It Grows. It Just Does Not Stay.

Khumalo measured hair length on four scalp regions in 1,042 South African schoolchildren and asked when each of them had last had a haircut.

Natural combed hair showed no difference in mean length between children who had gone two to five years since a cut and children who had gone more than five. No difference between one to two years and two to five either. Mean natural length, 5.1 centimeters.

The length stops changing after about a year. Not because the hair stops growing — it does not — but because breakage arrives at the same rate as new growth and the two cancel out. Steady state.

That single finding reframes an entire category of product marketing, a great deal of self-blame, and most of what gets said about growth oils.

The growth rate is slower, and that is not the interesting part

The numbers exist and they are small studies, so hold them loosely. Loussouarn, 38 young adults from central and western Africa, phototrichograms at three scalp sites, compared with Caucasian volunteers of similar age.

Density: 190 hairs per square centimeter against 227. Growth rate: 256 micrometers a day against 396. Telogen fraction: 18 percent against 14.

Roughly 65 percent of the growth rate and 84 percent of the density. The author noted the paradox in his own paper — those figures “might suggest a trend towards increased hair loss in Africans, even though it contrasts with a lower and slower incidence of the development of alopecia.”

The large replication, 2,249 young adults from 24 ethnic groups across five continents, confirmed the direction: lower density and slower growth for African hair, thicker and faster for Asian hair, high total density for Caucasian. The per-group numbers live in the paper body rather than the abstract, and any table of them circulating without a citation should be treated as invented until checked.

But a 35 percent slower growth rate does not produce a 5.1 centimeter steady state. Breakage does.

Where it breaks, and why

Khumalo’s 2000 study looked at more than 2,000 hairs from 12 African volunteers who had never chemically treated their hair, under light microscopy and scanning electron microscopy.

Knots were present on 10 to 16 percent of African hair shafts, against 0.15 percent in the comparison groups.

That is a seventy- to hundredfold difference, and it is the load-bearing fact in this entire subject. It is also a twelve-person study from 2000, and nothing since has replaced it.

The same work found shafts that appeared broken, electron microscopy features consistent with repeated breaks, and, examined in place on the head, interlocking of shafts into what the paper calls a mat of tightly interwoven hair — a tight coiled springlike structure. A knot is a stress concentrator. Pull a comb through a mat of them and the fiber fails at the knot.

The hair is not weak

The received wisdom collapses in two separate papers.

The first compared cystine-rich protein distribution across Black African, Caucasian and Asian hair, plus hair from patients with trichorrhexis nodosa and trichothiodystrophy, using transmission electron microscopy and silver staining. The pattern in Black African hair was normal — a normal electron-dense A layer, normal exocuticle, normally outlined cortical microfibrils — and unlike trichothiodystrophy. The authors’ conclusion, verbatim: the excessive structural damage observed in African hair shafts is consistent with physical trauma resulting from grooming rather than an inherent weakness due to any structural abnormality.

The second ran label-free shotgun proteomics on hair from Black African, Caucasian, Asian and Mixed-Ancestry donors, identifying more than 450 protein groups, and classified the samples by measured geometry as well as by race. No protein significantly differentiated racial or geometric groups. The authors describe racial classification as a social construct and its use as a biological endpoint as “imprecise and valueless in the era of precision medicine.” The paper is explicitly preliminary. It is also the only proteomic comparison there is.

So: normal composition, normal protein content, extraordinary knot frequency, and damage patterns that look mechanical. The fragility is topological, not compositional. The field’s own critical appraisal says exactly this — curly hair fragility is attributed to knotting and to crack and flaw formation, rather than to structural variations.

And the strength tests may have been measuring the wrong thing

A 2020 paper in the Journal of Investigative Dermatology found that the tensile strength of a curly fiber has two components, not one. There is the elastic region, which straight and curly fibers share, and a toe region — the springiness at the start of the stress-strain curve — which contributes meaningfully in curly fibers and is essentially zero in straight ones.

Then the authors dropped the methodological bomb. Standard tensile testing protocols straighten the fiber to mount it. Straightening it destroys the toe region before the measurement starts. In their words, de facto methods of tensile testing may erode curly fiber strength during preparation.

A follow-up using cyclic loading supports the mechanism: the toe region was present on the first cycle and disappeared as the fibers progressively straightened, until their behavior resembled straight hair. Consistent with hydrogen bonding that exists only while the coil does.

Decades of data showing curly hair is weaker may in part be measuring an artifact of how the sample was clamped. The authors ask for validation in larger studies and have not had it.

Combing out-damaged braiding

Fifteen frequent braiders — more than eight times a year — against fifteen occasional braiders, in Johannesburg, assessed by scanning electron microscopy, cuticle cohesion, tensile testing and traction alopecia scoring.

The frequent braiders had thinner hair, more cuticle damage, significantly weaker fibers and higher traction alopecia scores. Expected.

The unexpected result was in the combing arm. Surface damage rose linearly with the number of strokes, and cuticle damage from 480 combing strokes was more severe than that from frequent braiding.

Four hundred and eighty strokes is a fortnight of ordinary detangling. The gentle alternative out-damaged the thing it is an alternative to.

One disclosure: the authors include L’Oréal research affiliations. The finding cuts against nothing the company sells, so far as I can see, but it belongs in the sentence.

What I could not find, and looked hard for, is a study measuring combing force — in grams, in newtons — wet against dry, in high-curl hair. It does not appear to exist. Which means the single most universal piece of advice in textured hair care, detangle wet, rests on no force measurement anyone has published. It may well be right. Nobody has shown it.

Somebody did publish the benchmark — how often people comb, how fast, and what load a single fiber takes — and then built it entirely from Asian and Caucasian hair. That gap has its own piece.

Strokes are the currency

The length ceiling is a breakage ceiling. Anything that reduces breakage raises it; anything that promises faster growth is addressing the variable that is not limiting.

Not products, not oils, not the wash-day ritual — the number of times something is dragged through a mat of knotted fiber. Fewer sessions with more patience beats more sessions with better conditioner, and the linear damage curve says that is not a preference.

The hair is not defective. It is a coiled fiber with a hundred times the knot frequency of a straight one, of entirely normal composition, being groomed with instruments and protocols designed for something else. Every study that has gone looking for an intrinsic weakness has come back with normal cystine, normal proteome and normal microfibrils.

And the relaxed-hair figure from the same schoolchildren study is the one that lingers. Relaxed hair reached its steady state later and longer — 10.9 centimeters — but the authors called it “shorter than expected,” and it was significantly shorter on the occiput than everywhere else on the head. Khumalo flagged that occipital pattern as a clue to disease pathogenesis. Nineteen years on, nobody has followed it up.

This article is educational and is not a substitute for care from someone who can look at your skin. Anything new, changing, painful or spreading should be assessed rather than researched.