Menstrual cycle, trait estrogen level, and masculinity preferences in the human voice
Introduction
Masculine traits in men indicate long-term health (Rhodes et al., 2003), higher reproductive success (Mueller and Mazur, 1997, Mueller and Mazur, 1998, Pawlowski et al., 2000), but reduced commitment to relationships and offspring (Burnham et al., 2003, Gray, 2003, Gray et al., 2002, Gray et al., 2004). By contrast, feminine traits in men indicate a higher probability of relationship commitment and paternal investment (Burnham et al., 2003, Gray, 2003, Gray et al., 2002, Gray et al., 2004). Women exhibit stronger facial masculinity preferences during the most fertile phase of the menstrual cycle (late-follicular phase) than at other times (Johnston et al., 2001, Penton-Voak and Perrett, 2000, Penton-Voak et al., 1999). Menstrual cycle shifts in facial masculinity preferences have been observed when women evaluated men's faces for short-term relationships (Penton-Voak et al., 1999) and when relationship context was not specified (Johnston et al., 2001, Penton-Voak and Perrett, 2000, Penton-Voak et al., 1999).
Masculinity in men's face shape is preferred by women more for short-term than long-term relationships (Johnston et al., 2001, Little et al., 2002, Penton-Voak et al., 1999, Penton-Voak et al., 2003). Preferences for male facial masculinity are influenced by the attractiveness and femininity of the female judges (Little et al., 2001, Penton-Voak et al., 2003). While relatively unattractive and masculine women demonstrated stronger preferences for masculine males as short-term partners than as long-term partners, the effect of relationship context on masculinity preferences was weaker for attractive, feminine women (Little et al., 2001, Penton-Voak et al., 2003). This effect of own condition on women's masculinity preferences is thought to occur because more attractive, feminine women may be better able to obtain investment from masculine men during long-term relationships (Clark, 2004, Gangestad and Simpson, 2000, Little et al., 2001, Penton-Voak et al., 2003). Given that attractive and feminine women have more stable masculinity preferences across relationship contexts than unattractive and masculine women, attractive and feminine women should show less variation in their preferences for masculine males during the menstrual cycle than unattractive and masculine women. Fundamental frequency (an acoustic measure of voice pitch) in men is negatively related to testosterone throughout pubertal development (Butler et al., 1989, Harries et al., 1997, Harries et al., 1998) and during adulthood (Dabbs and Mallinger, 1999). Collins (2000) and Feinberg et al. (2005) found that low fundamental frequency and large apparent vocal-tract length (indicated by narrow spacing of formant frequencies) independently predicted perceived masculinity. Fundamental frequency correlated negatively with attractiveness (Collins, 2000) and correlated with perceived dominance (Tusing and Dillard, 2000) of men's voices. Enhancing masculine characteristics in voices (lowering fundamental frequency and increasing apparent vocal-tract length) using Praat audio software (Boersma and Weenink, 2001) also increased women's attributions of masculinity and attractiveness to male voices (Feinberg et al., 2005). Moreover, male vocal attractiveness is highly related to masculinity (Collins, 2000, Feinberg et al., 2005) and men with attractive voices have more mating success than men with unattractive voices (Hughes et al., 2004).
Testosterone enhances somatic tissue development (Notelovitz, 2002). Thus, vocal-tract length and testosterone are positively related (Fitch and Giedd, 1999). As vocal-tract length increases, formant dispersion decreases, closely spaced formants are associated with large body size in rhesus macaques (Macaca mulatta, Fitch, 1997), dogs (Canis familiaris, Riede and Fitch, 1999), and humans (Homo sapiens, Collins and Missing, 2003, Fitch and Giedd, 1999, Gonzalez, 2004). Other estimates of vocal-tract length predict body size in red deer (Cervus elaphus, Reby and McComb, 2003). Feinberg et al. (2005), Smith et al. (2005), and Fitch (1994) found that increasing apparent vocal-tract length in human voices increased perceived height. It is relevant here that Pawlowski et al. (2000) found that taller men had higher reproductive success.
Using voices manipulated in formant (vocal-tract length) and fundamental frequencies, we tested if women's preferences for masculine male and female voices were affected by menstrual cycle. In light of Feinberg et al. (2005) and Collins (2000), we predicted that masculinized men's voices would be preferred to feminized men's voices. Next we predicted that masculinity preferences for men's voices would be stronger when conception risk is high (late-follicular phase) than when conception risk is low (early-follicular and luteal phases). This would parallel findings for facial masculinity.
Gangestad et al. (2004) found that women's preferences for dominant behavioral displays in video clips (including voices perceived as dominant) are strongest during the late-follicular phase of the menstrual cycle. Gangestad et al. (2004) did not test for variation in women's preferences for dominant voices. Also, Gangestad et al. (2004) did not determine whether their observed cyclic shift in attraction to dominance in men was due to a change in sensitivity to dominance across the menstrual cycle (see Macrae et al., 2002), or a change in attraction to dominance across the menstrual cycle. We sought to address the above by asking women to assess attractiveness and dominance of voices across the menstrual cycle and determining if sensitivity and/or attraction to dominance change cyclically.
Penton-Voak et al. (2003) found that waist-to-hip ratio negatively predicted women's preferences for masculinity in men's faces. Waist-to-hip ratio is negatively related to estrogen level (Jasienska et al., 2004). Therefore, we can predict that between women, average levels of estrogen metabolites would positively correlate with women's masculinity preferences in men's voices.
As feminine and attractive women showed the least variation when evaluating attractiveness of masculinized faces in long-term and short-term contexts (Little et al., 2001, Penton-Voak et al., 2003), we predicted that women with high average (trait) estrogen (an index of femininity and reproductive health in women, Jasienska et al., 2004, Moran et al., 1999, Zaadastra et al., 1993) would have relatively stable preferences for masculinity across the menstrual cycle. By contrast, we predicted that women with low average estrogen would show the most marked masculinity preference change across the cycle.
We tested for menstrual cycle shifts in both men and women's voices. If cyclic shifts are linked to mate-choice then such shifts would be present for men's but not women's voices. If may be however that menstrual cycle shifts have no costs, in which case, they could occur for both sexes of voices.
Section snippets
Voice recordings
Four men's and 4 women's voices were recorded, speaking monopthong vowels “eh”, “ee”, “ah”, “oh”, and “oo” with an Audio-Technica AT4041 cardioid condenser microphone in a quiet room from a distance of approximately 20 cm. The voices were encoded directly onto computer hard disk in mono at 44.1 kHz sampling rate and 16-bit quantization using Sonic Foundry's Sound Forge 6.0. The voices, when manipulated, spanned the normal range of fundamental frequencies of adult men and women (average pitch
Results
Initially, tests for main effects of the vocal manipulations on attractiveness and dominance, collapsed across menstrual cycle phases, were evaluated using paired sample t tests. These results are displayed in Table 2. In both sexes, masculinized voices were rated more dominant than feminized voices. Women preferred masculinized men's voices to feminized men's voices. We also observed a significant preference for feminized women's voices.
Late-follicular and non-fertile E3G concentration ratios
Discussion
We found that masculinizing male voices increased their attractiveness, replicating findings by Feinberg et al. (2005). We also found that masculinized male and female voices were perceived as more dominant than feminized voices. This supports findings that pitch of voice was negatively correlated with attributions of dominance in men's voices and contradicts findings that pitch of voice and dominance ratings were not associated in women's voices (Tusing and Dillard, 2000). Our finding supports
Acknowledgments
The authors thank Lynda Boothroyd, Alexandra Boyden, D. Michael Burt, William Calderhead, Carolyn Cheetham, Sophia Durrani, Catherine Dutton, Fiona Elder, Lesley Ferrier, Suzanne Hall, Jennifer Hardingham, Laura Harper, Laura Johnson, Gillian McBride, Jennifer McChesney, Anne-Marie Morgan, Anne Perrett, Robert Pitman, Joanna Souter, Michael Stirrat, Katherine Timmins, and Susan Whiten for their assistance. We also thank the Reproductive Medicine Laboratory Staff and Martha Urquhart for their
References (50)
- et al.
Men in committed, romantic relationships have lower testosterone
Horm. Behav.
(2003) Self-perceived attractiveness and masculinization predict women's sociosexuality
Evol. Hum. Behav.
(2004)- et al.
Vocal and visual attractiveness are related in women
Anim. Behav.
(2003) Men's voices and women's choices
Anim. Behav.
(2000)- et al.
High testosterone levels predict low voice pitch among men
Pers. Individ. Differ.
(1999) - et al.
Manipulations of fundamental and formant frequencies influence the attractiveness of human male voices
Anim. Behav.
(2005) Formant frequencies and body size of speaker: a weak relationship in adult humans
J. Phon.
(2004)- et al.
Marriage and fatherhood are associated with lower testosterone in males
Evol. Hum. Behav.
(2002) - et al.
Sex differences in morphological predictors of sexual behavior—Shoulder to hip and waist to hip ratios
Evol. Hum. Behav.
(2003) - et al.
Ratings of voice attractiveness predict sexual behavior and body configuration
Evol. Hum. Behav.
(2004)