Eight glasses. Two litres. Half your body weight in ounces. Three competing rules, none of them from the same source, all of them quoted with total confidence. I followed the second one for a year, spent most of a flight looking for a bathroom, and eventually looked up where the number came from.
So let’s do the boring, useful version: what the actual recommendations are, where the myth came from, and how to tell whether you personally need more than your thirst says.
Where “eight glasses a day” came from
It’s a misreading. In 1945 the US National Research Council’s Food and Nutrition Board wrote that adults need about 1 ml of water per calorie of food – and then added the sentence everyone forgets: “The quantity is largely contained in prepared foods.” The 8×8 rule (eight 8-ounce glasses) dropped the second half and kept the first.
The current US reference intake from the National Academy of Medicine is about 3.7 litres/day for men and 2.7 litres/day for women – total water, meaning drinks plus the water in food. Roughly 20% of that arrives from food, not the tap.
What counts as fluid
Everything you drink counts, including coffee and tea. Habitual caffeine drinkers build a tolerance to its mild diuretic effect, so a morning coffee is mostly just… water with a better flavor. Alcohol is the exception – it increases fluid loss – and very high caffeine intake on top of a hot day can matter.
| Food | Water content |
|---|---|
| Cucumber | 96% |
| Zucchini / courgette | 94% |
| Tomato | 94% |
| Watermelon | 92% |
| Strawberries | 91% |
| Oranges | 88% |
| Plain yogurt | 85% |
| Cottage cheese | 80% |
| Cooked rice / pasta | ~60% |
| Bread | ~36% |
This is why a fit 70 kg person eating normally and drinking to thirst can be well short of “3.7 litres” and still be perfectly hydrated.
So how much should you actually drink?
- At rest, temperate climate: drink to thirst. In healthy adults the thirst mechanism tracks need closely – by the time you feel thirsty you’re about 1-2% down on body mass, which is still before performance suffers.
- During exercise: roughly 150-350 ml per 20 minutes, more in heat. Longer or harder sessions need electrolytes, not just water.
- Hot weather: sweat rates of 0.5-1.5 litres/hour are normal; pre-cool hydration matters more than mid-session gulping.
- Pregnancy: add about 300 ml/day; breastfeeding: about 700 ml/day – the biggest single reason a new mother feels chronically dehydrated.
- Over 60: the thirst signal blunts with age. Older adults need scheduled drinking, not thirst-based drinking.
- Kidney or heart conditions: the ceiling matters more than the floor – follow the fluid limit your clinician sets, not an internet rule.
How to check instead of guessing
Two methods that actually work:
- Morning body weight. Weigh yourself on waking before drinking. Losing more than ~1% across a day means you’re under-drinking; in hot training blocks I drink to close a 0.5 kg gap before a morning walk, which is the same logic at a smaller scale.
- Urine color. Pale straw is the target. Dark yellow suggests you’re behind – but ignore the reading if you take B-complex or riboflavin, which turn urine fluorescent yellow no matter how much you drink.
What doesn’t work: counting glasses while ignoring body size, climate and activity. A 55 kg office worker in air conditioning and a 95 kg courier in July are not on the same prescription.
What about the fancy water products
Three products get asked about constantly, and all three fail the same test – they cost money to change something your kidneys already handle.
Alkaline water. Drinking it raises stomach pH transiently, not blood pH. Blood pH is held between 7.35 and 7.45 by the lungs and kidneys, and food or water doesn’t move it in healthy people. The health claims attached to it are mostly in-vitro or industry-funded.
“Structured” or “hydrogen” water. No reproducible mechanism with human performance data behind it at retail prices. The hydrogen studies use generated water at concentrations far above bottled product.
Electrolyte powders. These do have a real use – sessions over roughly 60-90 minutes, hot environments, or sauna and heat exposure – because sweat sodium losses run 400-1,200 mg per litre. For a 25-minute office walk, plain water is correct and the powder is sugar plus marketing.
Why the 1-2% threshold matters more than the glass count
The reason hydration gets studied in athletes first is that the effects show up there: body-mass losses of 2% or more measurably reduce endurance performance, raise perceived effort, and degrade concentration and mood. The same threshold explains why a long hot commute leaves you unable to focus that evening – not dehydration as a disease, but a small deficit that costs you something.
That’s the practical reframe. Stop asking “am I drinking enough” in the abstract and start asking whether you’re losing more than about 1% of body mass across a day. That’s a two-minute check with a bathroom scale, and it answers the question for every body size, climate and activity level – no rule of thumb required.
Can you drink too much?
Yes, and it’s not just a joke. Endurance athletes who drink past thirst have developed hyponatraemia – blood sodium diluted below safe levels – which is why marathon guidance shifted from “drink as much as you can” to “drink to thirst plus known sweat losses.” For normal daily life the risk is trivial; for sessions over 90 minutes it’s real. The other practical cost of over-drinking is nocturia: getting up at night, which then wrecks the sleep you were trying to protect (see what muscle recovery actually requires for how sleep and training interact).
The protocol I actually use
500 ml on waking, then a bottle within reach while I work – usually two more by evening – plus whatever tea or coffee I want. In summer I add 300-500 ml around any walk longer than 30 minutes, and I stop large volumes two hours before bed. That’s it. No app, no tracking spreadsheet.
Two adjacent questions people get wrong the same way: what happens to your body when you walk 30 minutes a day, and how much caffeine is too much – the second one matters here because caffeine tolerance changes how much of your fluid intake actually counts.
Evidence note: The 3.7 L / 2.7 L figures are the National Academy of Medicine (US) adequate intakes for total water, and the 1945 Food and Nutrition Board wording – including the “largely contained in prepared foods” clause – is the documented origin of the 8×8 rule. Individual needs shift with body mass, climate, activity and medical conditions, so use the checks above rather than a fixed glass count.