Menstrual Synchrony [text]
Adina Sobo
so you can find them and read them yourself. The cite format is
modified a bit, and I hope it will make it easier for readers to
locate the papers they are interested in, but all things being equal
it may well backfire. I also had to chop the cites into a separate
post, because this is sorta long. Live with it.]
Let’s start with a little background….
While most of us learned in school that the menstrual (remember: we’re
talking
pre-menopausal post pubertal women here!) cycle is 28 days, it would be far
more accurate to say it’s 28 days plus or minus 3 or 4 days. A woman who
cycles from 25 - 31 days each month is considered to be regular, and about
67% of women will statistically fall into this category 85% or more of the
time. The remaining third of women are less regular than this – intervals
of 20 - 40 days are not that uncommon, and it’s also not rare for a woman to
have a 20 day cycle one month and a 40 day one the next. This pattern is
one of many examples of bodily cycles that can be timed. Arendt’s paper
on chronobiology would be a good resource for someone who wants
more examples of these types of things. It talks a bit about menstrual cycle
variations, and even human seasonality, as well as about many other
daily, weekly, or longer patterns. Some theories are that the menstrual
cycle may be directly affected by day length and, according to Law’s
research, even by the phase of the moon.
Menstrual cycles can be affected by a lot of different factors. First off,
there’s genetics. If your female relatives were irregular or had cycles
that were otherwise not the typical "28 days" the odds are you will
fall into the same pattern. The genes women inherit have a bit to
do with how much of each female hormone the body produces and
how sensitive the body is to those hormones. Body fat is another
obvious factor. We’ve all read about women athletes and anorectics
who stop menstruating when their body fat is too low – the body
essentially says (anthropomorphism alert!) "I’m starving to death,
so I’m not going to waste calories on non-essentials. Since I couldn’t
carry a fetus to term under these conditions, I’m not going to bother
preparing for a possible pregnancy." At the other end of the weight
spectrum, women who are carrying extra body fat often have irregular
or longer cycles, presumably because the female hormones are fat
soluble; they build up in the body and are released slowly over time.
This "confuses" the body’s physiology, and affects the cycles. Most
women already know that you can be knocked off your regular cycle
if you are stressed. I don’t know if this is also common knowledge for
men. ;-) Jarett’s work is only one of many papers that indicate how
psychosocial and personality traits, illness, etc. also predicted cycle
length and regularity. Matteo’s is another, adding in the way that
a regular sexual partner or even exposure to the opposite sex can
also affect menstrual cycles.
I’ll add in what any woman knows: it’s hard to say "my period started at
3:06 PM today." Sometimes you get cramps before there is any flow, sometimes
the reverse. It’s unusual for the flow to be heavy from the first – it
usually
starts with a lower flow, and builds to a peak. Some women may just spot a
tiny bit for a day or so before the heavier flow begins. And a small amount
of menstrual fluid is not instantly obvious – there are lots of folds in the
vagina and there are additional vulval folds so (to be a bit graphic) the
fluid doesn’t just go straight from the cervix to the great outside; it
takes
time for the flow to become visible. So there is the additional problem that
if you rely on self-reporting for when a woman’s period began, you’re
unlikely
to be able to say "the flow began at 3:06 PM." You need to use larger
chunks of time for reporting, often 12 to 24 hour intervals. This affects
the potential accuracy of the results.
So, obviously, there are a lot of potential statistical analysis pitfalls that
a researcher
would need to avoid in order to declare that there is a pheremonal component
to menstrual periods, or even that women living together tend to synchronize.
