appear to be context dependent.Extending the finding of Lie et al. (2008), we found that non-MHC diversity (d2) predicted female facial attractiveness inboth partner contexts, whereas MHC diversity (H) predictedattractiveness in male faces in both contexts. Importantly, ourresults provide some evidence for both the disassortativemating hypothesis and the “good-genes as heterozygosityhypothesis,” and are discussed below.

4.1. Genetic dissimilarity and facial attractiveness

Our findings that males prefer the faces of females that areMHC-dissimilar, and possibly non-MHC-dissimilar, implythat faces contain cues to genetic dissimilarity, both at theMHC and in general. They also add to the evidence that theMHC may play a special role in human sexual selectionbecause MHC dissimilarity influenced face preferenceswhen controlling for non-MHC-dissimilarity. Contrary toRoberts, Little, Gosling, Jones, et al. (2005) (see alsoRoberts, Little, Gosling, Perrett, et al., 2005), we found noevidence for an influence of genetic similarity on femalepreferences for male faces at either MHC or non-MHC loci.

The discrepancy in the previous findings and the currentfindings could potentially reflect differences in the designs ofthe two studies. First, Roberts et al. measured MHCsimilarity across three MHC genes (HLA-A, HLA-B, andHLA-DRB1), whereas we used 12 microsatellite markers inlinkage with at least one MHC gene (including HLA-A,HLA-B, and HLA-DRB1; Malkki et al., 2005), and that spana larger area of the MHC region than the three MHC genes.Second, importantly, we controlled for genetic similarity atnon-MHC loci, which Robert et al. did not. Third, rather than

prematching raters with a few faces (n=6) in terms of MHCsimilarity and dissimilarity, our subjects rated a large sampleof opposite-sex faces (n=79 or 80), thus allowing us toexamine continuous variation in the effect of geneticsimilarity on face preferences. However, our sample ofraters was smaller than the sample of Roberts et al. (n=32 vs.92), which might make a weak MHC-similarity preferenceharder to detect. Lastly, Roberts et al. used female raters whowere not taking oral contraception and at their most fertileperiod in their menstrual cycle, whereas we did not controlfor contraceptive use. However, if contraceptive useincreases females' preferences for MHC similarity assuggested by the odor preference study of Roberts et al.(2008), then this should increase, not hamper, our chances offinding assortative preferences.

Arguably, Roberts' method may be more sensitive thanour method to detect any MHC similarity preferences bycomparing extreme groups. However, we found significantdissimilarity preferences for males, suggesting that the lackof effect in females was not due to insufficient sensitivity.Thus, our results do not lend support to the conclusion ofRoberts et al. (2005) that face preferences in humans areMHC-assortative. Rather, we found evidence for MHCdisassortative preferences across two mating contexts.Altogether, the differences in experimental designs rendera direct comparison of results impossible. Further research isneeded to determine whether face preferences are assortativeor disassortative, and whether the apparent sex difference inpreferences is robust.

The lack of an effect of MHC similarity on femalepreferences is in line with two recent odor preference studiesthat did not find any preference for the odor of MHC similarmales (Roberts et al., 2008; Thornhill et al., 2003).Additionally, the male MHC dissimilarity preference forfemale faces is consistent with the findings that males preferthe scent of MHC-dissimilar females (Thornhill et al., 2003;Wedekind & Furi, 1997). It is also consistent with two studiesthat found actual couples to be more MHC dissimilar thanexpected by chance (Chaix et al., 2008; Ober et al., 1997).

The male preference for MHC-dissimilar females appearsat odds with reports that couples look more similar thanexpected by chance (e.g., Hinsz, 1989; Little, Burt, & Perrett,2006) and with familial imprinting studies reportingassortative facial preferences among couples (Bereczkeiet al 2002; 2004). However, the perceived similarity betweencouples may reflect similarity in facial characteristics that arenot associated with genetic similarity but rather withenvironmental similarity.

Why would males seek genetically dissimilar females?Because human males' reproductive success is constrainedby partner fertility, regardless of their own mating strategy(Buss & Schmitt, 1993), they should value fertility in theirpartners. A preference for MHC dissimilarity may serve toreduce fertility problems associated with increased MHCallele sharing, including longer birth intervals, recurrentspontaneous abortion, and fetal loss in humans (Beydon &

55H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

Saftlas, 2005; Ober, 1999; Ober, Elias, Kostyu, & Hauck,1992; Ober, Hyslop, Elias, Weitkamp, & Hauck, 1998, seealso Knapp, Ha, & Sackett, 1996).

One could speculate that MHC dissimilarity influencedmale, but not female, mate preferences because males maybe more sensitive than females to facial cues to MHCdissimilarity. Visual cues appear to be more important formale than female mate choice, whereas odor cues are moreimportant for female mate choice (Havlicek et al., 2008;Herz & Inzlicht, 2002). Interestingly, males also use facialresemblance to assess paternity, and the allocation ofparental investment to offspring (Platek et al., 2004), againsuggesting that facial cues are important for male reproduc-tive decisions.

4.2. Genetic diversity and facial attractiveness

Although the relative attractiveness of potential matesvaried across rating context (average rS=0.5 within indivi-duals), preferences for genetic diversity did not appear to becontext dependent. Genetic diversity was preferred in bothshort-term and long-term partners. Using the same malefaces as in Lie et al. (2008), these results extend previousfindings that MHC diversity is attractive in male faces whenno mating context is specified (Lie et al., 2008; Roberts,Little, Gosling, Perrett, et al., 2005). Here, MHC diversityexplained 10% of the variance in long-term attractivenessand 5% in short-term attractiveness. This difference isdifficult to compare formally, but the 95% confidenceintervals around the β weights of MHC-H when predictingshort-term versus long-term attractiveness overlap, thussuggesting that the difference in effect sizes may not bebiologically relevant. Females should prefer MHC diversityin both short-term and long-term mates because hetero-zygosity is somewhat heritable (Hoffman et al., 2007; Mittonet al., 1993). Moreover, provided MHC diversity isassociated with fitness, choosing an MHC-diverse matemay provide both resources and reduced risk of contagion inshort-term and long-term relationships as well as high-quality parental care in long-term relationships (Gangestad& Scheyd, 2005; Roberts & Little, 2007).

The finding that non-MHC-d2 significantly predictedfemale attractiveness extends the findings of Lie et al. (2008)who found a nonsignificant trend for non-MHC-d2 to predictattractiveness using the same female faces. Here, using adifferent sample of raters, we found that non-MHC-d2

explained 7% and 6% of the variance in attractiveness ratingsin the short-term and the long-term context, respectively.Consistent with previous findings (Coetzee et al., 2007; Lieet al., 2008; Thornhill et al., 2003), there was no evidence ofa male preference for MHC diversity in female faces. Thislack of an MHC preference is consistent with the malepreference for MHC dissimilarity in the same female facesbecause the most MHC-diverse female may not be the mostMHC-dissimilar female to a male (see Roberts, Hale, &Petrie, 2006, for further discussion). Combined, male

preferences for general genetic diversity and MHC dissim-ilarity could be adaptive as they would increase geneticdiversity in offspring, and reduce potential fertility-relatedproblems that are associated with reduced genetic diversityin females (Ober, Hyslop, & Hauck, 1999) and withincreased MHC allele sharing in couples (Beydon & Saftlas,2005; Ober, 1999).

It is not clear why standardized mean d2, and notheterozygosity, should be associated with male preferencesfor females (discussed in Lie et al., 2008). Mean d2 has beencriticized for being less sensitive to genetic diversity–fitnesscorrelations than heterozygosity (e.g., Slate et al., 2004;Tsitrone, Rousset, & David, 2001). However, mean d2 maybetter detect diversity–fitness associations in large, outbredpopulations where most individuals are heterozygous at mostloci, as is the case for our human sample (Höglund et al.,2002; Kretzmann, Mentzer, DiGiovanni, Leslie, & Amato,2006). It is therefore possible that non-MHC-d2 providesinformation regarding genetic variation outside the MHC notcaptured by heterozygosity, including more detailed infor-mation about the inbred–outbred continuum, which mayinfluence female fitness. Alternatively, non-MHC-H andnon-MHC-d2 differ in their underlying distributions, espe-cially when using a limited number of markers as here,which could explain some of the observed differencesbetween associations to facial appearance.

As for most other human sexual selection research, anobvious limitation of our study is that we examined matepreferences, not actual mate choice, in a laboratory setting.Although this approach has the advantage of examining matepreferences without the constraints of mate choice in the realworld (Roberts & Little, 2007), the challenge for futureresearch is to test whether these subtle preferences translateinto actual behavior. Moreover, here, partner preferenceswere judged from color photographs. Although this allowedus to establish that humans are sensitive to phenotypic cuesof genetic dissimilarity and diversity in faces, it mayunderestimate the cues available from real-life faces. Itwould be of interest to determine what other features mayalso signal MHC dissimilarity and MHC diversity, and therelative importance of these cues for mate choice. Futureresearch could use more complex, naturalistic stimuli thatcontain richer information about an individual, for example,dynamic video clips including faces, bodies, and voices, orreal-life interpersonal interactions such as in a speed-date-like setting or in actual couples.

In conclusion, our results add to the mounting evidencethat the MHC plays a special role in human mate preferences.We found that MHC dissimilarity was especially importantfor male mate preference and that MHC diversity wasespecially important for female mate preferences. Inaddition, males also showed a preference for faces offemales that are genetically diverse at non-MHC loci,suggesting that general genetic diversity may also play arole in human sexual selection. Moreover, both the male andfemale preferences appeared to be robust to the mating

56 H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

context in which partner preference judgments were made(e.g., short-term or long-term partner context). These resultssuggest that preferences vary between the sexes and that,combined, these preferences could work to enhance bothmale and female reproductive success by enhancing thegenetic diversity of offspring. The next challenge lies intesting these predictions in actual couples.

Acknowledgments

We thank Jason Kennington, Paco Garcia-Gonzalez,Maxine Beveridge, and Davina French for advice.This work was supported by the University of WesternAustralia, the Australian Federation of University Women,Western Australia (AFUW-WA), and the AustralianResearch Council.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.evolhumbehav.2009.07.001.

References

Aeschlimann, P. B., Haberli, M. A., Reusch, T. B. H., Boehm, T., &Milinski, M. (2003). Female sticklebacks Gasterosteus aculeatus useself-referencing to optimize MHC allele number during mate selection.Journal of Evolutionary Biology, 19, 558−569.

Amos, W., WorthingtonWilmer, J., Fullard, K., Burg, T. M., Croxall, J. P., &Bloch, D., et al. (2001). The influence of parental relatedness onreproductive success. Proceedings of the Royal Society of London BBiological Sciences, 268(1480), 2021−2027.

Apanius, V., Penn, D., Slev, P. R., Ruff, L. R., & Potts, W. K. (1997). Thenature of selection on the major histocompatibility complex. CriticalReviews in Immunology, 17(2), 179−224.

Bereczkei, T., Gyuris, P., Koves, P., & Bernath, L. (2002). Homogamy,genetic similarity, and imprinting; parental influence on mate choicepreferences. Personality and Individual Differences, 33, 677−690.

Bereczkei, T., Gyuris, P., &Weisfeld, G. (2004). Sexual imprinting in humanmate choice. Proceedings of the Royal Society of London - Series B:Biological Sciences, 271, 1129−1134.

Beydon, H., & Saftlas, A. F. (2005). Association of human leucocyte antigensharing with recurrent spontaneous abortions. Tissue Antigens, 65,123−135.

Bonneaud, C., Chastel, O., Federici, P., Westerdahl, H., & Sorci, G. (2006).Complex MHC-based mate choice in a wild passerine. Proceedings ofthe Royal Society of London B Biological Sciences, 273(1590),1111−1116.

Brown, J. L. (1997). A theory of mate choice based on heterozygosity.Behavioral Ecology, 8(1), 60−65.

Brown, J. L. (1999). The new heterozygosity theory of mate choice and theMHC. Genetica, 104(3), 215−221.

Brzustowski, J. Retrieved 25/11/2008 from http://www2.biology.ualberta.ca/jbrzusto/sharedst.php.

Buss, D. M., & Schmitt, D. P. (1993). Sexual strategies theory: Anevolutionary perspective on human mating. Psychological Review, 100,204−232.

Carrington, M., Nelson, G. W., Martin, M. P., Kissner, T., Vlahov, D., &Goedert, J. J., et al. (1999). HLA and HIV-1: heterozygote advantage andB*35-Cw*04 disadvantage. Science, 283(5408), 1748−1752.

Chaix, R., Cao, C., & Donnelly, P. (2008). Is mate choice in humans MHC-dependent? PLoS Genetics, 4(9), e1000184, doi:10.1371/journal.pgen.1000184.

Coetzee, V., Barrett, L., Greeff, J. M., Hanzi, S. P., Perrett, D. I., & Wadee,A. (2007). Common HLA alleles associated with health, but not withfacial attractiveness. PLoS ONE, 2(7), e640, doi:10.1371/journal.pone.0000640.

Coulson, T. N., Pemberton, J. M., Albon, S. D., Beaumont, M.,Marshall, T. C., & Slate, J., et al. (1998). Microsatellites revealheterosis in red deer. Proceedings of the Royal Society of London BBiological Sciences, 265(1395), 489−495.

Cunningham, E. J. A., & Birkhead, T. R. (1998). Sex roles and sexualselection. Animal Behaviour, 56, 1311−1321.

Doherty, P. C., & Zinkernagel, R. M. (1975). Enhanced immunologicalsurveillance in mice heterozygous at the H-2 gene complex. Nature, 256(5512), 50−52.

Duggal, P., Haque, R., Roy, S., Mondal, D., Sack, R. B., & Farr, B. M., et al.(2004). Influence of humans leukocyte antigen class II alleles onsusceptibility to Entamoeba histolytica infection in Bangladeshichildren. Journal of Infectious Diseases, 189, 520−526.

Freeman-Gallant, C. R., Meguerdichian, M., Wheelwright, N. T., &Sollecito, S. V. (2003). Social pairing and female mating fidelitypredicted by restriction fragment length polymorphism similarity at themajor histocompatibility complex in a songbird. Molecular Ecology, 12(11), 3077−3083.

Froeschke, G., & Sommer, S. (2005). MHC class II DRB variability andparasite load in the striped mouse (Rhabdomys pumilio) in the southernKalahari. Molecular Biology & Evolution, 22(5), 1254−1259.

Gangestad, S. W., & Scheyd, G. J. (2005). The evolution of human physicalattractiveness. Annual Review of Anthropology, 34, 523−548.

Havlicek, J., & Roberts, S. C. (2009). MHC-correlated mate choice inhumans: A review. Psychoneuroendocrinology, 34(4), 497−512,doi:10.1016/j.psyneuen.2008.10.007.

Havlicek, J., Saxton, T., Roberts, S. C., Jozifkova, E., Lhota, S.,Valentova, J., & Flegr, J. (2008). He sees, she smells? Male andfemale reports of sensory reliance in mate choice and non-mate choicecontexts. Personality and Individual Differences, 45(6), 565−570.

Hedrick, P. W., & Black, F. L. (1997). HLA and mate selection: No evidence inSouth Amerindians. American Journal of Human Genetics, 61(3),505−511.

Herz, R. S., & Inzlicht, M. (2002). Sex differences in response to physicaland social factors involved in human mate selection—The importance ofsmell for women. Evolution & Human Behavior, 23, 359−364.

Hill, A. V., Allsopp, C. E., Kwiatkowski, D., Anstey, N. M., Twumasi, P., &Rowe, P. A., et al. (1991). Common West African HLA antigens areassociated with protection from severe malaria. Nature, 352(6336),595−600.

Hinsz, V. B. (1989). Facial resemblance in engaged and married couples.Journal of Social and Personal Relationships, 6, 223−229.

Hoffman, J. I., Forcada, J., Trathan, P. N., & Amos, W. (2007). Female furseals show active choice for males that are heterozygous and unrelated.Nature, 445(7130), 912−914.

Höglund, J., Piertney, S. B., Alatalo, R. V., Lindell, J., Lundberg, A., &Rintamaki, P. T. (2002). Inbreeding depression and male fitness in blackgrouse. Proceedings of the Royal Society of London B BiologicalSciences, 269(1492), 711−715.

Hraber, P., Kuiken, C., & Yusim, K. (2007). Evidence for human leukocyteantigen heterozygote advantage against hepatitis C virus infection. He-patology, 46(6), 1713−1721.

Ihara, Y., Aoki, K., Tokunaga, K., Takahashi, K., & Juji, T. (2000). HLA andhuman mate choice: Tests on Japanese couples. AnthropologicalScience, 108, 199−214.

Jacob, S., McClintock, M. K., Zelano, B., & Ober, C. (2002). Paternallyinherited HLA alleles are associated with women's choice of male odor.Nature Genetics, 30(2), 175−179.

Jin, K., Speed, T. P., & Thompson, G. (1995). Test of randommating for highlypolymorphic locus—application to HLAdata.Biometrics, 51, 1064−1076.

57H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

Kempenaers, B. (2007). Mate choice and genetic quality: A review ofthe heterozygosity theory.Advances in the Study of Behavior, 37, 189−278.

Kirkpatrick, M., & Ryan, M. J. (1991). The evolution of mating preferencesand the paradox of the lek. Nature, 350, 33−38.

Knapp, L. A., Ha, J. C., & Sackett, G. P. (1996). Parental MHC antigensharing and pregnancy wastage in captive pigtailed macaques. Journalof Reproductive Immunology, 32(1), 73−88.

Kretzmann, M., Mentzer, L., DiGiovanni, R., Leslie, M. S., & Amato, G.(2006). Microsatellite diversity and fitness in stranded juvenile harpseals (Phoca groenlandica). Journal of Heredity, 97(6), 555−560,doi:10.1093/jhered/esl043.

Landry, C., Garant, D., Duchesne, P., & Bernatchez, L. (2001). ‘Good genesas heterozygosity’: The major histocompatibility complex and matechoice in Atlantic salmon (Salmo salar). Proceedings of the RoyalSociety of London B Biological Sciences, 268, 1279−1285.

Lechler, R., Warrens A., (Eds.), HLA in health and disease. 2nd ed. London:Academic Press.

Lie, H. C., Rhodes, G., & Simmons, L. (2008). Genetic diversity revealed inhuman faces. Evolution, 62(10), 2473−2486, doi:10.1111/j.1558-5646.2008.00478.x.

Little, A. C., Burt, D. M., & Perrett, D. I. (2006). Assortative mating forperceived facial personality traits. Personality and Individual Differ-ences, 40, 973−984.

Little, A. C., Jones, B. C., Penton-Voak, I. S., Burt, D. M., & Perrett, D. I.(2002). Partnership status and the temporal context of relationshipsinfluence human female preferences for sexual dimorphism in male faceshape. Proceedings of the Royal Society of London B BiologicalSciences, 269(1496), 1095−1100.

Malkki, M., Single, R., Carrington, M., Thomson, G., & Petersdorf, E.(2005). MHC microsatellite diversity and linkage disequilibrium amongcommon HLA-A, HLA-13, DR131 haplotypes: Implications forunrelated donor hematopoietic transplantation and disease associationstudies. Tissue Antigens, 66(2), 114−124.

McClelland, E. E., Penn, D. J., & Potts, W. K. (2003). Majorhistocompatibility complex heterozygote superiority during coinfection.Infection and Immunity, 71(4), 2079−2086.

Meyer-Lucht, Y., & Sommer, S. (2005). MHC diversity and the associationto nematode parasitism in the yellow-necked mouse (Apodemusflavicollis). Molecular Ecology, 14(7), 2233−2243.

Milinski, M. (2003). The function of mate choice in sticklebacks:Optimizing MHC genetics. Journal of Fish Biology, 63(Supp A), 1−16.

Milinski, M. (2006). The major histocompatibility complex (MHC), sexualselection and mate choice. Annual Review of Ecology, Evolution andSystematics, 37, 159−186.

Mitton, J. B., Schuster, W. S. F., Cothran, E. G., & De Fries, J. C. (1993).Correlation between the individual heterozygosity of parents and theiroffspring. Heredity, 71, 59−63.

Nordlander, C., Hammarström, L., Lindblom, B., & Edvard Smith, C. I. E.(1983). No role of HLA in mate selection. Immunogenetics, 18, 429−431.

Ober, C. (1999). Studies of HLA, fertility and mate choice in a humanisolate. Human Reproduction Update, 5, 103−107.

Ober, C., Elias, S., Kostyu, D., & Hauck, W. W. (1992). Decreasedfecundability in Hutterite couples sharing HLA-DR. American Journalof Human Genetics, 50, 6−14.

Ober, C., Hyslop, T., Elias, S., Weitkamp, L. R., & Hauck, W. W. (1998).Human leukocyte antigen matching and fetal loss: Results of a 10 yearprospective study. Human Reproduction, 13(1), 33−38.

Ober, C., Hyslop, T., & Hauck, W. W. (1999). Inbreeding effects on fertilityin humans: Evidence for reproductive compensation. American Journalof Human Genetics, 64, 225−231.

Ober, C., Weitkamp, L. R., Cox, N., Dytch, H., Kostyu, D., & Elias, S.(1997). HLA and mate choice in humans. American Journal of HumanGenetics, 61, 497−504.

Oliver, M. K., Telfer, S., & Piertney, S. B. (2009). Major histocompatibilitycomplex (MHC) heterozygote superiority to natural multi-parasiteinfections in the water vole (Arvicola terrestris). Proceedings of theRoyal Society of London B Biological Sciences, 276, 1119−1128.

Olsson, M., Madsen, T., Nordby, J., Wapstra, E., Ujvari, B., & Wittsell, H.(2003). Major histocompatibility complex and mate choice in sandlizards. Proceedings of the Royal Society of London B BiologicalSciences, 270(Suppl 2), S254−S256, doi:10.1098/rsbl.2003.0079.

Penn, D. (2002). The scent of genetic compatibility: Sexual selection and themajor histocompatibility complex. Ethology, 108, 1−21.

Penn, D., & Potts, W. (1998). MHC-disassortative mating preferencesreversed by cross-fostering. Proceedings of the Royal Society of LondonB Biological Sciences, 265(1403), 1299−1306.

Penn, D., & Potts, W. (1999). The evolution of mating preferences and majorhistocompatibility complex genes. American Naturalist, 156.

Piertney, S. B., & Oliver, M. B. (2006). The evolutionary ecology of themajor histocompatibility complex. Heredity, 96, 7−21.

Platek, S. M., Raines, D. M., Gallup, G. G., Mohamed, F. B., Thomson,J. W., & Myers, E. M., et al. (2004). Reactions to children's faces:Males are more affected by resemblance than females are, and so aretheir brains. Evolution and Human Behavior, 25, 394−405.

Potts, W. K., Manning, C. J., & Wakeland, E. K. (1991). Mating patterns inseminatural populations of mice influenced by MHC genotype. Nature,352, 319−621.

Reusch, T. B. H., Häberli, J. C., Aeschlimann, P. B., & Milinski, M. (2001).Female sticklebacks count alleles in a strategy of sexual selectionexplaining MHC polymorphism. Nature, 414, 300−302.

Roberts, S. C., Gosling, L. M., Carter, V., & Petrie, M. (2008). MHC-correlated odour preferences in humans and the use of oral contra-ceptives. Proceedings of the Royal Society of London B BiologicalSciences, 275, 2715−2722, doi:10.1098/rspb.2008.0825.

Roberts, S. C., Hale, M. L., & Petrie, M. (2006). Correlations betweenheterozygosity and measures of genetic similarity: Implications for under-standing mate choice. Journal of Evolutionary Biology, 19(2), 558−569.

Roberts, S. C., & Little, A. C. (2007). Good genes, complementary genesand human mate preferences. Genetica, 132(3), 309−321.

Roberts, S. C., Little, A. C., Gosling, L. M., Jones, B. C., Perrett, D. I., &Carter, V., et al. (2005). MHC-assortative facial preferences in humans.Biology Letters, 1(4), 400−403.

Roberts, S. C., Little, A. C., Gosling, L. M., Perrett, D. I., Carter, V., &Jones, B. C., et al. (2005). MHC-heterozygosity and human facialattractiveness. Evolution and Human Behavior, 26(3), 213−226.

Rosenberg, L. T., Cooperman, D., & Payne, R. (1983). HLA and mateselection. Immunogenetics, 17, 89−93.

Santos, P. S. C., Schinemann, J. A., Gabardo, J., & da Graca Bicalho, M.(2005). New evidence that the MHC influences odor perception inhumans: a study with 58 Southern Brazilian students. Hormones &Behavior, 47(4), 384−388.

Sauermann, U. U., Nürnberg, P. P., Bercovitch, F. F. B., Berard, J. J. D.,Trefilov, A. A., & Widdig, A. A., et al. (2001). Increased reproductivesuccess of MHC class II heterozygous males among free-ranging rhesusmacaques. Human Genetics, 108(3), 249−254.

Schwensow, N., Eberle, M., & Sommer, S. (2008). Compatibility counts:MHC-associated mate choice in a wild promiscuous primate. Proceed-ings of the Royal Society of London B Biological Sciences, 275,555−564, doi:10.1098/rspb.2007.1433.

Schwensow, N., Fietz, J., Dausmann, K. H., & Sommer, S. (2007). Neutralversus adaptive genetic variation in parasite resistance: importance ofmajor histocompatibility complex supertypes in a free-ranging primate.Heredity, 99(3), 265−277.

Schwensow, N., Fietz, J., Dausmann, K. H., & Sommer, S. (2008). MHC-associated mating strategies and the importance of overall genetic diversityin an obligate pair-living primate. Evolutionary Ecology, 22, 617−636.

Sefcek, J. A., Brumsbach, B. H., Vasquez, G., & Miller, G. F. (2006). Theevolutionary psychology of human mate choice: How ecology, genes,fertility, and fashion influence mating behavior. Journal of Psychologyand Human Sexuality, 18(2/3), 125−182.

Slate, J., David, P., Dodds, K. G., Veenvliet, B. A., Glass, B. C., &Broad, T. E., et al. (2004). Understanding the relationship between theinbreeding coefficient and multilocus heterozygosity: theoreticalexpectations and empirical data. Heredity, 93(3), 255−265.

58 H.C. Lie et al. / Evolution and Human Behavior 31 (2010) 48–58

Stewart, S., Stinnett, H., & Rosenfeld, L. B. (2000). Sex differences indesired characteristics of short-term and long-term relationship partners.Journal of Social and Personal Relationships, 17, 843−853.

Thornhill, R., Gangestad, S. W., Miller, R. D., Scheyd, G. J.,McCollough, J. K., & Franklin, M. (2003). Major histocompatibilitycomplex genes, symmetry, and body scent attractiveness in men andwomen. Behavioral Ecology, 14(5), 668−678.

Thursz, M. R., Thomas, H. C., Greenwood, B. M., & Hill, A. V. (1997).Heterozygote advantage for HLA class-II type in hepatitis B virusinfection. Nature Genetics, 17(1), 11−12.

Tregenza, T., & Wedell, N. (2000). Genetic compatibility, mate choice andpatterns of parentage: Invited review. Molecular Ecology, 9(8),1013−1027.

Trivers, R. (1972). Parental investment and sexual selection. In:B. Campbell (Ed.), Sexual selection and the descent of man(pp. 136–179). Aldine-Atherton: Chicago.

Tsitrone, A., Rousset, F., & David, P. (2001). Heterosis, marker mutationalprocesses and population inbreeding history. Genetics, 159(4),1845−1859.

Wedekind, C., & Furi, S. (1997). Body odour preferences in men andwomen: Do they aim for specific MHC combinations or simplyheterozygosity? Proceedings of the Royal Society of London BBiological Sciences, 264, 1471−1479.

Wedekind, C., Seebeck, T., Bettens, F., & Paepke, A. J. (1995). MHC-dependant mate preferences in humans. Proceedings of the RoyalSociety of London B Biological Sciences, 260(1359), 245−249.

Yamazaki, K., Boyse, E. A., Mike, V., Thaler, H. T., Mathieson, B. J., &Abbott, J., et al. (1976). Control of mating preferences in mice by genesin the major histocompatibility complex. Journal of ExperimentalMedicine, 144, 1324−1335, doi:10.1084/jem.144.5.1324.

Ziegler, A., Kentenich, H., & Uchanska-Ziegler, B. (2005). Female choiceand the MHC. TRENDS in Immunology, 26(9), 496−502.