Curl starts while the shaft is being made

A 2005 human-follicle study found retrocurvature of the bulb, asymmetry in the outer root sheath and more proliferating cells on the convex side of curly follicles. When dissected follicles produced new hair in culture, the curvature persisted.

That result supports a developmental origin below the skin rather than the idea that a strand becomes curly only after it exits the scalp. Styling can reshape bonds in the existing fiber, but the follicle continues to manufacture hair according to its own biology.

A strand is a changing 3D object, not one fixed oval

A 2025 micro-CT study examined low-, medium- and high-curl fibers from three donors. High-curl strands formed non-planar helices and averaged 12.33 twists and 6.81 waves per five centimeters. Mean cross-sectional ellipticity rose from 1.28 in low-curl fibers to 1.70 in high-curl fibers.

There was still substantial overlap within and between groups, and the donor sample was only three people. The useful lesson is geometric: ellipticity can contribute, but twist, curvature, diameter and how those properties change along the shaft matter together.

Highly heritable does not mean one gene

A twin analysis across three predominantly European-ancestry samples totaling 11,700 participants estimated broad-sense heritability of hair shape at 85% to 95%. That describes variation in those populations and environments; it does not mean one person's curl is 95% fixed or controlled by one switch.

A genome-wide study linked the TCHH region with hair morphology in European cohorts, but the variants explained about 6% of phenotypic variance. A later analysis of 28,964 people across nine cohorts identified many additional loci, and a 14-variant prediction model achieved only modest discrimination.

Genes, environment and styling operate at different levels

Genetic variants influence how a follicle develops and produces a shaft. Water, humidity, heat, mechanical force and chemical processing act on the keratin fiber after it has formed. Those styling effects can be temporary or structurally persistent without rewriting the follicle's inherited program.

Population labels are poor substitutes for an individual's hair. The genetic studies sampled specific ancestry groups, and fiber geometry varies within groups and within one scalp. The research explains biological contributors; it does not create a hierarchy of textures or a deterministic ancestry test.

Read the research

  1. Thibaut et al. — Human hair shape is programmed from the bulbBritish Journal of Dermatology (2005), PMID 15840091, DOI 10.1111/j.1365-2133.2005.06521.x
  2. van den Berg et al. — Three-dimensional variation of curly hair fibersJournal of Microscopy (2025), PMID 39564786, PMCID PMC11733847, DOI 10.1111/jmi.13365
  3. Medland et al. — Common variants in TCHH associated with straight hairAmerican Journal of Human Genetics (2009), PMID 19896111, PMCID PMC2775823, DOI 10.1016/j.ajhg.2009.10.009
  4. Liu et al. — Meta-analysis identifies loci involved in human head-hair shapeHuman Molecular Genetics (2018), PMID 29220522, PMCID PMC5886212, DOI 10.1093/hmg/ddx416

Editorial transparency

This article was prepared by the HairstyleChange Research Desk from the cited sources. It has not been presented as medical review by a named clinician. Corrections and qualified reviewer enquiries are welcome through our contact page.

How we research and correct articles ↗