One question holds two decisions: what goes on your skin, and what stays behind in the water.
Two products. Two yardsticks. One shared number.
By midsummer the beach bag holds both. The bottle carries a number everyone knows; the long-sleeved swim shirt carries another one almost nobody reads. Both were made for the same day of sun, and their protection is certified by entirely different methods — a difference that runs through the whole question.
Sunscreen has the outcome of a single randomised trial behind it; the swim shirt has a physically measured transmission. Their marine halves part company too, because one product leaves its load at the shoreline and the other at the factory and in the washing machine.
So this page places three measurements side by side. It shows how much of the protection the real-world dose actually delivers, maps how much surface the fabric takes over, and follows the routes by which what washes off us ends up in the water.
The two products are certified on two different yardsticks, and it is worth naming them at the outset. For sunscreen the yardstick is disease: people are followed for years and the count of new skin cancers is compared. For the swim shirt the yardstick is the material: a fabric sample is measured for the fraction of incident ultraviolet radiation that passes through it. The first shows what happened to people, the second what happens to a piece of fabric at the moment of measurement.
On the disease yardstick there is one randomised human trial, and only one. Sunscreen is otherwise an intensively researched subject, but the remaining studies are observational: they record what people used of their own accord and look for associations afterwards. In Nambour, Queensland, 1,621 adults were randomised in 1992 to daily or discretionary application. Nobody in the discretionary group was denied protection: they carried on using whatever sunscreen they normally used, and the trial's original question concerned the commoner keratinocyte cancers — whether sunscreen affects melanoma at all was still contested as late as 2011. Ten years after the trial ended the daily group had 11 new melanomas against 22 in the discretionary group. This is the source of the claim quoted everywhere, that regular sunscreen use roughly halves the risk of melanoma. The estimate is a hazard ratio of 0.50 — half the melanoma risk in the daily group. That is a single central value: the uncertainty range attached to it, the band within which the true value most likely lies, runs from 0.24 to 1.02, allowing anything from an effect stronger than halving to no effect at all. The p-value is 0.051. That means: if there were in truth no difference between daily and discretionary use, a difference as large as the one measured here, or larger, would appear in the sample with a probability of about 5.1 per cent. It sits a hair above the conventional 0.05 threshold, but the threshold alone does not settle whether the effect is real. Together with the wide uncertainty range it shows a promising result that does not confidently rule out chance variation. For invasive melanoma taken separately the difference is firmer. Nothing of comparable scale has repeated the trial since, and probably nothing will: assigning people to "carry on as you were" for several years is hard to justify now that the protective effect is broadly accepted.
On the materials yardstick the number is precise. UPF states the fraction of incident ultraviolet radiation that passes through the fabric: at UPF 50 it is one fiftieth, so the material holds back 98 per cent. The Australian and New Zealand standard specifies that this is measured on dry, unstretched, new fabric. The number therefore states exactly what the material does under the conditions of the test, and says nothing about how many melanomas are avoided by wearing it, because no such trial has been run on a garment.
The two numbers thus share no scale, and that is where the difficulty of the question comes from. For sunscreen we know what it achieved in people, with wide uncertainty; for fabric we know how much it holds back, with no human outcome evidence. For anyone looking for a single ranking, the choice of yardstick settles in advance which product wins. Two quantities remain that can be measured the same way for both: how much skin the protection reaches, and how long it stays there. The next two chapters measure exactly those.
Gambichler and colleagues measured 236 commercial summer garments, and 78 of them — a third — came in below UPF 15. UPF 15 means that roughly seven per cent of the radiation passes through the fabric, and below that value more still. Clothing protection is therefore item-dependent, and a large share of ordinary summer wear lets a meaningful amount of ultraviolet radiation reach the skin.
SPF is established in the laboratory with two milligrams of product per square centimetre. That layer is thicker than what most people apply: measurements put the real dose between 0.39 and 1.0 mg/cm², a quarter to a half of the certified amount. The number on the label describes a carefully weighed state, and your skin is in a different one.
Sunscreen research is divided on how much protection is lost when less product goes on, and the point at issue can be stated exactly: what relationship ties SPF to the applied amount. One calculation is proportional, so half the amount gives roughly half the protection and an SPF 50 product leaves about 25. The other is exponential, so half the amount gives the square root of the label and the same product leaves about 7. For one bottle and one layer thickness that is a three- to fourfold difference, which is why it matters which one the measurements support. The slider below draws both curves at once.
German ring tests found that at 0.5 mg/cm² the measured SPF is about a quarter of the labelled value and at 1.0 mg/cm² about half, while in a US study SPF 70 and SPF 100 products delivered 19 and 27 at half the dose. Both series support the proportional calculation. The exponential relationship spread from a 2010 measurement by Kim and colleagues on fifteen volunteers, and larger multi-laboratory ring tests have not confirmed it since.
The practical conclusion is the same on either curve: at the usual dose the protection lands between a quarter and a half of the label, so applying generously and reapplying early is a simple way to bring real protection closer to the expected value.
In the water, time is added to dose. Under the US FDA monograph a "water resistant" claim may only be made for 40 or 80 minutes, the test runs in twenty-minute immersion cycles, and the mandatory label text calls for reapplication immediately after towel drying. Anyone snorkelling for two hours is beyond the product's certified range.
The rule of nines is an estimating method from burn care: it divides the body surface into units of nine per cent, so the head is 9, each arm 9, each leg 18, and the trunk 18 at the front and 18 at the back. From this the trunk comes to roughly 36 per cent and the two arms to another 18, so a long-sleeved swim shirt hands about half the skin surface over to fabric. Precisely the half that is hardest to cream evenly and from which the product washes off most easily in the water.
The swim shirt's weak points sit elsewhere than a cotton tee's. For cotton the critical factor is wetness: the fibre takes up water and swells, the gaps between yarns open, and the already modest UPF of about 5 in a white cotton tee falls further when wet, by as much as half in published measurements. A synthetic knit largely avoids this, because polyester and polyamide yarns absorb almost no water and the geometry of the knit holds together. Even so, a one-way rule is unwarranted: Gambichler and colleagues measured 69 summer fabrics saturated with tap water and with salt water, and both the direction and the size of the effect differed between fabrics. Wetness is therefore a per-garment question, and the way to get a number for it is a UV Standard 801 certificate, which also measures the wet and stretched condition.
The second weakness is stretch, and this one belongs to the swim shirt. The garment is an elastane blend cut to sit close to the body, so it is under tension whenever it is worn: the fabric thins, the yarns move apart, and more radiation reaches the same patch of skin. Moon and Pailthorpe measured this on elastane fabrics and Kimlin and colleagues on a common garment fabric under stretch, and both found a substantial drop in UPF. Two practical consequences follow: a tighter size protects less well than one that merely sits against the body, and the label value does not measure this state at all, because it is taken on an unstretched sample.
The Australian and New Zealand AS/NZS 4399 and the US AATCC 183 both measure dry, unstretched, new fabric. The European UV Standard 801 also takes the value in worn, wet and stretched condition. The label number on a garment designed for water may therefore come from a state the garment is never in while it is used — which is a good reason to look for a certificate that covers the worn condition.
Ageing is silent as well. The bottle carries an expiry date and a reapplication rule; the shirt has neither, while chlorine, salt, washing and sunlight slowly work on the elastane and the finish.
The evidence on corals comes from the laboratory, and it can be stated precisely. In 2016 Downs and colleagues exposed coral planulae and cultured cells from seven coral species to oxybenzone and reported developmental deformity, DNA damage and death. In 2022 a Stanford group supplied the mechanism: corals and sea anemones convert the molecule into a glucoside that turns phototoxic in sunlight, and a bleached coral is more exposed to it because the symbiotic algae are what otherwise lock the toxin away. The bans in Hawaii, Palau, the US Virgin Islands and Bonaire cite these two works.
Measurements taken on actual reefs, however, are scarce, and this is where the chain weakens. The critical review by Mitchelmore and colleagues and the 2022 report of the US National Academies name the same two gaps: there are no long series of concentrations measured on real reefs, and there is no framework that sets the effect thresholds found in the laboratory against the amounts actually present in the water. What is known: concentrations at tourist sites sit in the lower microgram-per-litre range, and warming remains the dominant driver of reef decline, so the filters arrive alongside a far larger load. The story of the substitutes shows the same uncertainty: octocrylene degrades into benzophenone during storage, a finding met by two published rebuttals and an authors' response in the same issue, while the European consumer safety committee considers the ingredient safe up to ten per cent.
For each item, guess where it enters the environment.
A Central European swimmer needs one more translation. The bans come from reef jurisdictions, whereas Lake Balaton and the Adriatic are different systems, and far less freshwater data exists. In Swiss lakes the filters also arrived through wastewater treatment; a Queensland reservoir study measured around 20 nanograms per litre in summer, below local guideline values; and the risk quotient reported for benzophenone-3 spans two orders of magnitude. The question is open, and saying so is part of the answer.
The laboratory evidence is strong and the field risk assessment is incomplete, and those two things belong in separate columns. This page refrains from claiming that UV filters are harmless to aquatic life, and equally from claiming that the fate of the reefs turns on sunscreen.
One gap deserves to be stated: we found no measurement of how much microfibre a swim shirt releases directly into the water while swimming. The microfibre literature covers laundry, so the textile side is quantifiable at the washing machine and remains an estimate at sea.
Of the four activities below, pick the one that matters most to you right now. The suggestion starts from the fact that fabric covers surface while cream covers the remainder, and that time in the water consumes the cream's certified range.
Pick the activity that best describes you.
In the end the comparison converges on a single number. A long-sleeved swim shirt takes over roughly half of the body surface, holds that protection as the hours pass, and that same half is the part from which nothing washes into the water. On the remaining half the dose and the rhythm of reapplication decide the outcome, so apply generously and repeat early.
On fabric, surface is what counts. On cream, the dose and the clock.
Interference topic · Skin, sea and two yardsticks of evidence
Which precaution do we accept without evidence when it is convenient, and which one do we interrogate when it is not?