Human Ovarian Function and Reproductive Ecology: New Hypotheses
Author(s): Peter T. Ellison
Source: American Anthropologist, New Series, Vol. 92, No. 4 (Dec., 1990), pp. 933-952
Published by: Blackwell Publishing on behalf of the American Anthropological Association
Stable URL: http://www.jstor.org/stable/680653
Accessed: 26/01/2009 21:15
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=black.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].
Blackwell Publishing and American Anthropological Association are collaborating with JSTOR to digitize,
preserve and extend access to American Anthropologist.
http://www.jstor.org
PETERT. ELLISON
HarvardUniversity
Human Ovarian Function and Reproductive
Ecology: New Hypotheses
A seriesof eighthypotheses
is presented,basedon theresultsof currentresearch,concerning
the
to
constitutional
and
environmental
These
the
human
variables.
ovary
of
responsiveness
hypotheses
aremotivated
humanreproductive
positionthatseeksto understand
bya theoretical
physiologyas
theproductof naturalselection.Thehypotheses
are: (1) Ovarianresponsiveness
occursalonga
a final common pathway
of responseforms
graded continuum. (2) Thegradedcontinuum
for various"stresses."(3) Ovarianfunctiontracksenergy balance, not simplynutritional
status.(4) Ovarian
of energybalance.(5) Adfiunctiontracksaerobic activity independently
theinteraction
andenvironmentalfactors
ditive interactions characterize
modofconstitutional
maturationis synchronized with skeletal maulatingovarianfunction.(6) Reproductive
turation, especiallyof thepelvis. (7) Peakovarian
functionis notordinarilyachieveduntilthe
is
twenties.
Late
maturation
associatedwitha slower rise in indicesof
(8)
early
reproductive
ovarian
functionwith age, anda lower level of ovarian
functionin adulthood.Together,these
theoriesoffacultativemodulation
hypotheses
offemalereproductive
providefortwo,non-exclusive
to theprospects
effort.Onetheoryviewsovarianfunctionas responsive
for positivereproductive
outcome
as thesemaybeaffectedbymaternalage,maturation,
energybalance,andactivitylevel.
Thesecondtheoryviewsovarian
in a similarway to theneedto maintain
functionas responsive
maternalenergybalance.
long-term
T HAS BEEN TWENTY YEARS SINCE FRISCH AND REVELLE first published their contro-
versial hypothesis linking menarche to nutritional status in human females (Frisch and
Revelle 1970, 1971). The debate touched off by this hypothesis and its successors has
continued into the present, encompassing criticisms and responses on many levels, ranging from data analysis and interpretation (Johnston, Malina, and Galbraith 1971; Frisch,
Revelle, and Cook 1971; Johnston et al. 1975; Billewicz, Fellowes, and Hytten 1976;
Frisch 1976; Cameron 1976; Trussell 1978; Frisch 1978; Reeves 1979; Trussell 1980; Ellison 1981a, 1982) to health policy and international development (Scott and Johnston
1985; Ellison 1991). Both because and in spite of this debate, interest in the ecological
responsiveness of the human reproductive system has grown among anthropologists, human biologists, demographers, and physiologists, each group initiating data collection
and hypothesis testing by the methods of their respective disciplines. Demographic approaches have been particularly important in shaping current understanding of human
fertility regulation in natural fertility populations (Henry 1961; Bongaarts and Potter
1983; Wood and Weinstein 1988). Yet one of the most important advances in this area,
the recognition of the importance of the temporal pattern of nursing behavior in regulating natural birth intervals, was achieved by integrating an understanding of reproductive
physiology into demographic research (Delvoye et al. 1977; Konner and Worthman 1980;
Short 1983). As a result, the role of lactation as a natural birth-spacing mechanism, at
first the subject of cautious speculation by both demographers (Henry 1961) and cliniPETERT. ELLISONis ThomasD. CabotAssociateProfessor,Department
HarvardUniversity,Cambridge,
of Anthropology,
MA 02138.
933
934
AMERICAN
ANTHROPOLOGIST
[92, 1990
cians (Udesky 1950; Gioiosa 1955), has now virtually achieved the status of a principle
of human population biology (Howie and McNeilly 1982; Short 1987) and public health
policy (WHO/NRC 1983; Kennedy, Rivera, and McNeilly 1989).
During the same two decades, a wealth of information has also been accumulated concerning biological constraints on human ovarian function other than those associated
with breastfeeding. Significant nonpathological variation in ovarian function has now
been associated with age, nutritional status, energy balance, diet, and exercise. Much of
this information has come from clinical and physiological studies, although potential applications to a broader understanding of human ecology have not been lost to anthropologists (Bentley 1985; Peacock 1986, 1990). On the basis of this information it is now
possible to formulate more sophisticated hypotheses than those originally advanced by
Frisch and Revelle, which help to synthesize our current understanding, expose its weaknesses, and identify promising areas of future research.
This article provides such a set of hypotheses with a specific focus on the ecological
responsiveness of human ovarian function to factors other than lactation. My emphasis
is on physiological studies that pertain specifically to ovarian function, and not on demographic data that provide for second-order inferences only. This approach is a reflection both of a theoretical perspective and of more practical considerations. The theoretical perspective is evolutionary; that is, it is assumed that human reproductive physiology
as it now exists is the product of natural selection, encumbered, as always, with historical
and design constraints. From this perspective, variation in ovarian function may represent adaptive flexibility or responsiveness to conditions, rather than pathological dysfunction. Additionally, from this perspective one assumes that modern human environments have changed faster than our endowed physiology. Thus, the mechanisms modulating human ovarian physiology may still be true to their original design, even though
patterns of fertility may be the product of historically novel situations.
As a more practical matter, many of the results of demographic studies are still at variance with the results of physiological studies concerning the ecological regulation of human ovarian function. A reconciliation of these approaches is certainly still possible, even
inevitable, and indeed has begun to be realized'in some quarters (Campbell and Wood
1988; Ford and Huffman 1988; Leslie and Fry 1989; Ellison, Peacock, and Lager 1989).
My goal in this article, however, is not to forge such a reconciliation prematurely. As
study of the demographic importance of the physiology of lactation has taught us, it may
be valuable to push both forms of inquiry forward without ceding ascendancy to either,
in the expectation of benefiting from the dialectic that emerges.
I will first present a series of eight hypotheses concerning ovarian function as an element of human ecology, based on recent research results. These hypotheses are presented
largely in descriptive terms as an attempt to summarize and codify current information.
Yet they are still clearly hypotheses, some more strongly supported than others. As a
group, the hypotheses also provide a basis for initial steps toward a more synthetic theory
of the ecology of ovarian physiology in the second part of this article. Functional interpretations of the patterns described in the eight hypotheses are presented, for the most
part, in the context of that discussion. An ecological theory of ovarian function, as part
of a broader theory of human reproductive ecology, will have immense value, not only to
human biologists, but to students of human culture as well. Indeed, few areas of inquiry
have greater claim to the common ground that still exists among the various subfields of
anthropology than that which has as its center the phenomenon of human reproduction.
The Hypotheses
Hypothesis1. Ovarianfunction shows a graded responsecontinuumto importantecological, behav-
variables.Early studies of age, lactational, nutritional, and exercise
ioral,andconstitutional
effects on female reproductive physiology stressed the onset and maintenance of regular
menstruation. It is now apparent that disruptions in menstrual function represent only
935
0 VARIAN
ECOLOGY
FUNCTIONANDREPRODUCTIVE
Ellison]
the extreme end of a graded continuum of ovarian response to these variables (Figure 1).
Between the extremes of a fully fecund ovarian cycle and the interruption of menstrual
function are more subtle degrees of variation in ovarian function. The mildest form of
ovarian suppression involves a lower profile of progesterone secretion in the post-ovulatory, luteal phase of the cycle. A sufficiently suppressed luteal phase is termed "luteal
insufficiency"or "luteal inadequacy," referringto the compromised ability of the corpus
luteum to provide hormonal support to the endometrium during implantation and early
pregnancy (Riddick et al. 1983; McNeely and Soules 1988; Stouffer 1988). Suppression
of follicular development in the first half of the cycle may occur as well, and indeed is
often a precursor of luteal suppression in the same cycle (DiZerega and Hodgen 1981).
Ovulatory failure, which may occur without noticeable changes in menstrual pattern
(Poindexter and Ritter 1983), represents a yet more profound suppression of ovarian
function, necessarily dropping the probability of conception for that month to zero.
These various forms of ovarian suppression are not independent conditions, but occur
along an ordered gradient reflecting etiology. Mild luteal suppression may occur in isolation, but more profound luteal suppression is ordinarily a consequence of prior follicular suppression (DiZerega and Hodgen 1981). Follicular suppression, if profound
enough, can lead to ovulatory failure (Yoshimura and Wallach 1987). Continued
suppression can lead to menstrual irregularity, oligomenorrhea, and, if the suppression
is both profound and sustained, amenorrhea (Poindexter and Ritter 1983). When ovarian
suppression occurs in association with environmental, behavioral, or constitutional
"stresses," the degree of suppression along this gradient ordinarily correlates with the
intensity of the stress. While extreme thinness (Frisch and McArthur 1974), rapid or extreme weight loss (Graham, Grimes, and Campbell 1979; Pirke et al. 1985), heavy exercise (Bullen et al. 1985), early adolescence (Apter and Vihko 1983), and frequent, intense nursing characteristic of the early postpartum period (McNeilly et al. 1982; Howie
et al. 1982) are all associated with irregularor absent menstrual cycles, moderate thinness
(Bates, Bates, and Whitworth 1982), moderate weight loss (Lager and Ellison 1990),
moderate exercise (Ellison and Lager 1986), late adolescence/early adulthood (Ellison,
Lager, and Calfee 1987), and less intense nursing characteristic of the period of supplemented breastfeeding (Howie et al. 1981, 1982) are all associated with reduced ovulatory
frequency and luteal suppression, even though menses may be regular. Movement across
the gradient of response within the lifespan of a woman can be observed in all these cases:
toward diminished ovarian function with progressive weight loss (Ellison, Peacock, and
Lager 1989), extended exercise (Bullen et al. 1985), or increasing age after 35 years (Doring 1969); and toward augmented ovarian function with increasing age during adolescence (Apter and Vihko 1983; Read et al. 1984; Ellison, Lager, and Calfee 1987), weight
AMENORRHEA
+
OLIGOMENORRHEA
ECOLOGICAL,
OVARIAN
OVULATORY FAILURE
BEHAVIORAL, OR
FUNCTION
FOLLICULAR PHASE SUPPRESSION
CONSTITUTIONAL
LUTEAL PHASE SUPPRESSION
"STRESS"
FULLY COMPETENT CYCLES
+
t
Figure 1
The graded continuum of ovarian response to certain ecological, behavioral, and constitutional variables (cf. Prior 1985; Ellison 1991).
936
AMERICAN
ANTHROPOLOGIST
[92, 1990
gain or stabilization (Bates, Bates, and Whitworth 1982; Lager and Ellison 1990), or declining nursing intensity postpartum (Howie et al. 1981, 1982).
Ordinarily, only the extreme end of this gradient-that which involves disruptions of
menstrual function-lies above the clinical horizon, obvious to a woman or apparent to
her physician from a menstrual history. Detection of moderate suppression of ovarian
function requires more sophisticated and often more invasive procedures, and conditions
in this range often go unnoticed by both women and their physicians. Subtle expression
does not, however, imply benign effect. Luteal suppression of progesterone production,
for instance, is associated with failure of conception and repeated abortion (Lenton et al.
1988; McNeely and Soules 1988), and follicular suppression of estradiol with ova of diminished fertilizability (Yoshimura and Wallach 1987). More important, there is no instance yet reported of an ecological variable associated with disrupted menstrual function
that has not also been associated, at a more moderate level, with these more subtle
changes in ovarian function. Thus, a corollary of this first hypothesis would be that there
is no discrete, "one-zero" modulation of ovarian function short of pathology. Ovarian
function is on a "rheostat," not an "on/off switch."
Hypothesis2. The gradedcontinuumof ovariansuppression
forms a 'final commonpathway"
which
various
and
constitutional
variables
The
behavioral,
through
ecological,
affectfemalefecundity.
points at which weight loss, exercise, youth, age, and lactation intersect with the hypothalamic-pituitary-ovarian(HPO) axis may vary, and any given variable may have more
than one point of intersection. Elevated prolactin levels associated with lactation have
been shown both to be associated with altered hypothalamic regulation of pituitary gonadotropin release (Baird et al. 1979) and to inhibit gonadal steroid production (McNatty et al. 1979). Intense exercise disrupts the hypothalamic-pituitary axis (Cumming
et al. 1985), possibly through increased levels of endogenous opioids (Howlett et al. 1984;
Yen 1984), though testosterone and prolactin levels are raised acutely as well (Shangold,
Gatz, and Thyssen 1981). Weight loss is associated with decreased production of extragonadal estrogens (Siiteri, Williams, and Takaki 1976; Pintor et al. 1980; Ellison 1984)
and hypothalamic dysfunction (Warren et al. 1975; Vigersky et al. 1977). Adolescent
cycles are characterized by elevated androgen to estrogen ratios (Apter, Viinkka, and
Vihko 1978), strong circadian variation in pituitary activity (Boyar et al. 1974), and a
reduced capacity for steroid production (Apter et al. 1987).
Despite this diversity of intersection points with the reproductive endocrine axis, the
final impact of all these variables on female reproductive capacity is mediated by the
gradient of ovarian response described above. That is, changes in the activity of the HPO
axis take on functional significance primarily insofar as they affect the production of ova,
their fertilizability, or the ability of fertilized zygotes to implant and survive early gestation, ovarian function being the principal determinant of each of these. The gradient of
ovarian response thus serves as a "final common pathway" linking a broad array of ecological, environmental, and constitutional factors to human fecundity. Changes "above"
the level of the ovary in the HPO axis express themselves functionally in changes in the
production of ova and their quality (Stanger and Yovich 1984; Yoshimura and Wallach
1987; Eissa et al. 1987), which in turn are conditioned by and/or expressed as changes in
ovarian steroids, while many important events "below" the level of the ovary, such as
implantation and early gestation, are guided and regulated by ovarian steroid production
(Stouffer 1988; Fritz 1988; Eisenberg 1988). Ovarian function thus becomes the "linchpin" of the female reproductive system, with steroid levels being the most useful, readily
detectable reflection of ovarian function.
This hypothesis proves to be an asset to many research tasks of interest to anthropologists, since ovarian activity, reflected in patterns of steroid production, is much easier
to monitor under field conditions than hypothalamic or pituitary activity (Ellison 1988).
Indeed, even if it were a simple matter to monitor the hypothalamic-pituitary axis, it
would still be necessary to inferthe impact of changes at this level on ovarian function in
order to interpret their significance for human fecundity. For instance, it seems likely that
Ellison]
FUNCTIONANDREPRODUCTIVE
ECOLOGY
OVARIAN
937
aerobic exercise suppresses female reproductive function by disrupting the pattern ofhypothalamic secretion of gonadotropin-releasing hormone, possibly as a consequence of
altering endogenous opioid production within the central nervous system (as noted
above). Yet it is not necessary for an anthropologist, interested in the question of whether
women's workloads in a given society affect their reproductive physiology, to attempt to
assay hypothalamic activity. Changes at this level would only have significance for female
fecundity if they are sufficient to suppress ovarian activity to an appreciable degree. A
corollary, however, is that ovarian function integrates the effects of a potentially diverse
array of factors affecting female reproductive physiology, and in so doing loses the power
of discrimination. Our imaginary anthropologist may, for example, find that high workloads in his or her society are associated with significantly suppressed ovarian function,
and yet be unable to say whether that suppression is analogous in its mechanisms to the
suppression associated with aerobic exercise.
trendsin energybalance,notsimplynutritional
status.Even
3. Ovarian
Junctionreflects
Hypothesis
women of normal weight for height show evidence of luteal suppression and ovulatory
failure when losing weight (Bates, Bates, and Whitworth 1982; Pirke et al. 1985; Pirke et
al. 1989; Lager and Ellison 1990). The more precipitous the weight loss, the more severe
the ovarian suppression. Stabilization or reversal of the energy balance leads to a reversal
of the direction of progress through the gradient of ovarian suppression. In addition, a
lag time is observable between a period of weight loss and the full expression of ovarian
suppression (Lager and Ellison 1990). As a result, the cumulative effect of moderate
weight loss over an extended period can be comparable to that of a shorter period of more
precipitous weight loss. We observed ovarian suppression to intensify among a group of
horticulturalist women in the Ituri Forest of Zaire over a four-month period of gradual
weight loss (Ellison, Peacock, and Lager 1989), and to be most severe just after the peak
of a period of seasonal food shortage (Ellison, Peacock, and Lager 1986; Peacock and
Ellison 1984). This pattern of response has the effect of rendering the ovary more sensitive
to medium to long-term trends in energy balance-that is, those that extend for several
months or more-while responses to short-term variations may be somewhat buffered by
the graded nature of ovarian response and the lag time in its expression.
This hypothesis stands in contrast to the various versions of the hypothesis originally
advanced by Frisch and Revelle (1970, 1971), which have all stressed nutritional status
rather than energy balance. In its most recent form (Frisch 1987), the Frisch hypothesis
holds that a minimum percentage of body fat by weight is ordinarily required for the
maintenance or restoration of menstrual cycles in a mature woman, with a different level
of fatness required for the initial onset of menstrual cycles in adolescent girls. Considerable evidence supports the notion that extremes of underweight are associated with disruptions of hypothalamic-pituitary activity (e.g., Warren et al. 1975; Vigersky et al.
1977), but it is not clear from the research protocols whether those conditions of extreme
underweight are themselves associated with persistent negative energy balance. Similar
disruptions are in fact observable in association with simple weight loss as well (Graham,
Grimes, and Campbell 1979). Within the normal weight range, however, there is no evidence of variation in ovarian function that is associated with indices of nutritional status,
as opposed to measures of change in nutritional status (Ellison 1984; Ellison, Lager, and
Calfee 1987; Lager and Ellison 1990).
Hypothesis4. Ovarianfunctionreflectsaerobicactivityindependently
of energybalance.A substantial literature now documents the association of various forms of aerobic exercise with
suppressed ovarian function (cf. Howlett 1987; Cumming 1990; Warren 1990). This literature has drawn the attention of anthropologists interested in the potential impact of
physical activity on female reproductive function (Bentley 1985; Peacock 1986). Yet
questions have also been raised concerning the degree to which exercise-associated
suppression of ovarian activity is independent of other, potentially confounding variables,
such as nutritional status/body composition and other constitutional factors (Frisch, Wyshak, and Vincent 1980; Frisch et al. 1981; Malina 1983). Frisch (1987) has suggested
938
AMERICAN
ANTHROPOLOGIST
[92, 1990
that reproductive suppression in female athletes may be mediated by altered fat/lean ratios of body composition, while Malina (1983) has suggested that female athletes may
represent a biased sample of the female population selected for constitutional or developmental attributes-for example, high androgen profiles or late maturation-which
might be correlated independently with suppressed ovarian function.
By subjecting untrained women to rather vigorous exercise regimes while supplying
some with sufficient calories to maintain their weight, Bullen et al. (1985) were able to
show that exercise suppresses ovarian function independently of weight loss and of other
attributes self-selected in successful female athletes. Catherine Lager and I (Ellison and
Lager 1985, 1986) observed milder suppression of ovarian function in association with
moderate recreational running and stable weight in women of normal weight for height.
In addition to confirming that exercise-associated ovarian suppression is not limited to
trained or excessively lean athletes, this study also indicated that aerobic activity need
not be at an extreme level to measurably affect ovarian function, since the average running distance was only 12.5 miles per week. Moderate exercise of an hour or more a day
has also been linked to a 4.7-fold increase in the risk of infertility among nulligravid
women (Green et al. 1986), indicating that changes in ovarian function of this kind and
magnitude can have a significant effect on reproductive performance.
An important caution must be entered here, however, for those interested in the potential applicability of this hypothesis to other forms of physical activity, including those
more typical of the workloads and activity patterns of women in traditional societies. The
data on exercise and female reproductive function derive almost exclusively from studies
of periodic, aerobic activities, such as running, gymnastics, swimming, dancing, and rowing. There are few if any data linking patterns of low-level, chronic activity-such as
walking, carrying children, and cultivating gardens-with changes in ovarian activity.
Several ongoing field studies were designed specifically to address this question. Until
results of these or similar studies are in, however, one must be careful not to extend this
hypothesis beyond the limits of the data that give rise to it.
5. Additiveinteraction
thecombined
characterizes
Hypothesis
effectsof ecological,behavioral,and
constitutional
variableson ovarianfunction.When two or more of the factors capable of suppressing ovarian function are present simultaneously, they often interact additively, increasing the resulting degree of ovarian suppression. Lunn et al. (1980; Lunn, Austin,
and Whitehead 1984) have shown that prolactin levels (ordinarily strongly correlated
with the degree of ovarian suppression) of nursing mothers in The Gambia vary with food
availability and can be lowered by dietary supplementation. Bullen et al. (1985) have
shown that the combined effects of heavy exercise and weight loss produce more severe
ovarian suppression than exercise alone (Figure 2). Schweiger et al. (1987) have shown
that the degree of ovarian suppression associated with weight loss depends on age, being
greater in late adolescent/young adult women.
The catalog of possible interaction effects is by no means completely studied even for
dyadic combinations, much less for more complex suites of variables. Yet it is just such
combinations of factors that are likely to be encountered in many natural fertility populations: high workload in combination with poor energy balance, poor energy balance in
combination with intense and prolonged lactation, and so on. Particularly under-studied
at this point are interactions between age and lactation, although demographic data have
long led to speculation that lactational suppression of ovarian function is more profound
in older mothers (Henry 1961; Potter et al. 1965; Goldman, Westhoff, and Paul 1987). It
would be equally interesting to investigate further the interaction between energy balance
and age implied by the work of Schweiger et al. (1987). If it is true that young women
are particularly sensitive to the effects of energy balance, then the practice of feeding girls
in conjunction with puberty ceremonies, which has a relatively wide cross-cultural distribution, may take on new meaning as a cultural attempt to manipulate the fecundability
of newly marriageable females.
939
OVARIANFUNCTION AND REPRODUCTIVE ECOLOGY
Ellison]
EXERCISE
ALONE
EXERCISEPLUS
WEIGHTLOSS
10080--
80 -
On
B
10-Z
BIK
60- -
*
*
_*
*
~EFFECT
I---
40 20-
.
20 -
111
m|
None
Luteal suppression
H
o
ON OVARY
1^
*-
failure
Ovulatory
(Bullen et al. 1985)
MONTH I
MONTH II
MONTH I
MONTH II
Figure 2
with those of exercise and weight loss comalone
of
exercise
of
the
effects
Comparison
bined in inducing suppression of ovarian function in previously untrained women (data
from Bullen et al. 1985).
Hypothesis6. Female reproductivematurationis synchronizedwith skeletal maturation,especially
of thepelvis. During our evolutionary history the female pelvis has been subjected to two
contrasting sets of selection pressures: those concerning bipedal locomotion, favoring narrower pelves, and those concerning parturition, favoring a more ample birth canal. As a
consequence, the size and maturity of the pelvis act as important constraints on successful
reproduction. In the life cycle of an individual female, these mechanical constraints must
first be overcome before energetic constraints on reproduction become predominant. As
a result, the milestones of female reproductive maturation show a clear and close relationship to the progress of skeletal maturation (Shuttleworth 1937; Tanner 1962; Marshall 1974; Marshall and Limongi 1976), and pelvic maturation in particular (Moerman
1982; Ellison 1982). The relationship between menarcheal age and skeletal/pelvic maturation is stronger than that between menarcheal age and relative weight or fatness,
which are themselves correlated with skeletal maturation (Ellison 1981a, 1982), and
holds both within and between populations (Ellison 1982).
This hypothesis, like hypothesis 3, stands in contrast to the original Frisch and Revelle
hypothesis and its successors. The original Frisch-Revelle hypothesis (1970, 1971)-that
menarche was associated with the attainment of a "critical weight"-was
plagued by
faulty statistical analyses, and despite extended critique in the literature (ohnston, Malina, and Galbraith 1971; Frisch, Revelle, and Cook 1971;Johnston et al. 1975; Billewicz,
Fellowes, and Hytten 1976; Frisch 1976; Cameron 1976; Trussell 1978; Frisch 1978;
Reeves 1979; Trussell 1980; Bongaarts 1980; Menken, Trussell, and Watkins 1981; Ellison 1981a, 1981b, 1982; Frisch 1982) these failings continue to be underappreciated. Part
of the confusion stems from the publication of the original analyses in two journals, one
of which-Science-is
widely available and often cited, the other of which-Archives of Diseasein Childhood-is less available and less frequently cited. Yet it is in the second journal
that the most telling feature of the analysis, a plot of the data for weight and age at menarche that gave rise to the hypothesis, is featured (Frisch and Revelle 1971). As that plot
makes abundantly clear, the relationship between weight at menarche and age at menarche is a virtual textbook example of a zero correlation. For example, using data from
the Berkeley Growth Study (Tuddenham and Snyder 1954)-the largest of the three lonthe same relationship can
gitudinal growth studies used by Frisch and Revelle-virtually
be demonstrated between menarcheal age and the ordinal position of the 67 female sub-
AMERICAN
ANTHROPOLOGIST
940
[92, 1990
jects in the published version of the data as between menarcheal age and menarcheal
weight (Figure 3). Rather than use a correlational analysis, however, Frisch and Revelle
subjected the data to an analysis of variance-an unnecessary and inappropriate analysis
for examining the relationship between two continuous measurement variables-concluding that the mean weight at menarche does not change across arbitrary menarcheal
age groups. It is this analysis that is summarized, without a figure, in Science (Frisch and
Revelle 1970).
In fact, menarcheal age and menarcheal weight are as unrelated as two variables that
are both reflections of overall development can be. This point was belabored by early
critics of the hypothesis, who framed it in terms of the enormous variance about the mean
menarcheal weight in each menarcheal age class (Johnston, Malina, and Galbraith
1971). Subsequent amendments to the original hypothesis were essentially attempts to
escape from this trenchant criticism. For example, the reduction in the residual variance,
which at first seemed to be achieved by the introduction of estimates of percent body fat
(Frisch, Revelle, and Cook 1973), was later shown to be an ephemeral artifact of neglecting the error term in the regression equations used to form the estimates (Reeves 1979).
Finally, the notion of an averageweight or fatness at menarche was dropped altogether in
favor of the notion of a minimumfatness required (Frisch and McArthur 1974). Yet even
this concept is poorly supported by the data on menarche, since crossing the minimum
threshold of fatness does nothing to change the probability of experiencing menarche in
the next interval of time (Billewicz, Fellowes, and Hytten 1976; Trussell 1978, 1980; Ellison 1981a, 1982), a fact that is reflected in the persistent enormous variance in the fatness at which menarche actually occurs. One might as well propose a minimum shoe size
as necessary for the attainment of menarche on the basis of some empirical value below
which few if any women are observed to menstruate.
Skeletal maturation and pelvic size do meet the criteria of a functional minimum that
fatness and weight fail to meet (Ellison 1981a, 1981b, 1982), and thereforeprovide a more
credible working hypothesis. One can even question the Frisch hypothesis on theoretical
grounds and ask: Why, if fatness is the governing constraint on reproductive maturation,
do girls bother to grow "up"? Why not simply stop growing at an earlier age and gain
weight? At least part of the answer is simple: mechanical constraints on reproduction
exist prior to energetic constraints. Until a woman has achieved the physicalcapacity for
80 -
80 -
y = 55.343 - 0.41053x
R^2 = 0.003
y = 40.127 - 0.47923x
70-
0*
DS
60 '
<?
*
60-
*
**'**
r:**40_
0
C0
3450-
*,
,*
030
10
?h
oria
0**,
11
12
Age
elito
13
*
*
*
.
* .
.
14
If at Menarche
????e
?a
Growth
(Tudenham
tudy
-
6
.
*
?
*
?60
R^2 = 0.001
u3
*
15
;
(yrs)
addaa
-
0
?5
*.*:o
40
S
l
0
16
ujet
Snyde
*, ,
10
~-~
20
,ulse
-
11
12
auaio
Age
*
l
14
15
.
13
at Menarche
fteBr
16
(yrs)
19%00
Figure 3
Comparison of the relationship between menarcheal age and menarcheal weight among
subjects in the Berkeley Growth Study with the relationship between menarcheal age and
the ordinal position of those same subjects in the published tabulation of the Berkeley
Growth Study data (Tuddenham and Snyder 1954).
Ellison]
ECOLOGY
FUNCTIONANDREPRODUCTIVE
OVARIAN
941
successful reproduction-at least partially a function of size and scale-energetic capacity is irrelevant.
If the pelvic size hypothesis is at this stage more credible, it is by no means confirmed;
more data are surely needed. There are, however, two problems in particular contributing to confusion in this area that ought to be appreciated. First is simply the highly correlated nature of all physical measurements during maturation (Ellison 1981a, 1982); one
variable (e.g, weight) may appear to bear a significant relationship to reproductive maturation, which it in fact gains through its correlation with another variable (e.g., pelvic
size). Second is the developmental sequence in which limiting constraints may be arranged. That is, for example, pelvic constraints on successful reproduction may exist
prior to energetic constraints developmentally, and yet those energetic constraints may
be very real once the pelvic constraints have been overcome. Anorectic adolescents may
pass the pelvic threshold and yet be under an energetic threshold, obscuring the relevance
of the prior constraint. For this reason it is, I think, necessary to keep hypotheses concerning ecological influences on reproductive maturation (e.g., this hypothesis) distinct
from hypotheses concerning ecological influences on adult reproductive function (e.g.,
hypothesis 3), and not be tempted to conflate them into two manifestations of the same
hypothesis, as Frisch and McArthur (1974) have done.
untiltheearlytwenties.During the
functionis notordinarilyachieved
Hypothesis7. Peakovarian
first years after menarche, menstrual cycles are characterized by a high frequency ofovulatory failure and luteal phase defects (Apter, Viinkka, and Vihko 1978;Vihko and Apter
1984;Venturoli et al. 1987). Significant suppression of ovarian function is still observable
in women aged 18 to 22 years relative to women 23-35 years, after controlling the potentially confounding factors of exercise and energetic balance (Doring 1969; Ellison, Lager,
and Calfee 1987). This extended process of ovarian maturation is parallel to the final
stages of skeletal maturation and pelvic remodeling, which are not ordinarily complete
until the late teens/early twenties (Moerman 1982). The probability of success and the
quality of pregnancy outcome are both lower in women who are postmenarcheal but still
adolescent (Liestol 1980; Wyshak 1983), possibly reflecting the combined effects of ovarian immaturity and nutritional competition between the fetus and its still growing
mother.
A notion of "adolescent sterility" has long maintained a place in the literature on the
fertility of young women (Montagu 1946). More recently the concept has been revised
into a more sophisticated notion of "adolescent subfecundity," but empirical attempts to
measure its potency from demographic data are problematic (Foster et al. 1986). It is
probably best to recognize that several different factors may be contributing to patterns
of fertility in this age range, including psychosocial factors as well as biological ones, and
to attempt to understand each separately. In addition, any attempt to model evolutionary
influences on human sexual behavior and mating patterns based on assumptions about
the age-specific fecundability of females depends on an adequate understanding of age
variation in ovarian function.
In this context it is notable that accurate information on changes in ovarian function
with increasing age is presently inadequate for the formulation of a specific hypothesis.
If anything, the data support a decline in fecundity after age 35 without becoming precipitous until after age 40 (Doring 1969; Federation CECOS, Schwartz, and Mayaux
1982; Menken 1985; Menken, Trussell, and Larsen 1986), but it is difficult as yet to be
sure how significant this decline is, or the extent to which it represents ongoing attrition
through menopause versus declining ovarian function prior to menopause. It should soon
be possible, however, to generate a specific hypothesis about aging and ovarian function
to complement the picture in adolescence and young adulthood.
maturationis associatedwith a slowerrise in indicesof ovarian
Hypothesis8. Late reproductive
fiunctionwith age, anda lowerlevelof ovarianfiunctionin adulthood.The principal evidence for
this hypothesis comes from the work of Apter and Vihko (Apter and Vihko 1983; Vihko
and Apter 1984), although the hypothesis itself was foreshadowed by Frisch on the basis
942
[92, 1990
AMERICAN
ANTHROPOLOGIST
of historical data (Frisch 1978). In studying the menstrual cycles of adolescent girls, Apter and Vihko (1983) found that ovarian function increased with age at an absolutely
faster rate among girls with an earlier menarcheal age than among those with a later
menarcheal age. Girls who reached menarche at or before age 12 years, for example,
achieved a 50% ovulatory rate within 1 year after menarche, while girls who did not reach
menarche until age 13 years or older took 4.5 years to reach the same ovulatory frequency.
The evidence is equivocal over the question of when, if ever, late maturers catch up to
early maturers in ovarian function (cf. Apter and Vihko 1983; Vihko and Apter 1984).
Venturoli et al. (1987) have also found that an earlier establishment of regular cycles in
adolescence is associated with an earlier attainment of adult endocrine profiles of ovarian
function. In their study, persistently irregular but ovulatory cycles in adolescence are
associated with slower progress in ovarian function with age, though the trajectory eventually catches up with that associated with early establishment of regular cycles. Persistent anovulatory cycles, they suggest, may represent a basic physiological defect that
often continues into adulthood. To this may be added data from studies of linkages within
the gynecological histories of individual women (Gardner 1983; Gardner and Valadian
1983), which indicate that late maturers continue to have higher incidences ofoligomenorrhea and dysmenorrhea than early maturers well into their reproductive prime.
A good body of cross-cultural data does not yet exist to test whether the relationship
between tempo of growth and level of adult ovarian function also holds between populations. It is perhaps noteworthy that among the Lese of Zaire, where menarche does not
ordinarily occur until roughly age 16 years, the trajectory of ovulatory frequency by age
parallels that found among Boston women, but at a significantly lower level (Figure 4).
Nor can that difference in level be accounted for entirely by the suppressive effects of
weight loss and nutritional stress (Ellison, Peacock, and Lager 1989). It is important,
however, to remember that in any population variation in menarcheal age has both heritable and environmental components (Johnston 1974). We do not yet know whether variation in the level of adult ovarian function is correlated with one or both of these sources
100-
U* LESE
'
O
80O-
o
60- -
-J
40 -
!
BOSTON
.
|
"
1
U_
0ol
20-.
0,
K
|::
0
16- 22
23 -35
36-50
AGE GROUP
Figure 4
Lese and
for Lese
Ovulatory frequency
and Boston
Peacock, and
Boston women
women (Ellison,
and Lager
Ovulatory
(Ellison, Peacock,
frequency by
by age
age for
Lager
1989).
1989).
Ellison]
FUNCTIONANDREPRODUCTIVE
O VARIAN
ECOLOGY
943
of variance in menarcheal age. If the correlation holds for one but not the other of these
sources of variance, the potential exists for empirical studies to contradict each other depending on the relative weight given to environmental versus genetic variance in menarcheal age by different research designs.
This hypothesis is perhaps the most controversial of the set. Much of the research in
this area is not motivated by any concern to understand age-specific patterns of female
fecundity, but rather to understand an apparent association between early menarcheal
age and increased risk of breast cancer. Furthermore, contradictory data exist, particularly from earlier studies. For example, MacMahon et al. (1982) found no significant
relationship between menarcheal age and ovulatory frequency after controlling for years
since menarche. Their study, however, was based on urinary pregnanediol measurements
of samples originally collected for other purposes by a variety of protocols and representing a heterogeneous group of subject populations. These methodological problems
make their data difficult to interpret.
Although the precise form of the relationship may yet be unclear, the bulk of evidence
to date supports the notion that "defects" of ovarian function in adolescence are associated to a significant degree with similar "defects" later in adulthood. The thrust of this
hypothesis is to question the assumption of pathology that is often attached to patterns
of ovarian function that depart from the textbook typology of"normal" and to substitute
more neutral descriptive language based on the continuum of ovarian function described
in hypothesis 1. The introduction of such neutral language then opens the way for functional, as opposed to dysfunctional, interpretations of the patterns and autocorrelative
relationships in ovarian function across the lifespan of individual women that are so often
observed. Such a functional interpretation will be introduced in the discussion of an ecological theory of ovarian function, below.
Toward an Ecological Theory of Human Ovarian Function
Frisch (1984, 1988) took the first steps toward an ecological theory of human ovarian
function by suggesting that the purported relationship between fatness and fecundity had
evolved to prevent wasteful attempts at reproduction when a woman's energetic reserves
were inadequate to provide for a very high probability of success. Perhaps it has been the
power of this functional interpretation of variation in menstrual function that has sustained the Frisch hypothesis in the face of overwhelming empirical critique. Nor should
the value of this contribution be minimized; it has, in fact, transformed most subsequent
discussions of human reproductive maturation and adult reproductive function and injected a healthy respect for normal variation and its functional significance into the overly
typological realm of clinical research. I believe, in fact, that it will be this theoretical
contribution that will eventually remain as Frisch's enduring intellectual legacy.
The hypotheses reviewed here support two somewhat different theoretical positions,
not necessarily mutually exclusive, concerning the ecological responsiveness of human
ovarian function. One perspective views the pattern of ovarian responsiveness as an
evolved mechanism to optimize the probability of a successful reproductive outcome in
any individual instance of reproductive effort. The other perspective views the same pattern as a mechanism for ensuring long-term energy balance in reproducing females and
avoiding a downward spiral of physical depletion brought on by successive cycles of pregnancy and lactation.
Although many of the details of the empirical relationships between ecological, behavioral, and constitutional variables on the one hand and ovarian function on the other
differ from those specified in her hypotheses, it is possible to carry over Frisch's theoretical position in altered form. That is, the hypotheses enunciated here can still be incorporated into a theory stipulating that ovarian responsiveness has evolved to reduce investment of reproductive effort when the probability of successful outcome is sufficiently
low. The constraints on the probability of successful reproductive outcome are different,
944
AMERICAN
ANTHROPOLOGIST
[92, 1990
however. In adolescence, as discussed in detail above, physical constraints associated
with pelvic size are more predictive of a successful reproductive outcome than are energetic constraints. In adulthood, energetic constraints appear to become effective, but
more in terms of energy balance than in terms of stored energy supplies. Such an interpretation would depend on other linkages being in place, many of which have empirical
support. For instance, infant mortality risk is highly associated with low birth weight
(Shapiro, Schlesinger, and Nesbitt 1968). Full-term birth weight is associated with maternal pregnancy weight and pregnancy weight gain (Eastman andJackson 1968). If negative energy balance results in lower prepregnancy weights and lower weight gains in
pregnancy, then it could in fact be predictive of a poor reproductive outcome. Conversely,
positive energy balance would be predictive of a more positive reproductive outcome. The
fact that ovarian function appears more responsive to periods of negative energy balance
that extend over more than one month supports this view. Especially under marginal
nutritional conditions, weight gain in early pregnancy may provide a crucial supply of
stored energy for the demands of later pregnancy and lactation (Prentice and Whitehead
1987). The pattern of ovarian responsiveness to weight loss may have been designed to
minimize the probability of initiating a pregnancy in a state of negative energy balance
when the potential for this first trimester fat accumulation is compromised. From this
perspective the responsiveness of ovarian function to energy balance may be interpreted
as a mechanism to modulate reproductive effort in accord with its expected returns (i.e.,
the probability of a successful pregnancy). If the costs of a given level of reproductive
effort are fixed, in terms of the mortality risk incurred by the mother, and the other opportunities for reproduction foregone, then adjusting reproductive effort on the basis of
its probability of successful outcome would be favored by natural selection.
The effect of aerobic exercise on ovarian function can be incorporated into this theoretical framework if its anabolic consequences are taken into account. That is, regular
exercise shifts individual metabolism in favor of protein anabolism as part of a suite of
physiological adjustments to what is perceived by the body as a new chronic level of physical demand. These changes increase the metabolic requirements of the individual
woman even if her weight remains stable, effectively reducing the fraction of her energy
intake available for reproduction. From the point of view of the potential for gaining
weight in early pregnancy, this situation may simulate negative energy balance to a certain degree.
The most likely functional interpretation of hypothesis 8 in this context is that slow
physical development and reproductive maturation provide a kind of "bioassay" of
chronic qualities of the environment, particularly energy availability. Late maturation
relative to genetic potential indicates a chronically poor nutritional environment and
early development a chronically rich one. The correlation between adult ovarian function
and age at reproductive maturation suggests that the baseline for adult ovarian function
may be established in accordance with this assay of chronic environmental quality. One
can question, however, the wisdom of curtailing reproductive effort if there is little likelihood that currently unfavorable circumstances will soon improve. That is, a strategy to
avoid the costs of reproducing under conditions of poor energy balance only works if those
conditions can, in fact, be avoided. Shouldn't a woman living under chronically stressed
conditions set her baseline for ovarian function at the same level as a woman living under
chronically better conditions, while retaining the ability to make facultative adjustments
in ovarian function as environmental conditions fluctuate about their average levels? To
the extent that prepregnancy weight and pregnancy weight gain are independent correlates of reproductive outcome, the answer to this question may be "no." The time available for weight gain in pregnancy is fixed, so that a lower level of ovarian function prior
to conception does not alter the potential for first trimester weight gain. It does, however,
extend the opportunity for prepregnancy weight gain, an opportunity that may be particularly important if the previous sequence of pregnancy and lactation has resulted in a
net energy drain on the mother.
Ellison]
FUNCTIONAND
OVARIAN
REPRODUCTIVE
ECOLOGY
945
The consideration of long-term energy balance in reproducing women is central to the
second theoretical perspective that can be supported by the hypotheses reviewed here. It
is necessary, however, to shift one's thinking to a longer time frame: rather than consider
the effect of patterns of ovarian function on immediate reproductive outcomes, this perspective considers the effect on maternal morbidity and mortality, with benefits and costs
measured in terms of changes in net reproductive success over a woman's lifetime. According to this view, importance is attached to the maintenance of a long-term energy
balance in reproducing women (Prentice et al. 1981; Adair 1984; Whitehead, Paul, and
Ahmed 1986; Prentice and Whitehead 1987). Each period of pregnancy and lactation
constitutes an elevated energy demand over the nonpregnant, nonlactating state; under
marginal nutritional circumstances these states can even constitute a net energy drain,
which must be balanced by a net positive energy balance during the waiting time to conception separating periods of lactational amenorrhea and pregnancy in natural fertility
populations.
Short (1983) has emphasized the relative rarity of menstrual cycles in the reproductive
history of noncontracepting women. The majority of women's reproductive lives under
conditions of natural fertility are spent pregnant or lactating, with only a handful of menstrual cycles experienced as waiting time to conception every two to four years. Recent
models of proximate determinants of human fertility have also suggested that the variance in interbirth intervals attributable to variation in waiting time is only minimal to
moderate, and certainly far less than that attributable to variation in the length of lactational amenorrhea (Bongaarts and Potter 1983; Wood and Weinstein 1988). However,
the period of menstrual cycling during the waiting time to conception may constitute the
only opportunity for recouping maternal condition and achieving a net energy balance
over a woman's reproductive lifetime. The functional significance of this period increases
as chronic energy availability decreases and other opportunities to meet the increased
energetic demands of reproduction dwindle. Thus, the functional importance of the
length of the waiting time to conception may not lie in its effect on interbirth intervals,
but in its effect on maternal morbidity and mortality.
The hypotheses reviewed in this paper are compatible with this theoretical view as
well. Negative energy balance, aerobic exercise as indicative of both increased energy
expenditure and an adjusted internal energy budget, and slow physical and reproductive
maturation as indicative of chronically poor energy availability all decrease a woman's
ability to recoup her physical condition and reestablish a long-term energy balance. Suppressed ovarian function under these circumstances increases the waiting time to conception and hence the time available to achieve these goals. Although superficially similar
to the first theoretical perspective, this perspective views the pattern of variation in ovarian function as adjusting reproductive effort in accordance with its costs, rather than in
accordance with its expected returns. While the first view focuses on the rate of production of offspring, the second focuses on the morbidity and mortality probabilities faced
by the mother. There are practical implications as well. Adjustments in waiting time to
conception may have only a moderate impact on interbirth intervals, and thus not figure
strongly in analyses that use total fertility rate-the average number of offspring ever
born to post-reproductive women (Bongaarts and Potter 1983; Campbell and Wood
1988). If the major effects of facultative changes in waiting time to conception are on
maternal mortality, then an analysis using net fertility rate-average offspring ever born
to a woman, including allowance for reproduction lost through increased mortalitywould be called for. It is perhaps paradoxical that the functional significance of facultative variation in ovarian function might lie in its impact not on fertility, but on mortality.
These two theoretical perspectives are, of course, not mutually exclusive. The patterns
of variation in ovarian function summarized in the eight hypotheses presented here may
simultaneously meet two selective pressures and hence have dual functional significance.
A more robust theory will only evolve as the hypotheses themselves are confirmed, adjusted, or replaced. In the meantime, however, such a nascent theory can itself suggest
946
AMERICAN
ANTHROPOLOGIST
[92, 1990
further hypotheses to be considered. In particular, one area of potential importance in
the ecology of ovarian physiology is still largely unexplored, that being the relationship
between chronic disease burden and ovarian function. Many specific diseases can, of
course, have tremendous impact on fecundity, though they may not affect ovarian function in particular. Examples would include venereal diseases, which may cause lesions of
the reproductive tract, and malaria, which may compromise placental oxygen transport.
McFalls and McFalls (1984) provide a useful review of the catalog of such specific diseases and their effects on reproduction. Little is known, however, about the impact of
chronic disease morbidity per se on ovarian function. To the extent that high disease
burdens may divert metabolic energy to support immune function, tissue repair, and
other responses, the energy available for reproduction may be reduced. Hence one might
predict that ovarian function should vary inversely with disease burden by analogy with
the effects of weight loss and exercise. Another clear implication of these initial theoretical
formulations is the importance of a life-historical perspective on human reproductive
function. A new emphasis will have to be placed on long-term, prospective studies that
follow individual women through adolescence and reproductive maturity into old age in
order to properly understand the longitudinal relationships that probably obtain among
aspects of growth and maintenance, morbidity, and fecundity. Such an undertaking, akin
to the establishment of longitudinal growth studies in the early part of this century, requires vision on the part of funding agencies and patience and persistence on the part of
researchers, yet promises returns well worth the investment.
Most important, however, a theoretical perspective of this kind allows for a broader
view of human reproductive physiology than the overly eidetic view of most clinical textbooks. As with other areas of human biology, ovarian function demonstrates a considerable range of variability. Attempting to force that variability into a dichotomous framework of health and pathology obscures more than it enlightens our understanding. To
the extent that we are able to discover functional significance in the variability we observe, we improve our ability to incorporate human reproductive ecology into the broader
context of human evolution.
I am indebted to all my colleagues at Harvard and on the Ituri Project for their
Acknowledgments.
generosity of mind and spirit, especially Robert Bailey, Gillian Bentley, Ellen Bledsoe, Isabel Bradburn, Ben Campbell, Viviana Carro, Irven DeVore, Marion Eakin, John Fisher, Grazyna Jasienska, Alisa Harrigan, Mark Jenike, Catherine Lager, Susan Lipson, Mark Meredith, Mary
O'Rourke, Nadine Peacock, Helen Strickland, and Christine Thorne. This work is supported by
the National Science Foundation, grants BNS-83-19629, BNS-86-09116, and BNS-88-10931.
References Cited
Adair, L. S.
1984 Marginal Intake and Maternal Adaptation: The Case of Rural Taiwan. Current Topics
in Nutrition and Disease 11:33-55.
Apter, D., and R. Vihko
1983 Early Menarche, A Risk Factor for Breast Cancer, Indicates Early Onset of Ovulatory
Cycles. Journal of Clinical Endocrinology and Metabolism 57:82-86.
Apter, D., L. Viinkka, and R. Vihko
1978 Hormonal Pattern of Adolescent Menstrual Cycles. Journal of Clinical Endocrinology
and Metabolism 47:944-954.
Apter, D., I. Raisanen, P. Ylostalo, and R. Vihko
1987 Follicular Growth in Relation to Serum Hormonal Patterns in Adolescents Compared
with Adult Menstrual Cycles. Fertility and Sterility 47:82-88.
Baird, D. T., A. S. McNeilly, R. S. Sawers, and R. M. Sharpe
1979 Failure of Estrogen-Induced Discharge of Luteinizing Hormone in Lactating Women.
Journal of Clinical Endocrinology and Metabolism 49:500-506.
Bates, G. W., S. R. Bates, and N. S. Whitworth
1982 Reproductive Failure in Women Who Practice Weight Control. Fertility and Sterility
37:373-378.
Ellison]
FUNCTIONANDREPRODUCTIVE
ECOLOGY
OVARIAN
947
Bentley, G. R.
1985 Hunter-Gatherer Energetics and Fertility: A Reassessment of the !Kung San. Human
Ecology 13:79-109.
Billewicz, W. Z., H. M. Fellowes, and C. A. Hytten
1976 Comments on the Critical Metabolic Mass and the Age of Menarche. Annals of Human
Biology 3:51-59.
Bongaarts,J.
1980 Does Malnutrition Affect Fecundity? A Summary of Evidence. Science 208:564-569.
Bongaarts,J., and R. G. Potter
1983 Fertility, Biology, and Behavior: An Analysis of the Proximate Determinants. New York:
Academic Press.
Boyar, R. M.,J. Katz,J. W. Finkelstein, S. Kapen, H. Weiner, E. D. Weitzman, and L. Hellman
1974 Anorexia Nervosa: Immaturity of the 24-hour Luteinizing Hormone Secretory Pattern.
New England Journal of Medicine 291:861-865.
Bullen, B. A., G. S. Skrinar, I. Z. Beitins, G. vonMering, B. A. Turnbull, andJ. W. McArthur
1985 Induction of Menstrual Disorders by Strenuous Exercise in Untrained Women. New England Journal of Medicine 312:1349-1353.
Cameron, N.
1976 Weight and Skinfold Variation at Menarche and the Critical Body Weight Hypothesis.
Annals of Human Biology 3:279-282.
Campbell, K. L., andJ. W. Wood
1988 Fertility in Traditional Societies. In Natural Human Fertility: Social and Biological
Mechanisms. P. Diggory, M. Potts, and S. Teper, eds. Pp. 39-69. New York: Macmillan.
Cumming, D. C.
1990 Physical Activity and Control of the Hypothalamic-Pituitary-Gonadal Axis. Seminars in
Reproductive Endocrinology 8:15-24.
Cumming, D. C., M. M. Vickovic, S. R. Wall, and M. R. Fluker
1985 Defects in Pulsatile LH Release in Normally Menstruating Runners. Journal of Clinical
Endocrinology and Metabolism 60:810-812.
Delvoye, P., M. Demaegd,J. Delogne-Desnoek, and C. Robyn
1977 The Influence of the Frequency of Nursing and of Previous Lactation Experience on
Serum Prolactin in Lactating Mothers. Journal of Biosocial Science 9:447-451.
DiZerega, G. S., and G. D. Hodgen
1981 Luteal Phase Dysfunction Infertility: A Sequel to Aberrant Folliculogenesis. Fertility and
Sterility 35:489-499.
Doring, G. K.
1969 The Incidence of Anovular Cycles in Women. Journal of Reproduction and Fertility
(Suppl.) 6:77-81.
Eastman, N.J., and E.Jackson
1968 Weight Relationship in Pregnancy, 1: The Bearing of Maternal Weight Gain and prePregnancy Weight on Birth Weight in Full Term Pregnancies. Obstetrics and Gynecology
Survey 23:1003-1025.
Eisenberg, E.
1988 Luteal Phase Insufficiency in Endometriosis. Seminars in Reproductive Endocrinology
6:251-255.
Eissa, M. K., R. S. Sawers, M. F. Docker, S. S. Lynch, andJ. R. Newton
1987 Characteristics and Incidence of Dysfunctional Ovulation Patterns Detected by Ultrasound. Fertility and Sterility 47:603-612.
Ellison, P. T.
1981a Threshold Hypotheses, Developmental Age, and Menstrual Function. American Journal of Physical Anthropology 54:337-340.
1981b Prediction of Age at Menarche from Annual Height Increments. American Journal of
Physical Anthropology 56:71-75.
1982 Skeletal Growth, Fatness and Menarcheal Age: A Comparison of Two Hypotheses. Human Biology 54:269-281.
1984 Correlations of Basal Oestrogens with Adrenal Androgens and Relative Weight in Normal Women. Annals of Human Biology 11:327-336.
1988 Human Salivary Steroids: Methodological Considerations and Applications in Physical
Anthropology. Yearbook of Physical Anthropology 31:115-142.
948
A MERICAN
ANTHROPOLOGIST
[92, 1990
1991 Reproductive Ecology and Human Fertility. In Applications of Biological Anthropology
to Human Affairs. G. W. Lasker and C. G. N. Mascie-Taylor, eds. Cambridge: Cambridge
University Press. (In press.)
Ellison, P. T., and C. Lager
1985 Exercise-Induced Menstrual Disorders. New England Journal of Medicine 313:825-826.
1986 Moderate Recreational Running Is Associated with Lowered Salivary Progesterone Profiles in Women. AmericanJournal of Obstetrics and Gynecology 154:1000-1003.
Ellison, P. T., C. Lager, andJ. Calfee
1987 Low Profiles of Salivary Progesterone among College Undergraduate Women. Journal of
Adolescent Health Care 8:204-207.
Ellison, P. T., N. R. Peacock, and C. Lager
1986 Salivary Progesterone and Luteal Function in Two Low-Fertility Populations of Northeast Zaire. Human Biology 58:473-483.
1989 Ecology and Ovarian Function among Lese Women of the Ituri Forest, Zaire. American
Journal of Physical Anthropology 78:519-526.
Federation CECOS, F. D. Schwartz, and M.J. Mayaux
1982 Female Fecundity as a Function of Age. New England Journal of Medicine 306:404-406.
Ford, K., and S. Huffman
1988 Nutrition, Infant Feeding and Post-Partum Amenorrhea in Rural Bangladesh.Journal of
Biosocial Science 20:461-469.
Foster, A.,J. Menken, A. Chowdhurry, andJ. Trussell
1986 Female Reproductive Development: A Hazards Model Analysis. Social Biology 33:183198.
Frisch, R. E.
1976 Critical Metabolic Mass and the Age at Menarche. Annals of Human Biology 3:489-491.
1978 Population, Food Intake and Fertility. Science 199:22-30.
1982 Malnutrition and Fertility. Science 215:1272-1273.
1984 Body Fat, Puberty and Fertility. Biological Reviews 59:161-188.
1987 Body Fat, Menarche, Fitness and Fertility. Human Reproduction 2:521-533.
1988 Fatness and Fertility. Scientific American 258:88-95.
Frisch, R. E., A. V. Gotz-Welbergen,J. W. McArthur, T. Albright,J. Witschi, B. Bullen,J. Birnholz, R. B. Reed, and H. Hermann
1981 Delayed Menarche and Amenorrhea of College Athletes in Relation to Age at the Onset
of Training. Journal of the American Medical Association 246:1559-1563.
Frisch, R. E., andJ. W. McArthur
1974 Menstrual Cycles: Fatness as a Determinant of Minimum Weight for Height Necessary
for Their Maintenance and Onset. Science 185:949-951.
Frisch, R. E., and R. Revelle
1970 Height and Weight at Menarche and a Hypothesis of Critical Body Weights and Adolescent Events. Science 169:397-399.
1971 Height and Weight at Menarche and a Hypothesis of Menarche. Archives of Disease in
Childhood 46:695-701.
Frisch, R. E., R. Revelle, and S. Cook
1971 Height, Weight and Age at Menarche and the "Critical Weight" Hypothesis. Science
174:1148-1149.
1973 Components of Weight at Menarche and the Initiation of the Adolescent Growth Spurt
in Girls: Estimated Total Water, Lean Body Weight and Fat. Human Biology 45:469-483.
Frisch, R. E., G. Wyshak, and L. Vincent
1980 Delayed Menarche and Amenorrhea of Ballet Dancers. New England Journal of Medicine 303:17-19.
Fritz, M. A.
1988 Inadequate Luteal Function and Recurrent Abortion: Diagnosis and Treatment of Luteal
Phase Deficiency. Seminars in Reproduction Endocrinology 6:129-143.
Gardner,J.
1983 Adolescent Menstrual Characteristics as Predictors of Gynaecological Health. Annals of
Human Biology 10:31-40.
Gardner,J., and I. Valadian
1983 Changes over Thirty Years in an Index of Gynaecological Health. Annals of Human Biology 10:41-55.
Ellison]
FUNCTIONANDREPRODUCTIVE
ECOLOGY
OVARIAN
949
Gioiosa, R.
1955 Incidence of Pregnancy during Lactation in 500 Cases. American Journal of Obstetrics
and Gynecology 70:162-174.
Goldman, N., C. F. Westhoff, and L. E. Paul
1987 Variations in Natural Fertility: The Effect of Lactation and Other Determinants. Population Studies 41:127-146.
Graham, R. L., D. L. Grimes, and R. D. Campbell
1979 Amenorrhea Secondary to Voluntary Weight Loss. Southern Medical Journal 72:1259.
Green, B. B.,J. R. Daling, N. S. Weiss,J. M. Liff, and T. Koepsell
1986 Exercise as a Risk Factor for Infertility with Ovulatory Dysfunction. American Journal
of Public Health 76:1432-1436.
Henry, L.
1961 Some Data on Natural Fertility. Eugenics Quarterly 8:81-91.
Howie, P. W., and A. S. McNeilly
1982 Effect of Breast Feeding Patterns on Human Birth Intervals.Journal of Reproduction and
Fertility 65:545-557.
Howie, P. W., A. S. McNeilly, M.J. Houston, A. Cook, and H. Boyle
1981 Effect of Supplementary Food on Suckling Patterns and Ovarian Activity during Lactation. British MedicalJournal 283:757-759.
1982 Fertility after Childbirth: Infant Feeding Patterns, Basal PRL Levels and Post-Partum
Ovulation. Clinical Endocrinology 17:315-322.
Howlett, T. A.
1987 Hormonal Responses to Exercise and Training: A Short Review. Clinical Endocrinology
26:723-742.
Howlett, T. A., S. Tomlin, L. Ngahfoong, L. H. Rees, B. A. Bullen, G. S. Skrinar, and J. W.
McArthur
1984 Release of B Endorphin and met-Enkephalin during Exercise in Normal Women: Response to Training. British Medical Journal 288:1950-1952.
Johnston, F. E.
1974 Control of Age at Menarche. Human Biology 46:159-171.
Johnston, F. E., R. M. Malina, and M. A. Galbraith
1971 Height, Weight and Age at Menarche and the "Critical Weight" Hypothesis. Science
174:1148.
Johnston, F. E., A. F. Roche, L. M. Schell, and H. N. B. Wettenhall
1975 Critical Weight at Menarche: Critique of a Hypothesis. AmericanJournal of Diseases of
Childhood 129:19-23.
Kennedy, K. I., R. Rivera, and A. S. McNeilly
1989 Consensus Statement on the Use of Breastfeeding as a Family Planning Method. Contraception 39:477-496.
Konner, M., and C. Worthman
1980 Nursing Frequency, Gonadal Function, and Birth Spacing among !Kung Hunter-Gatherers. Science 207:788-791.
Lager, C., and P. T. Ellison
1990 Effect of Moderate Weight Loss on Ovarian Function Assessed by Salivary Progesterone
Measurements. American Journal of Human Biology 2:303-312.
Lenton, E. A., R. Sulaiman, O. Sobowale, and I. D. Cooke
1988 The Human Menstrual Cycle: Plasma Concentrations ofProlactin, LH, FSH, Oestradiol
and Progesterone in Conceiving and non-Conceiving Women. Journal of Reproduction and
Fertility 65:131-139.
Leslie, P. W., and P. H. Fry
1989 Extreme Seasonality of Births among Nomadic Turkana Pastoralists. American Journal
of Physical Anthropology 79:103-115.
Liestol, K. E.
1980 Menarcheal Age and Spontaneous Abortion: A Causal Connection? AmericanJournal of
Epidemiology 111:753-758.
Lunn, P. G., S. Austin, and R. G. Whitehead
1984 The Effect of Improved Nutrition on Plasma Prolactin Concentrations and Postpartum
Infertility in Lactating Gambian Women. AmericanJournal of Clinical Nutrition 39:227-235.
950
A MERICAN
ANTHROPOLOGIST
[92, 1990
Lunn, P. G., A. M. Prentice, S. Austin, and R. G. Whitehead
1980 Influence of Maternal Diet on Plasma-Prolactin Levels during Lactation. Lancet 1:623625.
MacMahon, B., D. Trichopoulos, J. Brown, A. P. Andersen, K. Aoki, P. Cole, F. deWaard, T.
Kauraniemi, R. W. Morgan, M. Purde, B. Ravnihar, N. Stromby, K. Westlund, and N. C. Woo
1982 Age at Menarche, Probability of Ovulation and Breast Cancer Risk. International Journal of Cancer 29:13-16.
Malina, R. M.
1983 Menarche in Athletes: A Synthesis and Hypothesis. Annals of Human Biology 10:1-24.
Marshall, W. A.
1974 Interrelationships of Skeletal Maturation, Sexual Development and Somatic Growth in
Man. Annals of Human Biology 1:29-40.
Marshall, W. A., and Y. de Limongi
1976 Skeletal Maturity and the Prediction of Age at Menarche. Annals of Human Biology
3:235-243.
McFalls,J. A., and M. H. McFalls
1984 Disease and Fertility. New York: Academic Press.
McNatty, K. P., D. M. Smith, A. Makris, R. Osathanondh, and K.J. Ryan
1979 The Microenvironment of the Human Antral Follicle: Interrelationships among the Steroid Levels in Antral Fluid, the Population of Granulosa Cells, and the Status of the Oocyte
in Vivo and in Vitro. Journal of Clinical Endocrinology and Metabolism 49:851-860.
McNeely, M.J., and M. R. Soules
1988 The Diagnosis ofLuteal Phase Deficiency: A Critical Review. Fertility and Sterility 50:115.
McNeilly, A. S., P. W. Howie, M.J. Houston, A. Cook, and H. Boyle
1982 Fertility after Childbirth: Adequacy of Post-Partum Luteal Phases. Clinical Endocrinology 17:609-615.
Menken, J.
1985 Age and Fertility: How Late Can You Wait? Demography 22:469-483.
Menken,J.,J. Trussell, and U. Larsen
1986 Age and Infertility. Science 233:1389-1394.
Menken,J.,J. Trussell, and S. Watkins
1981 The Nutrition-Fertility Link: An Evaluation of the Evidence.Journal of Interdisciplinary
History 9:425-441.
Moerman, M. L.
1982 Growth of the Birth Canal in Adolescent Girls. AmericanJournal of Obstetrics and Gynecology 143:528-532.
Montagu, A. M. F.
1946 Adolescent Sterility. Springfield, IL: Thomas.
Peacock, N.
1986 Women Athletes: A Model for Hunter-Gatherer Fertility? AnthroQuest 34:5-6.
1990 Comparative and Cross-Cultural Approaches to the Study of Human Female Reproductive Failure. In Primate Life Histories and Evolution. C.J. DeRousseau, ed. New York: WileyLiss. (In press.)
Peacock, N. R., and P. T. Ellison
1984 Endocrine Studies of Female Gonadal Function in the Northern Ituri Forest, Zaire.
American Association of Physical Anthropologists, 53rd Annual Meeting, Philadelphia.
Pintor, C., A. R. Genazzani, R. Pugoni, G. Carboni, A. Faedda, E. Pisano, T. Orani, G. D'Ambrogio, and R. Corda
1980 Effect of Weight Loss on Adrenal Androgen Plasma Levels in Obese Prepubertal Girls.
In Adrenal Androgens. A. R. Genazzani, J. H. Thijssen, and P. K. Siiteri, eds. Pp. 259-267.
New York: Raven Press.
Pirke, K. M., U. Schweiger, W. Lemmel,J. C. Krieg, and M. Berger
1985 The Influence of Dieting on the Menstrual Cycle of Healthy Young Women. Journal of
Clinical Endocrinology and Metabolism 60:1174-1179.
Pirke, K. M., U. Schweiger, T. Strowitzki, R.J. Tuschi, G. Laessle, A. Broocks, B. Huber, and R.
Middendorf
1989 Dieting Causes Menstrual Irregularities in Normal Weight Young Women through Impairment of Episodic Luteinizing Hormone Secretion. Fertility and Sterility 51:263-268.
Ellison]
FUNCTIONANDREPRODUCTIVE
OVARIAN
ECOLOGY
951
Poindexter, A. N., and M. B. Ritter
1983 Anovulatory Uterine Bleeding. Comprehensive Therapy 9:65-71.
Potter, R. G.,Jr., M. L. New,J. B. Wyon, andJ. E. Gordon
1965 A Fertility Differential in Eleven Punjab Villages. Milbank Memorial Fund Quarterly
43:185-201.
Prentice, A. M., and R. G. Whitehead
1987 The Energetics of Human Reproduction. Symposium on the Zoological Society of London 57:275-304.
Prentice, A. M., R. G. Whitehead, S. B. Roberts, and A. A. Paul
1981 Long-Term Energy Balance in Child-Bearing Gambian Women. American Journal of
Clinical Nutrition 34:2790-2799.
Prior,J. C.
1985 Luteal Phase Defects and Anovulation: Adaptive Alterations Occurring with Conditioning Exercise. Seminars in Reproductive Endocrinology 3:27-33.
Read, G. F., D. W. Wilson, I. A. Hughes, and K. Griffiths
1984 The Use of Salivary Progesterone Assays in the Assessment of Ovarian Function in Postmenarcheal Girls. Journal of Endocrinology 102:265-268.
Reeves, J.
1979 Estimating Fatness. Science 204:881.
Riddick, D. H., D. C. Daly, S. M. Rosenberg, and I. A. Maslar
1983 Progesterone and Endometrial Physiology in Human Fertility. In Progesterone and Progestins. C. W. Bardin, E. Milgrom, and P. Mauvais-Jarvis, eds. Pp. 231-245. New York:
Raven Press.
Schweiger, U., R. Laessle, H. Pfister, C. Hoehl, M. Schwingenschloegel, M. Schweiger, and K. M.
Pirke
1987 Diet-Induced Menstrual Irregularities:Effects of Age and Weight Loss. Fertility and Sterility 48:746.
Scott, E. C., and F. E.Johnston
1985 Science, Nutrition, Fat, and Policy: Tests of the Critical Fat Hypothesis. Current Anthropology 26:463-473.
Shangold, M. M., M. L. Gatz, and B. Thyssen
1981 Acute Effects of Exercise on Plasma Concentrations of Prolactin and Testosterone in Recreational Women Runners. Fertility and Sterility 35:699-702.
Shapiro, S., E. R. Schlesinger, and R. E. L. Nesbitt, Jr.
1968 Infant, Perinatal, Maternal, and Childhood Mortality in the United States. Cambridge,
MA: Harvard University Press.
Short, R. V.
1983 The Biological Basis for the Contraceptive Effects of Breastfeeding. Advances in International Maternal and Child Health 3:26-29.
1987 The Biological Basis for the Contraceptive Effects of Breastfeeding. InternationalJournal
of Gynaecology and Obstetrics (Suppl.) 25:77-109.
Shuttleworth, F. K.
1937 Sexual Maturation and the Physical Growth of Girls Age Six to Nineteen. Monographs
of the Society for Research on Child Development 2(5).
Siiteri, P. K.,J. E. Williams, and N. K. Takaki
1976 Steroid Abnormalities in Endometrial and Breast Carcinoma: A Unifying Hypothesis.
Journal of Steroid Biochemistry 7:897-903.
Stanger,J. D., andJ. L. Yovich
1984 Failure of Human Oocyte Release at Ovulation. Fertility and Sterility 41:827-832.
Stouffer, R. L.
1988 Perspectives on the Corpus Luteum of the Menstrual Cycle and Early Pregnancy. Seminars in Reproductive Endocrinology 6:103-113.
Tanner, J. M.
1962 Growth at Adolescence. 2nd edition. New York: Blackwell.
Trussell, J.
1978 Menarche and Fatness: Reexamination of the Critical Body Composition Hypothesis.
Science 200:1506-1509.
1980 Statistical Flaws in Evidence for the Frisch Hypothesis that Fatness Triggers Menarche.
Human Biology 52:711-720.
952
AMERICAN
ANTHROPOLOGIST
[92, 1990
Tuddenham, R. D., and M. M. Snyder
1954 Physical Growth of California Boys and Girls from Birth to Eighteen Years. University
of California Publications in Child Development 1(2).
Udesky, I.
1950 Ovulation in Lactating Women. AmericanJournal of Obstetrics and Gynecology 59:843851.
Venturoli, S., E. Porcu, R. Fabbri, O. Magrini, R. Paradisi, G. Pallotti, L. Gammi, and C. Famigni
1987 Postmenarchal Evolution of Endocrine Pattern and Ovarian Aspects in Adolescents with
Menstrual Irregularities. Fertility and Sterility 48:78-85.
Vigersky, R. A., A. E. Anderson, R. H. Thompson, and D. L. Loriaux
1977 Hypothalamic Dysfunction in Secondary Amenorrhea Associated with Simple Weight
Loss. New England Journal of Medicine 297:1141-1145.
Vihko, R., and D. Apter
1984 Endocrine Characteristics of Adolescent Menstrual Cycles: Impacts of Early Menarche.
Journal of Steroid Biochemistry 20:231-236.
Warren, M. P.
1990 Weight Control. Seminars in Reproductive Endocrinology 8:25-31.
Warren, M. P., R. Jewelewicz, I. Dyrenfurth, R. Ans, S. Khalaf, and R. L. Vande Wiele
1975 The Significance of Weight Loss in the Evaluation of Pituitary Response to LH-RH in
Women with Secondary Amenorrhea. Journal of Clinical Endocrinology and Metabolism
40:601-611.
Whitehead, R. G., A. A. Paul, and E. A. Ahmed
1986 Weaning Practices in the United Kingdom and Variations in Anthropometric Development. Acta Paediatrica Scandinavica (Suppl.) 323:14-23.
WHO/NRC (USA)
1983 Breastfeeding and Fertility Regulation: Current Knowledge and Programme Policy Implications. Bulletin of the World Health Organization 61:371-382.
Wood,J. W., and M. Weinstein
1988 A Model of Age-Specific Fecundability. Population Studies 42:85-113.
Wyshak, G.
1983 Age at Menarche and Unsuccessful Pregnancy Outcome. Annals of Human Biology
10:69-73.
Yen, S. S. C.
1984 Opiates and Reproduction: Studies in Women. In Opioid Modulation of Endocrine Function. G. Delitala, M. Motta, and M. Serio, eds. Pps. 191-209. New York: Raven Press.
Yoshimura, Y., and E. E. Wallach
1987 Studies of the Mechanism(s) of Mammalian Ovulation. Fertility and Sterility 47:22-34.