One food.
For how long?

As long as the first missing nutrient allows. With no vitamin C at all, the first slight gum and skin changes appeared from about day 38–60 in the 1969 Iowa trial, and the full clinical syndrome at 84–97 days in the 1971 Iowa trial; Sheffield volunteers showed their first signs at 17 weeks, with gum changes after 26 weeks (one man’s gums as early as 23 weeks) and skin bleeding from 26–34 weeks. On polished rice, beriberi appeared within four to six weeks at Changi and from day 87 in Fraser and Stanton’s trial. On Goldberger’s mainly cereal prison diet, pellagra came at about five months. The body’s vitamin B12 store—about 1–3 mg, much of it in the liver—lasts roughly 3–6 years with no intake.

What people lived through is just as well recorded: a year on fresh fatty meat with organs, 9–11 months on potatoes with margarine, 18 months on bread and potatoes with gruel, and potatoes with buttermilk as the staple of Irish labourers and cottiers, the poorest third of the country, from the 1780s until the blight of 1845. Most trials were stopped once disease was underway. Deaths are on record—25 sailors on the Ryūjō, many living almost entirely on the free white-rice issue, and a volunteer on day 77 of the 1912 Bilibid polished-rice trial—but no one was followed on a single food to a timed death. So each horizon is when disease began, not a survival time.

Yes—because they were not eating one food. The Royal Navy ration included biscuit, salt meat, peas, oatmeal, butter, cheese and beer, and ships took on fresh food in port. Scurvy came when that failed: under six weeks on one of Lind’s cruises, never in three months on another. Anson lost more than 1,300 of about 1,850 men in 1740–44 to scurvy, fever, dysentery and starvation together; his storeship, which found fresh greens, lost none.

Once lemon juice became standard issue in 1795, scurvy admissions at Haslar naval hospital fell from 1,457 in 1780 to one in 1806, and crews served for years. Cook’s Resolution sailed for three years with sauerkraut, wort and fresh greens and lost no man to scurvy. Whalers stayed out three to four years by resupplying in port.

On grain, mostly—not bread alone. Harvest workers’ rations put grain at up to about 80% of calories, but as bread, ale and pottage, with cheese, milk, fish and some meat. That is why bread-based diets lasted: something else supplied the vitamin C.

The strict version has been tried. William Stark lived 44 days on bread and water in 1769—with a little meat and wine on one day—without scurvy. Prisoners at Millbank on bread, gruel and a broth with a little beef and vegetables but no potatoes were first diagnosed with it six to seven months in. At Devizes, bread and potatoes with gruel kept prisoners free of scurvy for up to 18 months.

Arithmetic: food per day = energy target ÷ energy per 100 g, up to 6.4 kg (the heaviest daily intake of one staple on record, Irish labourers). Intake = grams × the USDA value. Protein is adjusted for its limiting amino acid; niacin counts tryptophan, and maize niacin is discounted because most of it is bound.

Thresholds: each of 19 needs has an average requirement (EAR), an RDA, and a deficiency floor drawn from the record: 10 mg vitamin C, 0.25 mg thiamin per 1,000 kcal, 4.4 mg niacin per 1,000 kcal, 0.6 mg riboflavin and 100 µg folate, where disease was recorded; and slower floors for B12, vitamin A, calcium, iron and zinc. The protein floor, 15 g of usable protein, is set just under the ~20–25 g on which people are recorded living for years, because no faster protein disease was timed in adults. A lean-meat diet—protein above 40% of energy with carbohydrate under 10%—counts as too lean; preformed vitamin A above 7,500 µg and selenium above 913 µg a day count as poisoning.

Order: failures are ranked by how quickly they appeared in people. The first is the horizon.

A healthy, nonpregnant man aged 19–50 who weighs 70 kg, eating 2,000 kcal a day unless you change it. Requirements are U.S. Dietary Reference Intakes for that man; where they differ within 19–50 (only magnesium), the higher 31–50 values are used. Changing calories changes how much food is eaten, not the requirements.

Not modeled: children, pregnancy, illness, starting stores, sunlight, individual absorption, food safety and contaminants. This is an educational explorer, not medical advice.

Dietary Reference Intake tables (Institute of Medicine; hosted by USDA) ↗

Each food is one USDA Standard Reference record for the preparation shown. The ten headline nutrients were checked field by field against SR28; the other 30 fields (B vitamins, minerals, fatty acids and amino acids) come from the same records in FoodData Central SR Legacy. Wholemeal bread and cheddar were added from that release.

One value is corrected from history: USDA lists no vitamin C in whole milk, but assays of pasteurized market milk found about 5.8 mg a litre (1946), which is used here. Where USDA has no value, it stays unknown.

Explore official USDA food data ↗

Eat the edible part in the preparation on the card—raw apples, baked potatoes, cooked rice—with no added salt, oil, seasoning, supplements or fortification. Bread and cheese contain the salt that is part of them. Butter is unsalted; milk is unfortified. Cook foods that need cooking, such as kidney beans and meat.

Dry oats and dry chia are weighed before cooking or soaking water.