Sexual Selection, Physical Attractiveness, and Facial Neoteny: Cross-Cultural Evidence and Implications [And Comments and Reply] Author(S): Doug Jones, C

Home , Evolution (journal), Facial (sexual act), Fitness (biology), Human tooth, Modesty, Neoteny

Sexual Selection, Physical Attractiveness, and Facial Neoteny: Cross-cultural Evidence and Implications [and Comments and Reply] Author(s): Doug Jones, C. Loring Brace, William Jankowiak, Kevin N. Laland, Lisa E. Musselman, Judith H. Langlois, Lori A. Roggman, Daniel Pérusse, Barbara Schweder, Donald Symons Reviewed work(s): Source: Current Anthropology, Vol. 36, No. 5 (Dec., 1995), pp. 723-748 Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological Research Stable URL: http://www.jstor.org/stable/2744016 . Accessed: 09/12/2011 18:16

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaborating with JSTOR to digitize, preserve and extend access to Current Anthropology.

http://www.jstor.org CURRENT ANTHROPOLOGY Volume 36, Number 5, December I995 ? I995 byThe Wenner-GrenFoundation for Anthropological Research. All rightsreserved ooII-3204/95/36o5-0004$3.oo

1JZVO Wl VX V.LIs X11 \s1sX L.... %J1%S XS 11%11a W1l a%.1lAl At4 1Vb13 1 Five Populations"(Human Nature 4:27i-96) and authorof the forthcomingPhysical Attractiveness and the Theoryof Sexual Se- Sexual Selection, lection:Results from Five Populations(Ann Arbor: University of MichiganMuseum ofAnthropology, in press).The presentpaper was submittedI4 XI 94 and accepted2o I 95; the finalversion Physical reachedthe Editor'soffice 24 II 95. Attractiveness, The firstpublication by Darwin to discuss human evolu- and Facial Neoteny tion at length bears the double title "The Descent of Man, and Selection in Relation to Sex" (i98I[i87i]) and consists of two works back to back. The firstwork dis- cusses human evolution and argues "that man is the Cross-culturalEvidence and modifieddescendant of some pre-existingform" (p. 9). The second presents the topic of sexual selection-a Implications formof natural selection resultingfrom "the advantage which certainindividuals have over otherindividuals of the same sex and species, in exclusive relationto repro- by Doug Jones duction" (p. 256). Darwin yoked human evolution and sexual selection togetherin a single volume because he believed that sexual selection had played a major role both in the descent of humans fromearlier forms and in the differentiationof human races. Physicalattractiveness and its relationto the theoryof sexual se- In the half-centuryafter Darwin publishedthis work, lectiondeserve renewed attention from cultural and biologicalan- thropologists.This paperfocuses on an anomalyassociated with a numberof authorstook up the topic ofphysical attrac- physicalattractiveness-in our species,in contrastto manyoth- tiveness across cultures. Many, such as Westermarck ers,males seem to be moreconcemed than females with the at- (i92i) and Ellis (i926), followed Darwin's lead in trying tractivenessof potential sexual partners,perhaps because hu- to relatethe developmentof standards of physical attrac- mans show farmore age-related variance in femalethan in male fecundity.The resultingselection for male attractionto markers tiveness in humans to the theoryof sexual selection. offemale youth may lead incidentallyto attractionto females Subsequently, however, the social sciences grew in- displayingage-related cues in an exaggeratedform. This paper creasinglydivorced from evolutionary theory and from reportscross-cultural evidence that males in fivepopulations the study of physical variation (Degler i99i). Almost 2o U.S. Americans, and (Brazilians, Russians,Ache, Hiwi) show an years ago, Berscheid and Walster (I974:i58) summarized attractionto femaleswith neotenous facial proportions (a combinationof large eyes, small noses,and fulllips) evenafter female the consequences of this divorcefor the studyof physi- age is controlledfor. Two furtherstudies show thatfemale mod- cal attractiveness:"Most social scientistshave shown a els have neotenouscephalofacial proportions relative to U.S. un- studied professional disinterest in . . . how our physical dergraduatesand thatdrawings of faces artificially transformed appearanceinfluence[s] our relationshipwith others." to make themmore or less neotenousare perceivedas corre- spondinglymore or less attractive.These resultssuggest several More recently,there has been a revival of interestin furtherlines ofinvestigation, including the relationshipbetween the topic of attractivenessand an explosion of social facialand bodilycues and the consequencesof attraction to neo- psychological research on the subject, demonstrating tenyfor morphological evolution. significantagreement across ratersin judgmentsof attractivenessand significantsocial consequences of at- DOUG JONES is VisitingScholar in Anthropologyat CornellUni- tractiveness.This literaturehas been surveyedat book versity(Ithaca, N.Y. I4853, U.S.A. [[email protected]]).Born in length by Patzer (i985), Hatfield and Sprecher(i986), i959, he was educated at Princeton University (B.A., i98i) and the University of Michigan (M.A., i989; Ph.D., I994). He has Bull and Rumsey (i988), and Jackson (i992). There is been conductingresearch on standardsof physical attractiveness also considerableethnographic material on standardsof in the UnitedStates, Paraguay, Brazil, and Russia since i989. He attractivenessin non-Westernsocieties, for example, Malinowski (i96i[i929]), Berndt (i95i), Weiner (I976), Gregor (i985), Boone (i986), Euba (i986), Munn (i986), i. I thankKim Hill ofthe University of New Mexicofor assistance Grinker (i990), and Jankowiak (I993). However, existing in collectingsome of the data reportedin this paperand Davic researchstill suffersfrom several limitations.First, the Buss,Conrad Kottak, and JohnMitani of the University of Michi. work of social is and de- gan,Carlos Alberto Caroso and Maria Hilda Paraisoof the Federa] psychologists heavilyempirical Universityof Bahia (Brazil),and BjarneForsterwald of the Puerta scriptive,with little in the way of theorythat would BarraMission (Paraguay) for help with different phases of the study explain why people findparticular features attractive or This researchwas supportedby NSF Doctoral DissertationRe. even why theyexperience physical attraction at all. Sec- searchImprovement Grant BNS-goo6394 and by grantsfrom the ond, culturalanthropologists have rarelymade research Universityof Michigan's Department of Anthropology and Evolu. tionand HumanBehavior Program. The Departmentof Anthropol. on standardsof attractivenessand theirconsequences a ogyat CornellUniversity provided access to libraryand computeJ major objectiveof fieldwork; the ethnographicliterature facilities. recordsfew if any attemptsto quantifyagreement be-

723 724 CURRENT ANTHROPOLOGY Volume 36, Number S, December I995 tween individuals about standardsof attractiveness,to the "new physical anthropology"of the I940S and 'sos, assess the social and life-historyconsequences of being and the neglect of sexual selection in the earlymodern perceived as attractiveor unattractive,or to compare synthesiswas inheritedby biological anthropology.Un- standards of attractivenessacross societies. Thus the til recently,relatively few serious quantitativestudies studyof attractivenessin the social sciences is underde- in physical anthropologypaid attentionto the possible veloped in importantrespects-undertheorized in psy- role of sexual selection in human evolution. Hulse's chology and both undertheorizedand underresearched (i967) studyof skin-colorvariation in modernJapan sug- in culturalanthropology. gested that sexual selection mightinfluence the evolu- In biological anthropologytoo, the topic of physical tion of this trait,and a numberof studies of assortative attractivenessand its possible evolutionarycauses and matingfor physical traits(Spuhler i968) suggestedhow consequences has been relativelyneglected. This neglect mating patternsmight influencegenotype frequencies. is partof a widerneglect of the theoryof sexual selection However, Hulse's work inspiredlittle comment or fol- in the field of evolutionarybiology between the I930S low-up, and the literatureon assortative mating was and the I970S-a period that West-Eberhard(i983:i56) largelysilent about the causes of the patternsobserved. calls "the ForgottenEra" of sexual selection theory(see Thus in spite of Darwin's argumentthat sexual selec- also Cronin i99i). In the I930S and '40s, evolutionary tion played a central role in human evolution,serious biologists formulatedthe "modern synthesis"-a syn- quantitativestudies of adaptationin physicalanthropol- thesis of Darwin's theoryof evolution and the new sci- ogy have focused overwhelminglyon adaptationto the ence of genetics. The pioneers of the modernsynthesis physical environment.Only recently,with the rise of had theirhands full investigatingadaptation to ecologi- human behavioral ecology,have anthropologistsbegun cal constraints.They were less concernedwith the evo- tryingto bringthe moderntheory of sexual selection to lution of social behaviorand had little use forDarwin's the studyof human evolution (Chagnonand Irons I979, theoryof sexual selection. Betzig,Turke, and BorgerhoffMulder i988). But to date, Why did the theoryof sexual selection take so long most such studies have focusedon sexual selection and to win acceptance in evolutionarybiology? Sexual be- mate choice more in relation to behavior than in rela- havior and social behaviorin generalpose special prob- tion to morphology. lems for evolutionarytheory. These problems result In conclusion, the studyof physical attractivenessin fromthe fact that in social evolution the fitnessof a the contextof the theoryof sexual selection,which was trait commonly depends on its frequency(Maynard givena centralplace by Darwin, Westermarck,and Ellis, Smith i982). How well a particularshape of tail serves has not been a major topic foranthropological inquiry a given bird in flightwill generallynot depend on the in the last half-century.It is not that investigationhas frequencyin the population of tails of various shapes- shown attractivenessto be ill-definedor inconsequen- thatis, its fitnessis frequency-independent.By contrast, tial; both social psychological studies and the ethno- the attractivenessof that tail to membersof the other graphicliterature suggest at least moderate agreement sex will depend on the preferencesof the other sex, in judgments of attractivenesswithin cultures and at which will oftendepend in turn on the frequenciesof least a moderatelyimportant role for attractivenessin various shapes of tail-that is, its fitnessis frequency- social and especially sexual interactions.Rather, partic- dependent. ular theoreticalpresuppositions have kept both cultural Frequency-dependentselection presentsa numberof and biological anthropologistsfrom giving these topics seeming paradoxes. Under frequency-dependentselec- theirfull attention.Cultural anthropologistshave been tion there is no guarantee that natural selection will reluctantto deal with the more "biological" side of hu- favorgenetic variants that maximize mean population man behavior-reluctant to considerhuman behavioras fitness.Instead, populations may attain an evolutionary the productnot just ofstrivings for individual gains such equilibriumin which no individualcan gain by adopting as material comfortand social status but of adaptations a differentstrategy, even thoughall would be betteroff forgenetic reproduction. Biological anthropologistshave if all acted differently.Or they may enter an endless oftenheld a view of adaptation that could not readily "arms race" in which each triesto get ahead of the oth- accommodate frequency-dependentselection, including ers withoutanyone's enjoyingany long-termgains. Fre- the possibility that one individual's reproductivesuc- quency-dependentselection may favor traits that in- cess mightcome at the expense of others. crease the reproductivesuccess ofindividuals but reduce The theoryof sexual selection has advanced so farin the viabilityof groupsand lower the productivityof eco- recentyears that it may be time forrenewed attention systems.It may favorthe evolutionof waste and extrav- to the relationshipbetween sexual selection and stan- agance, ratherthan efficiency,in sexual and other sig- dards of physical attractivenessin our species. The nals. It may result in coevolutionarypositive feedback theoryof sexual selection does not imply that people cycles that amplifyarbitrary traits. It was only begin- always maximize inclusive fitnessin choosinga mate- ningin the I96os,with the developmentof sociobiology adaptations are often imperfect,and in humans as in and evolutionarygame theory,that the problemsassoci- otheranimals estheticstandards may be partlythe prod- ated with frequency-dependentselection were addressed uct of nonadaptive "sensorybiases." Nor does it imply in a sustained fashion. that standardsof beauty are completely"hard-wired"; The modernsynthesis of the I930S and '40s inspired in humans as in otheranimals the developmentof stan- JONES Sexual Selection and Facial Neoteny | 725 dards of attractivenessis likely to involve a range of mammals (Daly and Wilson i983:279-3i2). Human mechanisms, from innate templates to imprintingto males are largerthan females.Human males attain sex- imitationand otherforms of social learning.It does im- ual maturityat a later age than human females and se- ply,however, that human beings,like otheranimals, are nesce more rapidly.In all societies with appreciablelev- likely to have genetic adaptations for assessing the els of violent conflict,male-male aggressionis more "mate value" of potential mates and that studying common thanfemale-female aggression. Violent compe- attractivenesswithout considering these adaptations tition is more common among human males than would be like tryingto understandthe eye without among females,and male sexual coercion of females is treatingit as an organof vision. far more common than the reverse.Polygyny is much more common than polyandry.And in most respects, human females seem to be more selective than human Sexual Selection and Physical Attractiveness: males in theirchoice of sexual partners. A Human Anomaly In one respect, however, human beings reverse the usual patternof differencesbetween more and less sex- Sexual selection occurs when some organismsgain an ually selected sexes-men are more concerned than edge in matingand fertilizationat the expense of others women with the physical attractivenessof a potential of theirsex. Sexual selection may take the formof con- sexual partner.This sex differencehas been found re- tests within one sex over chances to mate with and fer- peatedly in studies by social psychologists.A recent tilize membersof the othersex or courtshipof one sex meta-analysis(Feingold I 990) ofthe social psychological by the other.2 literatureon sex differencesin effectsof physical attrac- In many animal species, male reproductivesuccess is tiveness on romantic attractionshows consistent and more dependenton matingsuccess than is femalerepro- strong sex differences,including content analyses of ductivesuccess, so sexual selectioncommonly acts with "lonely-heartsadvertisements," studies of attractive- greater intensity on males than on females (Trivers ness and reporteddating success, reportsof interper- I97I, WilliamsI975, Clutton-Brockand Parkeri992, sonal attractionfollowing dyadic interaction,and sur- AnderssonI994). The resultis thatin manyspecies, veysof characteristics desired in a mate (see also Jackson males more than females show a syndromeof traitsas- i992:65-67). sociated with intense sexual selection. This "sexual se- This sex differenceis not limited to Westernsociety. lection syndrome"includes behavioraltraits: males are Buss (i 989) reviewssurvey data from37 populationsam- more likely than females to resortto violence against ples from33 countries and findsthat in every sample sexual rivals and to forcecopulations on resistingpart- males are more concernedthan femaleswith the physi- ners; males cpmmonly expend more time and energy cal attractivenessof a potential mate. The average sex and take greaterrisks than females in courtship;males differenceis more pronouncedamong the non-Western will generallycourt and attemptcopulation with a wider populations in his sample. While Buss's studyincludes rangeof partners then will females.The sexual selection no tribalpopulations, the same patternapparently holds syndromealso includes life-historytraits: males com- forthese as well. Fordand Beach (I 95 I:94), summarizing monly take longerthan females to attain sexual matu- evidence from nearly 2oo cultures, conclude that al- ritybecause of the sexual competitionthat they face though there is a great deal of cross-culturalvariation frommature males; males commonlyhave highermor- in standards of attractiveness,"in most societies the talityrates than females as a resultof intrasexualcom- physicalbeauty of the femalereceives more explicit con- petition; males commonly senesce more rapidlythan siderationthan does the handsomenessof the male. The femalesbecause highermortality rates reduce the selec- attractivenessof the man usually depends predomi- tion pressurefor longevity. Finally, the sexual selection nantlyupon his skills and prowessrather than upon his syndromeincludes morphologicaltraits: males are more physical appearance." Gregersen(i983) reportssimilar likely than females to display anatomical specializa- findingsin a more recentreview of nearly300 societies, tions for intra- and intersexual aggression,including mostlynon-Western and nonurbanized. horns,antlers, enlarged canine teeth,and body sizes in In otherwords, human beingsseem to be an exception excess of the ecological optimum; males commonly to the general rule among animals that male attrac- show greater development of sexual advertisements, tiveness mattersmore than female attractiveness.The both tactile (complexgenitalia) and visual (elaborateand importance attached to female (as opposed to male) brightlycolored adornments). physical attractivenessin our species stands in need of Among humans, considerableanatomical and behav- an explanation. ioral evidence suggeststhat males have been subject to strongersexual selection than females,although the dif- An Anomaly Explained? Fecundity,Age, --- 1r+lfe , nn"r%h"er han"r and Neoteny

Many anthropologistsbelieve that human behavior is 2. Smuts and Smuts (I993) argue that it is useful to distinguish a thirdform of sexual selection,sexual coercionof one sex by the so radicallydifferent in its ontogenyfrom that of other other. organismsthat the theoryof sexual selection is not ap- 726 CURRENT ANTHROPOLOGY Volume 36, Number S, December I995

plicable to human physical attraction. Polhemus How can we apply this principleto the anomaly of (i988:8) probably expresses the attitude of a whole female attractivenessin our species? Let us definethe school of anthropologyof "the body" concerningthe hu- mate value of a potential sexual partner,A, as the ex- man irrelevanceof the theoryof sexual selection when pectedreproductive success frommating with A divided he writes: by some baseline expected reproductivesuccess. The baseline expectedreproductive success mightbe the ex- A male baboon has a fixedidea of what a desirable pected reproductivesuccess frommating at random or female baboon should look like.... The same gen- from mating with an individual of maximum mate eral principleis true of any animal that reproduces value. As a general rule we expect that human beings, by sexual selection. But thereis an importantdiffer- and otheranimals, are likelyto have bothrelatively can- ence between baboons and ourselves.For otherani- alized, "hard-wired"responses to visual stimuli that mals the physical ideal is ioo% instinctivelyde- have been consistentlyassociated with high mate value termined.Thus all baboons of a particularspecies throughoutthe evolutionaryhistory of the species and pursue the same ideal.... For humans, on the other relativelyflexible learned responses to stimulithat have hand, ideals of beauty are learned.... In a world- been associated sometimes with high mate value and wide and historicalframework, there is no such sometimeswith low. In otherwords, standards of physi- thingas naturalhuman beauty. cal attractivenessare likely to have both species-typical Ifthis view ofthe differencebetween human and nonhu- and population-specificcomponents, and variation in man psychologywere correct,the anomaly of female these componentsmay be predictablegiven knowledge attractivenessin our species mightbe merelyone more of human biology and local circumstances (Symons consequence of our having freedourselves fromthe in- I979). For example, since fat stores may be selectively stinctiveconstraints that hobble the lives of otherani- advantageousin environmentssubject to episodic food mals. This view, however,is doublywrong. shortageand disadvantageousin environmentsrequiring First,learning often plays a large role in the acquisi- considerablephysical movement, one mightexpect that tion of standardsof attractivenessamong nonhuman an- estheticresponses to fatnesswould varybetween popu- imals. An immense literaturedemonstrates that early lations dependingon social learningand on individual experienceinfluences later mate choice via imprinting assessments of the consequences of being fat or thin, (Immelman I972). Imitation,too, plays a role in mate ratherthan developingin a uniformfashion within the choice among nonhuman animals, and social transmis- human species. sion of matingpreferences can even result in "fads" in By contrast,one mightexpect human beings to have mate choice that change fromone breedingseason to a relatively invariant, species-typical emotional re- the next (Pruett-Jonesi992). sponse to signs of aging, because age has a relatively Second, physical attractionin humans cannot be en- invariantassociation with fecundityand thuswith mate tirelya product of enculturation.This is shown most value. In a classic article Henry (i96i) reviews data on dramaticallyby the experimentsof Langlois et al. (i987). age-specificfertility rates in a wide range of "natural- In these experiments,infants between the ages of two fertility"(noncontracepting) populations. The levels of and three months were exposed to pictures of women fertilityin these populationsrange from a lifetimeaver- ratedattractive and unattractiveby adult raters;infants age of 6 to i i children per married female, but the spent more time looking at faces rated attractive.This shapes of the curves of fertilityversus age are remark- held even across racial/culturalboundaries: for Euro- ably similar across all populations. For all populations, pean-American infants looking at faces of African- female fertilityrates at age 30-34 are around 85% of American women rated by African-Americanmen and ratesat age 20-24, with furtherdeclines to around35 % for African-Americaninfants exposed to European- forwomen aged 40-44 and o% forwomen aged 50-54. Americanfaces rated by European-Americanmen. More recentwork suggeststhat the curve of naturalfe- Thus students of physical attractivenessare asking cundity(potential reproduction) differs somewhat from for trouble if they start out assuming that nonhuman the curve of natural fertility(actual reproduction)be- animals are creaturesof instinctand humans construc- cause the latter is influencedby such variables as age tions of culture. A better startingpoint regardingthe of spouse and frequencyof intercourse(James I979, role of learningin behavior is suggestedby several de- Menken,Trussell, and Larsen i986). Studiesthat control cades of researchin comparativepsychology: as a gen- [orthe lattervariables suggest that the decline in female eral rule,organisms have relatively"hard-wired" or can- Fecunditybetween 2o and 35 is less pronouncedthan alized responses to stimuli that have had relatively the decline in female natural fertility-butthe overall unvaryingfitness consequences over evolutionarytime shapes of the two curves are fairlysimilar. and relativelyflexible learned responsesto stimuli that The shape of the curve of fecundityversus age is very have been associated sometimes with positive fitness differentfor males. Goldman and Montgomery(i989), consequences and sometimes with negative. In other reviewingdata fromseveral traditionalsocieties, report words,given thatlearning entails costs,in termsof trial Eertilitydeclines to about 90% formen between 45 and and error,organisms are expectedto adapt to selectively 50, relativeto youngermen, and to about 8o% formen importantinvariants in theirenvironments with corre- Dver55, after controlling for age of wife and durationof spondingbehavioral, cognitive, or motivationalinvari- marriage. ances (Seligman I970, Johnston 1982). Fecundityversus age curves thus have two important JONES Sexual Selection and Facial Neoteny 1727

characteristicsthat may help to explain the anomaly of intermediatein juvenility.(It was judged to be 23 years female attractiveness:the curves (i) are relativelyin- old.) In other words, the level of juvenilitythat maxi- variantin shape across populationsand (2) show an ear- mizes perceived vulnerabilitydoes not maximize per- lier and more pronounceddecline in fertilityamong fe- ceived sexiness. Kenrickand co-workers(Kenrick I994) males than among males. Given the general rule that show that forteenage males the ideal sexual partneris organismscommonly have invariantresponses to stim- older than they are-again, more consistentwith the uli that have had relatively invariant fitness conse- hypothesisthat males are concernedwith cues to female quences over evolutionarytime, the firstcharacteristic fecunditythan with the hypothesisthat males prefer suggeststhat human beings are likely to have relatively younger,more easily dominatedfemales. Thus current invariantesthetic responses to signs of aging. The sec- evidence suggests that female attractivenesscannot ond characteristicsuggests that these responses are simply be equated with powerlessness and that some- likely to be strongerin males' evaluations of females thing more than changes in perceived vulnerabilityis than in females' evaluations of males. involved in age-related changes in physical attrac- This does not add up to a completetheory of physical tiveness.However, nothing in evolutionarytheory rules attractiveness,of course, or even a complete theoryof out the possibilitythat markersof female submissive- age-relatedchanges in physicalattractiveness. Fecundity ness may be attractiveto men, and the topic certainly is onlyone componentof mate value. Othercomponents deservesmore research. include the abilityand willingnessto provisionoffspring There may be room forargument about why attrac- and heritableviability or attractiveness("good genes"), tivenesschanges with age, but,in spite of a considerable and these componentsof mate value may also varywith literaturedevoted to the claim thathuman sexualityand age, while sensorybias will ensure that attractiveness standardsof physical attractivenessare culturallycon- does not trackmate value perfectly.Nevertheless, age- structed,there does not seem to be any evidence from related changes in fecundityare likely to be a particu- any society that seriously challenges the proposition larly importantcomponent of age-relatedchanges in thatphysical attractiveness is perceivedto decline from physical attractiveness,especially in females,both be- youngadulthood to old age, especiallyfor females. "The cause these changes have been relativelyinvariant over correlationof female age and sexual attractivenessis so the historyof the species and because othercomponents intuitivelyobvious that ethnographersapparently take of mate value such as provisioningability and inclina- it forgranted-as they do the bipedalism of the people tion may be more readilyassessable on the basis of be- theystudy-and the significanceof female age tends to havior than on the basis of physical appearance. be mentioned only in passing, in discussions of some- There is one alternative explanation for male at- thingelse" (SymonsI979:i88). Symonscites passingref- traction to youthfulfeatures in females that requires erences to the effectsof aging on female attractiveness a more extended treatment. Gowaty (I992:23I-40) in ethnographiesof the Kgatla,pre-revolutionary China, writes: the Yanomamo, and the Tiwi. Additionalreferences can be foundin ethnographiesof TrobriandIslanders (Mali- There should be strongselection on males to control nowski I987 [i929], WeinerI976) and Gawa females' reproductionthrough direct coercive con- (Munn I986) ofMelanesia, Mende (Boone i986) of SierraLeone, trol of females.... Evolutionarythinkers, whether and Mehinaku of Amazonia (Gregori985), to name just informedby feministideas or not, are not surprised a few. A number of social psychological studies (re- by one of the overwhelmingfacts of patriarchalcul- viewed in Jackson i992) have documented such age- tures,namely that men ... seek to constrainand related declines in physical attractivenessand demon- controlthe reproductivecapacities of women.... Ju- stratedthe expected sex differencesas well. venilization decreases the threatsome men may feel Let us summarizethe argumentup to this point. Hu- when confrontedwith women; many men are com- man beings are anomalous among sexually selected fortablearound women whom theycan clearlydomi- spe- cies in the importanceattached to female (relativeto nate and are profoundlyuncomfortable around male) appearance in mate choice. Human beings are women whom they cannot so clearlydominate. The anomalous in another respect as well: female fertility hypothesisthat femininitysignals abilityto be domi- commonly declines to zero long beforethe end of the nated throughjuvenilization is an alternativeto, but life span. As a resultof menopause thereis considerably not necessarilymutually exclusive of,other evolu- more age-relatedvariance in fecundityamong adult fe- tionaryhypotheses that posit that femininitysig- males than among adult males in our species. The sec- nals, sometimes deceptively,reproductive value and ond anomaly may explain the first: the importance fertility. attachedto female attractivenessin our species may re- Severalfindings seem to be at odds with thishypothesis. flect the operation of adaptations for assessing age- Berryand McArthur(i986) presentedsubjects with a se- related changes in fecundity,a component of female ries of outline profiledrawings representing individuals mate value. Whetherfor this reason or another,social rangingfrom juvenile to adult and collected ratingsof psychologicaland ethnographicevidence providesover- perceived social characteristicsof each drawing. The whelmingsupport for the propositionthat human beings drawing rated weakest and least threateningwas the have relativelyinvariant esthetic responses to signs of most juvenile-looking.(Subjects judged this drawingto males'agingand evaluations that these of responses females thanoperate vicemore versa. strongly in representa 4-year-old.)The drawingrated sexiest was 728 1 CURRENT ANTHROPOLOGY Volume 36, Number 5, December I995

Thus far we have been exclusively concerned with changes in attractivenesswith age ratherthan differences in attractivenessbetween individualsof the same age. However,if age-detectingmechanisms do not oper- k=-O.1 ( ate with perfectaccuracy, then adaptationsfor choosing a mate of a particularage may lead incidentallyto non- adaptivebiases in the choice of mates fromamong individuals who fall within a particularage-class. In other words,given that attractiveness varies with age, individuals may be more or less attractivethan othersof the same age in part because they have facial proportions associated with youngeror olderages. Because the reten- tion of traitsfrom early stages of the lifecycle into later stages,relative to ancestorsor to othermembers of the population, is known as neoteny ("holding on to youth"),the propositionabove may be rephrased:given varies be a that attractiveness with age, neoteny may - -| k=0 4 component of facial attractiveness.This proposition may hold with particularforce for female facial attractiveness: a by-productof the human male's attraction to markersof youthfulfecundity may be an attraction to adult femalespresenting markers of youth to an exaggeratedor "supernormal"degree. Beginningwith the anomaly of female attractiveness in our species, we are led to the hypothesisthat neoteny may be a componentof female facial attractiveness.The remainderof this paperwill be given over to testingand k=O.1 elaboratingthis hypothesis.

Cephalofacial Age Markers

To considerhow age-relatedchanges in physicalattrac- FIG. i. The effectsof negative (top row) and positive tiveness might result fromdiffering esthetic responses (bottomrow) cardioidal transformationson a 5 x 5 to differingfacial proportions,it is necessaryto review grid and on -afemale face. how the sizes and shapes of facial featureschange with age. It is convenient to divide these changes into changes in hard tissue (bone) and changes in softtissue (cartilageand connectivetissue). tive cardioidaltransformation are perceivedas olderand Hard tissue. In human beings as in othermammals, less cute; pictures subjected to a negative transforma- the neurocranium-the portionof the skull housingthe tion are perceivedas youngerand cuter.Full facial and brainbut also includingthe contiguousorbital region- profiledrawings of the heads ofbirds, monkeys, and dogs grows rapidly early in development,while the facial and even frontand side drawingsof VolkswagenBeetles skeleton proper-including the nasal and masticatory can be made to appear more or less "mature" or "cute" complexes-attains its maximum rate of growthonly by subjecting them to positive or negative cardioidal- later (Enlow iggo). As a result,juvenile mammals pre- strain.Figure i illustratesthe effectsof positive and neg- sent a characteristic"cute" appearance,with relatively ative cardioidal strain on a square grid and on a face. large eyes, high foreheads,and reduced snouts. The transformedgrids were producedwith the Mathe- Changes in the shape ofthe craniofacialskeleton asso- matica a2.2 software package, while the transformed ciated with aging are closely approximatedby a simple faces were redrawnfrom the originalface with the assis- mathematical transformationcalled cardioidal strain tance of polar coordinategraph paper. (Mark, Shaw, and Pittengeri988). Cardioidal strain of Attractionto "cute" proportionsmay be unlearned: degreek maps each point {x, y} onto {x', y'} = {x (i - even at 4 months of age infantsorient preferentially to- k . y/r),y (i - k y/r)}where r2 = x2 + y2. A shape ward pictures of infant rather than adult faces- subject to positive cardioidal strain (k > o) shows a althoughit is not knownwhether infantile facial propor- downward and outward expansion in featureslocated tions per se are the relevantcue (McCall and Kennedy towardthe bottomand a downwardand inwardcontrac- i980). I have cited research(Langlois et al. i987) show- tion in featureslocated towardthe top. Negative cardi- ing that infantsas youngas 2 monthsof age orientpref- oidal strain presentsthe opposite changes. Pictures of erentiallytoward attractive rather than unattrattivefe- the faces of childrenor youngadults subjectedto a posi- male faces. If, as I will argue, female attractivenessis JONES Sexual Selection and Facial Neoteny | 7e29 partly a matter of cephalofacial neoteny, then infant eastern Paraguay and the Hiwi (or Cuiva) of southern preferencesfor attractivefemale faces may be part of Venezuela, as well as three Westernized societies- a more general attractionto faces or facelike stimuli Brazil,the United States, and Russia. Each of the West- manifestinglow cardioidal strain.3 ernized populations is the subject of a large body of Softtissue. Skeletal growthslows down (but does not ethnographicliterature; below I provide a very brief stop [Behrentsi985]) with the attainmentof adulthood. summary of some facts relevant to physical attrac- However,other changes in facialproportions result from tivenessin the two indigenousSouth Americanpopula- the growthof cartilage and the atrophyof connective tions. The Ache and the Hiwi were firstpeacefully con- tissue. These affectthe relativesizes ofeyes, noses, ears, tactedby outsidersin the I96os and I970S; up to that and lips. "Beginningat age 25, the eyebrows steadily time theyhad lived as hunter-gatherers.Members of the descend froma position well above the supraorbitalrim two populations know what outsiders look like, but to a point far below it; saggingof the lateral aspect of most have little contact with outsiderson a day-to-day the eyebrows make the eyes seem smaller" (Larrabee basis. Both groups maintain a stronglyethnocentric and Makielski I993:I4). Cartilaginoustissues grow standardof physical attractiveness.For the Ache, Kim steadily throughout adulthood: ears get bigger, and Hill (personalcommunication) writes, "The Ache have noses get longer,wider, and more protrusivewith in- frequentlycommented on how ugly Europeans are par- creasingage. With the loss of connectivetissue, the ver- ticularlybecause of theirlong noses (theycalled us pyta milion or red zone of the lips gets thinner(Enlow I990, puku-long nose-behind our backs) and because they Larrabeeand MakielskiI993, SusanneI977). are so hairy." These responses are similar to those re- As a resultof changesin hardand softtissue with age, portedby Wagatsuma (i968) forthe firstgeneration of it is possible to estimate ages of adults using informa- Japaneseexposed to contactswith Westerners. Although tion about the relativesizes of eyes, noses, and lips. the Ache and the Hiwi have had little or no contact with Asians or Asian-Americans,they are curiousabout photographsof East Asian faces, generallyattracted to Neoteny and Attractiveness them, and aware of the similaritybetween these faces and their own. A previous study (Jonesand Hill I993) A youthfulor neotenous face is one that combines a showed much strongeragreement in ratingsof attrac- high ratio of neurocranialto lower-facialfeatures with tiveness among the threeWestern societies in the sam- a small nose and ears and full lips. The appendixgives ple (mean correlationsin ratingsof attractivenessr = a summaryof a numberof studies of neotenyand facial .64) than between Westernersand Ache and Hiwi. The attractiveness.Virtually all of themfind that neotenous lattercorrelations were still significantlypositive (r = facialproportions, as definedabove, contributeto female . I 8), however,suggesting a universalas well as a culture- attractiveness.Results formales are equivocal. One lim- specificcomponent to standardsof attractiveness. itationof these studiesis thatthey are confinedto West- Below I presentthree studies addressingthe topic of ern societies or societies stronglyinfluenced by Western neotenyand femalefacial attractiveness.Data formales ideals of physical attractiveness.For example, Wagat- are included forpurposes of comparison.Since previous suma (i968) shows thatcontact with Europeans and U.S. studies ofneoteny and facial attractivenessshow a posi- Americans has had a significantinfluence on Japanese tive relationshipfor females and an equivocal relation- standardsof attractivenessover the past hundredyears. ship for males, I use one-tailed statistical tests forfe- Because the line of argumentpresented above suggests males and two-tailed tests for males. The firststudy that attractionto female cephalofacialneoteny is a good includes data fromall five populations. The other two candidate fora human universal,it is importantto es- studies are more preliminary,but furtherfieldwork is tablishwhether such attractionis characteristicof non- planned to test the artificialstimuli of the thirdstudy Westernizedsocieties as well. in a wider rangeof populations. This paper reports results from an ongoing cross- cultural study of criteriaand consequences of physical STUDY I: AGE PREDICTORS AS ATTRACTIVENESS attractivenessone ofwhose aims is to investigatepossi- PREDICTORS ACROSS CULTURES ble universals of attractivenessby collectingdata from as wide a rangeof populations as practical.Populations BetweenI989 and i992 I collecteddata on standardsof studied to date include two relativelyisolated indige- physical attractivenessin fourpopulations: U.S. Ameri- nous South American groups,the Ache (or Guayaki) of cans, Brazilians,Russians, and Ache Indians. Kim Hill, currentlyat the Universityof New Mexico, assisted with data collectionamong the Ache and collectedaddi- 3. Neotenousfacial proportions in childrenmay be a releaserfor tional data among the Hiwi. The researchmaterial re- parental behavior (Lorenz I943). McCabe (I988) reviews studies ported on in this paper includes facial photographsof showingthat subjects report a greaterwillingness to nurtureand individualsin threepopulations and interviewdata and protectchildren with a highratio of neurocranialto lower-facial ratingsof attractivenessof facial photographsfrom five features.She also cites researchconducted at severallocalities showingthat abused childrenunder court protection (ages 3-6) populations.Facial photographsderive from undergrad- have lowerratios of neurocranial to lower-facialdimensions-that uates at the University of Michigan in Ann Arbor, is, are less cute-than age-matchednursery school control groups. studentsat the Federal Universityof Bahia in Salvador, 730 CURRENT ANTHROPOLOGY Volume 36, Number S, December I995

brazli, and natives ot two Actle villages. Katingsot phO (I993) and Jones(n.d.) providefurther discussion of pro- tographs and questionnaire/interviewdata were col ceduresinvolved in taking,rating, and measuringphoto- lected froma second sample of Universityof Michigai graphs. undergraduates,from middle- and lower-classresident For each photographin each population sample I cal- of Salvador,Brazil, from natives of anotherAche village culated the followingmeasures of facial proportions: eye fromstudents at the Russian State Universityof the Hu width (EW = mean of D [leftendocanthion, left exocan- manities in Moscow, and froma Hiwi settlement. thion] and D [rightendocanthion, right exocanthion]), U.S. Americanphotographic subjects and raterswere nose height (NH = D [glabella, subnasale]), and lip recruitedin introductoryanthropology and psycholog, height (LH = D [labiale superius, labiale inferius]), courses and by flyersposted on campus. For Braziliai where D (a,b) is the Euclidean distance between photo- and Ache photographicsubjects, photographicequip graphiclandmarks a and b, and landmarknames follow mentwas set up in public places, and interestedindivid definitionsin Farkas (I98I). For the analysis below I uals were invited to participate. Brazilian, Russian have divided each measure by face height(FH = D [gla- Ache, and Hiwi raterswere recruitedby goingfrom doo, bella, gonion])to correctfor differences in sizes offaces, to door and by approachingpotential ratersin publi4 producingthree indices of facial proportions:relative places. eye width (EW/FH),relative nose height(NH/FH), and U.S. photographicsubjects and raterswere largelyo relativelip height(LH/FH). These indices were selected European ancestry; attractivenessratings of and fron to measure the relative sizes of the three major facial Asian-Americansand African-Americansare omittedii features-eyes, nose, and lips. I demonstratebelow that this analysis because a restrictedsample providesa bet they do vary as expected with age-eye width and lip ter test of hypotheses. Brazilian subjects and raterc height decrease, and nose height increases. Two other largelyidentified themselves as beingof mixed ancestry featuresthat mighthave been included,ear heightand mostly Africanand European with some Indian ances cheek width,are omittedfrom this analysis: ear height try.(The issues raised forthe study of physical attrac was difficultto measure fromphotographs, and cheek tivenessby Brazil's combinationof race mixtureand ra width did not show a consistentrelation with age. cial stratificationcannot be treatedat any lengthher( Table i presents summarystatistics for these three but are discussed in Jones [n.d.].) Russian raterswere indices of facial proportionsand correlationsbetween largelyof Russian nationalitywith some othernational these indices (each subject to log transformation)and ities of the formerSoviet Union presentas well. Mear ratingsof physical attractivenessfor different combina- ages of photographicsubjects were 23, 20, and 29 fo: tions of photographicsubjects and raters.The table pro- Brazilian,U.S. American,and Ache femalesand 24, 2I vides some supportfor the hypothesizedrelationship be- and 32 for Brazilian, U.S. American, and Ache males tween neotenous facial proportionsand attractiveness, age ranges were I7-34, i8-25, I4-5I, I9-32, I8-30 especiallyfor relative eye width.However, simply piling and i6-6o. up a list of measures of facial proportionsand reporting Facial photographswere taken indoorsin the Unite( their correlationswith attractivenessratings provides States and outdoors in Brazil and Paraguay at a fixec only a weak test of the neoteny hypothesis.A better distance in a standardposition. Ratingsof photograph! test, combininginformation from different age cues, is were collected by having ratersrank subsamples of a presentedbelow. This analysis proceeds in two stages, photographicsample population.For Brazilian and U.S the firstresulting in several equations that can be used American photographicsamples, new subsamples were to calculate the predictedage of each photographicsub- drawn at random for each rater.Because Ache photo. ject on the basis of relative sizes of facial featuresand graphicsubjects spanneda widerrange of ages,individu- the second using these predictedages to produce an in- als fromthe Ache photographicsample were assignec dex of neoteny. to fixed subsamples with others of similar age. Each Age-predictorequations. I begin by using stepwise raterrated photographsof membersof the opposite sex multivariateregression to produce equations that pre- drawn froma single photographicsample population; dict age as a functionof relativesizes of facial features. ratings of all three photographicsample populations Forthis stageI use onlyphotographs of Ache, since Ache were collected in each of five populations of raters(ex. span a much widerrange of ages than eitherof the other cept forHiwi ratingU.S. Americans).The attractiveness two photographicsample populations. Using the step- ratingof a given photographrelative to a given popula. wise multiple linear regressionroutine from Systat s.o tion of ratersis the mean attractivenessrating of thai with bothp to enterand p to discard set at .I 5 produces photographacross ratersfrom that population,with age the followingequation: of photographicsubject partialled out. For Ache, re. ported correlationsbetween attractivenessratings anc PredictedAge i = - 141 log[EW/FH] (I; females) relevantvariables are means of correlationswithin sub- -62 log[LH/FH]- i28. samples (more exactly,means calculated using Fisher's z-transformationand its inverse [Sokal and Rolf I969: The variance accounted for(R2) is .23. 520-23]). An Apple Scanner connected with a Macin. Equation i predictsages ofAche femalesas a function tosh II was used to measure {x, y} coordinatesof a num. of relative eye width and relative lip height; relative ber of facial photographiclandmarks. Jones and Hill nose height drops out of the regression.However, an JO N E S Sexual Selection and Facial Neoteny | 7 3 I

TABLE I Correlationsbetween Measures of Facial Proportionsand Ratings of Attractiveness

Males RatingFemales Females RatingMales

Photographsof Ratedby EW/FH NH/FH LH/FH EW/FH NH/FH LH/FH

Brazilians Brazilians (i 9, I I) .39 .0I .24 + -.09 -.I9 .I7 N = 49, 20 U.S. Americans (i2, 20) .I9 -.I4 -.07 -.35 -*4I .42+ Russians (iI, I4) .32 .I3 .12 -.IO -.43+ .27 Ache (iI, I3) .43 -* .29* .23 -.07 -.I5 .IO Hiwi (4, 4) .28* -.I9 .2I .03 -47* 53* Mean .2IO .589 .I25 .I85 .585 .I35 S.D. .OI4 .030 .020 .oi8 .033 .027

U.S. Americans Brazilians(2o, 23) .23 -.26+ .I7 .00 .i8 .04 N = 5I, 35 U.S. Americans(ii, i8) .30* .04 -.04 -.03 .02 -.I5 Russians(I2, I4) .2 I -.20 .03 .07 .09 .05 Ache (2o, 2i) .25 + .IO .13 -.II -.04 -.0I Hiwi (o, o) n.a. n.a. n.a. n.a. n.a. n.a. Mean .225 .58I .I1I .202 .574 .ii8 S.D. .oi8 .024 .022 .OI3 .025 .022

Ache Brazilians(I7, I6) -.I2 .00 .IO .2 -. .26 .I9 N = 4I, 36 U.S. Americans(i2, i5) -.09 .I9 .o8 .22 -*99+ .I8 Russians (I2, I2) -.22 .3 I -.03 .I3 -.43 .I4 Ache (i5, I5) .I5 -.02 .00 .2I - .I7 .14 Hiwi (7, 4) .I9 - .32 .49 .3I+ -.II .0I Mean .I89 .599 .i28 .I77 .607 II8 S.D. .OI 3 .035 .023 .OI3 .032 .029

NOTE: EW/FH,eye width/faceheight; NH/FH, nose height/faceheight; LH/FH, lip height/faceheight. Numbers in parenthesesare numbersof raters for female and male subjects. +p < .I *p < -O5 equation can be produced with nearly the same pre- Indices of neoteny and facial attractiveness.A neo- dictivepower as Equation i by leavingrelative lip height tenous face is one thatretains youthful traits or presents out of the regression.The resultingequation is: markersof youth in an exaggeratedform relative to others of the same age. Equations I to 3 can be used to PredictedAge2= - io8 log[EW/FH] (2;females) produce indices of neoteny: I define facial neoteny as the differencebetween the actual age of a face and the + I39 log[NH/FH] - I7 age predictedby one of the equations above. In other words, The variance accounted for(R2) is .17. In otherwords, ages of Ache femalescan be predicted Neotenyi = Age - PredictedAge i (4; females) as a functionof relativeeye width in combinationwith eitherrelative lip heightor relativenose height. Neoteny2 = Age - PredictedAge 2 (5; females) For Ache males, stepwise multiple regressionof age relative on relativeeye width,relative nose height,and Neoteny3 = Age - PredictedAge 3 (6; males) lip heightproduces anotherequation: Thus a face with unusually large eyes, small nose, and PredictedAge 3 = - I46 log[EW/FH] (3 males) full lips in relation to face heightwill have a low pre- -6i log[LH/FH] - I36. dictedage accordingto Equations I to 3 and a highindex of neotenyaccording to Equations 4 to 6. The variance accounted for(R2) is .55. Table 2 shows correlationsbetween indices of neo- However, when relative lip height is excluded from teny and ratings of attractivenessfor differentcom- the regression,the remainingtwo variablesdo not yield binationsof photographic subjects and raters.When cor- anothersignificant predictor equation. Thus we are left relation coefficients for two differentsamples are with two alternativeage-predictor equations forfemales sufficientlysimilar, the samples can be treatedas a sin- and one formales. gle sample and the correlationcoefficients pooled (Sokal 732 CURRENT ANTHROPOLOGY Volume 36, Number S, December I995

TABLE 2 Correlations between Indices of Neoteny and Ratings of Attractiveness

Females Males RatingFemales Rating Males

Photographsof Ratedby NeotenyI Neoteny2 Neoteny3

Brazilians Brazilians (I9, II) 37* * .24* -.23 N = 49, 20 U.S. Americans(I2, 2o) .o6 .04 - .I7 Russians(ii, I4) .23 + .II -*05 Ache (II, I3) .38** 45 -.02 Hiwi (4, 4) .25* .25 *35 Mean - I.o yrs. - i.6 yrs. - I.3 yrs. S.D. 7.5 yrs. 5.5 yrs. 7.I yrs.

U.S. Americans Brazilians(2o, 23) .34* .36** 03 N = 5I, 35 U.S. Americans (ii, i8) .20+ .28* -.I3 Russians (I 2, I4) .23 + .26* .o8 Ache (2o, 2I) .25 + .i8 -.07 Hiwi (o, o) n.a. n.a. n.a. Mean -.2 yrs. -.I yrs. - i.6 yrs. S.D. 7.8 yrs. 4.5 yrs. 7.7 yrs.

Pooled samples Brazilians (39, 44) .36* * 30* * - .IO (Brazil + U.S.) U.S. Americans (23, 38) .I3+ .i6+ -.I5 N = Ioo, 55 Russians (23, 28) .23 + .I8* .02 Ache (3I, 34) 32 ** .32** -.05 Hiwi n.a. n.a. n.a.

Ache Brazilians (I 7, I 6) -.IO -.I3 .00 N = 4I, 36 U.S. Americans (i2, i5) -.05 -.i6 .I4 Russians(1I2, I12) -.I9 -.29 .0I Ache (i5, I5) .07 .12 .2I Hiwi (7, 4) *38* .24 + .07 Mean .2 yrs. -.2 yrs. - .4 yrs. S.D. 9.2 yrs. 9.6 yrs. 9.3 yrs.

Pooled samples Brazilians (56, 5o) n.a. n.a. -.07 (All) U.S. Americans (35, 53) n.a. n.a. -.07 N = I4I, 9I Russians(35, 40) n.a. n.a. .02 Ache (46, 49) .25** .26** .03 Hiwi (7, 5) .3I**I24** .23*

NOTE: Numbersin parenthesesare numbersof raters for female and male subjects. +p < .I *p < .05 * p < .0I

and Rolf I969:5 20-23); table 2 shows pooled data where two indices of neoteny,females are perceived as more appropriate.The rows reportingpooled correlationsfor attractiveto the extentthat theirpredicted ages, as cal- female subjects-for Brazilians, U.S. Americans, Rus- culated fromtheir facial proportions,are less than their sians, and Ache ratingBrazilian and U.S. photographs actual ages. and for Ache and Hiwi ratingall three sets of photographs-provide consistentsupport for the proposition tivenessconnection for Westerners rating Ache females is probably that neoteny is a component of female facial attrac- largelyan artifactof responses to eye shape.Ache have epicanthic tiveness.4Across five populations of raters,and across folds,and Ache witha highrelative eye width(log [EW/FH])also have relativelynarrow eyes. Westemraters all show an aversion to narroweyes-correlations between log (EW/EH)and ratingsof 4. The rangeof correlationcoefficients is such that,for Western femalefacial attractiveness = - .3 I, - .3 3, and - .44 forBrazilian, raters(Brazilians, U.S. Americans,and Russians),it is not legiti- U.S. American,and Russianraters. The apparentWestem aversion mateto pool coefficientsfor all threefemale photographic subject to Ache femaleswith wide eyesdisappears when eye shapeis con- samples. While Westemraters are attractedto neotenousfacial trolledfor. In addition,for all populationsof raters, the much wider proportionswhen rating U.S. Americanand Brazilian women, they rangeof ages amongAche photographicsubjects makes it more seemto showa mildaversion to neotenousfacial proportions when difficultto uncoverpossible correlates of attractiveness other than ratingAche women. This slightreversal of the neoteny/attrac-age. JONES Sexual Selection and Facial Neoteny | 733

STUDY 2: STUDENTS AND MODELS Table 3 also includes data on ages predictedfrom Equations I to 3 above. Female models have extremely A second test of the neoteny hypothesisincorporate( neotenous faces: predictedages based on facial propor- measurements of two new samples of facial photo tions are 6.8 years and 7.4 years. This does not mean graphs,of U.S. female and male models, on the theor thatthe facial proportionsof models match those of real thatthe distinguishingfeatures of attractivefaces migh 7-year-olds,since the ages predictedby these equations stand out especially clearly in comparisons betweei are based on naive linear extrapolationsof age-related these samples and Universityof Michigan undergradu changes in adult facial proportions. It does argue ates. The two model samples consist of (i) photographl strongly,however, that female models' faces represent of ten femalemodels displayedon the coversof Cosmo a "supernormal stimulus," presentingin exaggerated politan and Glamour magazines between i989 and I99 formthe featuresthat distinguishyoung women from (five from each magazine) and (2) photographsof tel old ones. male models displayedin advertisementsin Esquire an! GQ. Both samples were selected frommagazines on fil at the Ann ArborPublic Library.For the female mode STUDY 3: CARDIOIDAL STRAIN AND FACIAL samples,I began with the most recentissue ofeach mag ATTRACTIVENESS azine and workedthrough successively earlier issues un A thirdtest of the neotenyhypothesis employed draw- til I foundfive suitable faces fromeach magazine covei ings of faces made more and less neotenous by differing Models not facingdirectly toward the camera, model degrees of cardioidal strain. From the University of with theirmouths open, celebrities,and non-Caucasian Michigan photographicsample I selected nine photo- were excluded. Since Esquire and GQ normallyfeaturi graphs: three of females given high ratingsof attrac- celebrities rather than models on their covers, mall tiveness by University of Michigan undergraduates, models were selected fromconsecutive advertisement three of females given average ratings,and three of fe- a in these magazines, subject to the same restrictions males given low ratings (high-, medium-, and low- females.I used calipersto measureface height, right an( attractivenesssubgroups). I made a line drawingof each lefteye width,nose height,and lip heightand calculate( face and producedtwo new drawingsfor each ofthe orig- relative eye width (EW/FH),relative nose height (NH, inal nine by subjectingeach originaldrawing to negative FH), and relativelip height(LH/FH). The mean and stan and positive cardioidalstrain. For each trial,a raterwas darddeviation of these measurementsand the compara presentedwith one originaland two transformeddraw- ble figuresfor my sample of Universityof Michiganun ingsof a singleface (k = - o.i, O, O. i) andasked to rank dergraduate females are presented in table 3. Fo the threeversions in orderof attractiveness.Raters were females,all differencesbetween models and undergradu drawnfrom students in an intermediate-levelanthropol- ates are in the expected,neotenous direction-model ogy class; each set of drawingswas rated at least four have largerelative eye width,small relativenose height times. The entireprocedure was repeatedusing photo- and large relativelip height.T-tests (conducted on log graphsof nine males, divided as beforeinto high-,me- transformedvariables) show that all mean difference dium-,and low-attractivenesssubgroups. between female models and studentsare significant. For both males and females,faces subjected to positive cardioidal strain (k = O. i) were rated consistently TABLE 3 less attractivethan the originalfaces (fig.2). The results Facial Proportionsand PredictedAges of Studentsanc are highlysignificant (p < .oi, binomial test) forboth Models of Both Sexes sexes. Results fornegative cardioidalstrain (k = - O.i) are more complicated. For females, average attractiveness ratingswere higherfor neotenous faces than Students Models fororiginal faces. For males, attractivenessratings were lowerfor neotenous faces thanfor original faces. Results Mean S.D. Mean S.D. were marginallysignificant and nonsignificant,respectively(p = .o6, p = .ii, binomial test). Since male and female trendsare in the opposite direction,the differ- Females EW/FH .23 .0I9 .24 .OI* ence between the trendsfor the two sexes is strongly NH/FH .58 .024 .49 .034* significant5 LH/FH .2I.7 .022 OI55** For both sexes there seemed to be an interactionbe- Predicted Age i (yrs.) 20.2 8.o 7.4 3.I* Predicted Age 2 (yrs.) 20.2 4.5 6.8 3.8* * Males EW/FH .20 .OI3 .20 .OI4 5. The null hypothesisin thiscase is thatdf (the probability that NH/FH .57 .025 .47 .025** a neotenousfemale face, k = O.I, will be ratedmore attractive LH/FH .I2 .022 .I2 .oi8 thanthe originalface, k = o) is equal to dm(the probability that a Predicted Age 3 (yrs.) 22.8 8.3 23.I 5.8 neotenousmale facewill be ratedmore attractive than the original face).But forall possiblevalues of df= di,,the probability of get- tingthe observedresults both for females and formales-that is, *p < .05 the productof the probabilityfor males and the probabilityfor **p < .0I females-is less than.oi. 734 CURRENT ANTHROPOLOGY Volume 36, Number S, December I995

tive temales are apparentlyneotenous enough that mating theirfaces more neotenous via negative cardioidal

2.5 straindoes not make them more attractive(although it does not make them less attractiveeither), while unat- tractivefemales are rated more attractivewhen their 2 faces are made more neotenous. Attractivemales are apparentlyneotenous enough alreadythat makingtheir i~ .5-- facesmore neotenous via negativecardioidal strain actually lowerstheir attractiveness, while moderatelyattrac- I , [ Females tive and unattractivemales are apparentlynot so neotenous that making theirfaces more neotenous makes -0.1 0 0.1 them less attractive(but it does not make them more attractiveeither). Thus there is a limit to how far in- attractiveness creasing neoteny results in increasing attractiveness, 3 -- subgroup and that limit is reached sooner formales than forfemales.

2.5~~-- hi Summaryand Implications -' 2 ,/ X Di med. In humans, mate value declines with age beginningin -~~~ lo early adulthood. It declines more quickly for females than formales. Consequently,we expecthuman beings, especially males, to have adaptationsfor assessing age- 1 Males related changes in mate value. Regardless of whether -0.1 0 0.1 age-relatedchanges in physicalattractiveness (especially forfemales) result fromsuch adaptationsor fromother causes, it is possible thatparticularly attractive faces are those that presentage-related cues in an exaggeratedor supernormalform. Large eyes in relationto face height, small noses, and full lips are markersof youth.The re- cardioidalstrain, k sults presentedin this paper suggestthat neotenous fea- FIG. 2. The effectsof negative and positive cardioidal tures are indeed criteriaof female attractivenesseven transformationson the perceived attractivenessof when age is controlledfor. Specifically, faces of both sexes (includingfaces previouslyrated i. Women whose facial proportionsmake them look attractive,medium, and unattractive). youngerthan theiractual age (as measuredby regression equations predictingage as a functionof facial proportions) are perceived as more attractiveby male raters tween attractivenesssubgroup and the effectof negative fromfive populations (but see n. 4). cardioidal strain.For the threeunattractive females, i I 2. A sample of U.S. female models has significantly out of I 2 ratersfound neotenous faces more attractive more neotenous facial proportionsthan a sample ofU.S. than original faces, while for medium and attractive female undergraduatesand a strikinglylow predicted facesjust I 3 out of24 ratersfound neotenous faces more age, about 7 years,according to regressionequations pre- attractive.For the three attractivemales, II out of I2 dictingage as a functionof facial proportions. ratersfound neotenous faces less attractivethan non- 3. Cardioidal strain, a mathematical transformation neotenous faces, while for medium and unattractive shown by earlier researchto provide a good model for faces just i 6 out of 3I ratersfound neotenous faces less changesin facial proportionsduring the course of matu- attractive.In otherwords, negative cardioidal strain re- rationand to affectthe perceivedages of faces,also has sultedin a markedincrease in the attractivenessof unat- an effecton female facial attractivenessaccording to tractivefemale faces and a marked decrease in the at- U.S. raters.The effectis nonlinear,suggesting that neo- tractivenessof attractivemale faces but little changein teny is a componentof attractivenessonly up to a cer- the attractivenessof otherfaces. tain point. In summary,results from drawings of faces subjected 4. Results formale attractivenessin the above studies to cardioidal transformationssupport the hypothesis are weak and/orinconsistent. thatneoteny is a componentof facial attractiveness,at What are some possible futuredirections for research least forfemales, but theyalso suggestthat the relation- on neoteny and physical attractivenessand in the an- ship between neoteny and attractivenessis nonlinear. thropologyof physical attractivenessmore generally? For all attractivenesssubgroups and both sexes, making Researchto date providesstrong support for the hypoth- faces less neotenousvia positivecardioidal strain makes esis thatneoteny is a componentof femalefacial attrac- them less attractive.But highlyand moderatelyattrac- tiveness in Western societies. By contrast,the results JONES Sexual Selection and Facial Neoteny 1735

forthe two non-Westernizedsocieties presentedin this (I993a,b) notes that the waist-to-hipratio, another can- paper are only a beginning;much more researchwill didate fora universal componentof femalephysical at- have to be carriedout in a wide rangeof such societies tractiveness,varies not only with age but, even among with a varietyof research instrumentsbefore we can women of the same age, with endocrineprofiles associ- state with any confidencewhether the neoteny/attrac- ated withfecundity. The theoreticalcase thatneotenous tiveness connection is a human universal. Assuming cephalofacial featuresprovide such informationis not that this connection continues to be supportedby fur- very strong,because such features,unlike secondary therresearch, many additional topics will remainto be sexual characteristicssuch as enlargedbreasts, buttocks, addressed.These include the evolutionarybases ofmale hips, and thighs,do not appear suddenlyat pubertyun- attractionto femalecephalofacial neoteny and the possi- der the control of female sex hormones.However, the ble consequences of such attractionfor human morpho- topic clearly warrants investigation by reproductive logical evolution. ecologists and medical anthropologists. i. Evolutionary causes of attraction to neoteny- 2. Evolutionaryconsequences of attraction to neo- furtherpuzzles. While the connectionbetween neoteny teny.Darwin unitedhis discussionsof human evolution and female facial attractivenessproposed here may and sexual selection in a single volume because he partlyexplain the anomaly of female attractiveness,it thoughtthat the latter had played a major role in the raises otherpuzzles. An attractionto markersof youth former.The results of this paper suggestthat it is time may be adaptivefor men insofaras it leads them to find for renewed attention to the possible connection be- younger women more sexually attractivethan older tween sexual selection and morphologicalevolution in women, but it will not be adaptive if it leads men to our species. The evolution of modern Homo sapiens findjuveniles more attractivethan youngwomen. Why, over the past ioo,ooo years has been markedby a trend then,do males apparentlyfind markers of youthattrac- toward increasing craniofacial neoteny, including re- tive even among females at the age of maximum fertil- duced prognathism,increased brachycephaly,and gen- ity? Part of the answer may be that mate choice in the eral gracilizationin a number of populations (Weiden- real world involves attentionto more than just facial reich I945, Newman i962, Brace and Mahler I97I, cues. If real-worldmate choice involves both attention FrayerI98I). Biological anthropologistshave generally to markers of cephalofacial neoteny to discriminate invoked natural selection for ecological adaptation or youngadults fromold adults and attentionto secondary nonadaptive forces such as pleiotropyor biased muta- sexual characteristicsfrom the rest of the body to dis- tion to explain these trends.The analysis in this paper criminateadults fromjuveniles, then attractionto su- suggeststhat sexual selectionmay also be involved.Sex- pernormallyneotenous facial featuresis less puzzling. ual selection forneotenous featuresin femalesis likely This line of argumentsuggests that the neotenous fe- to have become a particularlypowerful force in human male faces generatedin Study 3 will be perceived as evolution once increases in lifeexpectancy had resulted more attractive, relative to nonneotenous faces, if in a largerfraction of the adult femalepopulation living attachedto drawingsof unequivocallyadult bodies. An- past the age of menopause, thus increasing the age- otherpart of the answercould be thatfemale mate value related variance in adult fecundity.Parallel trendsto- may be highest somewhat beforethe age of maximum ward neoteny in males could be a by-productof such fertility.Symons (I979:I89-90) arguesthat males are selection in females or a directproduct of sexual selec- most physicallyattracted to femalesof maximum repro- tion on males. Whetherongoing sexual selectionfor neo- ductive value, ratherthan maximum fecundity.Repro- tenous features(or other physical traits) can be mea- ductivevalue is a measureof expected lifetime reproduc- sured in living populations is a topic for future tion, and an individual choosing a partner with a investigation. long-termrelationship in mind should be adapted to The conventionalwisdom in the social sciences has take into account a mate's futurereproductive potential been that evolutionarytheory, including the theoryof as well as her currentfecundity.6 A testableimplication sexual selection,is more or less irrelevantin explaining ofthis line of argumentis thatthe attractivenessof neo- human social behavior.The successes ofsuch new fields tenous features(in comparisonwith markersof sexual as evolutionarypsychology and human behavioralecol- maturity)will vary dependingon whetherindividuals ogy,as well as improvedunderstanding of the physical are consideringlong-term or short-termrelationships. underpinningsof human behavior,are likely to forcea Finally,neotenous facial proportionsin females might reassessmentof this conventionalwisdom. The studyof provide informationabout levels of ovarian function physical attractivenessis a particularlypromising test above and beyond the informationthey provide about case forinvestigating the relationshipbetween biology age (Johnstonand Franklin I993). By analogy, Singh and culture.While this paper has emphasized the "biological" side ofphysical attractiveness, with the modern theoryof sexual selection as a startingpoint, this theory 6. Strictlyspeaking, reproductive value is not a perfectmeasure of will undoubtedlyhave to be expanded and revised to long-termmate value because a personcommitted to a long-term allow forthe unique importanceof social in our relationshipmay die orbecome infertile before the end ofa mate's learning reproductivecareer. When this possibilityis takeninto account, species (Laland I994). Whateverthe fateof the hypothe- matevalue becomesa functionboth of mate's age and ofego's age. sis that neoteny is a universal of female facial attrac- This topicis treatedin moredetail in Jones(n.d.). tiveness, this paper will achieve one of its aims if it 736 CURRENT ANTHROPOLOGY Volume 36, Numbers, December 1995 encouragesboth biological and culturalanthropologists smaller chins and jaws than typical female faces. Fe- to explore a topic that has lain fortoo long in the no- males' ideal male faces were similar to typical male man's-landbetween the two subdisciplines. faces.

Johnstonand Franklin(I993) Appendix: Research on Neoteny and Female Population. U.S. Americans. Facial Attractiveness Research methods. Researchersused a "genetic algo- rithm" to allow subjects to generateattractive female McArthurand Apatow (I983-84) faces by a process analogous to artificialselection. A Population. U.S. Americans. computerprogram generated a small population of fe- Research methods. Researchersused a police Identi- male faces froma set of random binarystrings ("geno- kit to produce"average" male and femalefaces and faces types") which specified shapes and positions of facial with differentcombinations of "baby-faced"and mature features. Subjects assigned attractiveness ratings to features: large and small eyes, low and high vertical faces. A new generationof faces was producedby "se- placementof features, small and largenoses and ears,all lecting" genotypesin proportionto theirattractiveness threebaby-faced features together, and all threemature and adding small random "mutations" to the binary featurestogether. strings.The trial-and-errorprocess continueduntil each Results. For both sexes, faces with large eyes were subject had "evolved" the most attractiveface. rated more attractivethan average faces, which were Results.Attractive faces had significantlysmaller eye- rated more attractivethan faces with small eyes. The chin lengths, smaller lower-facialproportions, fuller effectwas strongerfor females than formales. For other lips, and narrowermouths than averageundergraduates. traits,(i) forfemales only, average faces were ratedmore attractivethan faces with baby-facedfeatures, which Perrett,May, and Yoshikawa(i994) were rated more attractivethan faces with maturefea- Populations. English,Japanese. tures; (2) formales only, averagefaces were ratedmore Research methods. Researchersconstructed three fa- attractivethan faces with maturefeatures, which were cial images using graphics software: (i) a composite rated more attractivethan faces with baby-facedfea- (blend)of digitizedphotographs of 6o Caucasian English tures. females, (2) a composite of the most attractivei 5 females (as rated by English raters),and (3) an attrac- Fauss (I986) tiveness "caricature" that exaggeratedthe featuresdis- Population. Germans. tinguishingthe second composite fromthe first.They Research methods. Researcherasked subjects of both used the same techniquesto generatethree images based sexes to constructthe face of an ideal member of the on photographsof Japanesefemales as ratedby Japanese opposite sex using a police Identi-kit. males. Results. Ideal males and females had wide mouths Results. Images were rated3, 2, i (mostto least attrac- with full lips and a gracile lower jaw and chin. Ideal tive) by Japaneseand Caucasian ratersrating both own females also had gracile noses and high-archedeye- group and across groups. Attractivefaces had higher brows. cheekbones, a thinnerjaw, largereyes relative to the size of the face, and shortervertical distances between Cunningham(I986) jaw and mouth and between mouth and nose. Populations. U.S. Americans (raters),U.S. American and international(photographs). Research methods. Researcher collected measurements of relative size of facial featuresin a sample of photographsof 50 females.The sample included23 pho- Comments tographsof U.S. American college studentsand 27 photographs of Miss Universe contestants. Photographs were rated by a sample of U.S. American undergrad- C. LORING BRACE uates. Museum of Anthropology,University of Michigan, Results. Higherattractiveness ratings were associated Ann Arbor,Mich. 48109, U.S.A. 25 III 95 with larger and more widely separated eyes, wider cheekbones,narrower cheeks, and smaller noses. It is all verywell to show that males may rate females on various scales of "attractiveness"and that neote- Riedl (I990) nousness may be one of them, but in orderfor this to Population. Austrians. have any cumulative effecton the appearanceof future Research methods. Subjects were instructedto pro- generationsit also has to be demonstratedthat this is duce the most attractivefacial image using a computer somehow related to differentialreproduction (Ryan program which allowed them to manipulate sizes, I995). StephenJay Gould has arguedthat aspects of hu- shapes, and positions of facial features. man neoteny emerged because people find iA "cute" Results. Males' ideal femalefaces had largereves and (I977:35 O), and Joneshas expended considerableeffort JONES Sexual Selection and Facial Neoteny | 737

in demonstratingthat males evaluate femaleappearance WILLIAM JANKOWIAK to that effect.However, not only does this ignore the Departmentof Anthropology,University of Nevada, role of female choice in reproductivebehavior that was Las Vegas, Nev. 89154, U.S.A. 2 V 95 such an importantpart of Darwin's argumentin the first place (Mayr I972: 90-9 1)-a perspectivewhich has seen Joneshas provideda concise, lucid, and ultimatelyaes- a recentresurgence of interest(e.g., StrierI992, Cronin theticallypleasing response to one ofhuman evolution's I993)-but it does not take into account the fact that, most intriguingquestions: Why is the human male so in the available human examples, it is the males who interestedin a potential partner'srelative physical at- ''are characterizedby an appreciablyhigher variance in tractiveness?It should come as no surpriseto anyone their reproductivebehavior" while "women are uni- who has been partiallyconscious since the onset of pu- formlyexposed to the risk of pregnancyand rarelyfail berty that men around the globe have long been in- to reproduce"(Dobzhansky I972:77). And,in thepast, tensely preoccupied with the relative attractivenessof this was true whethermales thoughtthey were "cute" youthful-lookingwomen. Althoughthe existenceof the or not. sex-linkedpattem ofgazing is seldom contested,the rea- Evidently,if all femaleshave the same opportunityto sons forits persistenceaccount forsome of the liveliest reproduce,male choice is not going to have any effect and, at times, most emotionallycharged debates in the on the characteristicsof futurefemale form.Jones is historyof social science. quite rightthat human craniofacialform has undergone Cultural anthropologistsare more comfortableex- increasing gracilization over the past ioo,ooo years plaining the origins of men's keen interestin the nu- (Brace,Smith, and Hunt i99I), but he has not made a ances of femalebeauty as stemmingfrom an unresolved case thathuman choice had anythingto do with it. Fur- Oedipus complex,the persistenceof a patriarchaltradi- thermore,there are some veryimportant aspects of that tion,or the demands of a postcapitalistworld orderpro- gracilizationwhich cannot be detectedby readyvisual moting a consumption ethic that, in turn,encourages discriminationby eithermales or females.The decrease the sexual objectificationof the female, but not the in bone density and the thinningof the bones of the male, body. By failingto studythe phenomenonfrom a cranial vault cannot be visually discerned,and it would comparativeperspective, cultural explanations are prone be a most extraordinaryperson who would go to the to confusethe phenomenon'sorigins with the social fac- extent of assessing the relative degree of incisor shov- torsthat shape its culturalexpression or stylisticembel- eling and lingual-tubercledevelopment or third-molar lishment.From the perspectiveof Foucault and his ad- agenesis and the relative reductionof the hypoconulid mirers,for example, every social interactionis, firstand when consideringthe attractivenessof a potential sex- foremost,about power, and thereforemen's preference ual partner.That leaves "the neotenyhypothesis" right for youthfulnessis less, about aesthetics than about where it was some years ago, when it was judged to domination and control. It is not clear, however, if be "largely,if not totally,a bankruptconcept" whose men's preferencefor female youthfulnessarises froma persistence was due mainly to "anthropocentrism" will to dominate as much as froma general ability to (Shea I989:97)-in this case one could call it male chau- objectifya potentialpartner's relative fecundity, with an vinist anthropocentrism. unintendedconsequence beingthe transformationof the The emergence of "modern" human form over the female into an object of aesthetic contemplation. past ioo,ooo years and more is a consequence of reduc- It is importantto rememberthat youthfulnessis an tions from Middle Pleistocene levels of robustness importantresource of female power, which oftenresults which I have treated in considerable detail elsewhere in the heighteningof men's apprehensionand sexual (Brace,Smith, and Hunt i99i; BraceI995, n.d.).I sug- anxiety.While doing researchin the People's Republic gested that those reductionswere the effectsof muta- of China I was repeatedlystruck by the paradoxof men's tions that were not weeded out when selection forthe yearningfor while simultaneouslyfearing becoming in- maintenance of the formerlynecessary levels of ro- volved with a beautifulwoman. Althoughit was under- bustness was relaxed. The mutations that produce the stood that beautifulwomen (which in the Chinese con- trendobserved are not "biased" but just the most likely text meant women in theirearly twenties)were harder minimal kind of change that can occur. This is the to control,manage, or mold, when I broachedthe topic mechanism that I labeled the probablemutation effect of the ideal mate men invariablyacknowledged, at least (Brace I995). What it produces,in effect,is evolutionby in the realm of fantasy,that they preferredbeauty to entropy.When selection is reducedor suspended,every- submissiveness.Chinese men are not alone in being ap- thingthat had formerlybeen maintained simply tends prehensive of female beauty. Cautionary tales from to run down. In actual mechanisticterms, what we see around the globe repeatedlywarn men of the hidden is not really "neoteny," or the selective retentionof dangersand potentiallydire consequences of becoming youthfulform, but the increasingfailure of the develop- involved with a prettywoman who is a stranger.The mental process to produce the formerlynecessary adult femme-fatalearchetype and, to a lesser extent,the fatale configuration.This is the most likelyresult of the most status is a panhuman theme, suggestingthat cultures likely mutationsoccurring in the relaxationor absence everywherewam men and women to avoid becoming ofselection-which is the minimumworking definition overlyfixed on what is most desiredin the oppositesex. of the probablemutation effect. For men it is physical beauty,whereas forwomen it is 738 CURRENT ANTHROPOLOGY Volume 36, Number S, December 1995

accessing a man's relative social standing.The cross- are likely to have . . . species-typical. . . components, culturalpervasiveness of femme-fatale tales revealsthat and variation in these componentsmay be predictable beauty,which is always associated with youthfulness, given knowledge of human biology and local circum- is seen as conferringgreater power on its possessor. In stances." I wonder just how powerfulthis predictionis this context,youthfulness is seldom taken as evidence likely to be. We can only speculate about such local of submissivnessand thus powerlessness. circumstances,we don't know whetherthe matingpref- I find Jones'sexperimental design an ingenious way erences have been consistentover time, and we can be to bracket,for purposes of analysis,the possible effects less certainthan with otherspecies thatpreferences will of social learningwhile focusingon the impact of sex- translateinto greaterreproductive success forthe trait linked differencesin the perceptionof relativephysical carrier.At best,we are able to make non-specificpredic- attractiveness.It would be interestingto know what ef- tions such as "one mightexpect human beings to have fect sex orientationand thus potential erotic interest a . . . species-typicalemotional response to signs of would have on his researchfindings. If sex orientation aging." While I do not wish to denigratethe evolution- were controlledfor, lesbians and straightwomen might ary approach,which I considerboth importantand in- be expected to differin their appreciationof relative sightful,I believe it is essential that we be conscious of youthfulnessin females. I furthersuspect that lesbians the dangersinherent in this strategy. more than straightwomen will find relative maturity We would do well to ask ourselvesa numberof ques- aesthetically pleasing. Because straightwomen often tions: First,have we simplyfocused on the traitswhich come to value what men desirein the oppositesex, they support our hypothesis, ignoringthose that do not? will also find youthfulwomen more attractive.Con- Clearly,there are many age-indicativetraits. Would we versely,I predictthat comparedwith straightmen gays have got the same answers if we had focused on, say, will findrelatively youthful men aestheticallymore at- head shape, skin quality,and hair color? Secondly,have tractive.If my speculations-and that is all they are- we consideredalternative explanations for the findings? turnout to be accurate,then erotic preference may oper- Here I findmyself in disagreementwith Cosmides and ate as the triggerfor heightened or dampenedinterest in Tooby (I987), who writethat "learning is not an altema- the cultural objectificationof a potentialsex partner. tive hypothesis" to an evolutionaryexplanation. In a By groundinghis analysis in an evolutionaryframe- trivialsense this statementmust be true,since learning work, Jonesprovides a convincingexplanation for the must itselfhave evolved and its operationmust be con- human male's proclivityto focus on youthfulnessas an strainedand informedby processes operatingat other importantaspect of female attractiveness.His request levels and on othertime scales. However,this does not that cultural anthropologistsstudy the psychological negatethe factthat a matingpreference could be learned processes of sexual attractionharks back to the disci- or unlearned,and if learned it could be independentof pline's historical mission to study the particularand or influenced by the social environment.Nor, since universalaspects of human experience.It remainsto be there is empirical and theoreticalevidence that social seen how many anthropologistswill followhis lead and leamingmay allow maladaptivetraits to spread(Cavalli- enter the troubledwaters of documentingthe parame- Sforza and Feldman i98I, Boyd and Richerson I985, ters of our sex-linkedhuman nature. Durham i99I), should we assume that thereis an adaptive explanationfor all human behaviour.I believe that cross-culturalstudies such as Jones'sare invaluable,pre- KEVIN N. LALAND ciselybecause it is plausible thatattraction to facialneo- Sub-Departmentof Animal Behaviour,Madingley, tenyor any othertrait is a society-specificfad. Thirdly, Cambridge CB3 8AA, England. 23 Iv 95 are we just tellingstories? Human evolutionaryhistory began not in the Pleistocene but with the beginningof Jonesillustrates the value of an evolutionaryperspec- life. This means that there is no shortageof hominid, tive. I agree that the relationshipbetween human mat- ape, primate,or social carnivoremodels which we could ing preferencesand sexual selectiondemands further at- use as the basis foran adaptivestory. Evolutionary argu- tention,and cross-culturalcomparisons are particularly ments are so easy to constructthat empirical support insightful.I appreciatethe rigorof Jones's analysis, espe- should demand more than a cursoryreview of circum- cially since human sociobiologyand evolutionarypsy- stantialevidence or a quick questionnairehanded out to chologyhave oftenstrayed toward methodological lax- undergraduates.Jones's study is to be welcomed both ity. I am a little suspicious of his use of one-tailed for the modesty of its claims and for the rigorof its statisticsfor females (cf. two-tailedfor males), particu- analysis. larly since the findingsare used to justifya hypothesis Finally,we should be aware thathuman sexual selec- concerning a sex difference,but his conclusions are tion may operateby means, and at rates,atypical of ani- probablyfairly robust. His study touches on important mal populations. Classically, researchersinterested in points regardingthe nature and complexitiesof human human sexual selectionhave treatedcultural influences sexual selection, which I address below. First,I raise on mating preferencesas a confoundingfactor which some concernsabout the evolutionarypsychological ap- obscures understandingof how sexual selection has op- proach. erated.In contrast,a recenttheoretical analysislof mine Joneswrites that "standards of physical attractiveness has demonstratedthat mating preferencesdo not have JONES Sexual Selection and Facial Neoteny | 739

to be innate to generatesexual selection (Laland I994) cuses on the facial gestaltas the basis forperceptions of Ratherthan obscuringsexual selection,learned and so attractiveness.This view is consistentwith the litera- cially transmittedpreferences (for instance, for body ture demonstratingthat faces are perceived as wholes shape, hair color, or foot size) may themselvesgenerat( (e.g., Hosie, Ellis, and Haig I988, Morton and Johnson sexual selection, increasing the frequencyof the pre- I99I, Purcelland StewartI988, RhodesI986, Sergent ferredtrait. Since, in comparisonwith genetictransmis I984, Young, Hellawell, and Hay I987). Furthermore, sion, social transmission typically results in a morc this approach is more parsimoniousthan feature-based rapiddiffusion of a preferencethrough a population,cul- approaches; both Jonesand Cunninghamfind different turallygenerated sexual selectionmay be unusuallyfast, patternsof results formale and female faces,while we and the alleles underlyingfavored traits may be selected posit,and find,the same patternof results for both male to high frequencyin just a handfulof generations. and female faces. This analysis suggeststhat (i) there should be local, Accordingto Jones'stheory, faces with extremefea- society-specificcorrelations between favoredtraits and turesrepresenting neoteny will be perceivedas more at- mating preferences;(2) sexual selection may account tractivethan otherfaces because of selection pressures. for cross-culturalvariation in traitsunderlying attrac- As he points out, directionalselection favorstraits that tiveness;and (3) recentselection may have modifiedany are extremein theirdimensions (largeantlers, big tails, predilectionsfavored throughout the Pleistocene.These etc.). However, another form of selection-stabilizing theoreticalfindings renforce the importanceof empirical selection-is more prevalentthan directionalselection studies such as Jones'swhich exploremating preference and favorstraits that are the average of values in the patternsacross societies. But in focusingon those as- population(Barash i982, DobzhanskyI970). Thus,one pects of human matingpreferences which are universal, could just as easily predictthat selection would favor we should not neglect the fact that otheraspects show faces with average configurationsrather than extreme considerable cross-culturalvariability (Ford and Beacb features. I95 I, Rosenblatt I 974). Males in all societies may yearn 2. The relationshipof neoteny to attractiveness.Jones foran attractivemate, but in some societies "attractive" claims, as does Cunningham,that neotenyis an essen- means small feet, protrudingbuttocks, or pendulous tial component of facial attractivenessin females be- breasts.How can we account forsuch local preferences? cause it signals fecundity.However, we have shown And could they explain cross-culturalvariation in ana- that, although neotenymay be a componentof attrac- tomical or personalitytraits? tiveness,it is not essential to it. Empirically,if neoteny is fundamentalto attractiveness,judgments of neoteny and attractivenessmust be significantlyand highlycor- LISA E. MUSSELMAN, JUDITH H. LANGLOIS, AND related. However, we (Langlois, Roggman, and Mus- LORI A. ROGGMAN selmanI994) andothers (Berry i 99I) havedemonstrated Departmentof Psychology,University of Texas, that judgmentsof attractivenessand age are unrelated Austin, Tex. 78712, U.S.A. 9 v 95 in samples of college-agefemale faces, indicating that a neotenous appearanceis not requiredfor attractiveness. Jones'spaper is an interestingtreatise on the importance Furthermore,while of course Jonesis correctin pointing of physical attractivenessfor sexual selection,but sev- out that old faces are perceivedas less attractivethan eral points raised in it are in need of furtherconsider- young faces, it is also certainlypossible to think of ation. young-lookingfaces that are far fromattractive. Even i. The nature of attractiveness.Jones claims that at- infants,who are certainlyall neotenous, show the full tractivenessis "undertheorizedin psychology."In fact, rangeof facial attractiveness. thereare two currenttheoretical perspectives on the na- 3. Measuringfaces. Jones measures relative eye width, ture of facial attractivenessthat he does not consider. relativenose height,and relativelip heightfrom photo- One perspective,consistent with Jones's approach, graphsand uses these measurementsto produce equa- centerson the importanceof facial featuresfor defining tions that predictage on the basis of the size of these attractiveness.Cunningham and his colleagues (Cun- various features.These measurementsare problematic ningham I986, Cunningham, Barbee, and Pike I990) for several reasons. First,thousands of facial measure- suggestthat attractive faces are those thatpossess a con- ments are possible (Farkas I98I), and Jones'sselection stellation of mature, neotenous, and expressivefacial ofparticular features to measureseems to be guidedonly features.Their approachinvolves measuring the sizes of loosely by a prioritheoretical considerations related to particularindividual facial features and correlatingthese neoteny.His theoreticaldiscussion of changes in facial measurementswith overallfacial attractivenessratings. structureas a functionof age does not mention cheek We have offereda differenttheoretical approach in width,yet he measures it and then later omits it when which we define attractivefaces as those whose facial he findsthat it is not relatedto age. Furthermore,choice configurationsare closest to the average population of cheek width as a neotenous featureis not consistent configuration(Langlois and Roggman 1990, Langlois, with the featureschosen by Cunningham(Cunningham Roggman, and Musselman I994). This approach differs I986, Cunningham,Barbee, and Pike I990). According fromthat of Cunningham and Jonesin that it is not to Cunningham, "neotenous features" include eye concernedwith particularfacial features;instead, it fo- height,eye width, nose length,nose tip width,nostril 740 CURRENT ANTHROPOLOGY Volume 36, Number S, December 1995 width,forehead height, eye placement,and eye separa- may or may not serve as foodfor thought in this area of tion. Cheek width is a "mature" featureaccording to inquiry. Cunningham's criteria.If Jonesconsiders "neotenous" i. Does attractivenessnecessarily equate with sexual and "mature" featuresto be opposite ends of a contin- attractiveness?For sexual selection (or any kind of Dar- uum of age appearance,then he should also have mea- winian selection) to operate,the ultimate outcome has suredchin area, chin length,chin width,eyebrow thick- to be differentialfitness. In the case of femaleneoteny, ness, and cheekbone prominence,the other "mature" Jones argues that this process is actualized through featuresmeasured by Cunningham et al. We wonder males' beingmore attractedto youthful-lookingfemales what resultswould be obtained if these otherpotential because, on average,these females make forreproduc- measures were used and whether a neotenous-feature tively more valuable mates. Hence, attractivenessis approach can providemore than post-hocexplanations clearlyequated here with sexual attractiveness.We are and predictions.For neotenyto be useful as a theoryof told, however,that neotenyalso makes us perceivein- attractiveness,researchers who employ it should (i) be fants, animals, and even cars as "cute"-a quality consistentin theiruse of the same theoreticallydriven closely analogous to attractivenessbut withouta sexual measures ofneotenous featuresand (2) not conveniently component. The article does not specifyhow attrac- omit featuresthat theyfind to be unrelatedto age. tiveness was definedfor the raters.If Jones's subjects Second, measurements obtained from photographs had been given a series of picturesof children(or dogs) may not be accurate or reliable (Farkaset al. I980). Evi- to rate accordingsimply to their "attractiveness,"my dence indicatesthat when measurementsare takenfrom guess is thatwe would findan age gradientrunning from faces as opposed to photographs,attractive faces are youngerto older, with babies (or puppies) being rated more likely than less attractivefaces to have facial fea- more attractive. tureswithin + i standarddeviation of the mean (Farkas, The implication is that the female-attractiveness- Munro, and Kolar I987). Farkas et al. suggestthat "the neotenyassociation reportedhere may reflect,at least in face with most measurementsin the range of + i SD part,something other than an attractiveness-fecundity may be close to the 'ideal face"' (p. I28). Additionally, relationship.I am willing to take Jones'sword that the Jones'sestimation that his stimuliwould have the facial female models' predictedage of about 7 years does not proportionsof first- and second-graderssuggests that the mean that theirfaces are identicalto those of 7-year-old equations he has developed on the basis of measure- girls,but still, the models' "age" is considerablylower ments do not accuratelyestimate neoteny. than that of an average sample of 20-year-oldwomen 4. Coda. Jones claims that across five populations whose very high fecunditymen should have been se- more neotenous faces are perceivedas more attractive. lected to find extremelyattractive. Again, might neo- However,of the 42 correlationsmeasured (not including tenyelicit somethingother than just sexual attraction? the data frompooled samples), only i i (approximately The answer may lie in Jones'sn.3, where he mentions 26%) were significantat the .os level or greater.How that neotenous featuresprobably act as a release forpa- are we to account forthe 74% of the correlationsthat rentalbehavior. Among otherthings, such behavioren- indicated no significantrelationship between attrac- tails providingfor and givingprotection to individuals tiveness and neoteny?lAlthough Jones has providedus who are relativelydependent upon oneself.It could thus with interestingcross-cultural data, thus farthese data be argued that in the past neotenous adult females raise more questions than theyanswer. benefited disproportionatelyfrom male provision- ing-in which case neotenywould not be the result of sexual selection. DANIEL PERUSSE 2. Is female physical attractivenessreally a human Departmentof Anthropology,University of Montreal, anomaly? Jones develops his argumentfor a human Montreal, Quebec, Canada H3 C 317 anomaly by pointingout that (a) in most sexually se- ([email protected]. ca). 5 v 95 lected species males show greaterdevelopment of sexual advertisingthan females and (b) this patternis reversed Jonesis to be commendedfor a particularlyclear, well- in the case of humans, where men are more concerned designed,and interestingstudy. As he pointsout, sexual than women with physicalattractiveness. This seeming selectiontheory has been neglectedin relationto human exception is explained by (c) the human female age- morphology.This paperpresents some ofthe firstempir- related variance in fecundity.Hence, female physical ical tests ofexplicit evolutionary hypotheses concerning featureslinked to high fecundityand youthwould have femalefacial attractivenessand providesreasonable evi- evolved to be attractiveto males-that is, would have dence in supportof the author'sclaim that neotenyacts become criteriafor mate choice by males. If we really as a kind of "superstimulus"signaling female fecundity. want to findsome originalityin our species, I thinkwe Finding Jones's study generallysound and his results ought to look at a ratherthan b: as a sexually selected plausible, I will limit myselfto raisingtwo issues that species, human males are exceptional in the degreeto which male-male competitionis played out not in sex- I. If we restrictour comparisonto femalemeasures, of the 28 ual advertising but in control over resources (e.g., correlationsmeasured only i i, roughly40%, are significantat the PerusseI993, I994). Concerningb and c, however,the .os level or better. human situationfits strikingly well what seems a quasi- JONES Sexual Selection and Facial Neoteny I 74I universal pattern:females generallydo advertisetheir face. Even children,male or female,are more interested currentfecundity through a host of visual, olfactory, in children'sfaces than in adult ones. The typicalreac- postural,behavioral, and other cues. In nonhuman pri- tion to the Kindchenschema is a wish to protectand mates, forexample, female physical signalingof fecun- pamperthe object,whether it is Donald Duck, a kitten, dity has been observedin at least 55 species out of 79 or a baby. Since "nurturingbehaviour" is a component (reviewedin BlafferHrdy and WhittenI987). Such sex- of courtingin animals and humans (e.g.,birds feed each ual advertisingincludes vulval or labial swelling and otherduring courtship), this could have entailed a pref- reddening/pinkening/whitening,clitoral reddening, per- erencefor childlike craniofacial features. Konrad Lorenz ineal swellingand reddening,face reddening,chest blis- once said that in his opinion the Neanderthalswere not tering,and others.I fail to see the fundamentaldiffer- exterminatedbut simply found sapiens sapiens cuter ence between the human case and that of otherspecies. and preferredthem as mates. The preferencefor neote- The anomaly is only apparent,stemming from the fact nous-faced mates might thereforebe found in males that human females go into estrus with much greater and females,although it mightnot be so markedin the frequencyand asynchronythan their animal counter- tradition-boundfemales. Another approach to the ex- parts-to the extent that men are likely to have been planation of the preferencefor the Kindchenschemain selected to find physical featuresof repeatedlyfecund the male mate-selectionprocess is the trainedeye of the womenpermanently attractive. Ultimately, however, in hunter.Our perceptionof beauty and symmetryis based all these species males are attractedby those female on the necessityfor the hunterto assess his prey.Why physical traitsthat stronglycovary with fecundity. do we preferto bite into a crunchy,juicy apple thaninto A phenomenon that might come closer to a human a shriveledone that may have lost none of its flavor? anomaly would be the differentialattraction of men to Fromthe day we are bornwe are told thatnew things women who possessed attractivephysical featuresin are betterthan old ones and that we should exchange differentdegrees notwithstandingthe fact that they old fornew immediately.Fromm (I992) points out that were of equal fecundity(i.e., same-agebut differentially we regardour partnersas consumergoods and tryto get neotenousand thereforedifferentially attractive). Blaffer the best we can affordon the partnermarket in relation Hrdy and Whitten (I987), however, point out that at to our own market values. Since Sigall and Landy's least fourprimate species show variance in female sex- (I973) work we have known that men are judgedby the ual advertising.In threeof those species, the varianceis appearanceof theirspouses. This is why women in par- age-related.It may be of interestthat the firsttwo are ticular become status symbols; it goes without saying crab-eatingand Japanese macaques and the third is thata man can expresshis highstatus only with a young Homo sapiens. and good-lookingpartner. In nonhuman primates, as Goodall (i99I) points out, males preferexperienced females to young ones. The preferencefor childlike fea- BARBARA SCHWEDER tures seems to be a human characteristic.Perhaps this Tulbingerkogel56, A-3001 Mauerbach, Austria. matingstrategy has reproductivevalue only in monoga- 7 III 95 mous species. It would be interestingto know whether monogamous animals have developed a strategyfor The patternof the ideal partnerhas various components, judgingthe ages of potentialpartners. some innate,some imprinted,and some activelylearned fromthe environment.It is not easy to keep them apart and to focus on a single one. In everypopulation there DONALD SYMONS are tendencies toward maintainingtraditions and ten- Departmentof Anthropology,University of California, dencies toward trying new strategies. Females are Santa Barbara, Calif. 93I06, U.S.A. io iv 95 known to be more tradition-boundthan males in the mate-selectionprocess. Some female birds are said to Jonesmakes a lucid, succinct, and persuasive case for have an innate partnermodel whereas the males of their an adaptationistperspective on human physical attrac- species are imprintedafter birth (Daly and Wilson I983). tiveness. Although my comments are informedby the And in humans it seems that females are more content same perspective,I will proposesome alternativeinter- with the averagemale face thanmales are with the aver- pretations(see Symons I995 fora morethorough presen- age female face. Thelen (I983) speaks of minority-type tation and forreferences).' human mate preferenceon the partof men wherephysi- Jones's research is a significantcontribution to the cal appearanceis concerned.Jones's data leave no doubt "adaptationistprogram," the goal of which is to parti- that human males are more attractedby neotenouspro- tion organisms into functional components-that is, portionsin femalefaces than by the averagefemale face, into special-purposeproblem-solving devices. Neverthe- but I do not thinkit is sufficientto explain this in terms less, his main hypothesis-that the human male's pref- of its indication of fecundity. erence for "neotenous" female facial featuresis a by- The response to a childlike appearance (Kindchen- productof selection forpreferring youth-is in a sense schema) is innate. Eibl-Eibesfeldt(I984) points out that less adaptationistthan "constraintist."It does not imply its signal characteris based on the differencebetween its characteristicfeatures and those of the averageadult i. I thankNancy Etcoff and Yonie Harrisfor their metacomments. 742 CURRENT ANTHROPOLOGY Volume 36, Number S, December 1995

that in ancestral human populations neotenous facia exists in the population's gene pool that causes its bear- features per se indexed relatively high female mat ers to preferX on average. Now, the myriadvagaries value. of ontogeneticdevelopment essentially guarantee that If the human male's preferencefor neotenous facia individuals bearing this allele will, in fact, exhibit a featuresis merely a by-product,it presumablywoul range ofpreferences distributed around X. Selectionwill have entailed at least some costs in ancestral popula favorthe allele if deviations fromX in eitherdirection tions. For example, assuming that Jones'shypothesis i depressfitness to the same extent.But thisis not always correct, an ancestral male given the opportunityt the case. choose between two potential mates of the same agc An "asymmetricalfitness distribution" exists when- one of whom (A) had a more neotenous face than th ever a unit of deviationfrom the optimumin one direc- other(B), would have been willingto pay a higherbride tion depressesfitness more than does a unit of deviation pricefor A because of her more attractiveface, althoug] in the other.Extreme examples have been referredto as B, at a lower bride-price,would have representedbettc the "cliff-edgeeffect": The most nutritiousgrass (the value; or he might have failed to acquire B's superic theoreticaloptimum) may grow rightat the edge of a weaving skills, which would have contributedsomc steep cliff,but the cow that always tries to graze there thingto his fitness,and instead acquired A's more grac is unlikelyto have the highestfitness in the herd.When ile jaw, largereyes, smallernose, and fullerlips, whicl an asymmetricfitness distribution exists, selection can accordingto the by-producthypothesis, would have cor be expectedto favora preferencethat divergesfrom the tributednothing; or he mighthave chosen an older fe theoreticaloptimum (away fromthe cliffedge). For ex- male with neotenous featuresover a youngerfemale (c ample, perhapsin ancestralhuman populationsfemales highermate value) with averagefeatures. with relativelygracile jaws (and, hence, relativelyshort While many factors can constrain selection fror lower faces and large eyes) typicallyhad slightlyless achievingdesign perfection, it is nonethelessworth ask mate value than females with average jaws, but selec- ing, when consideringany given constraintisthypothc tion nonetheless favored males who preferredgracile sis, why selection didn't do better.If neotenous facio ratherthan averagejaws because the fitnesspenalty for featuresin themselvesdid not index relativelyhigh fe choosing females with robust jaws was comparatively male mate value, why didn't selection favor male large. whose preferenceswere a functionof veridical age cues As Jones's theoreticalpresentation makes clear, the And why are only certain facial features,rather thai general hypothesisthat human males evolved species- all, most attractivewhen neotenous?In short,althougl typical preferencemechanisms designed to assess fe- Jones'shypothesis may well prove to be correct,mor male mate value does not imply that males universally relentlessly adaptationist possibilities are worth ex will develop the same absolute standardsof female at- ploring. tractiveness.The ideal stimulus values of some female Perhapsthe facial proportionsthat Jonesinterprets ii phenotypicfeatures can be specifiedin an absolute sense terms of age cues also indexed some other aspect(s) c (e.g., unwrinkledand unblemished skin), and these are female mate value. One possibilityis hormonalstatus likelyto be perceivedas attractiveuniversally. The ideal which Jonesconsiders unlikely. Yet high androgenlev stimulus values of otherphenotypic features, however, els in women are positivelycorrelated with reproductiv, cannot be specifiedin an absolute sense because they system dysfunctions,and observable indices of higi varyamong human populations (e.g., skin color),hence androgenlevels-such as acne, hirsutism,and a higi the attractivenessof these featuresis likely to be as- waist-to-hipratio-seem to be systematicallyperceivec sessed relative to local phenotypesrather than abso- as unattractive.To my eye, the faces in Jones'sfigure lutely.For example,a psychologicalmechanism that in- appear to differmore in "masculinity"than in age. stantiatedan obligatepreference for a specificskin color Also, there is a clear adaptationistrationale for ex could not possibly have been universally adaptive pecting female mate value in ancestral populations t amonghumans. What could have been universallyadap- have been a negative functionof parity,and certainfa tive, however, is a mechanism that calibratedprefer- cial proportionsmay index parity.Maternal bone forma ences facultatively,using as input informationabout tion ratesare elevated duringpregnancy, which may per local skin colors (which in ancestral populations rep- manently lengthen the mother's face, and a growtl resentedadaptations to local conditions). hormone (hGH-V) is expressedin the placenta and se A species-typicalmale psychologicalmechanism that creted in large amounts into the maternal circulation instantiatesthe rule "Preferfemale skin that is a bit which may permanently"coarsen" her facial features. lighter than the adult female average" (in ancestral Jonesproposes that neotenous female facial feature populations relative lightnessprobably signified nubil- are supernormalstimuli, and he may be right.But ther, ity, nulliparity,and high estrogenlevels) would result is a more general (and more adaptationist)reason tha in very differentabsolute skin color ideals in Nigeria a "preferencemechanism"-in any species and in an, and Norway. Nigerian men would perceive Norwegian domain-might be shaped by selection to identifya women as much too light,and Norwegian men would ideal a stimulus other than the theoreticaloptimum perceive Nigerian women as much too dark. This line Suppose that, in a particularcase, the theoreticalopti ofreasoning also implies,however, that men m4ysome- mum stimulusvalue is X. Furthersuppose thatan allele times perceive women of anotherpopulation as more JONES Sexual Selection and Facial Neoteny | 743 attractive,in certainrespects, than women of theirowr tomical landmarks and demonstratemoderate correla- population.A classic example is reportedby Wagatsumc tions between these indices and ratingsof attractive- (in the paperJones cites). On firstcontact, Japanese mer ness. Althoughthe averagenesseffect is importantand perceivedwhite Westernwomen as less physicallyat deserves theoreticaltreatment in its own right(Koes- tractivethan Japanesewomen in most features,includ lag I990, Langloisand RoggmanI990), thereis also ing skin texture,facial hair, and eye color. But the mer abundantevidence, at least forfemales, that faces close perceivedWestern women's typical skin color as morn to the population average are not the most attractive attractive,because it was a bit lighterthan the adul possible faces (Alley and Cunningham i990 and refer- Japanesefemale average and, hence, close to theirideal ences in appendix). If thereis significantinterpopulation variation in fa A differentstrategy is necessaryto test the neoteny cial proportions,the perceptionof neoteny may be anal hypothesis.In this case it is appropriate,contra Mus- ogous to the perceptionof skin color. That is, humar selman et al., to "omit features... not correlatedwith males may have been selected to preferfemale facia: age." It also helps if informationfrom different features features that are relatively neotenous, by local stan can be combinedto providean overall index ofneoteny. dards,rather than to prefercertain absolute facial pro In studies i and 2 I focus on just threemeasures of facial portions.If so, males will not necessarilyprefer femalc proportions(relative eye width,relative nose height,and featuresthat are neotenous by the standardsof ever) relative lip height)because there are stronggrounds in human population. Surely it is possible fora woman'" the standardliterature on facial agingfor expecting the eyes to be too large,her lower face too short,her nose relative dimensions of the three major facial features too small, and her lips too full (imagineBetty Boop as - (eyes,nose, and lips) to change with increasingage and real woman). In fact,Jones's data imply a ceiling effec consistent evidence in my samples that the measures forthe attractivenessof facial neotenyeven withinpop do change with age in the expectedfashion. Alternative ulations. measures of the relativedimensions of the threemajor In sum, I propose that accurate predictionsabout a facialfeatures are less satisfactory.Eye heightand width given male's perceptionsof female facial attractiveness of mouth do not decrease significantlywith increasing can be derived only from(a) knowledge of the designs age. Eye heightgives some signs of beingpositively cor- of species-typicalpreference mechanisms in the human relatedwith ratingsof female attractiveness across sam- male brain and (b) knowledge of the female faces thai ples but may be as much an expressivefeature as a neo- the male has been exposed to (because informationcon tenous one. Nose widthis correlatedwith age; including tained in these faces will have calibratedsome of his it in age regressionsfor females produces an age pre- preferencemechanisms). Intrapopulation agreement and dictorequation similar to Equation 2.1 The correspond- interpopulationdisagreement in attractivenessratings ing index of neoteny predicts attractivenessabout as are not evidence that people "learn" standardsof attrac- well as Equation5 (JonesI994). tivenessfrom one another-any more than intrapopula- Pooling of correlationcoefficients, as in table 2, is a tion homogeneityand interpopulationheterogeneity in standardtechnique fortesting hypotheses across multi- skin color is evidence thatpeople "learn" theirskin col ple samples. What mattersis not the numberof samples ors fromone another. for which there are significantresults but the consis- tency with which males in five populations of raters preferneotenous featuresin the pooled samples. Whyneoteny? Age, fecundity, and sensorybias. I have Reply proposed not only that there is a connection between facial neoteny and female attractivenessbut that this connectionis a by-productof the universalmale percep- DOUG JONES tion thatyounger women are more physicallyattractive Ithaca,N.Y. 14853, U.S.A. I5 vi 95 than older women and that esthetic responsesto signs of aging are largely genetic adaptations to age-related Is therea link betweenneoteny and attractiveness?Vir- changes in fecundity.Several commentatorspoint out tually all commentatorsagree that the evidence pre- potential extensions or limitations of this line of ar- sentedin this papersupports the propositionthat female gument. facial neoteny is linked to male perceptionsof attrac- I argue that human beings are anomalous in that tiveness. The main dissent comes from Musselman, males are more concerned with female attractiveness Langlois, and Roggman,who point to evidence for aD than vice versa, but Perusse points out that males in averageness or "prototype" effect:faces with propor- many primatespecies are attentiveto cyclical and sea- tions close to those of the averageface are perceivedas sonal indicatorsof female fecundity.His point is well more attractive. In testing this hypothesis it makes taken. A more complete treatmentof the evolutionary sense to incorporateinformation from as many facial basis of attractionto markersof youth and fecundity measurementsas possible. Thus Jonesand Hill (I9931 TATt-VlIIA nr%-n.;i:Ar n a lt n -iPr. r e.hQP1 - ih p-lr. rofop-.t11.Q and Jones(n.d.) constructindices of the "averageness" of facial proportionswhich combine several hundred i. The equationis Age = - I68 * log[EW/FH]+ 87 * log[NH/FH] measurementsof distances between photographicana- + 9 5 *log[NW/FH]. 744 CURRENT ANTHROPOLOGY Volume 36, Number S, December 1995

Laland offerssome importantcaveats about applying they are also commonly attractedto secondarysexual evolutionarytheory to human psychology.Understand- charactersin the female figurethat appear at the time ing how the human body and brainhave been shaped by of puberty. natural selection oftenrequires inferences about selec- Finally,Symons proposes that the traitsI have labeled tive pressuresvery differentfrom those which most of neotenous may actually carryinformation about female us face today.At worst,researchers may select post hoc fecundityover and above the informationthey carry fromamong alternativemodels of Pleistocenehuman or about age. In my paper I expressmild skepticismabout otherape, primate,or social carnivorebehavior which- this possibility,partly because there is no evidence in evermodel best fitssome preferredadaptive story. How- my samples that indices of facial neotenyare correlated ever,we are on particularlysafe ground in makinginfer- with potential non-age-relatedindicators of fecundity ences about the selective pressureson mate choice with such as age at menarcheand waist-to-hipratio. But Sy- respectto age-relatedchanges in fecundity,because de- mons's (I995) recentwork on this subjecthas persuaded mographic evidence suggests that humans have faced me thatwe need directtests of the possibilitythat estro- much the same relationshipbetween age of mate and gen/androgenratios and parityhave effectson facial at- relative probabilityof conception not just throughthe tractivenessover and above the effectsof aging.2 Pleistocene but rightup to the demographictransition Attractivenessand morphological evolution. Brace of the past few centuries. suggestsseveral reasons fordoubting that attractionto Symons notes that according to my hypothesisat- neotenous featurescould have had any effecton human tractionto neotenous facial featuresis not adaptive per morphologyvia sexual selection. He claims that (i) in se but a by-productof attractionto youngerwomen. It the evolutionarypast all women had the opportunityto is a result,in otherwords, of "sensorybias." This is not reproduce,so therewas no scope forsexual selectionon an argumentagainst the hypothesisif we accept that females; (2) some changes in skull morphologyare not naturalselection will rarelybuild ioo% accuratepercep- readilyvisible and could not have been targetsof sexual tual systemsand thereforesensory bias will be ubiqui- selection; (3) Shea has demonstratedthat neotenyis "a tous. Sensory systems commonly respond to stimuli bankruptconcept" in the contextof human evolution; other than those to which they are adapted and often and (4) the "probable mutation effect"provides a suffi- respond especially stronglyto "supernormal" stimuli cient explanation for declining robustness since the thatpresent the distinguishingfeatures of a targetstim- Middle Pleistocene. ulus in exaggeratedform (Ryan I990, Enquist and Arak Brace's claim that all femaleshave the same opportu- I993). The argumentfrom asymmetrical fitness func- nity to reproduceis refutedby an immense body of re- tions is not an alternativeto the sensory-biasargument search in demographyand reproductiveecology (Bon- but a variationon it. It assumes evolutionaryconstraints gaarts I983, Ellison I990). Female fecundityshows such that natural selection cannot build sensory sys- considerablevariation both within and between socie- tems that always respond to stimulus X but never to ties; levels of ovarian functionand probabilitiesof con- nonadaptiveor less adaptive stimuli similarto X. ceptiondepend on a numberof environmentalvariables, Whyneoteny? Alternative explanations. Several com- includingfrequency of intercourse,nutrition, work ef- mentatorsconsider alternativeexplanations for the at- fort,and pathogenlevels. Additionalvariance in female tractionto neoteny.One popular explanationfor men's fitnessresults fromvariance in mortalityand offspring attractionto youthfulfeatures is that such featuresare survivorship.In turn,frequencies of intercourse,nutri- signs of powerlessness and submissiveness. I review tional levels, and infantmortality risks commonlyde- some reasons forbeing skeptical of this possibility,and pend in parton women's success in findingand keeping Jankowiakoffers cogent ethnographicgrounds for sup- desirablemates. Researchin modernindustrial societies portingthat the relationshipbetween beautyand power demonstratesan effectof attractivenesson a woman's is the opposite of that proposed by the hypothesis.My probabilityof marriageand on the social status and at- own observationsin Brazil corroboratehis account of tractivenessof her husband (reviewedin JacksonI992). sexuality in China. Far from supposing that old men Whether quantitative research in traditionalsocieties involved in relationshipswith youngwomen will typi- will show similar correlationsbetween attractiveness cally dominate theirpartners by virtueof theirage and and mating success is an open question; considerable experience,Brazilians often suggest that men in such qualitative evidence points in this direction. relationshipsare especially vulnerableto cuckoldryand Studies of assortativemating commonly show assort- economic exploitation. ment even forthe traitsnot easily assessable by poten- Perusse and Schwederpropose that attractionto neo- tial mates, such as lung volume and blood pressure tenous featuresmay be a by-productof the parentalfeel- (Spuhler I968). Presumablyassortment results not beings aroused by infantilefeatures, but this hypothesis does not seem to explainwhy men are moreconsistently to neotenous facial featuresthan women. Nor 2. None of the Brazilianor U.S. Americanfemale photographic attracted subjectshad had any childrenor reportedbeing pregnant when does it seem to explain why,although men are attracted photographed.Thus parityper se could not have contributedto to femalefaces thatpresent in exaggeratedform the fea- differencesin facialattractiveness in thesesamples, although sen- tures that distinguishyoung women fromold women, sorybias againstfacial porportions suggestive of parity cculd have. JONES Sexual Selection and Facial Neoteny 1745 cause people choose mates on the basis of lung volume anism" than shared culture, because the two popula- per se but because theychoose on the basis of correlated tions have similar facial proportionsbut little in com- traits. Choice of partnerson the basis of perceptible mon culturally(Jones and Hill I993). traitsis likely to have consequences forcorrelated im- While habituationto local physical featuresmay ac- perceptibletraits. count for a great deal of variation across populations, Shea (i988) criticizesthe hypothesisthat human be- Laland rightlynotes that many local preferencesfor ex- ings show an across-the-boardincrease in neotenyas a aggeratedcharacters cannot be explainedin thisfashion. result of a general slowdown in rates of morphological He has pioneeredmodels ofthe coevolutionof genes and change associated with delayed maturation.He also ar- culturallytransmitted standards of attractivenesswhich gues that allegedlyneotenous traitsin adult humans are are likely to illuminate cross-culturaldivergence in not generallyhomologous with similartraits in juvenile standardsof beauty. I would add only that the theoreti- nonhumanapes. This has no bearingon whetherspecific cal bare bones of these models will need to be fleshed anatomical regions have a neotenous appearance as a out with more empiricalwork on psychologicalmecha- resultof sexual selection (or othersocial selection). nisms and cultural contexts. Trends toward craniofacial neoteny in the past Future research will furtherclarify the relationship ioo,ooo years may result fromecological selection for between neotenyand facial attractiveness.More gener- increased mechanical efficiencyand reduced metabolic ally, I suggest that the study of sexual esthetics may cost,from sexual or othersocial selection,or fromdirec- develop in the same fashion as the study of language. tional mutation-Brace's probable mutation effect.Of Both language acquisition and the developmentof stan- these three possibilities, the last is probablythe least dards of sexual attractivenessare probablyregulated by plausible: a strongprediction of the last hypothesisis specialized "mental organs"shaped bynatural selection. increasingtrait variance over time, but Frayer(I977) But both are presumablyinfluenced as well by nonadap- findsthe opposite trend,toward reduced variance. tive sensoryand cognitivebiases and by social factors, Whethersexual selection has contributedto the evo- including perceptionsof what is popular and what is da atPAcl Qtdntl Q i n i tQ Q% 1Q 1 i la" lution of human craniofacialneoteny is an open ques- cwri wiAth nTVvi m .i nn tion. Brace providesno good reason to doubt it. Instead, his replydemonstrates a failureto take sexual selection seriously. Admittedly,this failure puts Brace in good company.When Darwin attributedthe showy plumage ReferencesCited of some male birds to female choice, Wallace proposed thatsuch plumage resultedinstead from the overflowof ALLEY, T. R., AND M. CUNNINGHAM. I99I. Averagedfaces surplus energy.Julian Huxley arguedthat the elaborate are attractive,but veryattractive faces are not average.Psycho- courtshipdances of some birds were not a product of logical Science e2:I 23-25. sexual selection but a by-productof emotional excite- ANDERSSON, M. I994. Sexual selection.Princeton: Princeton UniversityPress. ment.Mayr suggested that the extraordinarycomplexity BARASH, D. P. i982. Sociobiologyand behavior.New York: and diversityof male genitalia in a number of groups ElsevierNorth Holland. [LEM, JHL, LAR] resultednot fromfemale choice but frompleiotropy- BARTLEY, M. M. I995. Courtshipand continuedprogress: Julian that genes selected fortheir effects in otherparts of the Huxley'sstudies on birdbehavior. Journal of theHistory of Bi- body produced increased genitalic complexityas a side ology28:9i-i08. BEHRENTS, R. G. I985. Growthin the agingcraniofacial skele- effect.More recentresearch has made it clear that Wal- ton.Ann Arbor:Center for Human Growthand Development, lace, Huxley, Mayr, and many other major figuresin Universityof Michigan. evolutionarytheory had a blind spot about sexual selec- BERNDT, R. M. I95 I. Sexual behaviorin westernArnhem Land. VikingFund Publications in AnthropologyI6. tion(Eberhard I985, Cronini99i, BartleyI995). A simi- BERRY, D. S. I99I. Attractivefaces are not all createdequal: lar blind spot seems to have kept many biological an- Jointeffects of facial babyishness and attractivenesson social thropologistsfrom seriously investigatingthe sexual perception.Personality and Social PsychologyBulletin selection of human morphology. I7:523-3I. [LEM, JHL, LAR] Attractiveness,learning, and culture.Symons and La- BERRY, D. S., AND L. Z. MC ARTHUR. I986. Perceivingcharac- land consider the topic of variation in standardsof at- terin faces:The impactof age-relatedcraniofacial changes on social perception.Psychological Bulletin I00:3-I8. tractiveness.Symons notes that if people calibratetheir BERSCHEID, E., AND E. WALSTER. I974. Physicalattractive- standardsof attractivenessto the local population aver- ness. Advancesin ExperimentalSocial Psychology7:I57-2I5. age-for example, if the most attractiveface is a neote- BETZIG, L., P. TURKE, AND M. BORGERHOFF MULDER. Edi- nous version of the average face-then standardsof at- tors.I988. Human reproductivebehavior: A Darwinianper- spective.New York: CambridgeUniversity Press. tractivenesswill vary with population differencesin BLAFFER HRDY, S., AND P. L. WHITTEN. I987. "Patterningof facial proportions.Such variationwill involve learning sexual activity,"in Primatesocieties. Edited by B. B. Smutset but not culture. Consistent with this possibility,Ache al. Chicago:University of ChicagoPress. [DP] and Hiwi show strongeragreement with each otherthan BONGAARTS, j. I983. Fertility,biology, and behavior:An analy- with Brazilians, U.S. Americans, or Russians in stan- sis of theproximate determinants. New York:Academic Press. BOONE, S. A. I986. Radiance fromthe waters:Ideals offemi- dards of facial attractiveness.This agreementis more nine beautyin Mende art and culture.New Haven: Yale Uni- likelyto reflectthe operationof a "face-averagingmech- versityPress. 746 1 CURRENT ANTHROPOLOGY Volume 36, Number 5, December 1995

BOYD, R., AND P. J. RICHERSON. I985. Cultureand the evolu- ENLOW, D. H. I990. Facial growth.Philadelphia: Harcourt Brace tionaryprocess. Chicago: University of ChicagoPress. [KNL] Jovanovich. BRACE, C. L. I995. 5th edition.The stagesof human evolution. ENQUIST, M., AND A. ARAK. I993. Selectionof exaggerated EnglewoodCliffs: Prentice-Hall. [CLB] traitsby femaleaesthetic senses. Nature 36i:446-48. . n.d. Derivingthe quick fromthe dead: Bio-culturalinter- EUBA, T. i986. The humanimage: Some aspectsof Yoruba can- actionand the mechanismof mosaic evolutionin the emer- ons of artand beauty.Nigeria Magazine 54:9-2I. genceof "modern"morphology. American Anthropologist. In FARKAS, L. G. i98i. Anthropometryof thehead and face in press.[CLB] medicine. New York: Elsevier. BRACE, C. L., AND P. E. MAHLER. I97I. Post-Pleistocene FARKAS, L. G., W. BRYSON, B. TECH, AND J. KLOTZ. ig80. Is changesin the humandentition. American Journal of Physical photogrammetryof the face reliable? Plastic and Reconstruc- Anthropology34:I9I-203. tiveSurgery 66:346-55. [LEM, JHL, LAR] BRACE, C. L., S. L. SMITH, AND K. D. HUNT. I99I. "Whatbig FARKAS, L. G., I. R. MUNRO, AND J. C. KOLAR. i987. "Linear teeth you had Grandma! Human tooth size past and present," proportionsin above-and below-averagewomen's faces," in in Advancesin dentalanthropology. Edited by M. A. Kelley Anthropometricfacial proportions in medicine.Edited by L. G. and C. S. Larsen,pp. 33-57. New York:Wiley-Liss. [CLB] Farkas and I. R. Munro, pp. ii9-29. Springfield:Thomas. BULL, R., AND N. RUMSEY. Editors.I988. The social psychol- FAUSS, R. i988. Zur Bedeutungdes Gesichtsfur die Partner- ogyof facial appearance.New York:Springer-Verlag. wahl. Homo 37:I88-2oi. BUSS, D. M. I989. Sex differencesin humanmate preferences: FEINGOLD, R. B. I990. Genderdifferences in effectsof physical Evolutionaryhypotheses tested in 37 cultures.Behavioral and attractivenesson romanticattraction. Journal of Personality BrainSciences I2:I-49. and Social Psychology 59:98I-93. CAVALLI-SFORZA, L. L., AND M. W. FELDMAN. I98I. Cul- FORD, C. S., AND F. BEACH. i95i. Patternsof sexual behavior. turaltransmission and evolution:A quantitativeapproach. New York: Harper. Princeton: Princeton University Press. [KNL] FRAYER, D. W. I977. Metricdental change in the EuropeanUp- CHAGNON, N. A., AND W. IRONS. Editors.I979. Evolutionary per Paleolithicand Mesolithic.American Journal of Physical biologyof human social behavior:An anthropologicalperspec- Anthropology46:iog-2o. tive.North Scituate: Duxbury Press. . i98i. Bodysize, weaponuse, and naturalselection in the CLUTTON-BROCK, T. H.,AND G. A. PARKER. i992. Potential EuropeanUpper Paleolithic. American Anthropologist 83: reproductiverates and the operation of sexual selection. Quar- 57-73. terlyReview of Biology 67:437-56. FROMM, E. I992. 4th edition.Die Kunstdes Liebens.Muinchen: COSMIDES, L., AND J. TOOBY. I987. "From evolution to behav- Ullstein. [BS] ior: Evolutionary psychology as the missing link," in The lat- GOLDMAN, N., AND M. MONTGOMERY. i989. Fecundability est on the best: Essays on evolutionand optimality.Edited by and husband'sage. Social Biology36:I46-66. J.Dupre. Cambridge:MIT Press.[KNL] G OO D A L L, j. I 9 9 I. Ein Herz furSchimpansen. Reinbeck bei CRONIN, H. I99I. The ant and thepeacock: Altruismand sex- Hamburg:Rowohlt. [BS] ual selectionfrom Darwin to today.New York:Cambridge GOULD, S. J. I977. Ontogenyand phylogeny.Cambridge: Belk- University Press. nap Press of Harvard University Press. [CLB] . I993. Oh, those bonobos! Are fqmale primates the proto- GOWATY, P. A. i992. Evolutionary biology and feminism. Hu- typicalliberated women? Review of Female choices,by Mere- man Nature 3:2I7-49. dith Small (Ithaca: Cornell University Press, I993). New York GREGERSEN, G. i983. Sexual practices:The storyof human sex- TimesBook Review,August 29, p. I9. [CLB] uality. New York: F. Watts. CUNNINGHAM, M. R. I986. Measuringthe physicalin physical GREGOR, T. i985. Anxiouspleasures: The sexual lives of an attractiveness:Quasi-experiments on the sociobiologyof fe- Amazonianpeople. Chicago:University of ChicagoPress. male facialbeauty. Journal of Personality and Social Psychol- GRINKER, R. R. I990. Imagesof denigration: Structuring inequal- ogy50:9 25-35. [LEM, JHL, LAR] itybetween foragers and farmersin the IturiForest, Zaire. CUNNINGHAM, M. R., A. P. BARBEE, AND C. L. PIKE. I990. AmericanEthnologist I7:III-30. Whatdo womenwant? Facialmetric assessment of multiple HATFIELD, E., AND S. SPRECHER. i986. Mirror,mirror: The im- motivesin the perceptionof male facialphysical attractive- portanceof looks in everydaylife. Albany: State University of ness. Journalof Personalityand Social Psychology59:6I-72. New York Press. [LEM, JHL, LAR] HENRY, L. i96i. Some data on natural fertility.Eugenics Quar- DALY, M., AND M. WILSON. I983. Sex, evolution,and behav- terly8:8I-9I. ior. Boston: Willard Grant Press. HOSIE, J. A., H. D. ELLIS, AND N. D. HAIG. i988. The effect DARWIN, C. R. I98I (i871). The descentof man, and selection of feature displacement on the perception of well-known faces. in relation to sex. Princeton: Princeton University Press. Perception I7:46I-74. [LEM, JHL, LAR] DEGLER, C. N. I99I. In searchof human nature: The decline HULSE, F. S. i967. Selectionfor skin color among Japanese. and revivalof Darwinism in Americansocial thought.New AmericanJournal of Physical Anthropology 27:I43-5 9. York: Oxford University Press. IMMELMAN, K. I972. Sexual and otherlong-term aspects of im- DOBZHANSKY, T. I970. Geneticsof the evolutionaryprocess. printingin birdsand otherspecies. Advances in theStudy of New York: Columbia University Press. [LEM, JHL, LAR] Behavior 4:I47-74. . I972. "Genetics and the races of man," in Sexual selec- JACKSON, L. A. i992. Physicalappearance and gender:Sociobio- tionand the descentof man I87I-I97I. Editedby B. Camp- logical and socioculturalperspectives. Albany: State Univer- bell,pp. 59-86. Chicago:Aldine. [CLB] sityof New YorkPress. DURHAM, W. I99I. Genes,culture, and human diversity.Stan- JAMES, W. H. I979. The causes ofthe declinein fecundability ford:Stanford University Press. [KNL] withage. Social Biology 26:330-34. EBERHARD, W. G. I985. Sexual selectionand animalgenitalia. JANKOWIAK, W. I993. Sex, death,and hierarchyin a Chinese Cambridge: Harvard University Press. city:An anthropologicalaccount. New York:Columbia Uni- EIBL-EIBESFELDT, I. I984. Die Biologiedes menschlichenVer- versityPress. haltens.Miunchen: Piper. [BS] JOHNSTON, T. D. i982. Selectivecosts and benefitsin the evolu- ELLIS, H. i926. Studiesin thepsychology of sex. Vol. 4. Philadel- tion of learning.Advances in the Studyof Behavior i2:65- phia: F. A. Davis. io6. ELLISON, P. I990. Human ovarianfunction and reproductive JOHNSTON, V. S., AND M. FRANKLIN. I993. Is beautyin the ecology: New hypotheses. American Anthropologist92: eye of the beholder?Ethology and Sociobiology I4:i83-99. 933-5-2. JONES, D. I994. The evolutionary psychology of humfn physical JO NE S Sexual Selection and Facial Neoteny 1747

attractiveness:Results from five populations. Ph.D. diss.,Uni- aryhypotheses with behavioral data. Human Nature5:255-78. versityof Michigan,Ann Arbor,Mich. [DP] . n.d. Physicalattractiveness and the theoryof sexual se- POLHEMUS, T. I988. Bodystyles. Bedfordshire: Lennard. lection:Results from five populations. Ann Arbor: Museum of P RUE T T - JONES, S. I99 2. Independentversus nonindependent Anthropology,University of Michigan. In press. mate choice: Do femalescopy each other?American Natural- JONES, D., AND K. HILL. I993. Criteriaof physical attractive- ist I40:I000-I009. ness in fivepopulations. Human Nature 4:27i-96. PURCELL, D. G., AND A. L. STEWART. I988. The face- KENRICK. D. I994. Evolutionarysocial psychology:From sexual detectioneffect: Configuration enhances detection. Perception selectionto social cognition.Advances in ExperimentalSocial and Psychophysics43:355-66. [LEM, JHL, LAR] Psychology3I:75-I2I. RHODES, G. I986. Memoryfor lateral asymmetries in well- KOESLAG, J. I990. Koinophiliagroups sexual creaturesinto spe- knownfaces: Evidence for configural information in memory cies, promotesstasis, and stabilizessocial behavior.Journal of representationsof faces. Memory and Cognition I4:209-I9. TheoreticalBiology I44:I5-35. [LEM, JHL, LAR] LALAND, K. N. I994. Sexual selectionwith a culturallytransmit- RIEDL, B. I. M. I990. Morphologisch-metrischeMerkmale des ted matingpreference. Theoretical Population Biology mannlichenund weiblichenPartnerleitbildes in ihrerBedeu- 45:I-I5. tungfur die Wahl des Ehegatten.Homo 4I:72-85. LANGLOIS. J. H., AND L. A. ROGGMAN. I990. Attractivefaces ROSENBLATT, P. C. I974. Foundationsof interpersonal at- are onlyaverage. Psychological Science i:ii5-2i. [LEM, JHL, traction. New York: Academic Press. [KNL] LAR] RYAN, M. J. I990. "Sexual selection,sensory systems, and sen- LANGLOIS, J. H., L. A. ROGGMAN, R. J. CASEY, AND J. M. soryexploitation," in Oxfordsurveys in evolutionarybiology, RITTER. I987. Infantpreferences for attractive faces: Rudi- pp. I57-95. Oxford:Oxford University Press. mentsof a stereotype?Developmental Psychology 23:363-69. .I995. Offsettingadvantages. Review of Sexual selection, LANGLOIS, J. H., L. A. ROGGMAN, AND L. MUSSELMAN. by M. B. Andersson(Princeton: Princeton University Press, I994. Whatis averageand whatis not averageabout attractive I994). Science 267:7I2-I3. faces?Psychological Science 5:2I4-20. [LEM, JHL, LAR] S E L I G M A N, M. E. P. I970. On the generalityof the laws of LARRABEE, W. F., JR., AND K. H. MAKIELSKI. I993. Surgical learning.Psychological Review 77:406-I8. anatomyof theface. New York:Raven Press. SERGENT, J. I984. Configuralprocessing of faces in the leftand LORENZ, K. I943. Die angeborenenFormen moglicher Erfah- the rightcerebral hemispheres. Journal of Experimental Psy- rung.Zeitschrift far Tierpsychologie5:235-409. chology:Human Perceptionand PerformanceIO:554-72. [LEM, MC CABE, v. I988. "Facial proportions,perceived age, and care- JHL, LAR] giving,"in Social and appliedaspects of perceiving faces. Ed- SHEA, B. j. I989. Heterochronyin humanevolution: The case itedby T. R. Alley,pp. 89-95. Hillsdale,N.J.: Lawrence Erl- forneoteny reconsidered. Yearbook of Physical Anthropology baum Associates. 32:69-ioi. [CLB] MC CALL, R. B., AND C. B. KENNEDY. I980. Attentionof 4- SIGALL, H., AND L. LANDY. I973. Radiatingbeauty: The ef- month-oldinfants to discrepancyand babyishness.Journal of fectsof having a physicallyattractive partner on personpercep- ExperimentalChild Psychology29:I89-2oi. tion.Journal of Personality and Social Psychology28:2i8-24. MC ARTHUR, L. Z., AND K. APATOW. I983-84. Impressionsof [BS] baby-facedadults. Social Cognition e2:3I5-42. SINGH, D. I993a. Adaptivesignificance of female physical attrac- MALINOWSKI. B. i987 (i929). The sexual lifeof savagesin tiveness:Role ofwaist-to-hip ratio. Journal of Personality and North-WesternMelanesia: An ethnographicaccount of court- Social Psychology65:293-307. ship,marriage, and familylife among the natives of the Trobri- .I993b. Bodyshape and women'sattractiveness: The criti- and Islands,British New Guinea. Boston:Beacon Press. cal role ofwaist-to-hip ratio. Human Nature 4:297-32I. MARK, L. S., R. E. SHAW, AND J. B. PITTENGER. I988. "Natu- SMUTS, B. B., AND R. W. SMUTS. I993. Male aggressionand ral constraints,scales ofanalysis, and informationfor the per- sexual coercionof females in non-humanprimates and other ceptionof growing faces," in Social and appliedaspects of per- mammals:Evidence and implications.Advances in the Study ceivingfaces. Edited by T. R. Alley,pp. II-49. Hillsdale,N.J.: ofBehavior 22:I-63. LawrenceErlbaum Associates. SOKAL, R. R., AND F. J. ROLF. I969. Biometry:The principles MAYNARD SMITH, J. i982. Evolutionand the theoryof games. and practiceof statistics in biologicalresearch. San Francisco: Cambridge:Cambridge University Press. W. H. Freeman. MAYR, E. I972. "Sexual selectionand naturalselection," in Sex- SPUHLER, J. N. I968. Assortativemating with respect to physi- ual selectionand the descentof man I87I-I97I. Editedby B. cal characteristics.Eugenics Quarterly i5:128-40. Campbell,pp. 87-I04. Chicago:Aldine. [CLB] STRIER, K. B. I99 2. Faces in theforest: The endangeredmuru- MENKEN, J., J. TRUSSELL, AND U. LARSEN. I986. Age and in- qui monkeys of Brazil. New York: Oxford University Press. fertility.Science 233:I389-94. [CLB] MORTON, J., AND M. H. JOHNSON. I99I. CONSPEC and SUSANNE, C. I977. Individual age changes of the morphological CONLEARN: A two-processtheory of infant face recognition. characteristics.Journal of Human Evolution6:I8I-89. PsychologicalReview 98:I64-8I. [LEM, JHL, LAR] SYMONS, D. I979. The evolutionof human sexuality.Oxford: MUNN, N. D. I986. The fameof Gawa: A symbolicstudy of OxfordUniversity Press. value transformationin a Massim (Papua New Guinea) soci- I995. "Beautyis in the adaptationsof the beholder:The ety. New York: Cambridge University Press. evolutionarypsychology of human female sexual attractive- NEWMAN, M. T. i962. Evolutionarychanges in bodysize and ness," in Sexual nature/sexualculture. Edited by Paul R. Ab- head formin AmericanIndians. American Anthropologist ramsonand StevenD. Pinkerton,pp. 8o-ii8. Chicago:Univer- 64:237-57. sityof ChicagoPress. [DS] PATZER, G. I985. The physicalattractiveness phenomena. New THELEN, T. H. I983. Minority-typehuman mate preference.[BS] York:Plenum Press. TRIVERS, R. L. I97I. "Parentalinvestment and sexual selec- PERRETT, D. I., K. MAY, AND S. YOSHIKAWA. I994. Facial tion,"in Sexual selectionand the descentof man, I87I-I97I. shape and judgementsof female attractiveness. Nature 368: Editedby B. H. Campbell,pp. I36-79. Chicago:Aldine. 239-42. WAGATSUMA, H. I968. "The social perceptionof skincolor in PERUSSE, D. I993. Culturaland reproductivesuccess in modern Japan,"in Color and race. Editedby J.H. Franklin,pp. i29-65. societies:Testing the relationshipat theproximate and ulti- Boston:Houghton Mifflin. mate levels.Behavioral and BrainSciences i6:267-322. [DP] WEIDENREICH, F. I945. The brachycephalizationof recentman- I994. Mate choicein modernsocieties: Testing evolution- kind.Southwestern Journal of Anthropology I:I-54. 748 CURRENT ANTHROPOLOGY Volume 36, Number s, December 1995

WEINER, A. B. I976. Womenof value,men ofrenown: Newper- WESTERMARCK, E. i92i. The historyof human marriage. Vol. spectivesin Trobriandexchange. Austin: University of Texas I. London:Macmillan. Press. WILLIAMS, G. C. I975. Sex and evolution.Princeton: Princeton WEST-EBERHARD, M. J. I99I. "Sexual selectionand social be- UniversityPress. havior,"in Man and beast revisited.Edited by L. Tigerand YOUNG, A. W., D. HELLAWELL, AND D. C. HAY. I987. Con- M. Robinson,pp. I59-72. Washington,D.C.: SmithsonianInsti- figuralinformation in faceperception. Perception I6:747-59. tution. [LEM, JHL, LAR]

Calendar

I996 Fisica, Universitadi Genova, Via Balbi, 4, i6i26 Ge- nova, Italy (fax: [39] I0 2095987). January i-8. InternationalDevelopment Ethics Associ- me 13-I6. gth InternationalOral HistoryConfer- ation, 4th InternationalConference, Tamil Nadu, In- ence, Goteborg,Sweden. Theme: CommunicatingEx- dia. Theme: Globalization, Self-Determination,and perience.Deadline forabstracts December I. Write: Justicein Development. Write:Peter Penz, Faculty Sven B. Ek or BirgittaSkarin Frykman,Department of EnvironmentalStudies, York University,4700 of Ethnology,University of Goteborg,Vastra Hamn- Keele St., North York, Ont., Canada M3J IP3. gatan 3, S-4II I 7 Goteborg,Sweden. April I 8-20. Delta Blues Symposium:The Blues II, in- eptember8-14. Lithic Technology:From Raw Mate- terdisciplinaryconference, State University,Ark., rial Procurementto Tool Production,workshop in U.S.A. Deadline forabstracts January i5, I996. connectionwith the I 3th InternationalCongress of Write:Delta SymposiumCommittee, Department of the Union of Prehistoricand ProtohistoricSciences, English and Philosophy,P.O. Box I890, Arkansas Forli,Italy. Write: Sarah Milliken, c/o SegreteriaXIII State University,State University,Ark. 72467-I890, CongressoU.I.S.P.P., Via Marchesi, I, 47ioo Forli,It- U.S.A. (e-mail: [email protected] aly (fax: [39] 543 35805). [email protected]). )ctober19-20. Interface'95: 2oth Annual Humanities May 29-June 2. 3d European Ethnopharmacological and TechnologyConference, Atlanta, Ga., U.S.A. Colloquium and iSt InternationalConference of the Write:Julie Newell, Social and InternationalStud- Anthropologyand Historyof Health and Disease, ies, SouthernCollege of Technology,100 S. Mari- Genoa, Italy. Write: Comitato OrganizzazioneEtno- etta Parkway, Marietta, Ga. 30060-2986, U.S.A. farmacologiaAntropologia, Istituto di Antropologia

Errata

In Manzi and Passarello's report on the Neandertals size and shape values forthe leftthird molar of Breuil 3 fromGrotta Breuil in the April issue (CA 36:355-66), shouldhave been I I 9.9 and I I 3.9 respectively. two correctionsto table i (p. 36i) were overlooked;the