Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

An age-dependent ovulatory strategy explains the evolution of dizygotic twinning in humans


Dizygotic twinning, the simultaneous birth of siblings when multiple ova are released, is an evolutionary paradox. Twin-bearing mothers often have elevated fitness, but despite twinning being heritable, twin births occur only at low frequencies in human populations. We resolve this paradox by showing that twinning and non-twinning are not competing strategies; instead, dizygotic twinning is the outcome of an adaptive conditional ovulatory strategy of switching from single to double ovulation with increasing age. This conditional strategy, when coupled with the well-known decline in fertility as women age, maximizes reproductive success and explains the increase and subsequent decrease in the twinning rate with maternal age that is observed across human populations. We show that the most successful ovulatory strategy would be to always double ovulate as an insurance against early fetal loss, but to never bear twins. This finding supports the hypothesis that twinning is a by-product of selection for double ovulation rather than selection for twinning.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Declining prenatal survival and increasing double ovulation rates explain age-dependent twinning rates.
Fig. 2: Average per capita numbers of offspring surviving to 15 years (95% confidence intervals, n = 10) from simulations of the reproductive lives of 1,000 women.
Fig. 3: Results of mathematical modelling of the reproductive characteristics of different ovulatory strategies.

Data availability

The data have been uploaded to Dyad and are available at

Code availability

The computer code has been uploaded to Dyad and is available at This code is supplied by the authors with the request that future users of the code are aware that it represents an ongoing research programme and that they contact the authors in a collaborative spirit.


  1. 1.

    Maynard Smith, J. Evolution and the Theory of Games (Cambridge Univ. Press, 1982).

  2. 2.

    Mbarek, H. et al. Identification of common genetic variants influencing spontaneous dizygotic twinning and female fertility. Am. J. Hum. Genet. 98, 898–908 (2016).

    CAS  Article  Google Scholar 

  3. 3.

    Bulmer, M. G. The Biology of Twinning in Man (Clarendon, 1970).

  4. 4.

    Lummaa, V., Haukioja, E., Lemmetyinen, R. & Pikkola, M. Natural selection on human twinning. Nature 394, 533–534 (1998).

    CAS  Article  Google Scholar 

  5. 5.

    Sear, R., Shanley, D., McGregor, I. A. & Mace, R. The fitness of twin mothers: evidence from rural Gambia. J. Evol. Biol. 14, 433–443 (2001).

    Article  Google Scholar 

  6. 6.

    Forbes, S. A Natural History of Families (Princeton Univ. Press, 2005).

  7. 7.

    Robson, S. L. & Smith, K. R. Twinning in humans: maternal heterogeneity in reproduction and survival. Proc. R. Soc. B 278, 3755–3761 (2011).

    Article  Google Scholar 

  8. 8.

    Rickard, I. J., Courtiol, A. & Lummaa, V. Why is lifetime fertility higher in twinning women? Proc. R. Soc. B 279, 2510–2511 (2012).

    Article  Google Scholar 

  9. 9.

    Anderson, D. J. On the evolution of human brood size. Evolution 44, 438–440 (1990).

    Article  Google Scholar 

  10. 10.

    O’Connor, K. A., Holman, D. J. & Wood, J. W. Declining fecundity and ovarian ageing in natural fertility populations. Maturitas 30, 127–136 (1998).

    Article  Google Scholar 

  11. 11.

    Centers for Disease Control and Prevention, American Society for Reproductive Medicine & Society for Assisted Reproductive Technology 2011 Assisted Reproductive Technology National Summary Report (US Dept of Health and Human Services, 2013).

  12. 12.

    McCoy, R. C. et al. Common variants spanning PLK4 are associated with mitotic-origin aneuploidy in human embryos. Science 348, 235–238 (2015).

    CAS  Article  Google Scholar 

  13. 13.

    Atkinson, D. E. Frequency of dizygotic twinning. Nature 322, 780–780 (1985).

    Article  Google Scholar 

  14. 14.

    Buzatto, B. A., Buoro, M., Hazel, W. N. & Tomkins, J. L. Investigating the genetic architecture of conditional strategies using the environmental threshold model. Proc. R. Soc. B 282, 20152075 (2015).

    Article  Google Scholar 

  15. 15.

    Shipley, P. W., Wray, J. A., Hechter, H. H., Arellano, M. G. & Borhani, N. O. Frequency of twinning in California: its relationship to maternal age, parity and race. Am. J. Epidemiol. 85, 147–156 (1967).

    CAS  Article  Google Scholar 

  16. 16.

    Nylander, P. P. S. Ethnic differences in twinning rates in Nigeria. J. Biosoc. Sci. 3, 151–158 (1971).

    CAS  Article  Google Scholar 

  17. 17.

    Enders, T. & Stern, C. The frequency of twins, relative to age of mothers in American populations. Genetics 33, 263–272 (1948).

    CAS  PubMed  PubMed Central  Google Scholar 

  18. 18.

    Pollard, G. N. Multiple births in Australia, 1944–63. J. Biosoc. Sci. 1, 389–404 (1969).

    CAS  Article  Google Scholar 

  19. 19.

    Imaizumi, Y. & Hayakawa, K. Infant mortality among singletons and twins in Japan during 1999–2008 on the basis of risk factors. Twin Res. Hum. Genet. 16, 639–644 (2013).

    Article  Google Scholar 

  20. 20.

