In a study published in 1981, three dermatologists at the University of Pennsylvania gave 32 volunteers a single dose of light from a xenon lamp built to imitate sunlight. The lamp lit a circle of skin one centimeter across on the inside of each person's upper arm, and another on the outside of the forearm. Each circle got five times that person's burn threshold, the smallest dose of ultraviolet that leaves skin just visibly red a day later. Researchers call it the minimal erythema dose, erythema being the medical word for redness. Peter Wilson, Kays Kaidbey, and Albert Kligman then looked at the circles twice a week for eight weeks. 1
Fifteen of the volunteers always burned and never tanned. They were mostly of Celtic descent, "characterized by pale skin, blue or green eyes, blond or red hair and numerous blotchy freckles on sun exposed areas." The other 17 rarely burned and tanned easily. Four weeks after the single dose all 15 of the pale volunteers still had a red circle, against 1 of the other 17, and in 5 of the pale volunteers the redness was still there when the study ended at eight weeks. 1
None of the 32 blistered. Yet 73 percent of the pale volunteers said they had blistered from a sunburn in natural sunlight. 1
In 2024, 35.1 percent of American adults, an estimated 88.1 million people, said they had been sunburned at least once in the previous year, close to the 34.2 percent of 2005. Most of what is known about how a burn runs its course, what a tan is worth, and what years of sun do to skin comes from lamps shone on patches a few centimeters across, on the arms, backs, and buttocks of a few dozen volunteers, most of them pale. In those experiments a heavy burn on fair skin stayed red for weeks, and one dose of ultraviolet switched on enzymes in the skin that break down collagen. A tan built over a whole summer of real sun protected about as well as a sunscreen with a sun protection factor, or SPF, of 2. Where the same people were tested under a lamp and in the sun, as on a Sydney hospital roof in 2001, the sun took about a third more ultraviolet, measured on the same meter, to redden them. 2; 3; 1; 4; 5; 6
This page covers where the skin types came from, how long a burn takes and how long it lasts, what people mean by sun poisoning, what has been tested as treatment, how a tan forms and what it is worth, why dark skin burns less, and what sunlight does to collagen as skin ages. What happens inside a sunburn in its first three days, and how burns and tanning beds relate to melanoma, are on the sunburn and melanoma page. Each study is named as real sun, a lamp, a product put on people, food or pills eaten, cells, or a measurement in blood.
Key facts
- In 2024, 35.1 percent of American adults said they had been sunburned in the past year, and 55.1 percent of those said their most recent burn happened despite using sunscreen. 2
- After one lamp dose of five burn thresholds, all 15 volunteers of skin type I still had red skin at four weeks, against 1 of 17 of skin type IV. 1
- A tan built over three and a half months of Wisconsin summer sun protected 21 people against a lamp burn by a factor of 2.33. 5
- A whole-body lamp dose of four burn thresholds raised six men's body temperature by an average of 1.75 degrees Celsius at 12 hours. 7
- In real sun in Seoul, skin under black cloth that passed only heat still raised the gene activity of a collagen-cutting enzyme by 205 percent, against 1,085 percent in bare skin. 8
Doctors promoted ultraviolet light in the 1920s, as the suntan became fashionable
In the 1920s the discovery that ultraviolet light makes vitamin D in the skin "ushered in a period of enormous popularity for ultraviolet light exposure," in the words of Albert and Ostheimer, who traced the history in 2003. Sunlight and ultraviolet lamps were promoted as preventive medicine, and home sunlamps "were aggressively marketed to the public." A suntan, which had had "limited popularity" before, "now was viewed as de rigueur in the United States and Europe." The authors wrote that the part doctors played in starting the habit of sunbathing "is not commonly appreciated today." 9
Across the 1920s and 1930s, women's swimsuits went from baring an estimated 18 percent of the skin to 47 percent, and men's from 23 to 47. A team writing in 2014 made those estimates from bathing suits in Sears catalogs, using the rule of nines, the chart burn surgeons use to estimate how much of a body a burn covers. In 1930, its first full year, Jones Beach State Park outside New York had 1.5 million visitors. 10
The story most often told is that Coco Chanel started the fashion in 1923, when she fell asleep in the sun on a yacht on the French Riviera and came home burned and then brown. The Skin Cancer Foundation still tells it, beginning "Rumor has it." The 2014 history attributes to her, in a 1929 issue of Vogue, the line "the 1929 girl must be tanned." Writing in History Today in 2008, the historian Robert Mighall called the yacht story "the classic urban myth." 11; 10; 12
The skin types came from a 1972 sunscreen test and predict burn thresholds only roughly
In 1972 the Boston dermatologist Thomas Fitzpatrick ran an outdoor test of Australian, European, and American sunscreens on paid volunteers in Brisbane. His colleague Madhu Pathak recalled that Fitzpatrick "did not hesitate to comment to the Australian Press reporters on the poor photoprotection results of Australian sunscreens." From the way the fair-skinned volunteers reacted to midday sun, Fitzpatrick sorted them into three types: those who burn easily and never tan, those who burn easily and tan with difficulty, and those who burn moderately and tan moderately. Back in Boston the types helped set lamp doses for psoriasis patients, and they grew into the six-type scale described on the sunburn and melanoma page. 13
Between September 1987 and April 1988, 54 volunteers in Amsterdam, typed by what they said about how their skin burns and tans, had their forearms tested under a solar simulator, a lamp filtered to give a spectrum close to sunlight's. Their burn thresholds ran from 350 to 1,527.5 joules per square meter, and only types I and VI had ranges that did not overlap. "The skin type did not correspond well to the constitutional skin color," the authors wrote. In Porto Alegre, Brazil, 193 volunteers were tested with a lamp of UVB, the short-wave ultraviolet that does most of the burning, and by burn threshold they fell into two groups, types I to IV and types V and VI, with heavy overlap between neighboring types inside each. 14; 15
In 2004, 404 women in Beijing, Guangzhou, Shanghai, and Chengdu were typed twice, once by an interviewer and once on a form they filled in themselves, in a study sponsored by P&G (China). By interview 14.2 percent were type IV, and on the form 1 percent were. In Manipal, India, 100 healthy volunteers were typed by the color of their unexposed backs, because "none of our patients gave a history of sunburn." When a Copenhagen group compared skin type with a measured pigment factor in 84 volunteers in 2010, the measured factor predicted burn and tan thresholds better, and the authors noted that skin type, for types I to IV, has "only 4 grades." 16; 17; 18
Midday sun reddened pale skin in 20 to 60 minutes, and a strong lamp in 90 seconds
Albert Kligman's group in Philadelphia compared their xenon lamp with the sun in a 1972 study. Their volunteers' average burn threshold under the lamp was about 90 seconds, which they equated with about 40 minutes of midday summer sun at Philadelphia's latitude. Fifteen lamp doses, they wrote, held about ten hours of summer sun at its noonday peak, and added: "This illustration is fanciful, for the sun does not stand still." In Baltimore in 1989, a lamp matched to the ultraviolet of summer midday sun at 39 degrees north reddened 19 volunteers at doses equal to 30 to 60 minutes of that sun. 19; 20
In 1978 researchers at Schering-Plough in Memphis, the maker of Coppertone, tested ten fair-skinned volunteers under a xenon lamp and again in natural sunlight, to see whether the lamp could stand in for the sun in sunscreen tests. Scaled to each test's highest value, the burn thresholds averaged 0.67 under the lamp and 0.62 in the sun, and only two volunteers ranked differently between the two. 21
