Coming from zero background in biology, I am looking for expert opinion on evidence regarding Trivers-Willard hypothesis (1973) and humans.

In short the hypothesis states that if

  1. "Condition" of offspring during their whole life are correlated to that of their mother during pregnancy
  2. Males have better on average better reproductive success (more genes in next generation) than female when in "good condition", and vice versa if you compare male/females in "bad condition".
  3. Females objective is "eventual reproductive success" down the generations.

Then it makes sense for females to vary sex ratio, more males when in "good condition" and more females when in "bad condition".

I think it is safe to say that plenty of evidence exists on assumption one. The second assumption seems more nuance in todays environment, but one would imagine that during most human history it does hold true. On the third one I think is safe to assume it is true based on the theory of evolution by natural selection.

Any (rigorous) reference (articles, researchers, journals) or comments would be superb!

Reported association that I am already aware of are:

  1. Reduced gender ratio at times of war and famine (Ansari-Lari & Saadat, 2002; Song, 2012; Valente, 2015) and after natural disasters (Fukuda et al. 1998).

  2. Darwin (Descent of man); Norberg, 2004; Almond & Edlund, (2007); Hamoudi & Nobles, (2014) divorce or "unmarriedness" correlated with lowered gende ratio.

  3. Negative association between parents age and gender-ratio, Jacobsen, Møller, & Mouritsen, (1999).

  • $\begingroup$ Have you looked at ncbi.nlm.nih.gov/pubmed/… ? - although see @Remi.b's comments about statistical issues ... $\endgroup$
    – Ben Bolker
    Aug 14, 2016 at 18:54
  • $\begingroup$ Yes, I've looked around but have not found anything that seems to settle the debate. But most things I find point towards the same direction as Trivers-Willard. For example: humrep.oxfordjournals.org/content/13/8/2321.short $\endgroup$
    – snoram
    Aug 14, 2016 at 21:04
  • 1
    $\begingroup$ Interestingly Darwin had spotted a pattern: excess of male over female births is less when they are illegitimate than when legitimate - The Descent of Man $\endgroup$
    – snoram
    Aug 14, 2016 at 21:16
  • 1
    $\begingroup$ @snoram Be sure there has been a lot of work on the evolution of sex-ratio since Darwin made this claim. You could start with West and Sheldon 2002 for a review but you will very likely have to start with a very introductory course to evolutionary biology (such as Understanding Evolution for example) as you were unfamiliar with the term 'fitness'. $\endgroup$
    – Remi.b
    Sep 2, 2016 at 22:51
  • $\begingroup$ The paper you link to does not mentions humans specifically, but they state "the extent to which data on mammals fit this prediction is disputed". I wonder how much that has changes since 2002. $\endgroup$
    – snoram
    Sep 2, 2016 at 23:02

1 Answer 1


I am not aware of any such evidence in humans but I know little about this literature. If you happen to have data showing that, then it would probably be a relatively big thing! Below is a series of comments that might be of interest.

Evidence from Hyena

The evolution of the sex-ratio variation in hyena have been quite extensively studied. This very short article (not peer-reviewed) cite a number of key papers on the subject.

Logic behind the second assumption

In case clarification is needed. The variance in fitness is indeed often higher in males than in females. The reason is that females (by definition) produce the larger gametes, the one that don't leave the body and therefore often pay a much higher cost when reproducing than males do. As such, females are the limiting factors. As a consequence, males compete a lot to access females and "good males" receive a lot of mating opportunities while "bad males" receive little mating opportunity causing the higher variance in fitness among males than among females.


IMHO, there is a tradition of horrible statistics when dealing with human sex-ratio data. It is particularly the case as these papers are often published in journals where reviewer have little knowledge in statistics (e.g. medicine). Typically, "windowing" data is generally a very bad idea and previous papers in the field have reported very bad interpretation from data due to cherry picking specific ways to classify observations. Please be cautious in your statistical analysis!

About the post

The formulation of the hypothesis is clear enough but some sentences are misleading

it makes "natural selection" sense for females to vary sex ratio`

This sentence sounds very wrong. You should rather say "females that are able to vary the sex-ratio among their offspring will be have a higher fitness".

Note that I said "among their offspring" because while this would cause variation in sex-ratio in the population, genotypes are selected based on their ability to vary the sex-ratio among their offspring and not in the population. In fact, any given individual, would appreciate if all the other would produce the exact opposite sex-ratio than the one they actually produce among their offspring.

better chance of reproductive success

This sentence a little clumsy as well and should be replaced with higher fitness

  • $\begingroup$ Great. Thanks! But not sure I understand your last comment "replaced with higher fitness". The second assumption (as I understand) is that at some point on the "fitness scale" males start to have a reproductive advantage over females. $\endgroup$
    – snoram
    Aug 14, 2016 at 5:13
  • $\begingroup$ You are misusing the terms "fitness" and "reproductive success". reproductive success is the realized number of offspring of an individual for a chosen duration (often the whole life). Fitness is a function of survival and reproductive fitness and is generally defined as (but might differ slightly from one model to another) the expected lifetime reproductive fitness of an individual. Therefore "Better chance of reproductive success" translate into "higher fitness" and we usually use this expression. $\endgroup$
    – Remi.b
    Aug 14, 2016 at 5:41
  • $\begingroup$ Fitness just means good health. Agreed? (I checked the dictionary). By saying "better chance of reproductive success" I mean exactly higher expected number of offspring. Good health causes "Better chance of reproductive success" not the other way around. $\endgroup$
    – snoram
    Aug 14, 2016 at 17:50
  • 1
    $\begingroup$ You are getting confused with the use of the term fitness in the sense physical fitness and in the sense "biological fitness" (which is just termed "fitness"). In evolution, "fitness" does not mean "good health". In evolutionary biology we sometimes talk about "overall health" of an organism (which would mean something like "physical fitness") as an extremely multivariate trait that is eventually (but not necessarily) a good proxy for fitness. $\endgroup$
    – Remi.b
    Aug 14, 2016 at 18:07
  • $\begingroup$ I will do some small edits to my original posts based on your comments. But to clear I never used the word "fitness" (nor do Trivers and Williard) and use instead "condition" for something like "physical fitness". $\endgroup$
    – snoram
    Aug 14, 2016 at 18:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .