The Biology of Menstruation in Homo Sapiens: Total Lifetime Menses, Fecundity, and Nonsynchrony in a Natural-Fertility Population

by Beverly I. Strassmann
The Biology of Menstruation in Homo Sapiens: Total Lifetime Menses, Fecundity, and Nonsynchrony in a Natural-Fertility Population
Beverly I. Strassmann
Current Anthropology
Start Page: 
End Page: 
Select license: 
Select License

and varieties of "Neanderthal" man. Quarterly Review of Biol- ogy 32:330-47. -. 1958. "Upper Pleistocene men of the Southwest Asian Mousterian," in Hundert jahre Neandertaler. Edited by

G. H. R. von Koenigswald, pp. 185-98. Utrecht: Kemink en Zoon. JEL~NEK,T. 1969. Neandertal man and Homo sapiens in Central and Eastern Europe. CURRENT ANTHROPOLOGY 10:475-503.

KLEIN,R. G. 1969. Comment on: Neandertal Man and Homo sa- piens in Central and Eastern Europe, by J. Jelinek. CURRENT ANTHROPOLOGY IO:493-95.

KOLOSOV, 1972. Shaitan Koba, the Mousterian site in the

Crimea (in Ukrainian). Kiev.

MCKERN,T. w., AND L. KOZLIK. 1962. Palaeolithic man from the Crimea. American Journal of Physical Anthropology 20:403-6.

MARKS, A. E., YU. E. DEMIDENKO, AND V. USIK. 1994. New excavations at Starosele in the 1993 field season (in Russian). Archaeological Almanac 3:121-33.


M. KAY. n.d. The "chaine opCratoirel' in the Middle Palaeolithic of level I, Starosele, Crimea. Quaternaria Nova 4. In press.

RO GINSKY,YA.YA. 1954. The morphological features of the child's skull from the late Mousterian layer of Starosele cave (in Russian). Soviet Ethnography, no. I, p. 27-39.

ROGINSKY, YA. YA., M. M. GERASIMOV, S. N. ZAMYATNIN, AND A. A. FORMOSOV. 1954. The conclusion on the fossil man find at the cave site of Starosele, near Bakhchisarai (in Russian). Soviet Ethnography, no. I, pp. 39-41.

RON EN,A. Editor. 1982. The transition from the Lower to the Middle Palaeolithic and the origin of modern man. British Archaeological Reports International Series I 5I.

SMIRNOV,YU. A. 1991. Mousterian burials in Eurasia (in Rus- sian). Moscow: Nauka.

SOFFER,0. 1994. "Ancestral lifeways in Eurasia: The Middle and Upper Palaeolithic records," in Origins of anatomically modern humans. Edited by M. H. Nitecki and D. V. Nitecki, pp. 101-20. New York: Plenum Press.


C. k.,AND 1993. In search of the Ne- anderthals: Solving the puzzle of modern human origins. New York: Thames and Hudson.

THOMA, A. 1962. Le ddploiement dvolutif de 1'Homo sapiens. Anthropologica Hungarica 5(I-2).

-. 1971. "L'Cvolution polycentrique de 1'Homo sapiens," in Origine de l'homme moderne. Edited by F. Bordes, pp. 81-85. Paris: UNESCO. TILLIER,A.-M. 1989. "The evolution of modem humans: Evi- dence from young Mousterian individuals," in The human rev- olution: Behavioral and biological perspectives in the origins of modern humans. Edited by P. Mellars and C. Stringer, pp. 286-97. Edinburgh: Edinburgh University Press.

-. 1990. "NCanderthaliens et origine de l'homme moderne en Europe: Quelques rCflexions sur la controverse," in Paldolithique moyen rdcent et Paldolithique supdrieur ancien en Eu- rope. Edited by C. Farizy, pp. 21-24. Paris: MCmoires du MusCe de PrChistoire. TRINKAUS,E. 1989. "The Upper Pleistocene transition," in The emergence of modern humans: Biocultural adaptations in the later Pleistocene. Edited by E. Trinkaus, pp. 42-66. Cambridge: Cambridge University Press. ULLRICH,H. 195 5. Palaolithische Menschenreste aus der Sowjet- union: Das Mousterien-Kind von Staroselje (Krim). Zeitschrift fir Morphologie und Anthropologie 47(1):91-98.

-. I958. "Neandertalfunde aus der Sowjetunion," in Hundert Jahre Neandertaler. Edited by G. H. R. von Koenigswald, pp. 72-106. Utrecht: Kemink en Zoon.

URY SON,M. I. 1964. "Initial stages of the emergence of modem man," in The wellsprings of mankind: Basic problems in an- thropogenesis (in Russian). Moscow: University Publishers.

USPENSKY, I. 1969. Position of fossilized children from the


caves in Starosele and Teshik Tash in the evolutionaw homi- nid system according to neurocranial stereomorpholo~ical data (in Russian). Problems of Anthropology ~I:II~-24.

VALLOIS,H. V. 1955. Un nouvel homrne moustkrien en CrimCe:
Le squelette d'enfant de StaroseljC. L'Anthropologie 59:555-61.

VERESHCHAGIN,N. K. 1967. "Primitive hunters and Pleisto-
cene extinction in the Soviet Union," in Pleistocene extinc-
tions: The search for a cause. Edited by H. E. Wright, pp.
365-98. New Haven: Yale University Press.

WOLPOFF,M. H. 1980. Paleoanthropology. New York: Alfred A.

YAKIMOV,V. P. 1954. Problem of the correlation of fossilized
people of modem and Neandertal types (in Russian). Soviet Eth-
nography, no. 3, pp. 57-62.

-. 1969. Comment on: Neandertal Man and Homo sapiens
in Central and Eastern Europe, by J. Jelinek. CURRENT ANTHRO-
POLOGY 10:498

The Biology of Menstruation in Homo sapiens: Total Lifetime Menses, Fecundity, and Nonsynchrony in a Natural-Fertility Population'

Department of Anthropology, 1020 LSA Building, University of Michigan, Ann Arbor, Mich. 48109-1382,

U.S.A. ( VII 96

Studies of female reproductive biology in humans are almost entirely limited to women who spend the major- ity of their reproductive years in menstrual cycling. Given that human reproductive biology evolved when pregnancy and lactation were the usual reproductive states (Short 1976)~ it is important to consider reproduc- tive patterns in natural-fertility populations. In these populations, couples do not attempt to control their fer- tility in a parity-dependent fashion (Henry 1961). John- son et al. (1987) conducted a longitudinal investigation of menstruation in a noncontracepting population, the Gainj of Papua New Guinea, but their sample includes only 40 menstrual cycles in 36 women. Bentley, Harri- gan, and Ellison (1990) monitored 178 menstrual cycles among the Lese of Zaire, but because of endemic vene- real disease many Lese women were sterile and therefore displayed the Western pattern of cycling repeatedly without becoming pregnant. Here I present the first long-term, prospective data on menstruation in a true natural-fertility population, the Dogon of Mali. The

I. O 1997 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved ~~II-~~o~/~~/~~oI-~~~~$I.oo .
thank M. Brown, J. Warner, and D. Miller for help with the statisti- cal analyses, S. Moulin for assistance with data collection, and

R. D. Alexander, B. Gillespie, J. Warner, J. W. Wood, P. Smouse,

C. Kottak, C. Vincenz, and R. Wrangham for comments on the manuscript (or an earlier draft). I am also indebted to K. Sanogo and the Dogon themselves for their generous cooperation with the study. This work was supported by the L. S. B. Leakey Foundation, the National Science Foundation (BNS-8612291), and the Univer- sity of Michigan.


= Menstrualhut visit


0 10 20 30 40 50 60 70


FIG. I. A Dogon woman's visits to a menstrual hut in relation to her urinary pregnanediol-3 -glucuronide (PdG) profile. As expected, she went to the menstrual hut at the end of each luteal phase as her PdG levels were reaching baseline.

sample includes 477 untruncated menstrual cycles in

58 women. I focus on three specific questions: (I)From

menarche to menopause, how many menses do Dogon

women experience in a lifetime? (2) How are menses

patterned over the life span? and (3) Do Dogon women

synchronize their menstrual cycles?

The Dogon are millet farmers of the Sahel and have a total fertility rate of 8.6 * 0.3 live births per woman (Strassmann 1992). They provide a rare opportunity for monitoring female reproductive status in a preindustrial society because strict taboos require menstruating women to be segregated at night in special huts. The function of these taboos2 and of menstruation itself has been considered elsewhere (Strassmann 1992, 1gg6a, b). Through a nightly census (N = 736 days) of the women present at the two menstrual huts in the study village (July 1986 to July 1g88), it was possible to detect the onset of menstruation without interviews and thereby to avoid errors in recall or reporting3 A comparison of the timing of women's visits to the menstrual huts with the women's hormonal profiles (urinary pregnanediol-3- glucuronide and estrone-3-glucuronide) (fig. I)indicated that women (N = 70) went to the menstrual huts during 86% of all menses and stayed away from the huts when they were not menstruating (Strassmann 1gg6b). Pro- spective monitoring of 25 pregnancies indicated that all

2. In brief, the menstrual taboos of the Dogon are a male tactic for eliciting honest signals of female reproductive status. When a woman visits a menstrual hut, all members of her husband's patrilineage learn that she is neither pregnant nor in amenorrhea and that she will soon be ready to conceive. Information about the timing of conception is used in paternity assessments (Strassmann

1992, 1996bl.

3. Another advantage of the menstrual-hut census is that it permit- ted detection of women's first postpartum menses. Such detection is critical because women of high fecundity became pregnant on one of their first postpartum ovulations. If I had monitored only those women who were cycling at the start of the study, the sample would have overrepresented women of low fecundity.

25 births occurred approximately nine months after the

mothers' last visits to the menstrual huts (Str,, qf~mann

1992). These results confirm that if a Dogon woman vis-

its a menstrual hut she is in fact rnen~truating.~

The prolonged menstrual cycling of Western women

has been invoked as a risk factor for reproductive can-

cers (Short 1976, Eaton et al. 1994), a leading cause of

mortality. But how many menses did women typically

experience prior to the advent of contraception? Because

of the difficulty of collecting accurate information on

menstruation from interviews, few data are available to

answer this question. Eaton et al. (1994) calculated that

hunter-gatherer women who lived past menopause expe-

rienced roughly 160 ovulations in a lifetime, but this

estimate is not based on empirical data on either men-

struation or ovulation among hunter-gatherers.


I used objective records of menstrual hut attendance, as corroborated bv hormonal data, to determine the num- ber of menses experienced by Dogon women over the life span, assuming survival to menopause. In my sam- ple, the median age at menarche was 16 and the median age at menopause was 50. The mean number of hut vis- its per lifetime was IIO and the median was 94. These values were calculated as the sum of the mean (or me- dian) number of hut visits by women at each age (indi- vidual women who did not visit the huts at a particular age because of pregnancy and amenorrhea contributed a value of zero to the computation). To correct for the hormonally detected menses that were not signaled by a visit to the menstrual huts, I divided the total number of menses at each age by 0.86. After this correction, the estimated mean number of lifetime menses was 128 and the median was 109.

An American physician who recorded all of her men- ses throughout her lifetime and who had 3 births had a total of 355 menstrual cycles (Treloar et al. 1967). Her menstrual experience spans 32 years instead of the more typical 38 years for American women, so she may have had fewer menses than the average for American women. However, Treloar et al. report that up until her first pregnancy her cycle length was at least two days shorter than the mean, which would partially cancel the effect of a short menstrual experience. Eaton et al. (1994) estimate that American women have a total of about 450 ovulations. Their estimate assumes that from men- &he to menopause 96% of all menstrual cycles are ovu- latory. In the postmenarcheal and perimenopausal years women have many long anovulatory cycles (Baird 1g85), so Eaton et al.'s estimate of 450 ovulations per lifetime

4. I monitored the hormone profiles for 70 women for ten weeks and found only one occasion on which a woman went to a men- strual hut when she was not menstruating. This woman was preg- nant and appears to have been spotting. The data suggest that the primary reason women stayed home from the huts during 14% of all menses was simply to avoid an uncomfortable place (Strass- mann 1996bJ.

Age (years)
FIG. 2. Number of menses per woman per two years.

Error bars, standard errors of the mean. The sample

(N = 50) includes every woman in the study village

who had reached menarche (with the following

exclusions: 34 postmenopausal women, 17 women

who never menstruated during the study because of

pregnancy or amenorrhea, 13 women not resident

during the entire study, 4 infertile women, 3 women

not exposed to the risk of conception because of

separation or divorce, and 2 women who did not use

the menstrual huts).

is probably too high. For purposes of comparison, I as- sume that 400 menses per lifetime is not unusual for American women. The median of 109 menses per life- time among Dogon women is about a quarter of this value. Given that the Dogon are sedentary agricultural- ists, the high frequency of menstruation found in many contemporary human populations probably originated not with agriculture but with birth control.


The data also reveal striking features of the patterning of menstruation over female life histories. The fre- quency of menstruation among Dogon women had an approximately U-shaped relationship with age between menarche and menopause (fig. 2).To show this relation- ship more clearly, the graph excludes women who were continuously pregnant or in amenorrhea, and the full two years of data are presented. Women under zo had a mean of 10.8 ? 3.3 menses per two years, women aged 20-34 had 3.8 -c 0.6 menses, and women 35 and older had 12.9 2 2.2 menses. This result is consistent with reports that fecundity has an inverse U-shaped relation- ship with age (Bendel and Hua 1978, Jain 1969, Wood and Weinstein 1988). A potentially confounding variable is coital frequency, but the numerous menses among women <zo years are probably due to adolescent subfec- undity because women of these ages already had regular

Age (years)
FIG. 3. The percentage of women (N = so), by age,
who cycled during the study and who became
recognizably pregnant.


Variation in Reproductive Status with Age between Menarche and Menopause

Age (Years) N Cycling Pregnant Amenorrhea

NOTE:Variation is expressed as the mean proportion of time spent in each reproductive state over the two-year study period; 95% confidence limits are indicated. The sample includes all women in the study village who had reached menarche but not menopause, except for those who were not resident throughout the study or not exposed to the risk of conception because of in- fertility, separation, or divorce (N = 66).

sexual partners. In fact, it is normative for Dogon girls to begin sexual relations prior to menarche. The low menstrual frequency among women aged 20-34 years reflects a high pregnancy rate (fig. 3) and lactational sup- pression of ovulation (table I).Over the two-year study, cycling was the longest reproductive state for the youn- gest (<zo years) and oldest (>34 years) women, while lactational amenorrhea was the longest state for the women of prime reproductive age (20-34 years). After age 34, the frequency of menstruation increased dramat- ically and the pregnancy rate plummeted. This change can be attributed to reduced fecundity, lower coital fre- quencies at longer durations of marriage, and increased intrauterine mortality at the end of the reproductive years (Strassmann 1990, Wood 1989).~

5. Recent analyses indicate that husband's age is an important pre- dictor of fecundability among the Dogon (Strassmann, unpublished data).


FIG.4. Reproductive profile for the women (N = 122)

of the study village over 24months. Black, cycling;

white, pregnant; grey, in lactational amenorrhea;

diagonal striped, postmenopausal.

The reproductive profile of the women (N = 122) in the study village is shown in figure 4. Over the 736 days of the study the mean ( + standard deviation) proportion of women who were cycling on a given day was 0.25 + 0.01; 0.16 k 0.04 per day were pregnant, 0.29 + 0.04 were in lactational amenorrhea, and 0.31 + 0.006 were postmenopausal. (Cycling was defined as the interval from the first day of the first postpartum menses through the first day of the last menses before a recog- nized pregnancy; pregnancy was defined as the interval from the second day of the last menses through the day of birth or miscarriage; lactational amenorrhea was de- fined as the interval from the day after birth or miscar- riage through the day before the first postpartum men- ses.) On any given day, the subfecund women were overrepresented among the cycling women. The most fecund women conceived on one of their first postpar- tum ovulations and quickly dropped out of the pool of regularly menstruating women. For example, among the women aged 20-34 years, only two were sterile, and they had 23 and 29 menses each during the two-year study. The other women in their age-cohort had a mean of only 3.8 + 0.6 menses. Thus regular menses were a sign of sterility, not fecundity.

When menstrual cycle length was calculated from menstrual hut attendance, the median of the women's median cycle lengths was 30 days (lower and upper 95% confidence limits were 30.0 and 32.0 days, respectively; N = 58 women, 477 cycles; age-range was 15 to 53 years, mean age + standard deviation was 30.9 + 9.7 years). When cycles longer than 46 days (N = 73) or shorter than 17 days (N = 4) were excluded for being outliers as defined by Wilkinson (1990:550), the median of the women's median cycle lengths was 28.5 days

(lower and upper 95% confidence limits were 27.5 and

29.0 days, respectively; N = 54 women, 400 cycles). Es- timates of cycle length in any population reflect the age- structure of the sample, but cycle length among the Dogon is biologically indistinguishable from cycle length in Western populations (Chiazze et al. 1968,

Treloar et al. 1967, Vollman 1977).


McClintock (1971) reported that social interaction caused the menstrual onsets for groups of dormitory friends to grow closer together by two days over a four- to-six-month period. Neither McClintock nor subsequent investigators found a tendency for menses to be- come concordant, but the phenomenon is nonetheless referred to by the misleading term "menstrual synchrony." In careful analyses, Wilson (1987, 1992) showed that three statistical errors undermine McClin- tock's evidence as well as that of subsequent investiga- tors (Graham and McGrew 1980, Quadagno et al. 1981, Preti et al. 1986): (I)failure to correct for the conver- gence of menstrual onsets by chance, (2)exaggeration of the initial difference in onsets through miscalculation, leading to an erroneous impression of synchronization over time, and (3) possible sampling bias. Strassmann (1990) raised additional concerns having to do with the evolutionary novelty of the study populations (aggrega- tions of young, cycling women) and inconsistencies in the findings across studies, suggesting the possibility of random effects. Moreover, three studies found no evi- dence for synchrony in Western populations (Jarett 1984, Wilson, Hildebrandt Kiefhaber, and Gravel I99I, Treva- than, Burleson, and Gregory 1993). Despite the dearth of supporting data, many investigators continue to ac- cept the existence of menstrual synchrony rather than confronting the methodological issues Wilson raised (Graham 1991, Weller and Weller 1993).

Prior to the Dresent studv there had been no ~ublished tests for menstrual synchrony in natural-fertility popu- lations. Absence of synchrony in such populations would have two important implications: (I)it would se- riously weaken the hypothesis (Burley 1979, Turke 1984) that menstrual synchrony is adaptive, and (2)it would refute the widespread assumption among anthro- pologists that menstrual synchrony occurred in prein- dustrial societies (Buckley I988, Knight I988). In natu- ral-fertility populations, pregnancy and amenorrhea would prevent many women from cycling concurrently, but the possibility of synchrony can be tested among the remaining women. Among Dogon women of all ages who cycled during the present study, the median num- ber of menses per two years was 6.0. McClintock (1971) reported that most of the synchronization in her study set in over only four cycles. If that is the case, then many Dogon women had enough cycles to synchronize, particularly in the age-cohorts younger than 20 years and older than ?A vears.

In view of the above, I tested for menstrual synchrony among the Dogon. To avoid the methodological errors

TABLE 2 Test for Synchronization Using Cox Regression  
Variable  N Cycles  N Women  Coefficient  Standard Error  Exponentiated Coefficient  P-Value  

Village 477 Lineage 477 Economic unit 477 New moona 477 First quartet 477 Full moona 477

aThe first three lunar phases are shown contrasted with the omitted category, last quarter.

discussed by Wilson (1992)~ I introduce two new statisti- cal approaches. Using Cox regression (Cox 1972, Dixon 1992)~I asked whether, in any given cycle, a woman's risk of menstruating was influenced by the number of other women who were menstruating. The number of other women menstruating was calculated as three dif- ferent independent variables: "Village" refers to all other women in the village; "lineage" refers to all other women who lived with a particular lineage of related males; and "economic unit" refers to the women who habitually worked together and ate together. These vari- ables capture, with increasing degrees of closeness, the three main levels of social interaction among women. This analysis is modeled as follows: hilt) = ai(t).epx(t) where hi(t) is the ith woman's risk of menstruating at time t and ai(t) is the baseline hazard rate for the ith woman and x(t) is a time-varying explanatory variable tested individually. The three coefficients were nonsig- nificant, providing no evidence for synchronization (ta- ble 2).

If menstrual synchrony is caused by the lunar cycle (Pochobradsky 1974, Law 1986, Cutler et al. 1987) or other environmental rhythms (Little et al. 1989)) then it should occur villagewide. According to the null hypoth- esis of nonsynchrony, the menstrual onsets of different women should be independent, and thus the number of onsets per day should fit the Poisson distribution. Ac- cording to the alternative hypothesis of synchrony, men- strual onsets should be clustered. Under the null hy- pothesis, the expected number of onsets per day, E(N), equals the number of women at risk for menstruating on a given day (i.e., cycling and not pregnant or amenor- rheic) divided by the number of days of risk. Following this logic, I calculated the expected number of onsets E(Nt) for each day (t) as the number of women having menstrual onsets in a window centered on each day (t) divided by the number of days in the window. Over the 736 days of the study, the observed and expected number of days with o, I, 2, 3, and 4 onsets did not differ signifi- cantly = 1.50, d.f. = 4) (table 3). Thus the null hy- pothesis that the women's menstrual onsets were inde- pendent cannot be rejected.

Among the Dogon electric lighting is absent, so condi-

tions were favorable for testing specifically for lunar synchrony. I used Cox regression to ask whether a woman's risk of menstruating was influenced by lunar phase. Lunar phase was expressed as a time-varying cate- gorical variable with a dummy variable for each of the four lunar phases. The analysis was stratified by woman (N = 58)) and each cycle (N = 477) was an observation. None of the coefficients was significant (table z), provid- ing no evidence for lunar synchrony.

The lack of empirical support for menstrual syn- chrony among the Dogon and in Western populations does not establish the specieswide absence of the phe- nomenon, but it shifts the burden of proof onto those who argue that the phenomenon exists. Evidence for synchrony should also include an understanding of the mechanisms by which synchrony is achieved. At pres- ent, the pheromones or environmental influences that putatively cause synchrony in humans have not been identified. McClintock (1981) argued that menstrual synchrony is a functionless side effect of mechanisms that are adaptive in a different context. However, this context remains obscure. Synchrony does occur in other biological systems (hatching synchrony, firefly flashing, neuronal firing, to name a few), but in these cases the function of synchrony is better understood. Wallis (1985,

TABLE 3 Test for Independence of Menstrual Onsets

Number of Days Number of Menstrual Onsets per Day Observed Expecteda

Pearson's X2 = 1.50, d.f. = 4, n.s. 'If onsets were independent.


1992) reported synchrony of estrous swellings in chim-

panzees (Pan troglodytes), but her methodology is un-

clear and appears to violate the assumptions underlying

her statistical tests. For example, in the study of captive

chimps, her chi-square test violated the assumption of

independence of observations. Adaptive explanations for

menstrual synchrony in humans assume ovulatory con-

cordance or overlapping fertile periods (Burley 1979,

Turke 1984)) but if women's cycles get only two days

closer together over four to six months (McClintock

1971) the impact on the timing of conceptions is proba-

bly negligible.

If menstrual synchrony exists, then during synchroni- zation at least some of the women must shorten or lengthen their menstrual cycles so as to bring their men- strual onsets closer to those of the women with whom they are synchronizing. The plausibility of such adjust- ments needs to be evaluated in the light of other deter- minants of cycle length. In natural-fertility populations, many cycles include transient pregnancies which fail after only a few days or weeks, resulting in hormonal withdrawal, which then triggers menstruation. In these cycles, the timing of menstruation and menstrual cycle length are determined by the timing of pregnancy failure (Wilcox et al. 1988). Among noncontracepting American women, pregnancy occurred in 28% of menstrual cycles (N = 707 cycles), and 3 I% of these pregnancies ended in loss (Wilcox et al. 1988). Thus, in 28% of cycles, men- strual synchrony would have been precluded by preg- nancy.

Many other factors also impinge on cycle length and reduce the potential for menstrual synchrony. For exam- ple, long, irregular intervals between menses are associ- ated with anovulation in the years just after menarche and prior to menopause (Baird 1985). Among women of prime reproductive age, irregular, anovulatory cycles are associated with negative energy balance (Ellison ~ggo), lactation (Howie and McNeilly 1982)) and psychosocial stress (see Wasser and Barash 1983). In a study of 275,947 cycles in 2,702 American women, Treloar et al. (1967) concluded that the menstrual cycle is "characterized by variability rather than by regularity." At age 20, the me- dian cycle length was 27.8 days and the difference be- tween the 10th and 90th percentiles for person-year standard deviations was 6.3 days. Cycle length variabil- ity reached a minimum at age 36, when the median cy- cle length was 26.6 days and the difference between the two percentiles was 3.6 days. The inherent variability in cycle length has two components: (I)the different periodicities (cycle lengths) of different women, and (2) the substantial within-woman variability in cycle length (Treloar et al. 1967, Vollman 1977). Both are ob- stacles for synchrony.

Given the paucity of evidence, it is surprising that belief in menstrual synchrony is so widespread. I suggest that this belief arises, in part, from a popular misconcep- tion about how far apart one would expect the menstrual onsets of two women to be by chance alone (see fig. 5). If two women both have 30-day menstrual cycles, the maximum they can be out of phase is 15 days, and, on


FIG. 5. Menstrual onsets for two women. Top: Onsets are maximally concordant. Bottom: Onsets are maximally out of phase (i.e., half a cycle apart). If oone woman's onsets are independent of the other woman's onsets, then the difference in onsets for the two women is expected to be one-quarter the cycle length, on average.

average, their onsets are expected to be 7.5 days apart (Wallis 1985, Strassmann 1990, Wilson 1992). Fully half the time their onsets should be closer than 7.c davs. If

r., ,

menstrual synchrony among friends has psychological appeal, then onsets that happen to be close together may make a greater impression than onsets that happen to be disparate.

References Cited
BAIRD, D. T. 1985. "Endocrine basis of menstruation and its dis- orders," in Mechanism of menstrual bleeding. Edited by D. T. Baird and E. A. Michie, pp. 165-79. New York: Raven Press.

BENDEL, J.-P., AND C.-I. HUA. 1978. An estimate of the natu- ral fecundability ratio curve. Social Biology zj:z10-27.

BENTLEY, G. R., A. M. HARRIGAN, AND P. T. ELLISON. 1990. Ovarian cycle length and days of menstruation of Lese horticul- turalists (abstract). Amen'can Journal of Physical Anthropol- ogy, suppl.

BUCKLEY, T. 1988. "Menstruation and the power of Yurok women," in Blood magic: The anthropology of menstruation. Edited by T. Buckley and A. Gottlieb, pp. 187-209. Berkeley: University of California Press.

BURLEY, N. 1979. The evolution of concealed ovulation. American Naturalist 114:835-58.

CHIAZZE, L., F. T. BRAYER, J. J. MACISCO, M. P. PARKER, AND B. J. DUFFY. 1968. The length and variability of the hu- man menstrual cycle. Journal of the American Medical Associ- ation 203:377-80.

cox, D. R. 1972. Regression models and life tables. Journal of the Royal Statistical Society B 34:187-220.

CUTLER, W. B., W. M. SCHLEIDT, E. FRIEDMANN, G. PRETI, AND R. STINE. 1987. Lunar influences on the reproductive cy- cle in women. Human Biology 59:959-72.

DIXON, W. J. 1992. BMDP statistical software manual 2. Berkeley: University of California Press. EATON, S. B., M. C. PIKE, R. V. SHORT, N. C. LEE, J. TRUSSELL, R. A. HATCHER, J. W. WOOD, C. M. WORTHMAN,

N. G. BLURTON JONES, M. J. KONNER, K. R. HILL, R. BAILEY, AND A. M. HURTADO. 1994. Women's reproductive Can- cers in evolutionary context. Quarterly Review of Biology

69:353-67. ELLISON, P. T. 1990. Human ovarian function and reproductive ecology: New hypotheses. American Anthropologist 92:933-52. ELLISON, P. T., N. R. PEACOCK, AND C. LAGER. 1989. Eco~- ogy and ovarian function among Lese women of the Ituri For- est, Zaire. American Journal of Physical Anthropology 78:s 19-

526. GRAHAM, C. A. 1991. Menstrual synchrony: An update and re-

view. Human Nature z:z93-311. GRAHAM, C. A., AND W. C. MCGREW. Menstrual synchrony in

female undergraduates living on a coeducational campus. Psychoneuroendocrinology 5:245-52. HENRY,L. 1961. Some data on natural fertility. Eugenics Quar- terly 8:81-91.

HOWIE,P. w., AND A. S. MCNEILLY. 1982. Effect of breastfeed- ing patterns on human birth intervals. Iournal of Reproduction and Fertility 65:545-57.

JAIN,A. K. 1969. Fecundability and its relation to age in a sam- ple of Taiwanese women. Population Studies 23:69-85.

JARETT,L. R. 1984. Psychosocial and biological influences on menstruation: Synchrony, cycle length, and regularity. Psycho-neuroendocrinology 921-28.

MASLAR. 1987. Long ovarian cycles in women of highland
New Guinea. Human Biology 59:837-42.

KNIGHT,C. 1988. "Menstrual synchrony and the Australian rain- bow snake, " in Blood magic: The anthropology of menstrua- tion. Edited by T. Buckley and A. Gottlieb, pp. 232-55. Berkeley: University of California Press.

LAW, S. P. 1986. The regulation of menstrual cycle and its rela- tionship to the moon. Acta Obstetrica et Gynecologica Scandi- navica 65:45-48.

LITTLE, B. B., D. S. GUZICK, R. M. MALINA, AND M. D. RO CHA FERREIRA. 1989. Environmental influences cause synchrony, not pheromones. American Iournal of Human Biol- OgY 1153-57,

MC CLINTOCK, 1971. Menstrual synchrony and suppres-

M. K. sion. Nature 229:244-45. -. 1981. Social control of the ovarian cycle and the function of estrous synchrony. American Zoologist 21:243-256.

PO CHOBRADSKY,J. 1974. Independence of human menstrua-
tion on lunar phases and days of the week. American Iournal
of Obstetrics and Gynecology I 18:1136-38.

PRETI, G., W. B. CUTLER, C. R. GARCIA, G. R. HUGGINS, AND H. L. LAWLEY. 1986. Human axillary secretions influ- ence women's menstrual cycles: The role of donor extract of fe- males. Hormones and Behavior zo:474-82.

QUADAGNO, D. M., H. E. SHUBEITA, J. DECK, AND D. FRANCOEUR.1981. Muence of male social contacts, exercise, and all-female living conditions on the menstrual cycle. Psychoneuroendocrinology 6339-44.