In March 2001, 31 pale volunteers in Sydney had squares of their lower backs lit by a solar simulator and, later the same day, by the sun, "just after midday on the roof of the Department of Dermatology." The same UVB meter measured both doses. The volunteers reddened at an average of 68.7 millijoules per square centimeter of UVB from the lamp and 93.6 from the sun, about a third more, and the lamp threshold was lower on each of the four test days, though on the one windy day the gap was too small to be sure of. Skin color measured beforehand predicted each person's lamp threshold and not their sun threshold. Two of the authors worked at the Shiseido Basic Research Center in Yokohama. 6
In May 2013 a Danish surgical team laid 23 volunteers face down in Kusadasi, Turkey, from 1:22 to 2:02 in the afternoon, under a UV index of 9, on the scale weather services use to report the day's sunburning ultraviolet. They had chosen 40 minutes because earlier tests showed it reddened most people at that index, and still "many did not have a reaction." In Indonesia, 48 volunteers of skin types III and IV reddened after 18 to 23 minutes of sun, sooner at higher altitude. The World Health Organization's guide to the UV index advises against publishing burn times, because "people tend to interpret burn times to mean that there is a safe level of unprotected sun exposure." 22; 23; 24
A heavy burn on fair skin stayed red for weeks, and hot and tender for days
In the Philadelphia volunteers the redness was strongest at 24 hours. In the pale volunteers it stayed at its peak for an average of five days before it began to fade, and it lasted an average of 45 to 49 days, against 16 to 18 days in the volunteers who tan. The burn threshold was lower on the inner upper arm, which rarely sees sun, than on the forearm, yet once each site had five times its own threshold the redness ran the same course on both. The authors had "no explanation" for why it lasted so long in type I skin. 1
In a study published in 2017, a Guangzhou group gave 26 Chinese women one, one and a half, and two times their burn threshold from a solar simulator and measured the skin with meters for 168 days. Redness peaked within three days. At one burn threshold it returned to its starting reading at day 140, and at the two larger doses it had not returned by day 168. The meter's reading of melanin, the pigment of skin, peaked at day 14 and had not returned to its starting level at any dose by the end. 25
A Copenhagen group put a contact thermometer on lamp burns in 1998 and found barely visible redness 0.7 degrees Celsius warmer than untouched skin, bright red skin 3.5 degrees warmer, and one burn 9.5 degrees warmer. The rise was the same in 32 skin cancer patients as in 36 healthy people. The patches were less than a square centimeter across, and the authors thought larger burns would run hotter. 26
Harrison, Young, and McMahon measured the tenderness in London in 2004. Eighteen volunteers who burn easily took one, two, and three times their burn threshold from a solar simulator on patches of buttock skin, and the burned skin did not hurt unless something touched it. At three times the threshold, the temperature at which heat became painful fell by an average of 6.2 degrees from 45.8, "only a few degrees above normal skin temperature." A day after the dose, the pressure needed to cause pain had dropped from 95.2 grams to 26.7, and brushing the skin was enough to hurt most volunteers. AstraZeneca paid for the study, and the authors offered the burn as a model for testing painkillers. 27
Sun poisoning is not a diagnosis, and a whole-body lamp burn raised fever within 12 hours
The Cleveland Clinic uses the name sun poisoning for a severe sunburn with chills, fever, headache, or nausea, which a family doctor there says often mimics a flu or an allergic reaction, and it notes that the name "isn't a formal medical term." DermNet, a dermatology reference, lists sun poisoning as another name for polymorphic light eruption, a rash that appears hours to two days after sun in spring and early summer. The American diagnostic codes have entries for sunburn of the first, second, and third degree, and none for sun poisoning. 28; 29; 30
In Vienna in 1990, six men stood under an unfiltered bank of six blue-light lamps for less than two minutes, a whole-body dose of four times each man's burn threshold. Six hours later all six had "freezing, fatigue, burning sensations." At 12 hours their body temperature was up by an average of 1.75 degrees Celsius, and interleukin-6, a messenger that raises fever, had risen in their blood from 2.5 to 10.9 nanograms per milliliter. C-reactive protein, a marker of inflammation made by the liver, peaked at 48 hours, when the fever was already falling. The study was paid for by a science fund of the mayor of Vienna and by the fund of Austria Tabakwerke, the Austrian tobacco company. 7
In Turku, Finland, in 2005, a UVB lamp dose of three burn thresholds on a single 10 by 10 centimeter patch of back raised C-reactive protein by 42 percent a day later, on a high-sensitivity test, and it rose in 69 percent of the 16 volunteers. In volunteers who had first stood in a whole-body cabin of simulated sun, below the burn dose, every day for 10 or 30 days, it did not rise. 31
Britain's National Health Service (NHS) tells people to get urgent help after sun if they feel "hot, cold or shivery," dizzy or sick, or have a headache or muscle cramps, and warns that "severe sunburn can lead to heat exhaustion and heatstroke." The Merck Manual says fever, chills, weakness, and shock can follow when a large area of the body burns, "similar to those of a thermal burn." 32; 33
Laboratory burns rarely blistered, and burned skin peeled one to two weeks later
None of the 32 Philadelphia volunteers blistered from five burn thresholds, none of the 18 London volunteers from three, and 1 of 24 volunteers in a 2002 trial in Nice blistered from three, on the seventh day. The Merck Manual advises treating blistered sunburn like other partial-thickness burns, and the NHS says not to pop the blisters. We found no study that counted blisters after a burn in real sun. 1; 27; 34; 33; 32
In Philadelphia the burned skin peeled between one and two weeks after the dose, "especially between 10 and 14 days." In the volunteers who tan, the peeling left a pale center that stayed pale to the end of the study. The skin cells that die after a burn, called sunburn cells, move up through the skin as it renews. In twelve London volunteers given twice their burn threshold from a solar simulator, in a study published in 2000, sunburn cells were most common in the bottom layer of the epidermis, the outer layer of the skin, at 18 hours, in the middle at 24 hours, and in the upper layers at 36 hours. By 72 hours most had reached the surface layers. 1; 35
Kligman's group counted cells copying their DNA, a sign of cells about to divide, in the bottom layer of burned skin in three volunteers, all prisoners, in 1972: 621, 503, and 815 per 1,000 cells, against 60, 46, and 54 in their unburned skin. In the Guangzhou women, the water content of the skin fell to its lowest on day 7 at every dose. 19; 25
A Yokohama group collected the peeling flakes themselves in 1996, lifting them off with cellophane tape. After twice the burn threshold of UVB on the back, the ratio of an anti-inflammatory messenger to an inflammatory one in the flakes rose from about 5 to over 50 at 11 to 15 days, close to the level in the always-exposed skin of the face. The Merck Manual notes that "exfoliated skin can remain extremely sensitive to sunlight for several weeks," and the NHS advises against scratching or pulling off peeling skin. We found no study that measured why sunburned skin peels. 36; 33; 32
Steroid creams put on after a burn did little, and foods eaten before one helped only in small trials
In 1977 Malcolm Lane-Brown, a doctor at the Travelodge Medical Centre in Sydney, reviewed sunburn treatment for the journal Drugs. He began with the observation that Celtic skin "often burns readily in the sun, tans poorly and freckles," which "might best be summarised by saying 'If you're Irish, stay in the parlour'!" Of steroid creams he wrote, "No proper trials have yet been done." 37
In Nice in 2002, 24 volunteers had three patches of back burned with three times their threshold from a filtered xenon lamp, and two patches were treated with steroid lotions from six hours later. By day 8 redness had fallen 35.8 and 34.9 percent under the two steroids and 32.8 percent with nothing, and symptoms peaked on day 3 whatever was done. Schering AG paid for the trial and employed one of the authors, who concluded that both lotions were "safe and effective in the treatment of sunburn." 34
In 2008 a Copenhagen trial on 20 volunteers tested moderate and strong steroid creams put on 6 or 23 hours after a UVB lamp dose. It was double-blind, meaning neither the volunteers nor the assessors knew which cream was which. The authors concluded that the creams do not provide "a clinically useful decrease in the acute sunburn reaction." Only the strong cream put on half an hour before the dose made a difference they judged clinically relevant. 38
The one treatment trial we found that used real sun was run by Novartis Consumer Health, which employed its first author. It deliberately burned 172 volunteers with 2.8 times their burn threshold of sun on 15 percent of the body, then gave two applications of a 0.1 percent diclofenac gel, an anti-inflammatory painkiller like ibuprofen, or the same gel without the drug. The diclofenac group reported less pain. 39
A 70 percent aloe cream did nothing to prevent or treat lamp burns, compared with a dummy cream, in 20 volunteers in Thailand in 2005. In Germany in 2008, a 97.5 percent aloe gel kept under a dressing for two days after a UVB lamp burn reduced redness at 48 hours in 40 volunteers, more than 1 percent hydrocortisone in the same gel and less than 1 percent hydrocortisone cream. 40; 41
Beta-carotene, 90 milligrams a day for 23 days, more than doubled the carotene in the skin of 16 Boston women with an average age of 66 and did not change their lamp burns; a Mary Kay Cosmetics fellowship helped pay for the work. Nine volunteers in Germany who ate 40 grams of tomato paste with olive oil every day reddened 40 percent less under a lamp after 10 weeks than 10 who took the oil alone, and no less after 4, in a trial with an author from the Unilever Health Institute. Twelve women who drank high-flavanol cocoa daily reddened 15 percent less at 6 weeks and 25 percent less at 12, in a trial supported by Mars, which supplied the cocoa. 42; 43; 44
In Manchester in 2015, 50 adults took green tea capsules or dummy capsules for three months, and their burn thresholds did not differ. In Brescia in 2019, 1 gram of cocoa a day for a week did nothing in 10 people, and 4 to 6 grams raised the burn threshold. In Frankfurt in 1998, 2 grams of vitamin E with 3 grams of vitamin C a day for 50 days raised the burn threshold of 10 volunteers from 103 to 183 millijoules per square centimeter, and either vitamin alone did not; the volunteers taking the vitamins "could ascertain that they were taking" them. 45; 46; 47
Pale skin has to burn before it tans
Within minutes of UVA, the longer-wave ultraviolet that makes up most of what reaches the ground, skin darkens as melanin already in it changes chemically. This first change, called immediate pigment darkening, fades fast, and its "biological function remains unknown," a 2008 review says. A darkening that lasts at least three to five days, persistent pigment darkening, follows within hours. New melanin appears over the next days and is greatest from about 10 days to three or four weeks, and this delayed tanning is what people mean by a tan. 48; 49
In Copenhagen in 2010, the lightest-skinned of 84 volunteers needed up to 1.9 times their burn threshold from a solar simulator or UVB lamp before any tan appeared. The ratio was about 1.5 in medium and dark skin, and in very brown skin a tan and a burn began at the same dose. In Sydney only 16 of the 31 pale volunteers tanned measurably from both the lamp and the sun, and in them the tan began above the burn threshold in both. The Philadelphia volunteers of type I showed faint pigment only at three to four weeks, and only 60 percent of them. 50; 6; 1
A Japanese group lit the backs of 16 men with three filtered beams from one xenon lamp in 1993. With UVB alone, or UVB and UVA together, the burn came at a lower dose than the tan. With visible light added to the beam, the tan came first. One of the authors worked at the cosmetics laboratory of Kanebo. 51
In 1996 Barbara Gilchrest's laboratory put small fragments of DNA, like the pieces cells cut out when they repair ultraviolet damage, on pigment cells that had never been lit and on guinea pig skin, and got "a pigment response indistinguishable from UV-induced tanning." In 1987 her group in Boston had lit melanocytes, the cells that make melanin, growing alone in dishes; cells from donors of different ancestry started with about tenfold differences in melanin and raised it by similar proportions. The next step in the signal, found at Dana-Farber in 2007, is on the sunburn and melanoma page. 52; 53
A Beiersdorf team in Hamburg built matching tans on five volunteers in 2008 with simulated sun, UVB, or UVA, and measured the melanin in the roofs of small blisters raised over each patch. Under simulated sun eumelanin, the brown-black kind, rose 52 percent and pheomelanin, the red-yellow kind, 73 percent. Under UVA, patches tanned to the same visible depth held no more of either melanin than untanned skin. 54
A summer's tan protected about as well as SPF 2
In August, Derek Cripps of the University of Wisconsin tested 21 volunteers of skin types II to IV "who had deliberately sought sun exposure" through three and a half months of Wisconsin summer. Under a lamp, the tanned middle of the back needed on average 2.33 times the dose of untanned buttock skin to redden, and the figure hardly changed with skin type: 2.4 for type II, 2.45 for type III, 2.1 for type IV. The comparison set a tanned back against an untanned buttock, two different sites, and five people given UVB lamp courses over four weeks reached 8.01. The National Institutes of Health paid for the study. 5
Tans built with lamps over two weeks in London came out near 2 as well, as the sunburn and melanoma page describes. The World Health Organization's UV index guide says a dark tan on white skin gives protection "equivalent to an SPF of about 4," and Cancer Council Australia says "SPF3." 24; 55
In Copenhagen in 2004, two coats of a 20 percent dihydroxyacetone cream, the browning agent in self-tanners, gave the forearms of 10 volunteers a protection factor of 3.0 against a solar simulator a day later, 2.0 at day 5, and 1.7 at day 7. The Danish company Matas paid for part of the study. 56
In 1981 Robert Sayre's group found that in summer, people who spent time outdoors resisted burning more than people who stayed in, that in winter the two groups were the same, and that the benefit "can be lost within a month or two." Over six summer weeks in Uppsala, six volunteers took 24 to 98 standard erythema doses of real sun on the forearm, a standard erythema dose being a fixed amount of reddening ultraviolet. In the same volunteers, a lamp dose left thymine dimers, a common kind of ultraviolet DNA damage, in 17 percent of the cells of their habitually sunned forearms, against 32 percent in their never-sunned buttocks. In a Unilever-funded Copenhagen study of lamp doses given four days in a row, dark-skinned volunteers needed their daily dose cut by 40 to 50 percent by the fourth day to avoid a burn, while "for the most fair-skinned persons essentially no reduction in the daily UV dose was needed." 57; 58; 59
Dark skin took several times the dose to burn, and visible light alone darkened it
In Cripps's Wisconsin tests, the burn threshold of untanned skin under a lamp rose with skin type. Taking type I as 1, type II scored 1.67, type III 2.5, types IV and V nearly 4, and type VI 9.68, a figure that rests on two men. 5
In Utrecht in 2004, six white volunteers took twice their burn threshold from a solar simulator, and six Black volunteers took the highest of those doses, less than one burn threshold for them. The white skin showed DNA damage down into the dermis, the layer beneath the epidermis, along with incoming white blood cells and active tissue-dissolving enzymes. "Except for DNA damage in the supra basal epidermis, none of these changes was found in black skinned subjects," the authors wrote. 60
In 1997 a Copenhagen group tested 73 white adults under a solar simulator on never-sunned buttock skin. Pigment predicted the burn threshold closely, and the stratum corneum, the dead outer layer of skin, gave about 11 percent of the protection, 24 percent in the palest person and 5 percent in the darkest. A year earlier another Copenhagen study, of 14 people with vitiligo, a condition in which patches of skin lose their pigment, had put the outer layer's share at 57 percent in the normally pigmented skin beside the white patches. 61; 62
In the 2024 American survey, 9.1 percent of Black adults said they had been sunburned in the past year, and so had 37.8 percent of the Black adults the survey counted as sun-sensitive. On black or brown skin, the NHS says, "you may not notice a change in the colour of your skin." The Copenhagen thermometer study offered skin temperature as a way to grade burns in skin types V and VI, "where visual grading of erythema is difficult or impossible." 2; 32; 26
At Henry Ford Hospital in Detroit in 2010, a visible-light lamp on the lower backs of 20 volunteers of skin types IV to VI gave pigment "darker and more sustained" than long-wave UVA did, and skin type II showed no pigment at all. At Johns Hopkins in Baltimore in 2023, four daily doses of visible light, each about what 40 minutes of sunlight delivers, tanned 18 volunteers of types V and VI and left 13 of types I and II with "no discernable pigmentation and erythema." 63; 64
In 1978 Branda and Eaton in Minneapolis found that light at 360 nanometers destroyed 30 to 50 percent of the folate, a B vitamin needed for cell growth, in samples of human plasma within an hour. Ten pale patients having psoralen and UVA treatment had lower blood folate than 64 healthy people. Branda and Eaton proposed that dark skin evolved near the equator partly to protect folate. 65
In a 2001 study of people with and without skin cancer, carrying one or two variants of MC1R, a gene that helps decide which melanin skin makes, went with 3 and 11 times the odds of childhood freckles, and the authors put 60 percent of freckles in the population down to those variants. Sun spots, called solar lentigines, came with age: in 523 French women in 2013 they went with age, dark skin that tans well, a history of freckles, sun habits, and taking oral contraceptives or progestogens. The authors called solar lentigines "markers of photoaging," the aging of skin attributed to sun, "whereas freckles are mainly determined by genetic factors." 66; 67
One dose of ultraviolet switched on enzymes that break down collagen, and heat did too
In 1997 Fisher and colleagues lit the buttock skin of 59 white volunteers with fluorescent ultraviolet lamps, and one exposure raised three enzymes that cut collagen, the protein that gives the dermis its strength. The breakdown of collagen fibers rose 58 percent, and with four exposures two days apart, two of the enzymes stayed at 4.4 and 2.3 times their usual activity for a week. We found no study that followed people with measured sun doses for the decades that wrinkles take. 4
In Philadelphia in 1995, 23 volunteers took half their burn threshold from a lamp every weekday for 28 doses. At the same fraction of a burn, UVA thickened the skin, cleared out immune cells, and inflamed it more than simulated sunlight did, and an SPF 22 sunscreen worn under 11 burn thresholds a day did not prevent those changes. The study included authors from Procter and Gamble. 68
In 16 London volunteers in 2012, UVB and long-wave UVA doses that reddened skin equally raised matrix metalloproteinase 1, or MMP-1, a collagen-cutting enzyme, about equally. The authors, who included researchers from Beiersdorf, concluded that the small UVB share of sunlight, under 5 percent, "is likely to be the main cause of photoageing, at least in terms of MMP-1 expression." 69
In Seoul, between June and September, 16 volunteers of skin types III and IV lay with their buttocks bare to the sun between 11 in the morning and 2 in the afternoon, for 2 hours 15 minutes to 3 hours. One patch sat under a film that blocked ultraviolet and another under black cloth that blocked all the light and passed only heat. The bare skin warmed from 31.7 to 36.3 degrees Celsius, the filtered skin to 36.4, and the covered skin to 34.9. A day later the gene activity for MMP-1 was up 1,085 percent in bare skin, 233 percent under the filter, and 205 percent under the cloth, and the gene activity for new collagen had fallen to 14 percent of normal in bare skin and 56 percent under the filter. Only the bare skin drew in neutrophils, the white blood cells that arrive first at a burn, and the cosmetics company Amore-Pacific supported the study. 8
In Ankara, physiatrists, doctors of physical medicine, measured skin thickness by ultrasound in seven adults before and after a three-week seaside holiday. The upper dermis of the forehead thinned from 1.17 to 1.03 millimeters, and the skin at the navel thinned too. The study had no comparison group and no record of dose, and the authors called its results "preliminary." 70
The figure that the sun causes 80 percent of skin aging comes from one cosmetics study's grading scales
"An estimated 80 percent of visible skin aging in Caucasian women is caused by the sun," says the Skin Cancer Foundation. Its source is a 2013 study by researchers at L'Oréal Research and Innovation and Biotherm International, with a consultant and a dermatologist in private practice. They studied 298 women aged 30 to 78 in Montpellier, whom a dermatologist had sorted into 157 "sun-seeking" and 141 "sun-phobic" women, partly by grading the wrinkles, sagging, and pigment spots on their faces. Twelve experts then scored each face on 22 photographic scales, and the authors marked 18 of the scales as the signs most affected by ultraviolet. 71; 72
For each woman the authors added up her scores on the 18 marked scales and divided by the total on all 22. The average across all 298 women, sun avoiders included, was 80.3 percent, and the abstract states that "UV exposure seems to be responsible for 80% of visible facial aging signs." The same study found that pigment spots went strongly with sun, sagging went with age alone, and wrinkles went with both. The US Environmental Protection Agency (EPA) says "up to 90 percent of the visible skin changes commonly attributed to aging are caused by the sun," with no source given. 72; 73
At the Twins Day Festival in Twinsburg, Ohio, 186 pairs of identical twins filled in questionnaires and were photographed, and a panel judged how old each looked. More sun exposure went with an older look, more so with age, and so did smoking; hormone replacement went with a younger look, and a higher body mass index made twins under 40 look older and those over 40 younger. Among 1,826 Danish twins over 70, ten nurses estimated ages from photographs, and sun exposure predicted looking older in men but not in women, whose looks went with low body mass index and low social class. Both studies measured sun by what people remembered. A randomized trial of daily sunscreen and skin aging, in which chance decided who used sunscreen every day, ran in Nambour, Queensland, and is on the sunscreen page. 74; 75
About one American adult in three burns each year, the same share as in 2005
The share of American adults who told the National Health Interview Survey they had at least one sunburn in the past year was 34.2 percent in 2005, 37.1 in 2010, 34.1 in 2015, and 35.1 in 2024. The 2024 figure is an estimated 88.1 million people, 18.8 million of them burned four or more times. Young adults burned most, 46.0 percent of those aged 18 to 29, against 16.8 percent of those 65 and over. By group, 45.8 percent of white, 21.5 percent of Hispanic, 16.3 percent of Asian, and 9.1 percent of Black adults had burned. 3; 2
The 2024 survey also asked what people were doing when their most recent burn happened. Being in, on, or near water came first, at 60.6 percent, then exercising at 24.7, drinking alcohol at 17.6, deliberately tanning at 15.9, and working at 12.9. And 55.1 percent said their most recent sunburn happened "despite using sunscreen." Among American high school students in 2023, 54.7 percent had been sunburned in the past year, and 22.9 percent of those five times or more. 2; 76
Melbourne's summer surveys of 13,285 people from 1987 to 2017 tracked tanning through the SunSmart campaign, which began in the summer of 1988 to 1989. After adjustment, 43.1 percent said they did not like to get a suntan before the campaign, and 66.1 percent in the 2010s. Even before it, 82.3 percent disagreed that "A suntanned person is more healthy." The share burned on a summer weekend fell from 11.4 to 7.6 percent. 77
Misconceptions
What people believe
- A base tan protects you from sunburn.
- Dark skin does not sunburn.
- Sun poisoning is a kind of poisoning, or a diagnosis.
- The sun causes 80 or 90 percent of skin aging.
- A sunburn is over in a few days.
- Aloe or hydrocortisone cream heals a sunburn.
- UVB burns the skin and UVA ages it.
What the studies show
- A tan built over a Wisconsin summer protected against a lamp burn by a factor of 2.33, and UVA-built tans held no extra melanin. 5; 54
- In 2024, 9.1 percent of Black American adults reported a sunburn in the past year. Dark skin needed several times the lamp dose to burn. 2; 5
- Sun poisoning is not a medical term. It is used for a severe sunburn with fever or chills and for polymorphic light eruption, a rash. 28; 29
- The 80 percent is a ratio of grading scales in one study of 298 French women by L'Oréal researchers, and the EPA gives no source for its 90 percent. 72; 73
- After five burn thresholds from a lamp, fair skin stayed red for 45 to 49 days on average. 1
- Steroid creams put on 6 or 23 hours after a lamp burn gave no clinically useful benefit in a double-blind trial, and two aloe trials disagreed. 38; 40; 41
- At doses that reddened skin equally, UVB and long-wave UVA raised a collagen-cutting enzyme about equally, and repeated UVA below the burn dose changed skin more than simulated sunlight did. 69; 68
Questions people ask
What is sun poisoning?
Sun poisoning is a lay name, not a diagnosis. People use it for a severe sunburn with chills, fever, headache, or nausea, and for polymorphic light eruption, a rash that appears hours to two days after sun. In Vienna, a whole-body lamp burn raised six men's temperature by 1.75 degrees Celsius at 12 hours. 28; 29; 7
How long does a sunburn last?
In lamp studies, redness peaked within one to three days. After a heavy dose of five burn thresholds, fair skin stayed red for 45 to 49 days on average, and skin that tans for 16 to 18 days. 1; 25
How long does it take to get a sunburn?
It depends on the skin, the place, and the hour. A lamp matched to midday summer sun at 39 degrees north reddened 19 Baltimore volunteers at the equivalent of 30 to 60 minutes, and Indonesian volunteers of skin types III and IV reddened after 18 to 23 minutes of real sun. The World Health Organization advises against publishing burn times, because people read them as a safe amount of sun. 20; 23; 24
Why does sunburned skin peel?
Burned skin peeled one to two weeks after a lamp dose, as the dead sunburn cells rose to the surface and the cells beneath copied their DNA at more than ten times the usual rate. We found no study that measured the peeling itself. 1; 35; 19
Should you pop sunburn blisters?
The NHS says not to pop them, and the Merck Manual advises treating blistered sunburn like other partial-thickness burns. 32; 33
Does a tan protect you from sunburn?
A summer's tan in Wisconsin protected by a factor of 2.33, lamp tans in London by about 2, and a self-tanner by 3 for one day, falling to 1.7 after a week. 5; 56
Can Black people get sunburned?
Yes. In 2024, 9.1 percent of Black American adults said they had been sunburned in the past year. On dark skin the redness may not show. 2; 32
Does the sun age your skin?
One lamp dose raised enzymes that cut collagen, and real sun on the buttocks did the same, partly through heat. In twin studies, people who reported more sun looked older, though among Danish twins over 70 the link held for men and not for women. 4; 8; 74; 75
What helps a sunburn heal?
Steroid creams put on after a lamp burn gave no clinically useful benefit, aloe trials disagreed, and a diclofenac gel eased pain after a real-sun burn in a trial run by Novartis Consumer Health. The Merck Manual and the NHS advise cool water and painkillers such as ibuprofen. 38; 40; 41; 39; 33; 32
What we do not know
The Philadelphia authors had no explanation for why a heavy burn stayed red for about seven weeks in type I skin and under three in type IV, and the Guangzhou meters found pigment still above its starting level half a year after a single dose. We found no later study that explained either.
Almost everything about how a burn runs its course was measured on patches of a few square centimeters under lamps. We found no study that counted blisters after a burn in real sun, or that measured why burned skin peels. The one real-sun treatment trial was run by Novartis Consumer Health.
We found no long study of photoaging with measured doses. The twin studies asked people how much sun they had had, the Seoul study lasted one afternoon, and the Ankara ultrasound study had seven people and no comparison group. Which band of sunlight does most of the aging, UVB, UVA, or infrared heat, is still argued in the laboratory.
Visible light darkens dark skin and does little to pale skin. We found no study of what that pigment does for health.
The Philadelphia volunteers did not blister under five times their burn threshold, and 73 percent of the pale ones remembered blistering in the sun. None of the studies on this page tested why.
Key studies
- Wilson, Kaidbey, and Kligman, 1981, Philadelphia. One lamp dose of five burn thresholds left type I skin red for 45 to 49 days. 1
- Pathak, 2004. Fitzpatrick's 1972 Brisbane sunscreen test and the first three skin types. 13
- Westerhof and colleagues, 1990, Amsterdam. Skin type did not match measured skin color or burn threshold. 14
- Liu and colleagues, 2006, four Chinese cities. The same 404 women came out differently by interview and by form. 16
- Langner and Kligman, 1972, Philadelphia. A 90-second lamp burn equaled about 40 minutes of midday sun; burned skin copied DNA more than ten times as fast. 19
- Sayre and colleagues, 1978, Memphis. Lamp and sun burn thresholds ranked ten volunteers almost the same. 21
- Poon and colleagues, 2003, Sydney. The sun needed about a third more UVB than the lamp to redden 31 pale volunteers. 6
- Wan and colleagues, 2017, Guangzhou. Redness and pigment followed for 168 days after one dose. 25
- Harrison, Young, and McMahon, 2004, London. Burned skin hurt at a few degrees above skin temperature. 27
- Urbanski and colleagues, 1990, Vienna. A whole-body lamp burn raised fever and interleukin-6 at 12 hours. 7
- Laihia and colleagues, 2005, Turku. A small burn raised C-reactive protein; daily sub-burn doses beforehand prevented it. 31
- Sheehan and Young, 2000, London. Sunburn cells rose to the skin surface within 72 hours. 35
- Duteil and colleagues, 2002, Nice. Steroid lotions shortened a lamp burn little, in a Schering-funded trial. 34
- Faurschou and Wulf, 2008, Copenhagen. Steroid creams after a burn gave no clinically useful benefit. 38
- Magnette and colleagues, 2004. Diclofenac gel eased pain after a real-sun burn in 172 volunteers. 39
- Farrar and colleagues, 2015, Manchester. Green tea capsules did not change burn thresholds in 50 adults. 45
- Ravnbak, Philipsen, and Wulf, 2010, Copenhagen. Pale skin needed up to 1.9 burn thresholds to tan. 50
- Eller, Ostrom, and Gilchrest, 1996. DNA fragments alone started a tan. 52
- Wolber and colleagues, 2008, Hamburg. UVA tans held no extra melanin. 54
- Cripps, 1981, Madison. A summer's tan was worth a protection factor of 2.33. 5
- Faurschou and Wulf, 2004, Copenhagen. A self-tanner gave SPF 3 for a day. 56
- Rijken and colleagues, 2004, Utrecht. Black skin showed none of the burn chemistry at the same dose. 60
- Mahmoud and colleagues, 2010, Detroit. Visible light darkened skin types IV to VI and not type II. 63
- Bastiaens and colleagues, 2001. MC1R variants went with most freckles. 66
- Fisher and colleagues, 1997. One ultraviolet dose raised three collagen-cutting enzymes. 4
- Lavker and colleagues, 1995, Philadelphia. Repeated UVA below the burn dose changed skin more than simulated sunlight. 68
- Cho and colleagues, 2008, Seoul. Real sun, filtered sun, and heat alone all raised MMP-1. 8
- Flament and colleagues, 2013, Montpellier. The source of the 80 percent figure. 72
- Guyuron and colleagues, 2009, Twinsburg. Sun exposure went with an older look in 186 twin pairs. 74
- Holman and colleagues, 2026. One American adult in three sunburned in 2024. 2
Definitions
- Burn threshold. The smallest dose of ultraviolet that leaves a patch of a person's skin just visibly red a day later. Researchers call it the minimal erythema dose, or MED.