    Nybo Andersen, A. M., Wohlfahrt, J., Christens, P., Olsen, J. & Melbye, M. Maternal age and fetal loss: population based register linkage study. Brit. Med. J. 320, 1708–1712 (2000).

    CAS  Article  Google Scholar 

  21. 21.

    Donnet, M. L., Howie, P. W., Marnie, M., Cooper, W. & Lewis, M. Return of ovarian function following spontaneous abortion. Clin. Endocrinol. 33, 13–20 (1990).

    CAS  Article  Google Scholar 

  22. 22.

    Hazel, W., Smock, R. & Johnson, M. A polygenic model for the evolution and maintenance of conditional strategies. Proc. R. Soc. Lond. B 242, 181–187 (1990).

    CAS  Article  Google Scholar 

  23. 23.

    Hall, J. G. Twinning. Lancet 362, 735–743 (2003).

    Article  Google Scholar 

  24. 24.

    Fellman, J. & Eriksson, A. W. Stillbirths in multiple births: test of independence. Twin Res. Hum. Genet. 9, 677–684 (2006).

    Article  Google Scholar 

  25. 25.

    Johnson, C. D. & Zhang, J. Survival of other fetuses after a fetal death in twin or triplet pregnancies. Obstet. Gynecol. 99, 698–703 (2002).

    PubMed  Google Scholar 

  26. 26.

    Evron, E., Sheiner, E., Friger, M., Sergienko, R. & Harlev, A. Vanishing twin syndrome: is it associated with adverse perinatal outcome? Fertil. Steril. 103, 1209–1214 (2015).

    Article  Google Scholar 

  27. 27.

    Scaramuzzi, R. J. et al. Regulation of folliculogenesis and the determination of ovulation rate in ruminants. Reprod. Fertil. Dev. 23, 444–467 (2011).

    CAS  Article  Google Scholar 

  28. 28.

    Hoekstra, C. et al. Dizygotic twinning. Hum. Reprod. Update 14, 37–47 (2008).

    CAS  Article  Google Scholar 

  29. 29.

    Mayr, E. Cause and effect in biology. Science 134, 1501–1506 (1961).

    CAS  Article  Google Scholar 

  30. 30.

    Jean-Paul, S. & Tomas, F. First birth trends in developed countries: persisting parenthood postponement. Demogr. Res. 15, 147–180 (2006).

    Article  Google Scholar 

  31. 31.

    Pison, G., Monden, C. & Smits, J. Twinning rates in developed countries: trends and explanations. Popul. Dev. Rev. 41, 629–649 (2015).

    Article  Google Scholar 

  32. 32.

    Tomkins, J. L. & Hazel, W. The status of the conditional evolutionarily stable strategy. Trends Ecol. Evol. 22, 522–528 (2007).

    Article  Google Scholar 

  33. 33.

    R: A language and environment for statistical computing (R Foundation for Statistical Computing, 2018).

Download references


C. Lively, D. Dudle, L. Simmons, C. Smock and D. Hruschka provided comments on the manuscript drafts. We thank N. LeBas for detailed discussions of these ideas. J.L.T. was supported by a Future Fellowship (grant no. FT110100500) from the Australian Research Council. R.C.S. and W.N.H. were funded by DePauw University Faculty Development Grants.

Author information




J.L.T. proposed the study. W.N.H. and J.L.T. began developing the approaches. R.S. provided the demographic data and background on the biology of twinning. R.C.S. and W.N.H. finalized the mathematical model. J.L.T. and R.B. finalized the simulation model. W.N.H. wrote the first draft of the manuscript, and all authors contributed to the completed manuscript.

Corresponding authors

Correspondence to Wade N. Hazel or Joseph L. Tomkins.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Double ovulation as a threshold trait.

a, Three conditional (age-dependent) ovulatory strategies that differ in age of switching from single (dashed line) to double ovulation (solid line). b, Threshold trait depiction of a cohort of population with normally distributed variation in age at switching (mean age at switching from single to double ovulation is 30 yrs., SD of 3 yrs.). At age 27, the threshold (vertical line positioned at age 27) divides the population into individuals that double ovulate (those with switching ages less than 27) and single ovulate (those with switching ages greater than 27). c, As age increases, the fraction of double ovulators increases as a cumulative normal function as the threshold in b shifts to the right, increasing the fraction of the distribution of switching ages that is less than the threshold.

Extended Data Fig. 2 Age-dependent probabilities of live birth, and prenatal death from early losses, spontaneous abortion, and late losses.

a-c, points from Supplementary Table 3; lines are best fitting curves from least squared regression. a, Early loss of pregnancy (variable r in Supplementary Table 1); b. Spontaneous abortion (variable s in Supplementary Table 1); c, Late loss of pregnancy (variable t in Supplementary Table 1); d, Live birth (line from average of results in Fig. 1f; variable u in Supplementary Table 1).

Extended Data Fig. 3 Age-dependent probability of twin live births in double ovulators.

Probability of twin live births was calculated using equation 1, assuming declining probability of live birth with increasing maternal age as depicted in Extended Data Fig. 2d.

Supplementary information

Supplementary Information

Supplementary Figs. 1–3, Tables 1–3, methods and discussion.

Reporting Summary

Peer Review Information

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Hazel, W.N., Black, R., Smock, R.C. et al. An age-dependent ovulatory strategy explains the evolution of dizygotic twinning in humans. Nat Ecol Evol 4, 987–992 (2020).

Download citation


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing