Edward Miller (1996) Mate Attraction or Mate Confirmation:. Psycoloquy: 7(12) Sex Odor (2)

Volume: 7 (next, prev) Issue: 12 (next, prev) Article: 2 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 7(12): Mate Attraction or Mate Confirmation:

The Evolutionary Role of Pheromones
Book Review of Kohl on Sex-Odor

Edward Miller
Department of Economics and Finance
University of New Orleans
(504) 286-6913 (office)
(504) 286-6397 (fax)



The major absence from Kohl & Francoeur's (1995) book is an evolutionary perspective and an account of what function pheromones might serve in humans. For instance, there is a plausible function for menstrual synchrony. While Kohl and Francoeur's idea that pheromones might serve to attract mates is implausible, they could confirm the presence of a mate and adjust female fertility to whether any offspring was likely to benefit from male paternal investment.


behavioral development, genetics, gonadotropin, human sexuality, neuroanatomy, neuroendocrinology, odors, olfaction, pheromones, releasing hormone
1. Kohl & Francoeur's (1995) book overall is a good summary of what is known about the role of pheromones in humans, presenting much useful information.

2. The major weakness of the book is in the evolutionary area. While showing that pheromones in non-humans play an important role, and that therefore they could play a role in humans, they fail to consider exactly what role they could play that is not better served by other mechanisms.

3. In places, they imply that pheromones may serve a mate attraction function in humans as they do in other animals. They describe how humans have perfumed themselves, and how a few writers have spoken approvingly of the underarm odors of their beloved.

4. Yet, it is a matter of common observation that humans do not find underarm odors attractive, and that such odors do not serve to attract the opposite sex. Apparently, such odors repel.

5. Furthermore, there seems no real need for humans to use odors to attract the opposite sex. Humans have excellent eyesight and hearing and can recognize each other at long distances. The male and the female have distinctive shapes which make it easy to recognize them. For instance, the female has breasts, a narrow waist, and broad hips. Humans live in groups such that establishing initial contact with potential mates is easy.

6. Many of the animals which rely on odors and pheromones to recognize mates are either nocturnal, live in burrows, or live in thick brush, where finding and recognizing mates is difficult.

7. Yet it is very plausible that humans have (as Kohl & Francoeur point out) a pheromonal system. Humans have hair under the arms while they lack body hair in other areas more exposed to the atmosphere. Their apocrine glands feed into the hair follicles and appear designed to disseminate odors. Experiments with injection of radioactive steroids show that these are preferentially excreted from the axilla (the underarm area) (Brooksbank, 1970). The axilla secretions of males and females differ. The distinctive body odors of males and females appear only with puberty and the start of reproductive activity. There is a vomeronasal organ which is sensitive to different substances in males and females (Jenning-White, 1995). All of these taken together (as Kohl & Francoeur point out) make it very likely that humans have been subjected to selective pressures to emit and receive pheromones.

8. Let us consider a pheromonal system by which human males might attract females (such as Kohl & Francoeur suggest might exist). In many species, such a system could survive because it would be the most useful way for females to find mates. However, in human females being sexually attracted to a pheromone emitting male would be selected against. It is a commonplace of sociobiology that females should be careful in selecting their mates, and should choose ones with good genes who are willing to provision them (Ridley, 1993; Symons, 1979). For a human female to be sexually aroused by a male pheromone would be counterproductive, since it would make it hard to discriminate between males. This argument makes it unlikely that the observed male adaptations for emitting pheromones (underarm hair, apocrine glands, etc.) serve to attract females, since human females would be under selective pressure to ignore such signals. It is also well known that human females do not find the male underarm odor attractive, at least not at the conscious level.

9. A similar but weaker argument could be made for continued selection for males to be attracted by females' pheromones. The males that leave the most genes in later generations will seek to utilize every opportunity to mate with a fertile female. Potentially fertile females can be identified by such signs as an adult figure (but one not too old) and by the waist to hip ratio. Given that existing males are already attracted by visual cues, it is hard to see what pheromonal cues could add.


10. One place where Kohl & Francoeur conspicuously fail to take an evolutionary perspective is in discussing menstrual synchrony. They present evidence that the synchrony occurs but fail to discuss what purpose it might have served, or how it could have evolved. Without a plausible purpose for synchrony, it seems very implausible that it would have evolved, or could now exist.

11. Yet there is a plausible purpose for menstrual synchrony. It is in the female interest to have a male who helps provision her offspring. In turn, females seek to form pair bonds with males that will help them and provision them. However, the strongest, most dominant males try to father most of the children by preferentially mating with the females who are in their fertile periods. Females are better off to have the full support of a provisioning, but non-dominant male, than to have only partial support from a dominant male (who is probably dividing his support among several females). However, if these males are using their dominance status (greater strength and health based) to preferentially mate with ovulating females, the ovulating females cannot choose their own mates.

12. Females appear to have several adaptations to avoid dominant males and facilitate mating with non-dominant, but provisioning males. Concealed ovulation has been argued to be one adaptation (Knight, 1991; Miller, 1996). If the dominant male cannot tell when females are ovulating, his only alternative is to try to keep other males from them at all times. During the transitional period when only some female ovulation could be detected, the dominant males would have focused their attentions on the females they could detect ovulating. The females with concealed ovulation were then freer to form pair bonds with non-dominant but provisioning males. Benefiting from male provisioning, they out-reproduced the females whose ovulation was revealed. Thus genes for concealed ovulation gradually spread.

13. An even earlier adaptation by females was probably ovulatory (menstrual) synchrony. Kohl & Francoeur present the evidence that females that are in close contact with each other tend to have cycles that synchronize. Most of the studies that have looked for synchrony have found it, although a few have not. However, several of the studies that have failed to find it have had methodological problems that biased the studies against finding it. For instance, Jarett's (1984) study is internally inconsistent in that it announced there was only a trend towards synchrony, but then announced that there were five variables that had a statistically significant effect in determining synchrony. It appears that the problem was Jarett's transformation of the data before applying the Wilconxon test (Miller, 1996). Reference to the thesis from which the published report was taken shows that there are unreported results for the month of April which differ from the January results which were reported (Miller, 1996).

14. Wilson, Hildebrandt and Gravel (1991) have also reported a failure to find synchrony. However, their definition of synchrony (which admittedly has been commonly used) was limited to whether or not the nth cycle counted from a defined starting point for both women had become closer together (see Miller, 1996 for a detailed discussion of the methodological issues here). This will detect synchronization only if it takes the form of cycles becoming closer together.

15. However, synchronization can occur by one woman delaying a phase of her cycle (such as skipping ovulation, or delaying it). In this case the nth onset of one woman could become closer to the closest onset of another woman, even though it was further away from the nth onset of that woman. As an illustration, two women could have had their initial onsets 10 days apart. One woman's cycle might have been altered in such a way (by the other woman's pheromones) that her fourth onset was on the same day as the other woman's third onset was, but was 30 days later than that woman's fourth onset. By the ordinary definition the two women's cycles would have been synchronized, while by the equally numbered cycles method (used by Wilson et al. and others) the two women have become less synchronized (since they were 10 days apart on the first cycle and 30 days apart on the last cycle). When it is realized that the failure to find synchronization of Wilson et al. and Jarett may have been due to methodological mistakes and that their published reports contain evidence for synchrony, the case for synchrony is even stronger than Kohl & Francoeur reported.

16. Kohl & Francoeur fail to discuss what purpose menstrual synchrony could serve. A dominant male whose strategy was to try to monopolize ovulating females would have difficulty doing so if several females were ovulating about the same time. These synchronized females would have been better able to form pair bonds with non-dominant males whose mate acquisition strategy was provisioning. Even if the dominant male could defeat or intimidate any other male in a fight over an ovulating female, he could not monitor several females at once. If two ovulating females were to go in opposite directions, he would be forced to choose which one to follow. The other would be free to meet with the male who was provisioning her. Thus the synchronized females could more easily benefit from male provisioning.

17. In contrast, a non-synchronized female would frequently be the only female ovulating at a particular time. The dominant male (who was either non-provisioning, or who was provisioning but was dividing his provisioning among several females) would monopolize the female's fertility. Other provisioning males learned that their provisioning would not be rewarded with sexual access when it mattered, and probably gave up, shifting their provisioning to synchronized females who could reward them with sexual access. Thus, over time the non-synchronized females left fewer descendants than the synchronized females, and synchronization came to be the norm.

18. Notice this argument requires that the males be able to detect the females' ovulatory status so it must have evolved before, or concurrently with, the concealment of ovulation. However, since there appear to be no reproductive disadvantages associated with ovulatory synchrony, females have retained the synchronization mechanism even though it serves no current purpose.

19. One other peculiarity that Kohl & Francoeur draw attention to may have an evolutionary source. This is human's weak sense of smell and the human male's inability to detect female ovulation, even though females are apparently emitting chemical signals of their ovulatory state (as shown by the fact that other females are picking up signals that are used to synchronize ovulation). There is an evolutionary explanation for this (Stoddart, 1986). If males were differentially attracted to ovulating females they would devote most of their mating efforts to such females. As long as provisioning made little contribution to offspring survival, the males who only courted ovulating females would indeed be more successful in leaving descendants in the next generation.

20. However, suppose conditions changed so that the males who formed a pair bond with a particular female, and provisioned her enjoyed greater reproductive success. Males who were constantly distracted from their non-ovulating mate by a female who was more attractive because of pheromonal ovulatory signals would enjoy lower reproductive success. There would then be selection for a mutation which destroyed any previously existing mechanism for identifying and preferring ovulating females. This is an alternative explanation for the so-called concealment of female ovulation. The females did not evolve to conceal ovulation, the males evolved not to receive the message which females apparently continued to send out (as shown by the existence of menstrual synchrony). Of course, a third possibility is that females never did signal ovulation.


21. Kohl & Francoeur make a good case that humans are designed to emit and receive pheromones. The question is for what purpose. I would like to propose here what I have proposed elsewhere (Miller, 1996). This is that the purpose of these adaptations is not to attract the opposite sex (as Kohl & Francoeur suggest), but to inform the individual reproductive system of the continued presence of a mate.

22. In leading up to this conclusion, one should ask under what conditions are we close enough to smell (or otherwise perceive) the bodily emanations of another. Such smelling occurs when sleeping with or cuddling with mates or potential mates. Thus I would propose that the human pheromonal system serves to promote the establishment and continuation of pair bonds, or to increase the probability of reproduction occurring when the female is in such a relationship.

23. Let us start by looking at female reproductive success. As pointed out earlier, females benefit from male provisioning. Females would enjoy greater reproductive success if they did not get pregnant easily from a casual one night stand, but did get pregnant from sexual activity with a male who was in a pair bonded relationship with them, and likely to assist in raising the child.

24. Some of the effects presented in Kohl & Francoeur's book could be explained if the general mammalian pheromonal mechanisms had become specialized to tell the human female whether she was in an ongoing relationship with a male. For instance, Russel, Switz and Thompson (1980) found that placing male sweat on female's lips affected their cycles, making them more regular. Cutler (1991) has argued that such more regular cycles are more fertile. Exposure to men has been shown to influence the occurrence of ovulation in women (Veith, Buck, Getzlaf, Van Dalfsen & Slade, 1983). Several studies have shown that females (typically of college age) who have more contact with males have shorter and more regular cycles, which Cutler argues to be more fertile cycles.

25. Stanislaw and Rice (1987) have found evidence, in a data set involving over fifteen thousand cycles, that intercourse accelerates both the basal body temperature shift and the onset of menses, with the menses being accelerated more by a combination of follicular phase and luteal phase intercourse than by luteal phase intercourse alone.

26. There is some evidence suggesting that it is actually intercourse that makes the difference (Marconi, Auge, Oses, Quintana, Raffo & Young, 1989). Women who were having the fertility treatment of gamete intra-fallopian transfer (GIFT) were studied. Half were told to abstain from sex before and after the procedure. The other half were told to engage in sex close to the time of the hormone administration that started the procedure. Fifteen of the eighteen who had intercourse became pregnant, while only five of the eighteen abstainers became pregnant. This difference was highly significant, and does argue that sexual behavior makes a difference. (There were medical reasons to believe that the pregnancy came from the GIFT procedure, rather than from the accompanying normal intercourse). What purpose could all these mechanisms serve?

27. Kohl & Francoeur present many of these effects as merely interesting observations. I would propose that females with these mechanisms are more likely to become pregnant when they have a male around to assist in rearing the child than when they do not. Females with such mechanisms would be expected to leave more descendants than females without such mechanisms.

28. Kohl & Francoeur present strong evidence that in many mammals pheromones affect the hypothalamus, the release of gonadotropin releasing hormone, and through this mechanism the female's fertility. Evolution could easily modify this mechanism so that a male's presence enhanced the female's fertility. For instance, the presence of a male might lengthen the period which the lining of the womb was suitable for implantation, thus increasing the female's probability of becoming pregnant. Fertilization might occur in a female who had a one-night stand. However, if later she did not have close enough contact with a male to smell his pheromones the egg would be less likely to successfully implant and lead to pregnancy. However, if the female had been regularly sleeping or cuddling with a male, his pheromones would have led to her uterus being better prepared for implantation.