- C-reactive protein. A protein the liver makes during inflammation, measured in blood.
- Collagen. The protein fibers that give the dermis its strength.
- Delayed tanning. The tan people mean: new melanin that appears over days and is greatest from about 10 days to three or four weeks after sun.
- Dermis. The layer of skin beneath the epidermis, made mostly of collagen and elastic fibers.
- Dihydroxyacetone. The browning agent in self-tanners.
- Double-blind. A trial in which neither the volunteers nor the people assessing them know who got which treatment.
- Epidermis. The outer layer of the skin. Its surface is the stratum corneum, and its bottom layer holds dividing cells and melanocytes.
- Erythema. The medical word for redness of the skin.
- Eumelanin. The brown-black kind of melanin.
- Fitzpatrick skin type. A scale from type I, which always burns and never tans, to type VI, based on what people say about how their skin burns and tans.
- Freckle. A small pigmented spot that appears in childhood, mostly in people who carry MC1R variants. The medical term is ephelis.
- Immediate pigment darkening. A grayish darkening of melanin already in the skin within minutes of UVA, which fades within minutes to days.
- Interleukin-6. A messenger protein that raises body temperature during inflammation.
- MC1R. A gene that helps set which kind of melanin a person makes. Some variants go with red hair, freckles, and skin that burns easily.
- Melanin. The pigment of skin and hair. It absorbs ultraviolet.
- Melanocyte. The cell that makes melanin and passes it to the cells around it.
- MMP-1. A collagen-cutting enzyme, one of the matrix metalloproteinases, raised by ultraviolet and by heat.
- Neutrophil. The white blood cell that arrives first at an injury, including a burn.
- Persistent pigment darkening. A tan-to-brown darkening that begins within hours of UV, mostly UVA, and lasts at least three to five days.
- Pheomelanin. The red-yellow kind of melanin.
- Photoaging. The changes in skin, such as wrinkles and pigment spots, that are attributed to sun rather than to age alone.
- Polymorphic light eruption. A rash that appears hours to two days after sun, mostly in spring and early summer. It is one of the conditions people call sun poisoning.
- Psoralen. A plant chemical, found in citrus oils and some medicines, that makes skin more sensitive to UVA.
- Randomized trial. A trial in which chance decides who gets which treatment, so that the groups differ only in what the researchers gave them.
- Solar lentigo. A flat brown spot on sun-exposed skin that comes with age, often called a sun spot or age spot.
- Solar simulator. A lamp, usually a filtered xenon arc, built to give a spectrum close to sunlight's.
- Standard erythema dose. A fixed amount of reddening ultraviolet, 100 joules per square meter weighted by how strongly each wavelength reddens skin, used to report doses from meters.
- Stratum corneum. The dead outer layer of the epidermis, which peels after a burn.
- Sunburn cell. A skin cell dying after ultraviolet damage, seen under the microscope as a shrunken dark body.
- Thymine dimer. A common kind of ultraviolet DNA damage, in which two neighboring thymine letters are welded together.
- UV index. A scale of the day's sunburning ultraviolet, from 1 upward, with 11 and above called extreme.
- UVA. Ultraviolet from 315 to 400 nanometers, most of the ultraviolet reaching the ground. Long-wave UVA, or UVA1, runs from about 340 to 400.
- UVB. Ultraviolet from 280 to 315 nanometers. It does most of the burning.
- Visible light. Light from 400 to 700 nanometers, the part of sunlight people see.
- Vitiligo. A condition in which patches of skin lose their pigment.
Statistics
Four weeks after a solar simulator dose of five times their own burn threshold, all 15 skin type I volunteers in Philadelphia still had red skin, against 1 of 17 skin type IV volunteers. 1
In 5 of the 15 skin type I volunteers, redness from a single solar simulator dose of five burn thresholds was still present when observation stopped at 8 weeks. 1
After a solar simulator dose of five burn thresholds, the skin type I volunteers' redness stayed at peak intensity for an average of 5 days before it began to fade. 1
Across 54 Amsterdam volunteers of all six skin types, burn thresholds on the forearm under a xenon solar simulator ranged from 350 to 1,527.5 joules per square meter. 14
Of the 54 Amsterdam volunteers phototested under a solar simulator, only skin types I and VI had clearly different ranges of burn threshold; the ranges of the other types overlapped. 14
In never-sunned buttock skin of 73 Danish adults, the dead outer layer of the skin averaged 13.1 micrometers thick, with a range of 9.0 to 20.0. 61
In never-sunned buttock skin of London volunteers under a solar simulator, the burn threshold of skin type IV was about twice that of skin type I. 78
The burn thresholds of skin types I to IV overlapped most in the middle range of 2.9 to 4.0 joules per square centimeter of simulated sun. 78
In skin type I and II volunteers, the physical dose of long-wave UVA needed to redden skin was three orders of magnitude, about 1,000 times, greater than the dose of UVB at 300 nanometers. 78
In 12 skin type II volunteers given twice their burn dose, redness was still present at 72 hours after both simulated sun and long-wave UVA. 78
In 193 volunteers in Porto Alegre, Brazil, spread across all six skin types, the UVB lamp dose that just reddened unexposed skin ranged from 13 to 156 millijoules per square centimeter. 15
By burn threshold under a UVB lamp, the 193 Porto Alegre volunteers fell into two separate groups, skin types I to IV and skin types V and VI, with heavy overlap between adjacent types inside each group. 15
In 49 Copenhagen volunteers of skin types II to V given lamp or solar simulator doses on four days in a row, dark-skinned people needed their daily dose cut by 40 to 50 percent by day four to avoid a burn, while the fairest needed essentially no cut. 59
Interviewed about how their skin burns and tans in the sun, 71.4 percent of 404 women in four Chinese cities were skin type III, 14.7 percent type II and 14.2 percent type IV. 16
When the same 404 Chinese women filled in the sun-reaction skin-type questionnaire themselves, 74.3 percent came out type III, 25.6 percent type II and 1 percent type IV. 16
In the same 26 women, redness peaked within 3 days and faded completely within the 168 days only at the lowest dose, one burn threshold of simulated sun. 25
The burn thresholds of the 26 Guangzhou women under a solar simulator ranged from 29.2 to 112.6 millijoules per square centimeter. 25
After 6 weeks of Swedish summer sun, sunscreen cut the p53 damage signal in forearm skin by a mean 70 percent and blue denim by 74 percent, compared with bare skin. 58
After a lamp dose, 32 percent of cells in never-sunned buttock skin carried thymine dimers against 17 percent in the chronically sunned forearm skin of the same 6 Uppsala volunteers. 58
In a Boston placebo-controlled trial of 16 women with a mean age of 66, 90 milligrams of beta-carotene a day for 23 days more than doubled the carotene in their skin but did not lessen the redness from three burn doses of simulated sun. 42
Three burn doses of simulated sun produced a mean 20 sunburn cells after 3 weeks of a low-carotene diet and 19 six days after a single 120 milligram dose of beta-carotene, in 8 women each time. 42
After 10 weeks of eating 40 grams of tomato paste a day with olive oil, 9 German volunteers reddened 40 percent less under a solar simulator than 10 who took olive oil alone. 43
Women who drank high-flavanol cocoa every day reddened 15 percent less after 6 weeks and 25 percent less after 12 weeks from 1.25 burn doses of simulated sun, while a low-flavanol group did not change. 44
Of three melatonin creams tested in real sun on 23 volunteers, only the 12.5 percent cream reduced redness compared with placebo; the 0.5 and 2.5 percent creams did no better than placebo. 22
The Philadelphia volunteers' average burn threshold under the xenon lamp was about 90 seconds, which the authors equated with about 40 minutes of midday summer sun at Philadelphia's latitude. 19
When the burn thresholds of 10 fair-skinned volunteers were scaled to each test's highest value, they averaged 0.67 under a xenon solar simulator and 0.62 in natural sunlight, and only two volunteers ranked differently between the two tests. 21
Indonesian volunteers of skin types III and IV reddened after a mean 22 minutes 40 seconds of real sun at 0 to 300 meters above sea level, 20 minutes 34 seconds at 300 to 600 meters, and 18 minutes 14 seconds above 600 meters. 23
The mean real-sun burn dose of the Indonesian volunteers was 137.92 millijoules per square centimeter at 0 to 300 meters altitude, 131.87 at 300 to 600 meters and 121.64 at 600 to 900 meters. 23
Four hours after the solar simulator equivalent of an hour of midday summer sun, 9 Baltimore volunteers had lost a mean 80.2 percent of their thymine dimers, which the authors equate to a half-time of about 100 minutes. 20
Twenty Chinese women of skin types III and IV, tested in winter under a solar simulator, had burn thresholds of 25 to 40 millijoules per square centimeter, an average of 32.46. 79
In 10 healthy Italians, one week of 1 gram a day of high-flavanol cocoa did not change the redness from simulated sun, while 4 to 6 grams a day raised the visible burn threshold. 46
A day after a lamp dose, the share of skin cells showing the p53 damage signal rose from about 3 percent in unexposed skin to about 30 percent on the buttock and 33 percent on the forearm of 6 Uppsala volunteers. 58
Long-wave UVA also gave the London volunteers an immediate redness that lasted less than 3 hours, which the authors could not separate from the heat of the lamp. 78
After three burn doses of simulated sun, brushing the burned skin caused pain in most of the London volunteers. 27
Six Vienna volunteers given a single whole-body UVB lamp dose of four times their burn threshold all felt cold, tired and burning 6 hours later, and their body temperature was significantly raised at 12 hours. 7
The whole-body sunburn dose in the Vienna study came from a UVB lamp in under 120 seconds, and the volunteers' individual burn thresholds ranged from 190 to 450 joules per square meter. 7
C-reactive protein, a liver-made marker of inflammation, peaked 48 hours after the whole-body sunburn, when interleukin-6 and fever were already falling. 7
The tenderness of burned skin began within a few hours of the dose, peaked about 24 hours later and lasted through the full 72 hours of testing. 27
None of the 32 Philadelphia volunteers blistered from five burn thresholds of simulated sun, although 73 percent of the skin type I volunteers recalled a blistering sunburn from natural sunlight. 1
In the Philadelphia skin type IV volunteers, burned skin peeled between 1 and 2 weeks after a solar simulator dose of five burn thresholds, leaving a pale center with a red, darkened margin. 1
No patch of London volunteer skin blistered after up to three burn doses of simulated sun or UVA, and the only lasting mark was a tan. 27
Skin moisture in the 26 Guangzhou women fell to its lowest 7 days after simulated sun and was fully back by day 28 only at 1.5 times the burn dose. 25
On the simulated-sun sunburns of 24 skin type III volunteers in Nice, redness fell by 35.8 percent over 8 days under a methylprednisolone milk, 34.9 percent under a hydrocortisone emulsion, and 32.8 percent with no treatment. 34
Steroid-treated sunburn patches that healed completely did so on days 7 and 8, and untreated patches on day 8, in the 24-volunteer Nice trial. 34
Only one of the 24 Nice volunteers developed blisters after a severe simulated-sun burn, on day 7, and no burned patch wept. 34
In three Philadelphia prisoner volunteers sunburned under a xenon lamp, cells copying their DNA along 1,000 basal cells rose from 60, 46 and 54 in unexposed skin to 621, 503 and 815 in burned skin, more than tenfold. 19
In the dead outer skin of Japanese volunteers, the ratio of the anti-inflammatory IL-1 receptor antagonist to IL-1 alpha was 8 on the unexposed inner upper arm and over 100 on the sun-exposed face. 36
Twice the burn dose of UVB reddened the volunteers' backs from 6 to 48 hours after exposure, and pigmentation and peeling followed. 36
In most of the 24 Nice volunteers, the combined score for redness, swelling, burning and itching of a simulated-sun burn peaked on day 3, whether or not the patch was treated. 34
After a solar simulator dose of five burn thresholds, faint tanning appeared in the skin type I volunteers only at 3 to 4 weeks, and only in 60 percent of them. 1
Among the same 54 volunteers, the doses that started a tan overlapped even between skin types I and VI. 14
Among 84 Copenhagen volunteers of skin types I to V, the lightest-skinned needed up to 1.9 times their burn dose from a solar simulator or UVB lamp before any tan appeared. 50
In the same study, medium and dark-pigmented volunteers needed about 1.5 times their burn dose to start a tan, and very brown-pigmented volunteers tanned at exactly the dose that reddened them. 50
Only 16 of the 31 pale Sydney volunteers developed a measurable tan from both the solar simulator and real sunlight. 6
In 26 southern Chinese women followed for 168 days after one dose of simulated sun, the melanin reading peaked at day 14 and had not returned to baseline by day 168 at any dose. 25
Over 6 summer weeks of real sun in Sweden, each of 6 volunteers received 24 to 98 standard erythema doses, a mean of 48. 58
Six weeks of Swedish summer sun tanned all 6 volunteers' forearm skin both with and without sunscreen, while a patch under blue denim stayed clearly paler. 58
After 6 weeks of summer sun in Sweden, fewer than 1 percent of forearm skin cells carried thymine dimers, in both protected and unprotected skin. 58
A tan built from 6 or 12 UVA lamp sessions on the backs of 12 Scandinavian and 12 Indian or Pakistani volunteers took 5 to 6 months to fade, whatever the starting skin color. 80
A UVA tan reached its full depth 2 to 3 days later after 12 lamp sessions than after 6, and ended darker. 80
Two weeks of sub-burning doses raised eumelanin, the brown-black pigment, by 52 percent under simulated sun and 32 percent under UVB, and not at all under UVA, in blister roofs from 5 Hamburg volunteers. 54
Free 5-S-cysteinyldopa, a building block of red-yellow pigment, rose 4.1-fold in skin tanned with simulated sun and 3.7-fold with UVB, while UVA had little effect. 54
UVA lamps added the same absolute amount of pigment to fair and dark skin, so the percentage gain in color was larger in the fair-skinned volunteers. 80
On the forearms of 10 Copenhagen volunteers, two coats of a 20 percent dihydroxyacetone self-tanning cream gave a sun protection factor of 3.0 against simulated sun one day later, 2.0 at day 5 and 1.7 at day 7. 56
Twenty-four hours after twice the burn dose of simulated sun, messenger RNA for POMC, the parent protein of the tanning hormone, was up to 6.2 times its level in unexposed skin, and a broad-spectrum sunscreen held it at 0.5 times. 81
In 8 skin type II volunteers tanned daily under a solar simulator through a sunscreen containing the tan accelerator 5-methoxypsoralen, protection against DNA damage began after 3 to 5 daily exposures and peaked after 8. 82
After 10 daily exposures, the protection from a psoralen-boosted tan declined by about 5 percent a week and some still remained 14 weeks later. 82
Skin tanned through a 5 percent PABA solution with three doses of 15 burn thresholds each held 78, 62 and 106 DNA-copying cells per 1,000 basal cells, close to unexposed skin, against 621, 503 and 815 in sunburned skin. 19
After three and a half months of summer sun, the tanned skin of 21 Wisconsin subjects needed a mean 2.33 times the burn dose of their unexposed buttock skin, a natural SPF of 2.33. 5
A 365 nanometer UV lamp at 78 to 117 millijoules per square centimeter raised the same Gujarat volunteers' melanin reading by at most 7.79 units, with less consistent results than sunlight. 83
In 6 adults in Gujarat, India, 20, 35 and 50 minutes of direct sunlight raised the skin's melanin reading by a mean 37.78, 46.22 and 59.20 units by day 3. 83
Daily sub-burning whole-body doses of simulated sun for 10 or 30 days beforehand completely prevented that rise in C-reactive protein, though only the 30-day course lessened the redness. 31
After 10 daily sub-burning whole-body doses of simulated sun, the Turku volunteers' baseline C-reactive protein fell by 35 percent. 31
Over a climbing expedition on the north side of Everest, the melanin-hemoglobin reading of sun-exposed skin rose by a median 83.6 percent in 15 expedition members and 24.7 percent in 10 Sherpas. 84
Measured against constitutive pigment, the color of skin that never sees the sun, the dose to start a tan stayed at about 1.3 times the burn dose across the 84 Copenhagen volunteers. 50
Under a solar simulator, the dead outer skin layer supplied on average 11 percent of the 73 Danish volunteers' total protection against just-visible redness. 61
The dead outer skin layer gave 24 percent of the burn protection in the least pigmented of the 73 Danish volunteers and 5 percent in the most pigmented. 61
Pigment measured by skin reflectance correlated with the solar simulator dose needed for each grade of redness at coefficients of 0.78 to 0.89 in the 73 Danish volunteers. 61
Freckles and acquired moles, which the authors read as signs of repeated acute sun exposure, were found in 57 percent of the Porto Alegre skin type I volunteers, 26 percent of skin type II and 23 percent of skin type III. 15
In Utrecht, 6 white volunteers of skin types I to III were given 12,000 to 18,000 millijoules per square centimeter of simulated sun, twice their burn dose, while 6 Black volunteers of skin type VI were given 18,000, less than one burn dose for them. 60
Twice the burn dose of simulated sun caused DNA damage in epidermal and dermal cells, an influx of neutrophils and active tissue-dissolving enzymes in all 6 white Utrecht volunteers; in the 6 Black volunteers the only change found was DNA damage above the bottom layer of the epidermis. 60
After 4 minutes of simulated sun spanning ultraviolet to near infrared, the skin of 6 Berlin volunteers of skin types IV and V formed 60 percent of the free radicals formed in 6 volunteers of skin type II. 85
Under near-infrared light alone, the skin of the Berlin skin type IV and V volunteers formed significantly more free radicals than skin type II, while under ultraviolet it formed significantly fewer. 85
The same two weeks raised pheomelanin, the red-yellow pigment, by 73 percent under simulated sun and 23 percent under UVB, with no rise under UVA. 54
In the skin of 5 Hamburg volunteers of skin types II and III, eumelanin made up more than 90 percent of all melanin, both before and after a two-week course of lamp tanning. 54
On 10 Detroit volunteers of skin types IV to VI, 480 joules per square centimeter of visible light containing a trace of UVA1 reddened the skin at the two highest intensities, and the redness faded within 24 hours. 86
The visible-light dose of 480 joules per square centimeter in the Detroit study carried about 2.4 joules of UVA1, a dose the authors judged too small by itself to tan skin types IV to VI. 86
Measured by burn threshold under a xenon solar simulator and scored against skin type I as 1.0, skin type II had a built-in protection factor of 1.67, type III 2.5, types IV and V nearly 4, and type VI 9.68. 5
Measured skin color explained 70 percent of the variation in the simulated-sun dose that first produced sunburn cells in 42 skin samples. 87
Even at two and three times the simulated-sun dose that first made sunburn cells in dark skin, no pyrimidine dimers appeared in the bottom layer of the epidermis or in the dermis of three dark skin samples. 87
Four daily doses of 120 joules per square centimeter of visible light on the buttocks gave immediate darkening and a delayed tan in 18 Baltimore volunteers of skin types V and VI, and no visible pigment or redness in 13 of skin types I and II. 64
Cultured melanocytes from donors of different racial origin started with about tenfold differences in melanin but raised it by similar proportions after daily solar simulator doses. 53
On the backs of 23 young volunteers given 28 daily doses at half the burn threshold, UVA caused more thickening, lost immune cells, inflammation and lysozyme deposits on elastin fibers than simulated sun at the same fraction of a burn. 68
An SPF 22 sunscreen worn under 11 burn doses of simulated sun a day for 28 days did not prevent the cumulative skin changes, from thickening to lysozyme deposits on elastin, in the Philadelphia volunteers. 68
After twice the burn dose of simulated sun, total decorin, the protein that shields collagen fibers from cutting enzymes, fell by nearly 25 percent at 8 hours and 40 percent at 24 hours in the skin of Ann Arbor volunteers. 88
Fragments of broken decorin in the same irradiated skin rose nearly 11-fold at 8 hours and 7-fold at 24 hours. 88
One burn dose of UVA1 raised expression of MMP12, an enzyme that digests elastin, 15-fold at 6 hours and about 30-fold at 24 hours in London volunteers' skin. 89
Measured by quantitative PCR, UVA1 raised MMP12 messenger RNA 19.2 times more than an equally reddening dose of UVB at 6 hours. 89
At doses that reddened skin equally, UVB left about 3 times more thymine dimers in the dermis of London volunteers than UVA1 did. 69
After simulated sun on the skin of Ann Arbor volunteers, messenger RNA for CCN1, a protein that builds up in aged dermis, peaked at 10.5 times normal at 8 hours, and the protein itself stayed raised for at least 48 hours. 90
After one burn dose of simulated sun 5 days a week for 6 weeks on the buttocks of 12 volunteers, messenger RNA for the collagen-cutting enzyme MMP-2 rose up to 2.8-fold in unprotected skin, and a broad-spectrum daily cream prevented the change. 91
After one solar simulator dose of five burn thresholds, redness lasted an average 45.3 days on the inner upper arm and 48.7 days on the forearm of 15 skin type I volunteers, against 17.5 and 16.3 days in 17 skin type IV volunteers. 1
Bright red lamp burns on never-sunned buttock skin were on average 3.5 degrees Celsius warmer than the skin beside them, and the largest rise measured in 68 people was 9.5 degrees. 26
One ultraviolet lamp exposure increased the breakdown of type I collagen in volunteers' buttock skin by 58 percent. 4
Carrying one or two MC1R gene variants went with 3 and 11 times the odds of childhood freckles, and 1.5 and 2 times the odds of severe solar lentigines. 66
In the Montpellier study of 298 women, the 18 photographic scales the authors marked as sun-related made up an average 80.3 percent of each woman's total score across 22 aging scales. 72
The share of American adults reporting at least one sunburn in the past year was 34.2 percent in 2005, 37.1 percent in 2010 and 34.1 percent in 2015, in national household surveys of 25,159 to 31,399 adults. 3
In the 2024 National Health Interview Survey, an estimated 88.1 million American adults, 35.1 percent, reported at least one sunburn in the previous 12 months, and 18.8 million, 7.5 percent, reported four or more. 2
Among American adults sunburned in 2024, 60.6 percent said their most recent burn happened in, on or near water, 24.7 percent while exercising, 17.6 percent while drinking alcohol and 12.9 percent at work. 2
In 2023, 54.7 percent of American high school students reported at least one sunburn in the past year, and 22.9 percent of those reported five or more. 76
In Melbourne summer surveys, the adjusted share of people who did not like to get a suntan rose from 43.1 percent in 1987 to 1988 to 66.1 percent in the 2010s, and the share sunburned on a summer weekend fell from 11.4 to 7.6 percent. 77