The vehicle has become a guilt-ridden trap.
Somewhere between the bathroom scale and the self-help aisle, a story was written about your body. You did not write it, but you have been living inside it so long that you probably believe it is yours. The story says: you are responsible. Your weight, your skin, your fatigue, your digestion, your mood — yours to manage, yours to fix, yours to blame when the fix does not hold. The story is told so universally and so early that most people mistake it for a natural law. But it is worth asking who is protected by such a story. Who benefits when the organism blames itself? In clinical psychology, when a person absorbs guilt that belongs to a system, it is recognised as a pathology. In the food industry, it is recognised as a marketing opportunity. This chapter is about what happens when you stop looking at the animal and start looking at the enclosure.
In the autumn of 2018, in a laboratory at the Hawkesbury campus of Western Sydney University, I watched a woman in a white coat feed a eucalyptus leaf into a machine that cost more than my car. The machine was a high-performance liquid chromatograph — HPLC — and it was doing something that would have seemed absurd to anyone unfamiliar with the field: it was running a full toxicological profile on a single leaf, from a single tree, to determine whether that leaf was safe to feed to a koala.
I was twenty-seven, two years into a doctoral project on soil mites in the Acari collection, and my laboratory was two doors down from the eucalyptus analysis facility. I had walked past it every day for weeks without registering what was happening inside. Then one morning the door was propped open, and I looked in, and I did not leave for two hours.
The facility was analysing eucalyptus browse for the zoological programme. The process was this: field technicians collected fifty grams of mature leaf tips from tagged trees on the university's research plantations — each tree individually numbered, GPS-mapped, and chemically fingerprinted. The samples were brought to the lab in sealed bags, frozen at minus twenty degrees Celsius, then freeze-dried, then ground in a mill to particles of one millimetre or less. This powder — which had been, that morning, a living leaf on a living tree — was then subjected to a sequence of analytical procedures that would not have been out of place in a pharmaceutical quality control laboratory. HPLC with reversed-phase gradient elution for the formylated phloroglucinol compounds. Gas chromatography-mass spectrometry for the volatile terpenes. Near-infrared reflectance spectroscopy for nitrogen availability. In vitro digestion assays using polyethylene glycol, pepsin, and cellulase to estimate apparent digestibility. The complete panel measured forty-nine distinct chemical compounds in a single leaf.
Forty-nine. In a leaf. For a koala.
I asked the woman — Dr. Sarah Chen, a plant chemist who had been running the facility for six years — why this level of analysis was necessary. Her answer restructured the way I thought about every living system I had ever studied. "Because," she said, "two trees of the same species, growing side by side, in the same soil, receiving the same rainfall, can have completely different toxin profiles. One is safe to eat. The other will make the animal sick. You cannot tell by looking. You cannot tell by species. You can only tell by testing."
This is worth pausing on, because the implications extend well beyond koala nutrition. The eucalyptus genus contains over seven hundred species. Koalas eat approximately fifty of them. Within any given region, they specialise in one to three primary species. But even within an acceptable species, individual trees vary wildly in their concentrations of formylated phloroglucinol compounds — the primary class of feeding deterrents — as well as terpenes, tannins, and cyanogenic glycosides. A tree that is perfectly edible in March may be toxic by November. A tree that is safe at the base canopy may be dangerous at the crown, where UV exposure increases secondary metabolite production. The chemistry is not merely complex. It is individually variable, seasonally unstable, and structurally hostile to generalisation. Think it doesn't relate to people? I wish.
Wild koalas navigate this chemical landscape using their noses. Ben Moore and William Foley at ANU demonstrated that formylated phloroglucinol compounds are the primary determinants of feeding. At concentrations above forty-five milligrams per gram of dry matter, leaf intake drops by half. The animals sniff each leaf before eating, using volatile terpene compounds as olfactory cues to predict the concentration of non-volatile toxins deep in the leaf's tissue. The animal is, in effect, running a rapid field assay on every mouthful. Its nose — a biological chromatograph operating through two hundred million olfactory receptor neurons — detects what no human eye could distinguish. When a leaf sniff indicates excessive toxin load, the animal moves to the next tree. If no acceptable tree is available, koalas do not eat. They prefer starvation to consuming a leaf their nose has flagged as dangerous. As the father of a fussy eater, I can relate to the plight of Dr. Chen. In captivity, when the animal can no longer simply move to the next tree, preventing starvation depends on ensuring the correct chemistry of the individual leaves — and that chemistry can only be determined with analytical instrumentation that costs, per unit, roughly what a small apartment costs in western Sydney.
There is more.
The koala's capacity to process even acceptable levels of plant toxin depends on a specialised gut microbiome — a consortium of bacterial species, predominantly Lonepinella koalarum and members of the Synergistaceae, that metabolise phloroglucinol compounds and terpenes into excretable byproducts. If this microbiome is disrupted — by stress, antibiotics, or an abrupt dietary change — the animal loses its ability to properly process food. It can be surrounded by edible leaves and still starve, because the chemical processing chain between ingestion and nutrition has been broken. The facility monitored this too, through faecal glucuronide assays — measuring the metabolic byproducts of toxin processing in the animal's waste to determine whether its gut was functioning.
I asked Dr. Chen what happened when the chemistry was wrong. I expected her to describe something dramatic — seizures, perhaps, or organ failure. What she described was worse: nothing visible. The animal sits in the tree. It loses weight — three to ten percent of body mass — but koalas are small and furred, and the loss is not obvious from the ground. Over time, the coat roughens. Muscle strength may reduce slightly. The demeanour flattens. The faecal pellets soften and smear — keepers call this "dirty tail," and it signals that the hindgut fermentation on which the animal's entire nutritional strategy depends has been disrupted. The animal does not cry out. It does not collapse. It sits where it has always sat, doing what it has always done, looking — to any observer without instrumentation — fine.
The data from Cape Otway, published by Whisson and colleagues in PLOS ONE in 2016, illustrate what this decline looks like at population scale. Researchers radio-collared twenty-one koalas in a manna gum forest that was undergoing progressive defoliation. Through early September 2013, seventy-one percent of the animals maintained good body condition scores — seven or above on a ten-point scale. Within two months, fifteen of the twenty-one were dead or euthanised. The trajectory was not a gradual slide. It was a threshold event: the animals compensated, compensated, compensated, and then they could not. Cortisol levels in chronically stressed populations — measured by Narayan and colleagues through faecal glucocorticoid metabolites — ran ten to twenty-five times above baseline. The organism was in allostatic overload. It had been in allostatic overload for months, possibly years, while appearing to function normally. This is what chronic subclinical failure looks like in a thirteen-kilogram marsupial. It is also, as I would come to understand, what it looks like in a seventy-kilogram primate who seemed fine at his last checkup.
The mechanism underlying this decline is not simply nutritional insufficiency. It is a biochemical paradox. The terpenes in eucalyptus — particularly 1,8-cineole and p-cymene — must be detoxified by the koala's liver through cytochrome P450 enzymes. This detoxification is metabolically expensive. But the terpenes that enter the bloodstream during processing do not merely drain energy. Maher and colleagues at the University of Sydney demonstrated in 2019 that these compounds, at the concentrations naturally present in koala blood after eating eucalyptus, directly suppress immune function — specifically IFN-gamma, IL-6, IL-10, and IL-17A, the cytokine pathways required for mucosal immunity. The very act of eating eucalyptus suppresses the immune system the koala needs to fight infection. This is not a design flaw. It is an evolutionary trade-off that works in the wild, where the animal can regulate its intake by moving between trees. In captivity, or in a degraded habitat where the available browse is uniformly poor, the trade-off becomes a trap: more eating to compensate for poorer nutrition, which means more terpene absorption, which means more immune suppression, which means greater vulnerability to chronic infections.
I spent that morning watching a woman in a white coat run pharmaceutical-grade chemistry on a eucalyptus leaf so that a thirteen-kilogram marsupial could eat lunch. The animal's food was tested for forty-nine compounds. Its gut was monitored for bacterial function. Its chronic stress was tracked through hormone metabolites in its waste. Its immune suppression was understood as a consequence of its diet. Its treatment was assessed for downstream effects on its microbiome. Everything that entered the animal was measured against the animal — not against a regulatory standard, not against an economic threshold, but against the specific biology of the specific organism. And I remember thinking, very clearly: is this what care looks like?
I left the laboratory at noon and walked across the campus to get something to eat. The Hawkesbury campus sits in the flat agricultural country west of Sydney, and its food options at the time consisted of a cafeteria, a vending machine corridor, and a small food court near the student union. I bought a cheese and salad sandwich on wheat. Sat at a metal table in the sun watching a cheer meet.
It was not until several years later — standing in the penguin house at Rotterdam, developing the framework that would become this book — that I understood what I had witnessed that morning, and what I had done at noon, and the relationship between the two.
A civilisation deploys its most sophisticated analytical chemistry to ensure that the food entering one animal's digestive system is compatible with that animal's specific biology. Forty-nine compounds measured. Individual trees tested. Seasonal variation tracked. Gut microbiome monitored. Immune consequences modelled. The entire apparatus exists so that the animal's food nourishes rather than harms it. And then I ate an unimposing, eucalypt-adjacent cheese and salad sandwich on wheat. I would not like to apply Dr. Chen's analytical framework to what I ate.
The bread was labelled "wheat" — which, in Australian food terminology, means it contained some proportion of wholemeal flour. The ingredient list read, in compressed regulatory language: wheat flour, water, yeast, salt, emulsifiers, vegetable oil, wheat gluten, malted barley flour, vitamins. This list was not complete. Under current food labelling regulations in both Australia and the UK, substances classified as "processing aids" are exempt from declaration. The bread contained, in addition to its listed ingredients, between eight and twelve enzyme processing aids — fungal alpha-amylase, maltogenic amylase, xylanase, lipase, phospholipase, protease, transglutaminase, hemicellulase — none of which appeared on the label. These enzymes are classified as "processing aids" rather than "ingredients" on the grounds that they are destroyed during baking. Research from the University of Bochum has demonstrated that up to twenty percent of the allergenicity of fungal alpha-amylase survives in bread crust.
The word "vitamins" on the label referred to mandatory fortification: calcium carbonate, iron, niacin, and thiamin, added to white and brown flour in Australia and the UK. The word "fortified" tells you what was taken. During industrial milling, the bran and germ are removed from the wheat grain, and with them approximately sixty to eighty percent of the grain's zinc, magnesium, manganese, chromium, vitamin B6, vitamin E, folate, selenium, and fibre. Four nutrients are added back. The calcium carbonate was never in wheat to begin with — it was mandated during wartime rationing in 1943 and never removed. The enrichment does not restore what was lost. It patches four holes in a wall with twenty-five.
The emulsifiers were mono- and diglycerides of fatty acids and DATEM — diacetyl tartaric acid esters of mono- and diglycerides. These are standard in industrial bread. A 2021 study published in Microbiome found that DATEM causes significant, non-reversible reduction in gut bacterial density and decreases the abundance of Faecalibacterium — one of the primary anti-inflammatory commensal bacteria in the human gut. Two doors down from where I was eating, Dr. Chen's facility monitored the koala gut for precisely this kind of microbial disruption. My gut was not monitored. The emulsifier was not tested for its effect on my microbiome. It was tested for its effect on dough.
The wheat itself deserves attention. Fan and colleagues at Rothamsted Research in England — where the world's longest-running agricultural experiment has been tracking crop composition since 1843 — published data in 2008 showing that zinc, iron, copper, and magnesium concentrations in wheat grain were stable from 1845 to the mid-1960s, then declined significantly. The decline coincided precisely with the introduction of semi-dwarf, high-yield Green Revolution cultivars — varieties bred to produce more grain per hectare. They succeeded. The grain contained measurably fewer minerals per gram. The bread in my sandwich was made from wheat that had been, by any historical measure, bred to produce more of itself with less of what the organism eating it requires.
The salad was iceberg lettuce. Iceberg lettuce is ninety-five point six percent water. It contains negligible protein, negligible iron, negligible calcium, and less than three milligrams of vitamin C per hundred grams — roughly one-tenth the amount in romaine, which costs no more to grow. The British and Irish Association of Zoos and Aquariums discourages iceberg lettuce in herbivorous animal diets because the animals fill up on water-weight and fail to consume adequate nutrients. In my sandwich, the iceberg was the healthy part.
If that sandwich had been submitted to Dr. Chen's facility for the same analytical rigour applied to the koala's eucalyptus, the panel would have identified approximately forty-nine compounds requiring assessment — eight undeclared enzyme residues in the bread, two emulsifiers with documented effects on gut bacteria, heat-induced acrylamide and hydroxymethylfurfural in the crust, pesticide residues on the lettuce, and mineral oil hydrocarbons migrating from the recycled cardboard packaging. Of these, roughly twenty-five to thirty would have no evolutionary precedent in the primate diet — compounds the human digestive system has never encountered in two million years of the genus Homo. A zoo nutritionist would have rejected the sandwich as feed for a captive chimpanzee. I ate it for lunch and believed I was making a sensible choice.
Nobody tested any of this for compatibility with my biology. Nobody profiled my gut microbiome to determine whether the DATEM would disrupt my bacterial community. Nobody checked whether the wheat from which the bread was milled retained sufficient minerals to justify calling it food. Nobody tested the lettuce for the pesticide cocktail that the UK Expert Committee on Pesticide Residues finds on nearly half of all salad samples surveyed. The sandwich had been tested — rigorously, expensively — but not for me. It had been tested for shelf stability, for flavour consistency, for manufacturing efficiency, and for compliance with regulatory standards that define "safe" as "unlikely to cause acute illness in a statistically average adult within the period of observation." The koala's food was tested for the koala. My food was tested for the supply chain.
The koala does not feel guilty about needing forty-nine compounds tested. Dr. Chen does not blame the koala for being hard to feed. The system is designed around the animal's biology and its instincts, not around the animal's willpower. If the animal fails to thrive, the browse is examined. The feed is adjusted. The assumption, always, is that the system must bend to the organism — not the other way around.
This is the first and most visceral dimension of the human enclosure failure, and it is the one that most people encounter multiple times every day without recognising it as a failure at all. The food system that feeds Homo sapiens was not designed around the species' nutritional biology. It was designed around the economics of production, distribution, and sale. The animal is not the client. The animal is the market.
Koala care and eucalyptus analysis may sound expensive to the reader. And many may assume it is unrealistic to scale this kind of precision to eight billion humans. I would encourage the reader to perform a quick mental arithmetic on the cost of the level of detail and precision we put into marketing — the advertising, the packaging design, the behavioural research that tells food companies exactly which shelf height maximises impulse purchases, which colour palettes trigger appetite, which label claims bypass critical thought. Whatever figure you arrive at, double it and you are still not close. Now add the cost of the Therapeutic Goods Administration. The diet industry. The wellness influencers. The government food pyramids that are redrawn every decade and never quite manage to address why the population keeps getting sicker. It costs far more to convince an organism to go against its biology than it would to simply feed it correctly.
To understand how far the human food supply has drifted from the organism it ostensibly serves, it helps to start with the organism.
Homo sapiens is an omnivorous primate with a digestive system calibrated, over roughly two million years of evolution, to process a varied diet of wild plants, fruits, nuts, seeds, tubers, insects, fish, and intermittent meat. The gut is intermediate in length between a dedicated carnivore's and a dedicated herbivore's — shorter than a gorilla's, longer than a cat's — reflecting the species' evolutionary strategy of exploiting a wide range of food sources depending on seasonal and regional availability. The microbiome — the community of approximately thirty-eight trillion bacteria that inhabit the gastrointestinal tract — co-evolved with this dietary pattern and is, like the koala's, exquisitely sensitive to what it is asked to process.
The research is extensive and, at this point, unambiguous. Tim Spector at King's College London has demonstrated through the PREDICT studies — the largest ongoing nutritional science programme in the world, tracking tens of thousands of participants — that individual metabolic responses to identical foods vary by as much as tenfold between people. A meal that produces a moderate blood sugar response in one person produces a diabetic-range spike in another. A fat source that is efficiently metabolised by one gut microbiome is inflammatory in another. Spector's conclusion, published across multiple peer-reviewed papers, is that there is no universally "healthy" diet — that nutrition is individual, microbiome-dependent, and impossible to generalise at the population level.
This finding, had it emerged in zoological science, would have been considered unremarkable. Dr. Chen could have told you in 2009 that two koalas of the same species respond differently to the same tree, because their microbiomes differ, because their stress histories differ, because their mothers' pap — which seeds the infant gut with its founding bacterial community — differed. The individuality of nutritional response is a foundational principle in captive animal management. It is treated as a revelation in human nutrition because human nutrition has, until very recently, been a field that studied populations rather than organisms.
The consequence of this population-level approach is visible in every supermarket on the planet. The food supply is standardised. A loaf of bread in Perth is chemically identical to a loaf of bread in Portland. A chicken breast in Leiden is raised, fed, processed, and packaged by the same methods as a chicken breast in Lagos. People with ulcerative colitis are encouraged to try again to eat the same foods as people without. The system is optimised for consistency, not for the individual biology of the animal consuming it. A zookeeper who fed every gorilla in a troop the same diet, in the same quantities, at the same times, without adjusting for individual metabolic needs, stress levels, or microbiome health, would be failing a first-year husbandry assessment. The human food system does this to eight billion animals simultaneously and calls it a supply chain.
The ancestral diet of Homo sapiens is not a mystery. Archaeological, isotopic, and coprolite evidence provides a detailed picture of what the species consumed for the roughly two million years preceding agriculture. Daniel Lieberman at Harvard — the same researcher whose work on barefoot running appeared in Chapter 1 — has synthesised this evidence in The Story of the Human Body, and the picture is consistent across geographic regions and time periods.
The diet was varied: between two hundred and three hundred different food sources per year in most hunter-gatherer populations studied. It was seasonal: different foods at different times, with long periods of scarcity punctuating shorter periods of abundance. Eaton and Konner, who founded the field of evolutionary nutrition in a landmark 1985 paper in the New England Journal of Medicine, initially characterised the ancestral diet as universally high in fibre — seventy to one hundred and fifty grams per day, compared to the average modern intake of fifteen. But this generalisation does not survive contact with the full data. The Inuit thrived for millennia on a diet that was almost entirely animal-based, with negligible fibre. The Masai consumed primarily milk, blood, and meat. Both populations, prior to Western dietary contact, showed the same absence of metabolic disease as populations eating tubers and fruit. The pattern that holds across every pre-agricultural and early-agricultural population ever studied is not any particular macronutrient ratio — not high fibre, not low carbohydrate, not high fat. The pattern is this: unprocessed food from the local environment, eaten by the organism that evolved alongside it. The composition varied enormously. The principle did not.
The diet was low in sugar — fructose was available only in seasonal fruit, and honey, which constituted an estimated two to three percent of total energy intake, was rare and required significant effort to obtain. It was high in micronutrient diversity, because the variety of plant and animal sources provided a broad spectrum of vitamins, minerals, and phytonutrients that no single food category could replicate.
There is a reason the modern diet delivers so much sugar, and it is not human weakness. It is a fifteen-million-year-old mutation. Approximately fifteen to twenty million years ago, during a period of Miocene global cooling that reduced fruit availability across the forests where ancestral hominoid apes lived, the gene encoding uricase — the enzyme that breaks down uric acid — mutated into a non-functional pseudogene. Gaucher and colleagues reconstructed the evolutionary history of this mutation in PNAS in 2014. Every great ape carries the broken gene. Every human carries it. Nearly every other mammal on the planet has a working copy. The consequence, documented by Richard Johnson and Peter Andrews in Evolutionary Anthropology in 2010, is that when a human eats fructose, the resulting elevation in uric acid triggers a metabolic cascade that Johnson calls the "survival switch": increased lipogenesis, blocked fatty acid oxidation, and induced insulin resistance. The switch converts dietary fructose into stored body fat with extraordinary efficiency. When fruit was seasonal — available for weeks, not months — the switch was adaptive. The ape ate the fruit, stored the fat, survived the winter. The switch turned on and turned off.
The switch no longer turns off.
In 1700, the average English person consumed approximately four pounds of sugar per year — roughly five grams per day. By 1800, the figure had reached eighteen pounds. By 1900, sixty. By the year 2000, the average American consumed one hundred and fifty pounds of sugar annually — one hundred and eighty-five grams per day, much of it fructose from high-fructose corn syrup in processed food. A thirty-seven-fold increase in three centuries. The survival switch, which evolved to activate briefly each autumn when fruit ripened, now operates continuously, fifty-two weeks a year, in an organism that has not changed biologically since the Pleistocene. The obesity epidemic is not a failure of willpower. It is a survival mechanism performing exactly as designed, in an environment that no longer provides the off signal. This is not a Western problem because Western people are weak. It is a Western problem because the Western food supply is the one that delivers the signal without the off switch. No society that has adopted this food supply has escaped the outcome. Not one.
The modern diet inverts nearly every ancestral parameter. The average Western adult consumes products derived from four species — wheat, corn, rice, and soy — in various processed configurations. Where the ancestral diet delivered variety, the modern diet delivers monotony. Where the ancestral diet delivered intermittent sugar in a matrix of fibre and water, the modern diet delivers concentrated fructose stripped of everything that slowed its absorption. Where the ancestral diet changed seasonally, the modern diet is unchanging — the same products, from the same supply chains, in the same supermarket, fifty-two weeks a year. The microbiome, which evolved to process variety, is asked to process monotony. The results are measurable and they are not subtle.
Even the fruit has changed. Migicovsky and colleagues at Dalhousie University measured ten phenotypes across more than a thousand apple accessions and found that modern cultivated apples are three point six times heavier than their wild ancestor, Malus sieversii, and contain sixty-eight percent less phenolic content. The polyphenols — the compounds that protect the organism against oxidative stress, the compounds that were, in the ancestral fruit, part of the nutritional package alongside the sugar — were bred out because they cause enzymatic browning when the fruit is cut. A brown apple does not sell. A nutritionally depleted one does. The fruit was bred to look better and work worse.
The decline extends beyond fruit. Donald Davis at the University of Texas analysed USDA composition data for forty-three garden crops between 1950 and 1999 and found statistically reliable declines in protein, calcium, phosphorus, iron, riboflavin, and vitamin C — six, sixteen, nine, fifteen, thirty-eight, and twenty percent respectively. Mayer, Trenchard, and Rayns, extending the analysis to eighty years of UK data in the International Journal of Food Sciences and Nutrition in 2022, found that iron in vegetables had fallen by fifty percent, copper by forty-nine percent, and sodium by fifty-two percent since 1940. Individual extremes are difficult to read without pausing: watercress has lost eighty-eight percent of its iron. Cauliflower has lost sixty percent. Oranges have lost seventy-five percent. Benton and Thacker, reviewing the complete body of evidence in Foods in 2024, noted that eighty percent of this decline occurred in the last thirty to forty years. The degradation is not historical. It is accelerating.
The air itself is contributing. Loladze, in a meta-analysis published in eLife in 2014 that covered over seven thousand observations across one hundred and thirty plant species, found that elevated atmospheric carbon dioxide — the same increase that is warming the climate — reduces mineral concentrations in plant tissue by an average of eight percent. Myers and colleagues confirmed this in Nature the same year: wheat, rice, barley, and soybeans grown under the CO2 levels projected for mid-century show reduced zinc and iron. Zhu and colleagues found that rice under elevated CO2 loses ten percent of its protein. The mechanism is straightforward: more carbon dioxide accelerates plant growth, and the same quantity of soil minerals is distributed across more tissue. The plant grows bigger. It does not grow more nutritious. The organism that eats it receives more volume and fewer of the micronutrients that volume was supposed to deliver.
The Hadza people of Tanzania — one of the last remaining populations practising a predominantly hunter-gatherer subsistence pattern — harbour a gut microbiome approximately forty percent more diverse than the average Western adult's. This diversity is not a trivial statistic. Microbiome diversity is the single strongest predictor of metabolic health identified in nutritional science. Low diversity is associated with obesity, type 2 diabetes, inflammatory bowel disease, depression, autoimmune conditions, and cardiovascular disease. The human gut, like the koala's, requires the right inputs to function. Unlike the koala, the human receives no analytical assessment of whether the inputs are right. The assumption, embedded so deeply in the food system that it is invisible, is that if the food is on the shelf, it is fine.
The Hadza are not alone. In 1989, a Swedish physician named Staffan Lindeberg travelled to Kitava, a small island in the Trobriand archipelago of Papua New Guinea, and conducted what would become one of the most comprehensive dietary health surveys of a non-Western population ever published. Over the course of multiple visits spanning a decade, Lindeberg and his colleagues examined approximately twelve hundred islanders. The Kitavan diet was sixty-nine percent carbohydrate — a figure that, by the logic of every low-carb bestseller in the airport bookshop, should have produced a population riddled with insulin resistance, obesity, and metabolic disease. The carbohydrates came from tubers — yam, sweet potato, taro — along with fruit, coconut, and fish. Zero dairy. Zero refined sugar. Zero cereals, margarine, or oils. Less than one percent of caloric intake came from Western foods, roughly three US dollars' worth per year. Nobody had designed this diet. Nobody had optimised it. It was simply what grew on the island, prepared the way it had always been prepared, eaten the way it had always been eaten.
The results, published across a series of papers in the Journal of Internal Medicine and Metabolism between 1993 and 2001, were not merely good. They were, by Western epidemiological standards, impossible. Lindeberg and Lundh examined two hundred and thirteen adults and performed a hundred and seventy-one electrocardiograms. They found no case corresponding to stroke. No sudden cardiac death. No angina pectoris. ECG abnormalities were minimal even in subjects in their eighties and nineties. Blood pressure showed no association with age — a finding so contrary to Western clinical assumptions that it bears repeating: in Kitava, getting older did not raise your blood pressure. Fasting insulin in Kitavan men averaged 3.08 international units per millilitre, compared to 6.98 in age-matched Swedish men. Kitavan women: 3.37, against 6.65 in Swedish women. At ages fifty to seventy-four, Kitavan insulin levels were half those of their Swedish counterparts. And here is the detail that should dismantle any remaining confidence in simple dietary arithmetic: insulin in Kitavans decreased with age, while in Swedes it increased. The statistical significance was beyond dispute — P less than 0.001. The organism eating sixty-nine percent carbohydrate from unprocessed tubers was running cooler, metabolically, than the organism eating a modern Scandinavian diet.
Cordain, Lindeberg, and colleagues then published, in the Archives of Dermatology in 2002, a finding that no dermatologist has satisfactorily explained without reference to diet. Of the twelve hundred Kitavans examined, three hundred were aged fifteen to twenty-five — the age window in which acne vulgaris is most prevalent. The number who had acne, of any grade, was zero. Not low. Zero. Among the Aché of Paraguay — a second non-Western population examined in the same study, a hundred and fifteen subjects over eight hundred and forty-three days — the figure was also zero. In Western populations, the prevalence of acne in adolescents is seventy-nine to ninety-five percent. The paper's authors noted, with the restraint characteristic of peer-reviewed understatement, that the difference "cannot be solely attributed to genetic differences" and "likely results from differing environmental factors." The environmental factor was the food. Between seventy-six and eighty percent of Kitavan adults smoked daily — a habit that, in any Western risk model, would predict catastrophic cardiovascular outcomes. They had none. The browse was right. The organism was fine.
Eight thousand kilometres to the west, and twenty-four years later, a team led by Hillard Kaplan published in The Lancet the results of a study that approached the same question with different technology. Kaplan and colleagues performed coronary CT scans on seven hundred and five Tsimane adults in the Bolivian Amazon, aged forty to ninety-four. The CT scan measures coronary artery calcium — calcified atherosclerotic plaque — and is the most reliable non-invasive predictor of heart attack risk available to modern cardiology. The Tsimane results were the lowest ever recorded in any population on earth. Eighty-five percent of those scanned had a coronary artery calcium score of zero. At age seventy-five and above — an age at which, in the American MESA cohort study, roughly eighty-six percent of adults show measurable plaque — sixty-five percent of Tsimane elders still had perfectly clean arteries. The Tsimane diet was seventy-two percent carbohydrate, predominantly from plantain, manioc, rice, corn, and wild game, with fish providing most of the animal protein. Fourteen percent protein. Fourteen percent fat. No processed food. No refined sugar. No dairy.
What makes the Tsimane data particularly difficult to dismiss is the confound that should, by every Western biomedical model, have destroyed the result. Fifty-one percent of the Tsimane adults scanned had elevated C-reactive protein — a marker of systemic inflammation, driven in their case by chronic parasitic and bacterial infections endemic to the Amazonian lowlands. Chronic inflammation is the mechanism by which, according to three decades of cardiovascular research, atherosclerotic plaque forms and destabilises. The Tsimane had the inflammation. They did not have the plaque. Their arteries were cleaner than those of any Western population ever measured, including populations with access to statins, cardiac rehabilitation, and the full apparatus of preventive cardiology. The organism was inflamed, infected, and free of the disease that inflammation is supposed to cause — because the substrate on which inflammation acts, the metabolic environment created by the diet, was not there.
Three populations. Three continents. Three different diets — the Hadza eating wild tubers, baobab, and honey in the East African savanna; the Kitavans eating yam, coconut, and reef fish on a Melanesian island; the Tsimane eating plantain, manioc, and river fish in the Amazonian basin. Different macronutrient ratios. Different climates. Different gene pools. The same outcome: negligible cardiovascular disease, negligible metabolic syndrome, negligible obesity, and biomarkers that Western physicians would classify as belonging to a younger species. Nobody in any of these populations was following a programme. Nobody was counting macros or measuring glycemic index or taking supplements to compensate for what their food lacked. The food did not lack anything, because it had not been altered to lack anything. It was the browse of the local environment, eaten by the organism that evolved in concert with it. The match between food and biology was not optimised. It was never broken.
What happens when it is broken is not a matter of speculation. It has been documented, repeatedly, with the precision of a controlled experiment — except the experimenters were governments and corporations, and the subjects did not consent. The people of Nauru, a Pacific island nation, ate a traditional diet of fish, coconut, and fruit for millennia. In the mid-twentieth century, phosphate mining brought wealth, and with wealth came imported processed food — white rice, sugar, tinned meat, soft drinks. Within a single generation, the prevalence of type 2 diabetes exceeded forty percent, the highest rate ever recorded in any population on earth. A twenty-fold increase. Same island. Same gene pool. Same people. Different food. The Pima Indians of Arizona and the Pima Indians of the Sierra Madre in Mexico share recent common ancestry and near-identical genetic profiles. The American Pima, eating a standard American diet, have a diabetes prevalence of thirty-eight percent. The Mexican Pima, eating a traditional diet of beans, corn, and squash, have a prevalence of 6.9 percent. The genome did not change. The feed changed. Every documented case of a non-Western population transitioning to processed food shows the same trajectory: the diseases of civilisation — obesity, diabetes, cardiovascular disease, acne, dental caries, depression — appear within one to two generations. Trowell and Burkitt catalogued this pattern across dozens of populations in Western Diseases: Their Emergence and Prevention in 1981. The pattern has not changed. It has only accumulated more data points.
A zookeeper reviewing this evidence would not find it surprising. A zookeeper knows that when you change the browse, you change the animal. Not in theory. Not over millennia. Within years. Within a single generation. The question a zookeeper would ask is not why the Kitavans and the Tsimane are so healthy. That requires no explanation. The animal is eating what it evolved to eat. The question a zookeeper would ask is the one that Giles, at his desk in the enclosure, eating his sandwich from the supply chain, has never been trained to formulate:
What happened to my browse?
It is not fine. A zoologist would know this. A zoologist would look at the inputs, look at the organism's evolutionary dietary profile, look at the health outcomes, and conclude that the food supply is producing chronic subclinical malnutrition across the entire captive population. Not starvation — the animals are not thin. Many are overweight, which is itself a diagnostic signal: an animal gaining excess fat in captivity is typically consuming energy-dense, nutrient-poor food in the absence of adequate movement, which is precisely the behavioural profile of the modern Western adult. The food system is not failing to feed the animal. It is failing to nourish it. The distinction is everything.
The pattern that emerged in Chapter 1 — a well-intentioned intervention that severs the animal from its own biology — recurs across every dimension of the Vehicle. The shoe silenced the foot. The food supply silenced the gut. And the built environment silenced the body.
Homo sapiens is, as established in the previous chapter, a persistence predator — an endurance specialist whose physiology is calibrated for sustained aerobic movement across varied terrain. The cardiovascular system, the musculoskeletal architecture, the thermoregulatory apparatus, the neurotransmitter systems that regulate mood, sleep, appetite, and immune function — all of these were shaped by, and remain dependent on, regular physical movement of a specific type: low-to-moderate intensity, sustained, outdoors, on uneven ground, in social groups.
The modern enclosure provides a chair.
The average office worker in the developed world sits for nine to eleven hours per day. This figure — documented across multiple large-scale studies, including the UK Biobank cohort of over five hundred thousand participants — exceeds the amount of time the organism spends in any other single posture, including sleep. The species that evolved to move sits. The species that evolved to run is stationary. The consequences, again, are not subtle. A 2012 meta-analysis published in The Lancet by I-Min Lee at Harvard estimated that physical inactivity is responsible for approximately 5.3 million deaths per year globally — roughly the same as tobacco. The organism is not diseased. It is sedentary, in an enclosure that was designed for the convenience of the institution rather than the biology of the inhabitant.
Some humans attempt to compensate through dedicated exercise — the gym, the run, the fitness class. The impulse is correct; the execution recapitulates the pattern. The gym is a fluorescent box in which the animal performs repetitive movements on machines that constrain the body to linear planes of motion, isolated from weather, terrain, social bonding, and sensory input. The treadmill is the most precise metaphor for the human enclosure that I have encountered: an organism running at full capacity, going nowhere, in a controlled environment, while a screen mounted on the wall provides a simulation of the outdoor world it is not permitted to inhabit. I run on a treadmill three times a week. I noted in the previous chapter that this amounts to performing the activity the organism evolved to perform in precisely the manner that severs it from every signal its body was designed to receive. I have not stopped doing it. The enclosure makes the alternative inconvenient.
One more dimension of the Vehicle requires attention, because it is the one most consistently violated and least consistently recognised as a welfare issue.
Homo sapiens is a diurnal species with a circadian biology calibrated to solar light cycles. The suprachiasmatic nucleus — a cluster of approximately twenty thousand neurons in the hypothalamus — uses light input from specialised retinal ganglion cells to synchronise the organism's internal clock with the external day-night cycle. This clock regulates not merely sleep-wake patterns but hormone secretion, immune function, metabolic processes, body temperature, cognitive performance, and emotional regulation. It is not a convenience. It is the organism's master timing system, and it requires a specific environmental input to function: bright light during the day, darkness at night. The species evolved in equatorial Africa, where this input was reliably provided by the sun.
The modern enclosure provides electric light, shift work, screens that emit blue-spectrum light at wavelengths optimally calibrated to suppress melatonin production, and an economic system that treats sleep as an obstacle to productivity. Matthew Walker at the University of California, Berkeley, has documented in exhaustive detail — across two decades of research compiled in Why We Sleep — that chronic sleep restriction produces impairments in immune function, emotional regulation, memory consolidation, cardiovascular health, and metabolic stability. Walker's summary is blunt: "The shorter your sleep, the shorter your life." The average adult in the industrialised world sleeps six hours and thirty-one minutes per night. The biological requirement, established through sleep laboratory studies, is seven to nine hours. The gap — between sixty and one hundred and fifty minutes per night — is not a lifestyle choice. It is an environmental deficit, imposed by an enclosure whose economic rhythms are misaligned with the organism's biology.
A zookeeper who systematically restricted the sleep of a captive animal by ninety minutes per night — who installed lights that disrupted the animal's circadian biology, who structured the animal's daily routine around an economic schedule rather than its natural sleep-wake cycle — would face professional sanction. For Homo sapiens, it is called the working week.
The pattern is now visible across the entire Vehicle dimension. The food severs the animal from its nutritional biology. The environment severs the animal from its movement biology. The light severs the animal from its circadian biology. In each case, the mechanism is the same: a system designed for institutional convenience rather than species-typical functioning, producing chronic subclinical impairment that the organism experiences as normal because it has never known anything else. The koala at Cape Otway maintained good condition scores right up until the crash. So does the modern Western adult.
The woman I described in the previous chapter — the statistically median human who sleeps ninety minutes short, commutes in isolation, eats alone at her desk — is not failing to maintain her Vehicle. She is maintaining it with the tools the enclosure provides. She is doing everything the system tells her to do. She is eating the food on the shelf, sleeping the hours the schedule permits, exercising in the manner the built environment allows. The Vehicle is not neglected. It is serviced by a system that does not know what the animal needs, has never tested for it, and would not restructure itself if it did.
In a laboratory two doors down from where I used to work, a woman in a white coat is testing a eucalyptus leaf to make sure a koala can eat lunch. Nobody is testing what you had for breakfast.
The body is the first enclosure. But a well-maintained body in a barren environment is not flourishing — it is surviving. What does the animal do after its needs are met? Chapter 3 examines the dimension that every zookeeper checks first and every human system ignores: play.