29. Such a mechanism would require only a slight modification of the mechanism that Kohl & Francoeur show that other mammals already had (i.e., mechanisms for emitting pheromones into the atmosphere, mechanisms for detecting their presence, neurons linking these to the relevant parts of the brain, and mechanisms by which the brain could influence the reproductive organs). All that would be required would be a change in the "program" that transformed pheromonal inputs into signals to the reproductive organs, such that male pheromones increased fertility. Given the general mammalian model, and that in some mammals the male presence has been shown to induce ovulation or to prepare the female for pregnancy, the evolutionary obstacles to the proposed mechanism appear minor.

30. For a mechanism to remain in existence, it is necessary for those that possess the mechanism to out reproduce those that lack it.

31. It was discussed earlier why humans would not have been selected to use pheromones to attract mates. However, if the effect of male pheromonal cues was to make the female more fertile, there would indeed be selection for male ability to emit the pheromones, and for female ability to receive them. Males who lacked the apparatus for emitting the pheromones (less apocrine glands, less underarm or genital hairs, lack of the relevant hormones) would produce fewer offspring from their matings, even with pair bonded females. Likewise, females who lacked the mechanisms for receiving the male pheromonal signals would not be able to adjust their fertility to whether or not a male was present, and would be out-reproduced by those females who did receive the signals.

32. It might be argued that there are other mechanisms by which females could discover whether they had a mate, such as whether they felt in love, or whether they were being regularly provisioned. This is true, but to adjust her fertility to these signals the female would need to evolve a whole new mechanism to transmit feeling in the brain into endocrine signals that affected fertility. Male provisioning and human pair bonding appear to be relatively recent events in human evolution, and there probably has not been time for such a new mechanism to emerge. Much simpler was to change the standard mammalian pheromonal mechanism (or possibly to retain it) so that it was used to inform the female reproductive system of whether this was a good time to become pregnant.

33. If males are sending pheromonal signals while sleeping with or cuddling with females, a number of features of human behavior could be explained. Humans are known to be unusually sexual animals engaging in much non-reproductive sex (including copulations when the female is not fertile). If such sex serves to increase female fertility as the evidence from Cutler and others suggests, the selection for male and female desires for frequent sex would be explained.

34. Likewise a female who did not desire frequent sex (perhaps because her body sought it only when fertile) would have fewer offspring, or at least fewer offspring that had a male available to serve as father. The same would apply to a female who did not enjoy cuddling. Very likely she would enjoy less reproductive success because many of the children she bore would be at times when there was no male available to assist in rearing them.

35. If females are affected by male pheromones, are males affected by female pheromones? Females do have underarm hair, and do appear to be designed to emit pheromones. The males' vomeronasal organ responds to female emitted chemicals. This, along with the female adaptations for emitting pheromones, suggests that males should be affected by female pheromones. What purpose could such signals from females to males serve?

36. It would be in the male's interest to take risks until he acquired a mate and children, but then to be more cautious since his disability or death would reduce the survival of the children he had already borne. Married men do seem to be more conservative than single men (as seen in crime and automobile accident rates). A plausible mechanism would be for regular exposure to female pheromones such as result from sleeping with a female, to result in the male becoming more cautious.

37. Kohl & Francoeur failed to mention one of the most interesting of pheromonal effects. This was reported by Cowley and Brooksbank (1991). They had subjects and controls wear necklaces that emitted androstenol overnight. The next morning questionnaires were filled out regarding social interactions (conversations) with other individuals that morning. No effects were found for males, or for female interactions with other females. However, female interactions with males were much more frequent and more intense. Neither the original researchers nor Kohl & Francoeur give any explanation for this effect. It is very likely the increased interaction is explained by a change in female behavior since the wearers of the necklaces got much higher and more prolonged exposure to androstenol than any males they came into contact with, and male wearers of androstenol necklaces showed no changes in frequency of interactions.