The seminal (you’ll excuse the term) paper in the field of Menstural
Synchrony
was that of MK McClintock, in 1971. This initial report of menstrual
synchrony
indicated that social interactions among groups of women could regulate their
ovarian cycles, as demonstrated by their menstrual periods’ appearance. This
paper appeared quite a while ago, and I don’t think it’s online (at least,
I didn’t
find it), and was written back before abstracts were archived to the UC library
system, so I don’t know all the details about it. I do know that McClintock
didn’t
have another article published about menstrual synchrony until 1988. Or, at
least,
nothing that was archived by Medline/Healthstar. At that time, she wrote:
"Menstrual synchrony was similar to an archeologist finding a fossilized
tooth, which demonstrated the existence of a prehistoric creature.
Menstrual
synchrony could have turned out to be like the chronodonts, prehistoric
creatures for which we still have only their fossilized teeth as evidence
for their existence. Fortunately, after almost 3 decades of work, we have
excavated the site and been able to unearth more about the structure of
this
particular creature. It is social regulation of ovulation throughout the
lifespan--a creature made up not only of menstrual synchrony, but various
forms of the timing of spontaneous ovulatory cycles in adults. It also
includes
the social regulation of ovulation at other points during the reproductive
lifespan: puberty, inter-birth intervals and reproductive senescence.
Menstrual synchrony is but one indicator of the phenomenon; it is now clear
that there is a great deal more to it than was seen at the time of the
original
report."
What was she getting at? Your guess is as good as mine. As an ethologist, I
know that puberty is largely controlled by genetics and nutrition, there are
hormonal stimuli from nursing that affect the recurrence of ovulation. I know
that senescence is apparently a complex matter, which some ethologists put
down to the increasing risk of maternal mortality and fetal abnormalities
selecting for individuals who reach menopause earlier. They are fighting
with the ethologists who hold out that women who stopped producing their
own young to start nurturing grandchildren had better fitness from the
Darwinian standpoint. As to the rest, many animals there *is* a strong
correlation between pheromones and the onset or deferral of estrus.
The obvious conclusion from the last is that since many mammals have
estrus controlled in part by pheromones (as well as by genetics, nutritional
status, day/night length, etc.) it’s possible that human menstrual cycles
(despite being very different from estrus cycles from a biochemical point
of view) might also be controlled at least in part by pheromones. In any
case, the knowledge of animal pheromones and the initial McClintock
report led to a number of researchers looking at human menstrual
synchrony as a real scientific problem.
There were some initial problems. Pheromones were not proven to exist
in humans until very recently; indeed, the presence of a vomeronasal organ
(not exactly the same as your smelling apparatus, it shares your nasal
cavities but senses the chemical signals of pheromones, rather than scents)
was only proven in humans a few years ago. There is no known pheromone
in humans that relates to female reproductive readiness.
[minor aside #1: Pheromones are fascinating, and I strongly recommend
Jim Kohl’s book "The Scent of Eros" as a great introduction on the topic
for the average lay public. Jim is, iirc, a member of Las Vegas Mensa. He
& I have gone a few rounds over the book – I do not agree with ALL his
conclusions, but it’s still a super thing to read.]
[minor aside #2: You know those perfumes and after shaves you can buy
that claim to be full of pheromones that will attract the opposite sex?
They are usually using the pheromones from pigs or minks or civets.
If you want to date within your species, don’t bother – scientists don’t
yet know what the human sexual attractant pheromone is, nor even
have they verified the existence of one that still effects the human brain]
That all said, if there is a pheromonal signal from humans about their
reproductive status, we do know where it would be produced and where
it would be most readily available: the armpits and the groin. Based on
theory alone, Preti et al swabbed some women’s armpits, and set things
up so that women were breathing the collected chemicals from there
or one of several other possibilities. Their research group believed that
their results showed that "constituents from the axillary region of donor
females can shift the time of menstrual onset of another group to conform
with the donors' cycle."
That sounds super, right? Like McClintock is validated.
Unfortunately, I don’t agree. I thought there were problems with methodology
and their analysis. Remember what I said above about variability in cycles?
That’s one big factor. Add in the short term of the experiment – this was
only done for a few months; so Preti’s researchers could only indicate what
an apparent trend was from month 1 to month 3. Preti’s experimental
group was only ten women, and they reported that they had 26 - 33 day
menstrual cycles to start with. So, we have a small sample size and a lot
of inherent variability in that sample. That means that for the results to
be mathematically significant, we have to have a fairly high standard. My
own impression of the Preti group data analysis was that it was not
sufficiently controlled.
In another post, I gave the example of left turn signals blinking – you
can see how, just because of the mathematics of the situation and the
way your brain wants to make patterns, if you watch two cars blinkers
simultaneously, they will seem to synchronize, then dance apart, and
then come back together. This is just a side effect of the mathematics
of their two cycles, and it doesn’t mean that they really ARE synchronizing.
Wilson (1987). Came to similar conclusions to my own. Wilson wrote:
"Reanalysis of the data indicates that four subjects in the extract sample
synchronized with the extract cycles because of "errors" in the extract
applications and another four synchronized as a result of experimental
design, mathematical properties of cocycling menses onsets, and chance
variations. After these factors are accounted for, no evidence suggests
that the cycles of the subjects in the extract sample were modulated by
the female-derived axillary secretion."
Later, in 1991, Wilson et al wrote a paper on the results they got when
they tried to replicate McClintock’s work. The title says it all:
"Two studies of menstrual synchrony: negative results."
The Wellers are probably the strongest supporters of menstrual synchrony.
They’ve been principal authors on at least 9 papers since 1992. They’ve
looked
at different types of groups of women in each study they’ve written – moms
&
daughters, roommates, co-workers, team-mates, lesbians, and Bedouin.
Some of their conclusions over time have supported the synchrony, while others
have not. As you can see, despite their interest in the theory, they
themselves
can’t support it wholeheartedly. Their very first study looked good on the
surface. They studied lesbian couples and found "Synchrony was very frequent,
with half the subjects menstruating within 2 days of their partner." [1992]
However, good science is replicable (the whole point of the 1991 paper
by Wilson I refer to above), and other people came up with different
conclusions when they tried studying lesbians. For example, Trevathan
et al found there was not any evidence for menstrual synchrony in lesbian
couples.
Wilson’s 1992 paper was harsher:
"All experiments and studies were based on the methods and research design
introduced by McClintock (1971). Three errors are inherent in research
based
on her model: (1) an implicit assumption that differences between menses
onsets of randomly paired subjects vary randomly over consecutive onsets,
(2)
an incorrect procedure for determining the initial onset absolute
difference
between subjects, and (3) exclusion of subjects or some onsets of subjects
who
do not have the number of onsets specified by the research design. All of
these errors increase the probability of finding menstrual synchrony in a
sample. One or more of these errors occurred in the experiments and studies
reporting synchrony; no significant levels of menstrual synchrony occur
when
these errors are corrected. Menstrual synchrony is not demonstrated in any
of
the experiments or studies.
The Wellers continued, in the same vein, for several years, with less and less
success…
"Mothers and daughters living in the same domicile displayed a significant
degree of synchrony. Roommates in private residences were also synchronous,
although not significantly more than roommates residing in dormitories."
[1993]
"None of the samples showed significant synchrony. However,
social interaction variables were found to be important factors
contributing
to the relative degree of synchrony. Friendship, mutual activity, and
length
of time working together were significantly related to the degree of
synchrony." [1995a]
"While many studies have confirmed McClintock's (1971) finding of
human menstrual synchrony, it is also clear that menstrual synchrony
does not always occur." [1995b]
"Synchrony was not found among the teams nor among best friends. No
meaningful relation was found between synchrony and menstrual-related or
social-interaction variables." [1995c]
It was at this point that the Wellers changed their experimental approach.
They
were not getting the results they wanted to see, so they came up with a new
plan of attack. Their next paper, published in 1997, was on how such studies
would ideally be done. They wrote:
"On the basis of this analysis, […] we recommend the approach in
which women record their onset dates after they have been together
for a prolonged period of time. The observed onset differences between
the women are then compared to expected onset differences, based
on the assumption of random onset occurrence."
Their next paper [1997b], touted by them as "definitive" involved data were
collected for each woman for a total of 3 months, from 27 Bedouin
nuclear families living under conditions optimally conducive for synchrony:
These were women who had been living together for many years in a highly
sexually segregated society, where each woman had very similar living
conditions,
including "minimal use of oral contraceptives." They apparently thought
that the living conditions of the Bedouin made them better subjects than even
the lesbian lovers they had looked at a few years previously. I do applaud
them
for going to the trouble of looking at the comments others had made, and
trying to come up with a protocol that would satisfy their critics as well as,
they hoped, give them an affirmative answer to the question of whether
women synchronize cycles.
Their findings, in this case were
"A 20%-25% shift toward synchrony was found for sisters-roommates,
sisters-roommates who are close friends, and the family (all women in the
family between 13 and 50 years of age)."
Many people, including myself, did not find this study "definitive." To allow
for cycle variability, they used the criterion of synchrony that it occurred in
2
out of 3 months. The underlying assumption of the state of menstrual synchrony
is that cycle length variability is minimal. The paper cited below by Arden
& Dye gives a very clear example of how cycle variability can affect the
apparent results. Another problem with this is their assumptions about male
contact. Bedouin society is segregated, but that is not the same as being
sexually inactive. Many studies have shown that regular coital activity with
a single regular partner (and the Bedouin are very much officially against
women having more than one partner!) can affect menses. Nor were they
able to verify that none of the women were using contraceptives, as they
had to rely on what the women would tell them.
For whatever reason, the Wellers then went back to looking at non-Bedouin
women, apparently changing their minds about the definitive nature of this
paper. ;-) They stopped making projections about menstrual periods based
on long term observations and demanding 2 months out of 3 for an analysis
of synchronicity, but rather went to looking only at the last month.
"Menstrual onsets of close friends tended to occur on the average within 3.5-
4.3 days of each other while onsets of co-workers who were not close friends
were significantly more broadly ranged (7.7-9.0 days of each other)."
[1999a]
"A very high degree of synchrony was found among families (51%) and among
sisters (51%), and a relatively high degree of synchrony was found among
close
friends (30%). The results suggest that synchrony can be attained without
sleeping
in the same bedroom or even in the same house. The findings also suggest
that an
effective range of mutual exposure is an attribute of human menstrual
synchrony
below or above this range, the phenomenon may not occur." [1999b]
As you’d expect – especially after that last sentence, above – there was
a lot of
discussion and disagreement over this. Both of the papers below by Schank have
a fair bit to say about mathematical analysis of cyclic phenomena and the
Weller’s conclusions.
My own conclusions are probably fairly clear from the above. But, to summarize
them: IMO, if there is any effect toward menstrual synchrony based on
pheromones
or behavior of women sharing housing, it has not been proven. I would insist
on
a fairly high standard of evidence, with cycles converging over time and then
staying converged; none of the studies thus far have been long enough term to
demonstrate this. I further believe that this sort of proof would be very
difficult
to show, as it is likely that a cohabitation effect would be far smaller than
the
effects of the women’s own inherent patterns given to them by genetics and
early development and the environmental effects of circadian rhythms, seasons,
stress, illness, nutritional status, mating status, etc. I’d expect a
certain amount
of synchronicity in families because of the genetic links, and I’m very
suspicious
of the idea that there’s some strange "optimal" level for a synchronicity
effect,
where there is a bell-shaped curve rather than a dose-dependent one.
Of course, there’s still room to continue studying this, but given the poor
showing
that the researchers have made thus far, I’m not sure who would fund it.
"The fox, when he cannot reach the grapes, says they are not ripe."
-- George Herbert, _Jacula Prudentum_ [1640]