Professor of Economics and Finance

University of New Orleans

New Orleans, LA 70148

emmef@uno.edu(E-Mail)

November 4, 1995

Cowley & Brooksbank in this journal provided evidence that female exposure to androstenol increases the number and depth of female interactions with males. However, in nature, the only females customarily close enough to males to be affected by male emitted androstenol would be sleeping and cuddling partners, i.e. mates or potential mates. Thus, the observed behavioral response is probably a mechanism for promoting bonding with mates, and evaluation of potential mates. Recognizing an evolutionary purpose for the female reaction to androstenol makes one feel more confident that the observed effects are real.

Cowley & Brooksbank (1991) have provided evidence that human females exposed to 5a-16-androsten-3a-ol (androstenol) overnight interact more with males than unexposed females do. The experimental procedure involved having students wear overnight a necklace designed to release androstenol. Next morning the subjects filled in a questionnaire regarding who they had interacted with (i.e. who they had spoken with, including greetings), who initiated the interaction, and the depth of the interaction ("i.e. the extent to which he/she felt the exchange encompassed personal involvement."). No effect was found for males, or for female interaction with females, but the females reported a significant increase in interactions with males. For instance, the control females reported 3.08 interactions that morning with males, while those females exposed to androstenol overnight reported 5.42 interactions with males, a difference that had probability of occurring by chance of .024. The measure of depth rose from 34.2 to 97.4, significant at .011. The measure of duration rose from 6.09 to 26.7, significant at .001. [para 4]

The authors of the article offer very little explanation for what function the rather strong effects of androstenol serve, saying at the end merely that "it appears that exposure to androstenol induces 'approach' response in females." Earlier (p. 657), the authors had stated "clearly from a sociobiological point of view the benefit would be to the male signaler as the source of the pheromone." The purpose of this paper is to provide a plausible reason for the emergence and continuation of female sensitivity to androstenol, and for male ability to emit it. [para 5]

It does appear that human males are designed to emit androstenol (Kohl & Francoeur, 1995; Stoddart, 1990). It is produced in apocrine glands that are typically at the follicles from whence hair grows, especially under the arm and in the public regions. They appear designed to wick up the odors from the apocrine glands and distribute them to the atmosphere. This is shown by the axillary hairs being associated with more apocrine glands than hairs in other parts of the body. The musky odor produced from the axillae is caused by the presence there of androstene and 5a-androstenol. Androstenol is known to be a sexual pheromone in the pig which causes the sow to assume the position required for mating. Men produce much more androstenol than do women in their urine (Brooksbank, 1962) and more androsterone in their axillae (Bird, & Gower, 1982, as cited by Stoddart, 1990, p. 66). [para 6]

While some human hairs can be argued to serve various purposes, such as protecting the head from the sun, or sexual signaling, or reducing friction during sex (pubic hairs), the underarm hairs are not located to keep us warm, nor to protect us from the sun, or even to be visible at a distance. The only plausible purpose appears to emit odors or pheromones. They do this efficiently. Only one in ten shaved armpits are described as odorous 24 hours after washing, while nine out of ten unshaved armpits are described as odorous (Shelley, Hurley & Nicols, 1953, as cited in Stoddart, 1990, pp. 66-67). The concentration of hairs here, served by multiple apocrine glands, must be an adaptation for the production and dissemination of odors. [para 7]

Humans do appear designed to receive pheromones, since they have functioning odor detection and apparently functioning vomeronasal systems (Garcia-Velasco& Mondragon, 1991; Monti-Bloch, Jennings-White, Dolberg, & Berliner, 1994, Moran, Jafek, & Rowley, 1991) with two of the substances reported in this journal (Monti-Bloch & Grosser, 1991) as affecting the human vomeronasal organ subsequently revealed (Jenning-White, 1995) to be estra-1,3,5(10),16-tetraen-3-ol (estratetraenol) and androsta-4, 16-dien-3-one (androstadienone), with the later affecting females more strongly and being structurally related to androstenol. Humans are not noted for their ability to detect odors over long distances. [para 8]

Thus, I would propose that in considering human pheromones that we abandon the analogy from insects and some mammals of an effect operating at long distances, such as the attraction or identification of mates. Humans can clearly recognize and identify mates, potential mates, and children at long distance by sight and sound. A pheromonal mechanism would add little to what can be accomplished by these other means. [para 9]

However, pheromones in many mammals do play a role in reproduction, most prominently in attracting a mate and preparing the mate for intercourse and reproduction. I would suggest that the human pheromone, androstenol, (and possibly other human pheromones) play a role in reproduction, but that this role relates to the formation and maintenance of the pair bond. In essence, detection of a pheromone emitted by the opposite sex, and only detectable at very close range (such as occurs when sleeping together) is a sign of the presence of a mate. The pheromone serves to promote behavior that leads to the formation or continuation of a pair bond, or to reproductive behavior which is optimal in the presence of a pair bond (having a baby). [para 10]

If human females are affected by androstenol in nature, one would ask where and how they get sufficient exposure for a small but functional main and accessory olfactory systems to detect it. The obvious answer is in sleeping with their mates. Bonded humans regularly sleep together, frequently with the female cuddled in the arms of the mate in a position that places her nose near the axillae (armpits) of the male. Because typically women are shorter than men, their heads are often in the vicinity of their mate's shoulders, and underarm regions. [para 11]

Cowley & Brooksbank demonstrated that females exposed to androstenol engage in many more, longer duration, and deeper interactions with men. In the experiment these were a wide range of men they met presumably ranging from family members, those met on public transport, those met in halls etc. Because they were wearing an androstenol emitting necklace they would have been under the influence of the androstenol. [para 12]

However, in natural conditions the androstenol exposure would have probably been during the night. The exposure would have disappeared when the female arose and went about her daily business. The only male who benefitted from her greater "approachability" would normally be either her mate, or a mate to be. [para 13]

Much of the androstenol induced interaction would be what is referred to in our society as "pillow talk", discussion with the mate or potential mate while lying with him, or cuddling with him. Notice human males and females have desire to "cuddle" with each other which will normally place the females nostrils close enough to the males chest and underarm regions to be affected by any pheromones being emitted. It is easy to imagine how such pillow talk creates and strengthens the bond between male and female. [para 14]

It may also be noted that the pillow talk may help the female evaluate the male. Engaging in much longer and deeper interactions than normal, she may learn things from the conversation that help the female to evaluate the suitability of that male as a long term mate, and his willingness to assist in rearing children. [para 15]

It is useful to ask if a system by which males are designed to emit androstenone to promote pillow talk, and females to receive it would be evolutionarily stable. By evolutionarily stable is meant that individuals of either sex that lacked the trait would enjoy less reproductive success than ones that had it (Maynard-Smith, 1982). [para 16]

First consider a female who did not respond to androstenol. Engaging in less interaction with her mate, she might form a less strong pair bond, and would have less male assistance in raising her children. Identifying and attracting a male to provide sperm is not a female problem in humans (hence pheromones probably do not play a role), but inducing one to help raise the children is a problem (Symons, 1979; Hrdy, 1981; Ridley, 1993). Thus, anything that promotes a pair bond probably helps the female. [para 17]

Also, a female who did not experience an increased desire to interact with a male she had slept with, or cuddled with, would have less information about him. Such information assists her in deciding whether to form or maintain a pair bond with him. The androstenol promoted interaction should facilitate the bonding between the male and female, which is critical to female reproductive success. [para 18]

Consider a female who interacted very readily with all males, regardless of whether she was cuddling or sleeping with them. Such interaction would be very time consuming. Thus, females who did not adjust their willingness to interact to whether the male had already been granted sufficient access to cuddle or sleep with (which normally indicates either the existence of a pair bond, or a pair bond in the process of formation) would waste much energy, and increase their risk of forming inappropriate pair bonds. [para 19]

The same argument would apply to a female who had a very sensitive system for recognizing androstenol or other pheromones. She would be affected by the pheromones emitted not only by mates and potential mates, but by other males in the vicinity. She would lose the benefit of disproportionate interaction with mates and potential mates. Thus, this theory can explain not only why females have retained the ability to be affected by the pheromone, androstenol, but also why their sensitivity to it is relatively weak (at least when compared with the sensitivity exhibited in other species to pheromones). [para 20]

What about males? Consider a male who did not emit androstenol, or who had relatively few hairs to disseminate it. Females would not be induced to interact with him when he cuddled or slept with them. He would be less likely to form the sort of strong pair bond that gave him continued sexual access and which led to reproductive success. [para 21]

Thus, a role for androstenol as a human pheromone that promotes "pillow talk" seems to be one that could emerge and survive. Since, in other species pheromones signal between males and females, it is likely that it earlier played a role in the relations between males and females. Once pair bonding became important in humans, this mechanism could have been adapted, merely by forming a few new connections in the brain, for differentially interacting with mates and potential mates. Mean while, other uses for the mechanism might have gradually disappeared if they were no longer needed, or were counterproductive. [para 22]

Admittedly, there is an alternative interpretation to the Cowley & Brooksbank finding. Since the subjects' reports on interaction were based on questionnaires filled in after the necklaces were turned in, and it is hard to recall all interactions, it is possible that the number of interactions were not affected by the androstenol, but their ability to recollect interactions was. This explanation is unlikely. Since forgotten interactions would probably be the least in depth and shortest ones, such as a casual greeting, remembering more interactions would lower the average depth. However, androstenol raised the depth. However, even if the effect was due to androstenol affecting memory for interactions with male, the effect could still be part of the pair bonding mechanism. Something that improved a woman's memory for interactions with a man she slept or cuddled with would be a most useful adaptation. It would assist her in choosing whether to continue the relationship, and increase her retention of information that would assist her in maintaining the pair bond. [para 23]

Other effects of androstenol and related compounds on behavior can be interpreted as adaptations that facilitate pair bonding (Cowley, Johnson, & Brooksbank, 1977; Filsinger, Braun, Monte & Linder, 1984; Kirk-Smith, Booth, Carroll & Davies, 1978). For instance, Cowley, Johnson, & Brooksbank (1977) found that women rate men more positively when exposed to androstenol. Benton (1982) found that women exposed to anrostenol on the upper lip rated themselves as more submissive around mid-cycle, a feeling that encourages impregnation. Cowley and Brooksbank (1991, p. 658) mention that Jackson, working in their laboratory had found evidence that the effects of an intial exposure to androstenol can persist for at least two weeks. This is very consistent with a pheromone that acts to inform the individual of the regular presence of a member of the opposite sex (i.e. a mate), but not of one whose only purpose is to facilitate a single sexual interaction. [para 24]

Of course, androstenol and other pheromones released by males could have other purposes. The author has argued elsewhere (Miller, 1995) that the evidence that female fertility is promoted by the proximity to a human male is a logical adaptation designed to increase the relative probability that women become pregnant by "dads" who will stay around and assist in raising the offspring, rather than by "cads" who will after a "one night stand" abandon the women. If one of the effects of physical proximity is to promote fertility, perhaps by changing the hormonal environment so that implantation was more like, it would be a useful adaptation. [para 25]

Several, otherwise puzzling effects in the literature could be explained such as the tendency for women who regularly saw men to have more menstrual cycles of a length more conducive to fertility (Cutler, 1991; Cutler, Garcia, & Krieger, 1979; Cutler, Garcia, & Krieger, 1980; Cutler, Preti, Huggins, Erickson,& Garcia, 1985; Cutler, Garcia, Huggins,& Preti, 1986), the ability of an extract from males to affect the female cycle (Cutler, Preti, Krieger, Huggins, Garcia, & Lawley, 1986; Russel, Switz, & Thompson, 1980), exposure to men influencing the occurrence of ovulation in women (Veith, Buck, Getzlaf, Van Dalfsen, & Slade, 1983), and the ability of sexual intercourse to increase the probability of success in gamete intrafallopian transfer (Marconi, Auge, Oses, Quintana, Raffo, & Young, 1989). However, discussing this possible effect of male pheromones is the subject of another paper (Miller, 1995). [para 26]

As reported, Cowley & Brooksbank just have described an interesting effect. However, if female exposure to androstenol serves the purpose hypothesized, it could be important. For instance, if human pair bonding is promoted by exposure to axillary pheromones, perhaps antiperspirants serve to reduce a useful form of chemical communication from males to females. This seems an important enough question to justify further research. [para 27]

An obvious next step in the research is to do a double blind study in which the males in a married couple use either an antiperspirant or a deodorant on a random basis and then fill in questionnaires on the extent of their interactions (conversations), both that morning and at other times. Since there is a good chance that female emitted pheromones affect males, a similar experiment should be done with females. [para 28]

Bird, S. & Gower, D. B.: The validation and use of a radioimmuno-assay for 5a-androst-16-3-one in human axillary collections. The J. Steroid Biochem 14, (1982) 213-219. [para 30]

Benton, D.: The influence of androstenol-a putative human pheromone- on mood througout the menstrual cycle. Biol. Psychol. 22, (1982) 141-147. [para 31]

Brooksbank, B. W. L.: Urinary excretion of androst-16-en-3a-ol. Levels in normal subjects, and effects of treatments with trophic hormones. J Endocr. 24, (1962) 435-444. [para 32]

Cowley J. J., Johnson, A. L. & Brooksbank B. W. L.: The effect of two odorous compounds on performance in an assessment of people test. Psychoneuroendocrinology, 2, (1977) 159-172. [para 33]

Cowley J. J. & Brooksbank B. W. L.: Human exposure to putative pheromones and changes in aspects of social behavior. The J. Steroid Biochem. Mol. Biol. 39 (1991) 647-659. [para 34]

Cutler W. B.: Love Cycles: The Science of Intimacy. Villard Books, New York (1991) [para 35]

Cutler W. B., Garcia C. R. & Krieger A. M.: Luteal phase defects: A possible relationship between short hyperthermic phase and sporadic sexual behavior in women. Horm. Behav. 13 (1979) 214-218. [para 36]

Cutler W. B., Garcia C. R. & Krieger A. M.: Sporadic sexual behavior and menstrual cycle length in women. Horm. Behav.14 (1980) 163-172. [para 37]

Cutler W. B., Preti G., Huggins G. R., Erickson B. & Garcia C. R.: Sexual behavior frequency and biphasic ovulatory type menstrual cycles. Physiol Behav. 34 (1985) 805-810. [para 38]

Cutler W. B., Preti G., Krieger A., Huggins G. R., Garcia C. R. & Lawley H. J.: Human axillary secretions influence womenUs menstrual cycles: The role of donor extract from men. Horm Behav. 20 (1986) 463-73. [para 39]

Cutler W. B., Garcia C. R., Huggins G. R. & Preti G.:Sexual behavior and steroid levels among gynecologically mature premenopausal women. Fertil.Steri. 45 (1986) 496-502. [para 40]

Filsinger E. E., Braun J. J., Monte W. C. and Linder D. E.: Human (Homo spaiens) responses to the pig (Sus scrofa) sex pheromone 5-alpha-androst-16-en-3-one. J. Comp. Psychol. 98 (1984) 219-222. [para 41]

Garcia-Velasco J. & Mondragon M.: The incidence of the vomeronasal organ in 1000 human subjects and its possible clinical significance. The J. Steroid Biochem. Mol. Biol. 39, 4B (1991) 561-563 [para 42]

Hrdy S. B. The Women that Never Evolved, Harvard University Press, Cambridge (1981). [para 43]

Jenning-White, C .: Perfumery and the Sixth Sense. Perfumer & Flavorist, 20 (July/August, 1995). [para 44]

Kirk-Smith, M., Booth D. A., Carroll D. & Davies, P.: Human social attitudes affected by androstenol. Res. Commun. Psychol. Psychiat. Behav. 3 (1978) 379-384. [para 45]

Kohl J. V. & Francoeur R. T.: The Scent of Eros. Continum, New York (1995). [para 46]

Marconi G , Auge L, Oses R. Quintana R., Raffo F. & Young E.: Does sexual intercourse improve pregnancy rates in gamete intrafallopian transfer? Fertil. Steril. 51 (1989) 28-34. [para 47]

Maynard-Smith J.: Evolution and the Theory of Games. Cambridge University Press, Cambridge (1982) [para 48]

McClintock, M.: Estrous synchrony: Modulation of ovarian cycle length by female pheromones. Physiol. Behav. 32 (1984) 701-705. [para 49]

Miller, E.: Male signals of his continued presence and female receptivity. manuscript (1985). [para 50]

Monti-Bloch L., & Grosser B. I.: Effect of putative pheromones on the electrical activity of the human vomeronasal organ and olfactory epithelium. J. Steroid Biochem. Mol. Biol. 39, 4B (1991) 573-583. [para 51]

Monti-Bloch L., Jennings-White C., Dolberg D. S., & Berliner D. L.: The human vomeronasal system. Psychoneuroendocrinology 19 (1994) 673-686. [para 52]

Moran D. T., Jafek B. W. & Rowley J. C.: The vomeronasal (Jacobson's) organ in man: ultrastructure and frequency of occurrence. J. Steroid Biochem. Mol. Biol. 39 (1991) 545-552. [para 53]

Ridley M.:The Red Queen: Sex and the Evolution of Human Nature. Macmillan, New York (1993). [para 54]

Russel M. J., Switz G. M. & Thompson K.: Olfactory influences on the human menstrual cycle. Pharm. Biochem. Behav. 13 (1980) 737-738. [para 55]

Stoddart, D. M. (1990). The Scented Ape: The Biology and Culture of Human Odor. Cambridge: Cambridge University Press. [para 56]

Symons D.: The Evolution of Human Sexuality, Oxford University Press, New York (1979). [para 57]

Veith J. L., Buck M., Getzlaf S., Van Dalfsen P. & Slade S.: Exposure to men influences occurrence of ovulation in women. Physiol Behav. 31 (1983) 313-315. [para 58]