SHORT,R. V. 1976. The evolution of human reproduction. Proceedings of the Royal Society of London B 195:3-24. STRASSMANN, I. 1990. The reproductive ecology of the Dc-

B. gon of Mali. Ann Arbor: University Microfilms. -. 1992. The function of menstrual taboos among the Do- gon: Defense against cuckoldry? Human Nature 3:89-131. -. 1996a. The evolution of endometrial cycles and menstrua- tion. Quarterly Review of Biology 7 I: I81-220.

-. 1996b. Menstrual hut visits by Dogon women: A hor- monal test distinguishes deceit from honest signaling. Behavioral Ecology 7:304-I 5.

TRELOAR, A. E., R. E. BOYNTON, B. G. BEHN, AND B. W. BROWN. 1967. Variation of the human menstrual cycle through reproductive life. International Iournal of Fertility 1x77-126.

TREVATHAN, W. R., M. H. BURLESON, AND W. L. GREGORY. 1993. No evidence for menstrual synchrony in lesbian couples. Psychoneuroendocrinology 18:425-3 5.

TURKE,P. W. 1984. Effects of ovulatory concealment and syn- chrony on protohominid mating systems and parental roles. Ethology and Sociobiology 5:33-44.

VOLLMAN,R. F. 1977. The menstrual cycle. Philadelphia: W. B. Saunders.

WALLIS, J. 1985. Synchrony of estrous swelling in captive group-living chimpanzees (Pan troglodytes). International Jour- nal of Primatology 6:335-50.

-. 1992. "Socioenvironmental effects on timing of first post- partum cycles in chimpanzees," in Topics in primatology, vol. I. Edited by T. Nishida, W. C. McGrew, P. Marler, M. Pick- ford, and F. B. M. de Waal. Tokyo: University of Tokyo Press.

WASSER, S. R., AND D. P. BARASH. 1983. Reproductive suppres- sion among female mammals: Implications for biomedicine

and sexual selection theory. Quarterly Review of Biology 38:513-37.

WELLER,L., AND A. WELLER. 1993. Human menstrual syn- chrony: A critical assessment. Neuroscience and Biobehavioral Reviews 17:427-39.

WILCOX,A. J., C. R. WEINBERG,J. F. O'CONNOR,D. D. BAIRD, J. P. SCHLATTER, R. E. CANFIELD, E. G. ARMSTRONG,AND B. C. NISULA. 1988. Incidence of early loss of pregnancy. New England Journal of Medicine 319:189-94.

WILKINSON, L. 1990. SYSTAT: The system for statistics. Evanston: SYSTAT.

WILSON,H. C. 1987. Female axillary secretions influence women's menstrual cycles: A critique. Hormones and Behav- ior 21:536-46.

-. 1992. A critical review of menstrual synchrony research. Psychoneuroendocrinology17: 565-91. WILSON, H. C., HILDEBRANDT KIEFHABER, S., AND GRAVEL,

v. 1991. Two studies of menstrual synchrony: Negative re-
sults. Psychonezzroendocrinology 16:35 3-59.
WOOD, J. w. 1989. Fecundity and natural fertility in humans.
Oxford Reviews of Reproductive Biology I I :61-109.
WOOD,J. w. 1994. Dynamics of human reproduction. New
York: Aldine de Gruyter.

WOOD, J. w., AND M. WEINSTEIN. 1988. A model of age-
specific fecundability. Population Studies 42:85-I 13.

Population Increase and Sex-Biased Parental Investment in Humans: Evidence from I8th- and I9th-Century Germany1

Zentrum fur Philosophie und Grundlagen der Wissenschaft, Universitat Giessen, Otto-Behaghel-Str. ro C, 0-35394 Giessen, Germany (Vo1and)I Department of Psychology, University of Liverpool,

P.O. Box 147, Liverpool L69 3BX, U.K. (Dunbar) Centro de Supercomputaci6n de Galicia, Avda. de Vigo, sln, E-15706 Santiago de Compostela, Spain (Enge1)lBahnstrasse22, D-03946 Ditfurt, Germany

(Stephan).22 IV 96

Accordng to Fisher (1930)) parents are shaped by natural selection to invest more in the sex of offspring that is expected to yield the greatest net fitness gains for them. Which of the two sexes is favoured depends on two fac- tors: the differences in the per capita costs that parents incur when producing sons and daughters and the differ-

I. @ 1997 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved 0011-3204/97/3801-0010 $1.00. We thank Arthur E. Imhof for generously making the data from his life-expectancy project available. Financial support came from the Deutsche Forschungsgemeinschaft and from the U.K. Biotechnol- ogy and Biological Sciences Research Council. Monique Borgerhoff Mulder provided valuable comments on the manuscript.

  • Recommend Us