38. However, if we ask how in nature women would have had a heavy exposure to androstenol (emitted by males), we realize it would have been through sleeping or cuddling with males (an idea the author of this review has developed in a paper titled "Androstenol, a Pillow talk Pheromone"). These would normally be either a mate, or a potential mate. It is then easy to imagine that females would benefit from increased interaction with the male. One effect of increased conversation and exchange (pillow talk) might be to strengthen the pair bond. Another benefit of pillow talk would be to learn more about the male, thus gaining information that the female could use to decide whether to continue the relationship, or to manipulate the male. A male who did not respond to the female's androstenol inspired attempts to talk would give an adverse signal to the female (suggesting he had something to hide or was not committed to her). However, if he engaged in extensive conversation he would reveal more about himself, information it was in the interest of the female to know.

39. Thus, in summary Kohl & Francoeur's book represents a very interesting summary of what is known about possible human pheromones. Since so little is known about human pheromones, the book raises more questions than it answers. It provides a source of material for interesting speculations, some of which have been presented above. Without the factual basis developed in the book, these ideas would not have been developed.


Brooksbank, B.W. (1970) Labeling of steroids in axillary sweat after administration of 3H-F5 pregnenolone and 14C-progesterone to a healthy man. Experientia, 26, 1012-1014.

Cowley, J.J. & Brooksbank, B.W.L. (1991) Human exposure to putative pheromones and changes in aspects of social behavior. The Journal of Steroid Biochemistry and Molecular Biology 39: 647-659.

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

Cutler, W.B. & Garcia, C.R. (1980) The psychoneuroendocrinology of the ovulatory cycle of woman: A review. Psychoneuroendocrinology 5: 89-111.

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

Cutler, W.B., Garcia, C.R., & Krieger, A.M. (1980) Sporadic sexual behavior and menstrual cycle length in women. Hormones and Behavior 14: 163-172.

Cutler, W.B., Preti, G., Huggins, G.R., Erickson, B., & Garcia, C.R. (1985) Sexual behavior frequency and biphasic ovulatory type menstrual cycles. Physiology and Behavior 34: 805-810.

Cutler, W.B., Preti, G., Krieger, A., Huggins, G.R., Garcia, C.R., & Lawley, H.J. (1986) Human axillary secretions influence women's menstrual cycles: The role of donor extract from men. Hormones and Behavior 20: 463-73.

Jarett, Laura R. (1978) Psychosocial and biological influences on menstruation: Menstrual synchrony, menstrual cycle length, and menstrual regularity. Dissertation, New York University.

Jarett, Laura R. (1984) Psychosocial and biological influences on menstruation: synchrony, cycle length, and regularity. Psychoneuroendocrinology 9: 21-28.

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

Knight, C. (1991) Blood Relations: Menstruation and the Origins of Culture. New Haven: Yale University Press.

Kohl, J.V. (1995) Precis of: The Scent of Eros. PSYCOLOQUY 6(33) sex-odor.1.kohl.

Kohl, J.V. & Francoeur, R.T. (1995) The Scent of Eros. New York: Continuum.

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

Miller, E.M. (1995a) Comment on Jarett's menstrual synchrony study. Manuscript.

Miller, E.M. (1995b) Female unattractiveness and concealed ovulation as a strategy for increasing per capita paternal investment. Manuscript.

Miller, E.M. (1995c) Androstenol as a pheromone promoting pillow talk. Manuscript.

Miller, E.M. (1995) Menstrual Synchrony: A Methodological Comment.

Miller, E.M. (1996) Fertility and a Mate's Signals of Continued Presence. Manuscript.

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

Russel, M.J., Switz, G.M. & Thompson, K. (1980) Olfactory influences on the human menstrual cycle. Pharmacology Biochemistry and Behavior 13: 737-738.

Stanislaw, H., & Rice, F. (1987) Acceleration of the menstrual cycle by intercourse. Psychophysiology 2: 714-7.

Stoddart, D.M. (1986) The role of olfaction in the evolution of human sexual biology: An hypothesis. Man, 21: 514-20.

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

Veith, J.L., Buck, M., Getzlaf, S., Van Dalfsen, P. & Slade, S. (1983) Exposure to men influences occurrence of ovulation in women. Physiology & Behavior 31: 313-315.

Wilson, H.C., Hildebrandt, S. & Gravel, V. (1991) Two studies of menstrual synchrony research: Negative results. Psychoneuroendocrinology 16, 353.

Volume: 7 (next, prev) Issue: 12 (next, prev) Article: 2 (next prev first) Alternate versions: ASCII Summary