Corrections: what the first edition got wrong

The first edition said it was. The Tsimane of Bolivia, forager-horticulturalists with no office chairs, report more pain than several comparison populations, and chronic pain at rates within the range of industrialized countries. What changed is what comes after the pain: disability.

In Swedish national registries, starting medication was associated with lower death rates over two years.

and the link between night shifts and breast cancer rests on more than the single study the first edition used.

In humans, the evidence for inheriting a grandparent's experiences through epigenetic marks is weak.

There is no validated way to measure autophagy in a living person.

Gut serotonin does not cross into the brain.

The first edition blamed them for inflaming the body and making the gut lining leak. The fats people eat did change, enormously; that seed oils make the gut leak is not established.

The study cited compared garden crops in 1950 and 1999.

People at South African sites lived through it.

The systems are linked, and inflammation is one interacting cause among several, not a single root.

That pathway is a proposal with observational support.

The evidence is narrower: touch of the right kind is followed by lower cortisol and heart rate.

Early small trials were promising; larger placebo-controlled trials in Crohn's disease found no benefit.

The pain system can stay turned up a long while, but the change is reversible.

Many people lived on with lasting damage, cared for by others.

The first edition told readers to keep the bedroom at 18 to 19°C (64 to 66°F). Neither paper it cited contains that number.

Anna Lembke recommends 30 days away from a compulsive behavior, not 24 to 72 hours.

The trial the first edition cited cut days absent from work by 29%, not days of illness, which did not change.

Not "more than 600 times": most of the body's cells are red blood cells with no DNA. Dozens of round trips.

About 1 to 2% outside Africa, not 1 to 4%.

Permanent settlement is conventionally dated to about 3,600 years ago, with seasonal use far earlier, not "roughly 25,000 years".

Shared descent over about 15 million years, not 90 million.

27.5 in 2023, not 30.4.

Not only humans, orcas and pilot whales: five toothed whales, and wild chimpanzees at Ngogo in Uganda.

No single "10%" figure exists in the study cited.

After six weeks of daily mild cold, the extra heat it produced rose by roughly 180 calories a day, not 200 to 400.

In the national survey cited, 82.1% of Black Americans and 30.0% of white Americans were deficient.

The "700,000 to 2.1 million" figure had no source; about 350,000 are registered for commerce.

Brain cells do not shrink by 60%; in mice, the space between them grows by about 60%, and even that model is disputed.

The estimate depends heavily on the analysis used to make it.

About 806,000 Americans from 1999 to 2023, not "over 500,000".

Adults who survive childhood often live into their late sixties and seventies, a decade or so short of industrialized populations, and far fewer children reach adulthood. Nor were heart attacks unknown: about a third of ancient mummies scanned for the purpose showed hardened arteries.

The first edition said generalized anxiety disorder affects about 6% of the world. The World Health Organization puts anxiety disorders of every kind together at about 5.8%; generalized anxiety disorder alone is far rarer, about 1.8% of adults in a given year.

Twelve since 2023, not nine.

The famous "9 to 17 extra years" of telomere aging compared the most and least stressed mothers in one small study, not caregiving mothers with childless women.

About half of Americans eat less than the average requirement, which is not the same as clinical deficiency, and standard water treatment generally leaves minerals in tap water.

It does not "paradoxically impair cellular hydration": it dilutes the blood and makes cells swell.

The first edition said it raises oxygen uptake by 10 to 15%. The one small study behind the figure found skin oxygen readings about 10% higher in six of eight volunteers.

Americans average about 5,100 steps a day by step counter, not "fewer than 4,000". No study supports the first edition's forecast that self-driving cars would end walking within 20 years.

Lactase persistence took thousands of years to spread, not "about 300 generations", and its timing tracked famine and disease rather than milk drinking.

US tweens averaged 5 hours 33 minutes and teens 8 hours 39 minutes of entertainment screen media a day in 2021.

The first edition's "50 to 150%" figure was tied to no dataset anyone could check.

The figures were attributed to the wrong studies.

The similarity tracks living together and closeness, not shared meals.

Published in 2015, with the tablet's light peaking at 452 nm.

One short night cut natural killer cell activity by about a quarter, not 70%, and a normal night restored it.

The "four to seven hours a day" figure could not be traced to a primary study, and neither could the claim that outdoor play fell by half between 1955 and 2005.

Visits of 20 to 30 minutes gave the biggest cortisol drop per minute spent, not the biggest drop overall.

The "96 times a day" figure comes from a three-day survey run for a device-insurance company, not from measurement.

Heart-rate variability appears never to have been measured in them, so the ancestral level is unknown.

No primary source exists for the first edition's comparison of hunter-gatherer and modern body-fat percentages.

The first edition tied its drowning to the Storegga Slide tsunami; a 2024 model suggests much of it was lost earlier.

No trace could be found of the 2005 paper the first edition credited with a rise in aerobic capacity in six to eight weeks.

Daniel Lieberman popularized the idea. S. Boyd Eaton and Melvin Konner proposed it in 1985 and, with Marjorie Shostak, called it a "mismatch" in 1988.

attributed to Lieberman cannot be found in his work and has been dropped.

David Carrier proposed it; Louis Liebenberg documented the hunts.

was presented as fact. A hominin roughly seven million years old, found in Chad far west of the Rift Valley, undermined it.

Four-legged animals do have the ligament the first edition said "barely exists" in them: in dogs, injuring it is among the most common orthopedic problems.

The first edition credited Holly Dunsworth's team with the pelvis-versus-brain explanation of early birth; their paper argues close to the opposite, that the mother's energy budget sets the timing.

does not absorb more ultraviolet light; it lets more through.

is the food scientist Steven Witherly's phrase, and his example was Cheetos.

compared Guahibo Amerindians, rural Malawians and US city dwellers, not the Hadza and the Yanomami.

was about bowel movements, not hip angles in chairs.

had two participants, the author James Nestor and Anders Olsson.

is folklore with no primary source.

the first edition cited was three different papers merged into one, and the minutes it recommended come from modern cohorts, not from any ancestral pattern.

could not be traced to the World Health Organization source it was credited to.

The claim that women's stress systems react more to social threats and men's to achievement rested on one study of 50 students. Pooled across 34 studies, men's cortisol ran higher.