You Have to Drink Hot

The building had a service corridor on the ground floor, and at the end of it a room about the size of a closet with a hot plate, a percolator, and a folding chair that had been repaired with wire. The man who kept the building running sat in that chair most afternoons. He was older than my father, Black, raised in the South, and he had been doing the work long enough that tenants asked him about the boiler before they called the managing agent. On the worst afternoon of that summer I came through with an iced coffee going soft in my hand, the cup so wet it kept trying to leave my fingers, and he was sitting there holding a mug with steam coming off the top of it.

I stopped walking. He watched me stop. Then he set the mug down and told me, with no air of revelation at all, that to cool down you have to drink hot.

I said nothing worth repeating. Standing in front of a man who had been managing his own body through American summers since before I was born, holding a beverage designed by a marketing department, the only argument I could assemble was the melting cup in my hand. Upstairs I went, back to work. That was thirty years ago, and I have carried his sentence around ever since without once checking it.

So let me check it.

Start with the day. A hundred and ten degrees. I have always said it that way, in the same spirit as everyone else who tells a New York heat story, because a hundred and ten is the municipal number for unbearable, the figure a body files when the thermometer’s opinion has stopped mattering. Nobody ever checks a number like that. I checked. Central Park has never recorded a hundred and ten. The all-time high there is 106, set on July 9, 1936, in a record that runs back to 1869, and the city’s coastal humidity is part of why our extremes stay below the inland numbers. What the middle nineties delivered was a run of murderous days that never touched the record: 102 degrees on July 15, 1995, with an overnight low of 84, which is the figure that kills people, since a city that refuses to cool at night gives nobody a chance to recover. Heat index readings that week ran well above the thermometer, which means the joke had a measured core. The air said 102 and my body said 110, and my body was reading the instrument built for what an afternoon does to a person. I cannot date the afternoon any closer than that, and after thirty years I have stopped apologizing for the number. It was the one part of the story I had right.

Now the claim itself. In 2012 a team at the University of Ottawa’s Thermal Ergonomics Laboratory put nine men on stationary bicycles for seventy-five minutes, wired them to skin sensors and a metabolic mouthpiece, and gave them water at four temperatures: 1.5, 10, 37, and 50 degrees Celsius. The paper, by Bain, Lesperance, and Jay, appeared in Acta Physiologica under a title that gives away the ending, reporting that body heat storage during physical activity is lower with hot fluid ingestion under conditions that permit full evaporation. The conditions were chosen to permit exactly that: a room kept desert-dry, so that nothing the riders sweated out could sit on the skin. Those who drank the 50-degree water, roughly 122 Fahrenheit, finished with less heat stored in their bodies than they did at any of the other temperatures, water at body temperature included. The result deserves its actual size. That is nine men in one laboratory, all of them pedaling, all of them gulping measured water on a schedule with a thermometer running well past the sphincter, which sits a long way from a man sipping coffee in a folding chair. What the lab produced was a mechanism demonstrated where it had room to work, and it has been quoted every summer since as an instruction.

The mechanism runs through the gut. The paper’s own guess pointed higher, at thermosensors lining the throat and mouth, so the laboratory went looking. Two years later the same group ran hot and cold water straight into the stomach through a nasal tube, and in a second pass had riders only swill it and spit. Swilling changed nothing. The stomach load moved sweat rates within minutes, at every site they measured, which placed the sensors in the abdomen and ruled the mouth out. Somewhere down there, receptors report the arrival of a heat load to the brain, which answers by opening the sweat glands wider than the actual load justifies. A small volume of hot liquid adds a small quantity of heat and buys a disproportionate quantity of sweat. If every drop of that sweat evaporates off the skin, the evaporation carries away more heat than the drink delivered, and the body finishes the exchange ahead.

That condition carries the whole case. If every drop evaporates. Ollie Jay, who ran the study and now directs the Heat and Health Research Centre at the University of Sydney, has put the practical version plainly: when the extra sweat drips on the ground, go get something cold. Sweat sitting on skin does no cooling; sweat wiped off with a towel does no cooling; the phase change is the entire transaction, and a humid August afternoon in New York is an environment ill equipped to complete it. Jay warns against his own drink on humid days and under clothing that holds the sweat, which is a rare thing to hear from a researcher whose most famous result gets quoted every summer as a license.

The verdict on the man in the folding chair splits cleanly. His rule is effective because the physiology underneath it is real and measured: hot input, oversized sweat response, net heat loss. His rule is not effective because he was applying it in a coastal city at high humidity, wearing work clothes, in dead air between two buildings, where the extra sweat he bought with that coffee had nowhere to go and ran down his back. He was carrying a true rule outside the conditions that make it true.

There is a symmetrical joke pointed the other way. In 2016 Jay’s group found that an icy slush suppresses sweating so efficiently that the lost evaporative cooling outweighs the heat the body dumps into melting the ice. My iced coffee had a problem of its own, with better advertising.

Where did his sentence come from? Nobody handed it to him with a citation. Hot drinks in hot places are close to universal: tea poured three times over in the Sahara, tea at noon in the Gulf, chai in Rajasthan in May, coffee on a Mississippi porch in a July when nobody in the county owned an air conditioner. Practices like that get assembled where air is dry, or moving, or both, and where evaporation finishes what the sweat starts. Then they travel, the way people travel, and land in a wet harbor city that voids the warranty. The Great Migration carried an enormous freight of practical knowledge north, most of it calibrated to the country it came from, and this is one small piece of that cargo.

Which brings me to the question that has been sitting under the coffee the whole time. What else did all of us inherit with the confidence of law and the evidence of nothing?

Call the category the totem. A totem is a rule handed down with a mechanism attached, where the rule gets repeated and the mechanism never gets opened. The mechanism is the tell. “Wear your coat” is a preference. “Wear your coat or you’ll catch your death” is a totem, because it carries a causal claim about cold air and infection that the speaker has not tested and cannot defend, and that the listener will pass along to a child in one sentence, in one tone, with the certainty fully intact.

Some totems are factless from the beginning. Feed a cold, starve a fever traces to a 1574 dictionary, where John Withals wrote that fasting is a great remedy of fever. It came out of a medicine that read fever as a fire and food as fuel, and it survived the collapse of that entire model of the body without losing a single adherent. The best part is that four and a half centuries later the proverb’s own speakers still cannot agree on what it says. In 1853 an essay in the Dublin University Magazine argued that the saying was never a prescription, and that the correct reading ran the other way: feed a cold and you will have a fever to starve. Here is a rule with no evidence, no settled meaning, and a survival record that would embarrass most of what we teach in school.

Other totems eventually get tested, and the testing can read as comedy. Donald Unger, a California allergist, spent his boyhood hearing from his mother and several aunts that cracking his knuckles would give him arthritis, and later heard the same sentence from his mother-in-law. He cracked the knuckles of his left hand at least twice a day and left his right hand alone as a control. That went on for fifty years, by his own count at least 36,500 cracks on the one side, and then he compared the two hands and published the result as a letter in Arthritis and Rheumatism in 1998: no arthritis in either hand, no observable difference between them. He collected the Ig Nobel Prize in Medicine in 2009 at Sanders Theatre, the citation by then crediting him with more than sixty years on the protocol. The joke stops being funny on the second pass. The only controlled trial his mother’s claim ever received cost a man his entire adult life, ran with a sample size of one, and got published as correspondence. That is the ordinary rate at which household rules receive an audit.

A third kind turns out to be right, and the vindication arrives sideways. Stephen Rennard, a pulmonologist at the University of Nebraska Medical Center in Omaha, spent years watching his wife Barbara make her Lithuanian grandmother’s chicken soup whenever the house came down with something. In 1993 he took the recipe into the laboratory. The 2000 paper in Chest, with Barbara Rennard as first author, ran the soup through a Boyden blindwell chamber and found that it inhibited neutrophil chemotaxis, the migration of the white cells that drive the inflammatory misery of a cold, and did so in proportion to concentration. Honesty matters more here than the headline: this was cells in a dish, with no patient involved anywhere, and many a supermarket soup inhibited every bit as strongly as Grandma’s did. The finding is effective because it proposes a real anti-inflammatory mechanism for a remedy that a century of clinical medicine had never bothered to test. The finding is not effective because it cannot tell you whether eating the soup shortens a cold by so much as an hour. Grandma’s recipe came out of that chamber with its category intact and its special standing gone.

Then there are the totems that carry weight, where pulling out the ritual pulls out the reason people are alive.

Maize came out of Mesoamerica with a procedure wrapped around it. You cook the dried kernels in water with lime or wood ash, let them steep, wash them, and grind the result into masa. The alkali loosens the hulls and makes the dough behave, which is what a cook would notice and what a cook would teach. It also releases the niacin, which in untreated corn sits chemically bound and passes through the body unused. It adds calcium, improves the amino acid balance, and cuts mycotoxin contamination. Nobody in the Valley of Mexico knew about vitamin B3. Nobody had to. The step lived in the recipe, the recipe lived in the family, and the family lived.

Corn crossed the Atlantic without the lime pot. Europe imported the seed and the yield and left behind the one step in the sequence with no visible purpose. Polenta, grits, cornbread, mush. Pellagra followed the crop through Italy, Spain, Egypt, and the American South, arriving in the twentieth century as a disease of the poor with four signature symptoms and a reputation for madness. South Carolina alone reported thirty thousand cases by 1912, with a case fatality rate near forty percent. At its peak the disease killed seven thousand people a year across fifteen southern states. When the Public Health Service sent Joseph Goldberger into the region in 1914, he found mill villages and tenant farms living on what he called the three M’s, meal and molasses and meat, where the meal was cornmeal and the meat was fatback. He proved the case in both directions, curing orphanage children by feeding them and producing the disease in prisoners by restricting them, and when the profession refused the answer he ran what he called filth parties in the spring of 1916, swallowing and injecting material taken from pellagra patients to demonstrate that the disease would not transmit. He and his colleagues stayed free of the disease. The profession stayed unconvinced for years, partly because a privately funded commission had already tied pellagra to bad sanitation, a real pattern in its own surveys, and read that pattern as contagion, which it was not, and partly because a Jewish New Yorker informing the South that its diet was killing its poor was not a message with a hearing.

The two stories side by side give the shape of the problem. A Mexican grandmother could not have told you what niacin was, and her family did not get pellagra. A trained American commission with funding and a survey design could tell you a great deal about sanitation, and it got the cause backward while thousands died. The unexamined totem was carrying real load. The tested hypothesis was wrong. Neither result is what the folklore of science predicts.

The lime pot does not carry the whole account, and I will not pretend otherwise. A diet of meal, molasses, and fatback runs short of the protein that supplies tryptophan, which a body converts into niacin on its own, so a tenant family in Alabama could have limed every kernel it owned and walked into the disease anyway on the strength of what else was missing from the table. Goldberger landed where his evidence pointed, on poverty. The lime step raises the floor without building the house. What it does show is how a procedure can hold a piece of biochemistry in trust across three thousand years and several thousand miles, and how casually the holding gets broken when the food travels and the method stays home.

One further turn here should keep everyone modest. In 1998 Jennifer Billing and Paul Sherman at Cornell published a survey in the Quarterly Review of Biology arguing that heavy spice use in hot countries was cultural evolution at work, since many spice compounds are antimicrobial and food spoiled faster before refrigeration. The explanation was satisfying, it was repeated everywhere, and it became the standard answer to why the food is hot where the weather is hot. In 2021 Lindell Bromham’s group at the Australian National University tested it against 33,750 recipes from seventy cuisines using ninety-three spices, controlling for the awkward fact that neighboring cultures share ancestors as well as climates, and reported in Nature Human Behaviour that the temperature relationship largely dissolves under those controls, that other antimicrobial ingredients such as vinegar and alcohol show no matching pattern, and that spice use runs with poverty and with health outcomes that have nothing to do with infection, car crashes included. A scientific explanation for a folk practice spent twenty-three years operating as its own unexamined totem, passed hand to hand through textbooks and Sunday supplements with a citation stapled to it, which is the credential the coffee man lacked, and no better checked.

The plains version of this is the one I grew up inside. Everybody in Nebraska knows that a green sky means hail. The color is real, and that part has been settled with instruments: Frank Gallagher and his colleagues carried a handheld spectrophotometer through the storm seasons of the middle nineties, on the ground and aboard research aircraft, and measured genuine green light coming off thunderstorms. They recorded green storms over ground that was not green and blue storms over ground that was green, which killed the theory that the sky was reflecting the fields. What the measurements did not support was the causal story every farmer attaches to the color. Gallagher’s group recorded green and nongreen thunderstorms alike and watched single storms shift from dark blue to bluish green, and most storms that drop hail never turn green at all. The eye was right about the sign, and the mouth was wrong about the reason. Reading weather off a sky is a survival literacy with a real track record, and the explanations bolted onto it afterward are the part that goes bad in storage.

A totem survives because it is cheap social software, and accuracy has little to do with what keeps it running. Saying “to cool down you have to drink hot” in a service corridor performs membership and instruction at the same moment, and the man saying it is giving you something. The longest-running rules carry no feedback loop: nobody who cracks a knuckle ever receives the disconfirming letter, so the warning propagates undefeated for a century. The next tier has a feedback loop that feels like confirmation, which is where hot coffee sits, since the sweat does arrive, promptly and unmistakably, and a body reads any strong response as an answer. And welded into procedures sit the rules that outlast everything, where the reason has fallen off and the step remains, because a step in a recipe asks nothing of your beliefs.

That last category is where the danger lives, running in both directions. Take the audit seriously enough and you will eventually meet a step you cannot justify and cannot afford to remove. Refuse the audit and you keep every error your grandparents made, forever, at compounding interest. My own rule, having finally checked this one, is to ask what a step costs before asking what it proves. Cracking your knuckles costs nothing and threatens nothing, so the warning against it can go in the bin this afternoon, and Unger paid for that certainty so the rest of us would not have to. Boiling corn in lime costs an hour and a pot of ash, and anyone proposing to drop it should be made to explain the entire chemistry first, out loud, in front of witnesses. Pellagra is what the second conversation looks like when nobody bothers to hold it.

Which returns me to the corridor. I have had this backward for thirty years, thinking of the man as somebody who recited a saying. He was reporting. He had spent decades running an experiment on the only subject available to him, in weather that will kill an unacclimated person, and he had arrived by feel at a real physiological mechanism years before a laboratory in Ottawa confirmed it with nine cyclists in a dry room. His mistake was geographic. He brought a good rule north into a wet harbor town where the warranty runs out, and he was still closer to the physiology than I was, standing there with my iced coffee and my expression.

There is a second thing I had backward, and it is less comfortable. I did not weigh his claim on the merits, because I had already filed the source. He was the man who fixed the boiler, and I was a young man with a college degree and a cold drink, and the filing happened before any thinking did. That is the quiet engine under most of what passes for skepticism. We price the claimant long before we get anywhere near the claim. A commission with letterhead got pellagra wrong for years while a step in a Mexican kitchen had the answer built in and no letterhead at all.

I never asked him his name. I never asked who taught him. Standing in front of a man holding sixty years of evidence, I said nothing: I was young, he was building staff, and the cold cup in my hand had already told me I knew. What I do now, in August, is buy a coffee at the deli, stand in the shade until the first sweat comes up, and then walk. It works less well than he promised and considerably better than I expected. The cup, at least, has stopped sweating all over my hand.