Skip to main content
Log in

An Early History of the Heritability Coefficient Applied to Humans (1918–1960)

  • Historical Essay
  • Published:
Biological Theory Aims and scope Submit manuscript


Fisher’s 1918 paper accomplished two distinct goals: unifying discrete Mendelian genetics with continuous biometric phenotypes and quantifying the variance components of variation in complex human characteristics. The former contributed to the foundation of modern quantitative genetics; the latter was adopted by social scientists interested in the pursuit of Galtonian nature-nurture questions about the biological and social origins of human behavior, especially human intelligence. This historical divergence has produced competing notions of the estimation of variance ratios referred to as heritability. Jay Lush showed that they could be applied to selective breeding on the farm, while the early twin geneticists used them as a descriptive statistic to describe the degree of genetic determination in complex human traits. Here we trace the early history (1918 to 1960) of the heritability coefficient now used by social scientists.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others


  1. While Nathaniel Comfort’s (2012) history of medical genetics and Aaron Panofsky’s (2014) history of behavior genetics are both rich historical treatments, neither author focuses on the history of the heritability concept, our specific focus in this article.

  2. Fisher (1919) is explicit about this aim in the opening paragraph. Donald MacKenzie (1981) argues that Fisher’s support of eugenics had a decisive influence on all of his work in statistics. Eric Johnson (2021) provides some insights into some of the implications of Fisher’s lifelong support for eugenics on the occasion of Fisher’s window being removed from his college at Cambridge.

  3. Fisher did not use this term; rather, as Visscher and Walsh note, Fisher introduced “the ‘percentage of variance due to heritable factors’ (i.e., what we now call heritability)” (Visscher and Walsh 2019, p. 11) and “Fisher defined quantities that are now called narrow sense and broad sense heritability (variance due to ‘essential genotypes’ and ‘genotypes’, respectively)” (2019, p. 11). Lush, discussed below, was the first to use the term “heritability” for these variance ratios.

  4. Note here another use of a quasi-causal linkage phrase.

  5. Burks is an underappreciated figure in the early development of social science genomics. Trained at Berkeley and Stanford by Louis Terman, Burks’s dissertation was a pioneering attempt to use the adoption design to estimate proportions of variance attributable to “heritable” and “environmental” causes. A 1927 meeting organized by Terman at the National Society for the Study of Education, titled “Nature and Nurture Part 1: The Influence upon Intelligence (National Society For The Study Of Education 1927),” marked the beginning of modern social science genomics. At that meeting, Ms. Burks (as she was called by Terman, who referred to her as his “secretary”), 25 years old, dominated the proceedings. She gave two papers and a comment, among them one titled “Statistical Hazards in Nature-Nurture Investigations,” in which she sternly corrected many of the common practices of the time. Sadly, she was never able to find a stable academic career and committed suicide in 1943.

  6. Newman, Freeman, and Holzinger drew from Fisher’s Statistical Methods for Researchers (1950 [first edition published 1925]) rather than his 1918 paper.

  7. Clark (1956) presents h2 via this equation \({h}^{2}= \frac{{r}_{M - }{r}_{D}}{1-{r}_{D}}\), which he says comes from Holzinger (1929). As we show here, Holzinger does not introduce the notation h2 until 1937 in Newman et al. (1937)). Clark defines h2 as follows: “Heritability, designated by h2, may for our present purposes be defined as the proportion of the variance within like-sexed dizygous twin pairs which is attributable to genetic factors” (1956, p. 51).

  8. We also note that Thomas Kelley (1930) used the word “heritability” in an unquantified sense to refer to behavioral differences.

  9. David Marks (2019) documents scientific misconduct carried out by Eysenck. We introduce Eysenck here as he was one of the first psychologists to use a version of the heritability measure in the 1950s, and in particular one of the first postwar investigators to misinterpret it as a measure of something called “genetic determination.” Also, we note that he was drawing on Holzinger rather than Lush. We are not endorsing Eysenck’s work nor are we attempting to rescue him from any of the well documented charges of scientific misconduct.

  10. Kang et al. (1978) include the equation in their, at the time, exhaustive summary of mathematical methods for estimating heritability in human behavior genetics.

  11. Note the use of similar terms in this equation and Holzinger’s Eq. (7) discussed above.


Download references


The authors would like to thank both Jim Tabery and Jonathan Kaplan for their many helpful comments on earlier drafts of this article. We also thank Stuart Newman for his comments on our penultimate draft and Deborah Klosky for her copyediting.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Stephen M. Downes.

Additional information

Publisher's Note

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

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Downes, S.M., Turkheimer, E. An Early History of the Heritability Coefficient Applied to Humans (1918–1960). Biol Theory 17, 126–137 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: