The Bortle scale rates a sky from class 1 to class 9 by what can be seen in it
John Bortle, a comet observer, published the scale in the February 2001 issue of Sky & Telescope. He judged that the faintest star an observer can see, the naked-eye limiting magnitude, varies too much from one pair of eyes to the next to rate a sky on its own, so each class is described by sights: how far the zodiacal light reaches, how much detail the Milky Way shows, whether the galaxy M33 can be seen by eye, how bright clouds look. Class 1 is the darkest sky on Earth; class 9 is a downtown sky. The naked-eye and 32 cm telescope limits below are Bortle's; the SQM ranges are later third-party values, not his.
| Class | Name | Naked-eye limit | 32 cm telescope limit | Approx. SQM | What you see |
|---|---|---|---|---|---|
| 1 | Excellent dark-sky site | 7.6 to 8.0 | 17.5 | 21.75 to 22.0 | Zodiacal light, gegenschein and zodiacal band all visible; M33 an obvious naked-eye object; the Scorpius and Sagittarius Milky Way casts faint shadows; airglow visible |
| 2 | Typical truly dark site | 7.1 to 7.5 | 16 to 17 | 21.6 to 21.75 | Zodiacal light yellowish and bright enough to cast shadows at dusk and dawn; clouds show only as dark holes in the stars; summer Milky Way highly structured; M33 easy by eye |
| 3 | Rural sky | 6.6 to 7.0 | 16 | 21.3 to 21.6 | Some light pollution on the horizon; clouds lit near the horizon, dark overhead; Milky Way still complex; M4, M5, M15 and M22 visible by eye; M33 with averted vision |
| 4 | Rural/suburban transition | 6.1 to 6.5 | 15.5 | 20.3 to 21.3 | Light domes in several directions; zodiacal light reaches less than halfway to the zenith; Milky Way impressive overhead but without detail; M33 hard even with averted vision |
| 5 | Suburban sky | 5.6 to 6.0 | 14.5 to 15 | 19.25 to 20.3 | Zodiacal light faint at best; Milky Way weak near the horizon and washed out overhead; clouds brighter than the sky |
| 6 | Bright suburban sky | 5.5 | 14.0 to 14.5 | 18.5 to 19.25 | No zodiacal light; the sky within 35° of the horizon glows grayish white; Milky Way only near the zenith; M33 invisible, M31 modest |
| 7 | Suburban/urban transition | 5.0 | 14 | 18.0 to 18.5 | Whole sky light gray; Milky Way invisible; M31 and the Beehive barely glimpsed; through a telescope the brightest Messier objects are pale |
| 8 | City sky | 4.5 | 13 | below 18.0 | Sky whitish gray or orange, bright enough to read headlines; some constellation stars missing; M31 and the Beehive only for an experienced observer |
| 9 | Inner-city sky | 4.0 or less | - | below 18.0 | Many constellation stars invisible; beyond the Pleiades no Messier object shows by eye; the Moon, planets and a few bright clusters remain |
M33, the Triangulum Galaxy, is the scale's main test object. It is a large, faint glow, and diffuse objects are harder to see than stars of the same magnitude, so many observers find Bortle's descriptions of the darkest classes optimistic.
Sky brightness is measured in magnitudes per square arcsecond, and a higher number is a darker sky
A reading of 21 magnitudes per square arcsecond means each square arcsecond of sky glows as brightly as a star of magnitude 21. Because the magnitude scale runs backward, a larger number is a darker sky, and every 2.5 magnitudes is a factor of 10 in brightness. The natural moonless sky at the zenith, lit by airglow, zodiacal light and starlight, reads about 22.0. The same reading can be given as a luminance in candelas per square meter, which is what lighting engineers use: luminance = 108,000 × 10-0.4 × reading. This table converts readings to luminance and to how many times brighter than the natural sky they are:
| SQM, mag/arcsec² | Luminance, mcd/m² | Times the natural sky | Share that is artificial | Approx. Bortle class |
|---|---|---|---|---|
| 22.00 | 0.171 | 1.0 | none | 1 |
| 21.75 | 0.215 | 1.3 | 21% | 1 or 2 |
| 21.50 | 0.271 | 1.6 | 37% | 3 |
| 21.00 | 0.430 | 2.5 | 60% | 4 |
| 20.50 | 0.681 | 4.0 | 75% | 4 |
| 20.00 | 1.08 | 6.3 | 84% | 5 |
| 19.50 | 1.71 | 10 | 90% | 5 |
| 19.00 | 2.71 | 16 | 94% | 6 |
| 18.50 | 4.30 | 25 | 96% | 6 or 7 |
| 18.00 | 6.81 | 40 | 97% | 7 to 9 |
| 17.50 | 10.8 | 63 | 98% | 8 or 9 |
| 17.00 | 17.1 | 100 | 99% | 8 or 9 |
| 16.00 | 43.0 | 251 | over 99% | 8 or 9 |
The artificial share is the part of the sky's glow that comes from lights on the ground, taking the natural sky as 22.0. A reading of 21.0 is already 2.5 times the natural sky, with three-fifths of its glow artificial.
Classes 1 to 3 differ by less than double in sky brightness; classes 4 to 7 cover a 20-fold range
The classes are not equal steps. On the usual SQM ranges, the whole span from class 1 to the bright edge of class 3 is a sky 1.9 times the natural sky. Class 4 alone spans a full magnitude, and from the dark edge of class 4 to the bright edge of class 7 the sky brightens 21-fold. Small changes matter most at the dark end, where the eye is working at its limit: going from class 3 to class 2 means a sky only slightly darker, but the zodiacal light, the gegenschein and the Milky Way's dust lanes come out of it. At the bright end, a few magnitudes of extra glow change little, because only bright objects are left to see.
Galaxies and nebulae are seen by surface brightness, not catalog magnitude, and large galaxies behave about a magnitude fainter than listed
The magnitude listed for a galaxy or nebula is its integrated magnitude: all its light added up, as if it came from a single star. A star puts that light into a point; a galaxy spreads it over an area that can be larger than the full Moon. What the eye responds to is its surface brightness, the light per square arcsecond, in the same units as a Sky Quality Meter reading. For an object of magnitude m covering A square arcseconds, surface brightness = m + 2.5 log10(A). Averaged over their listed sizes, which are ellipses here, these Messier objects compare with the sky like this. Surface brightness is in magnitudes per square arcsecond; the last two columns give each object's brightness as a multiple of the sky behind it, for a natural sky reading 22.0 and a class 5 suburban sky reading 20.0.
| Object | Mag. | Size | Surface brightness | Vs. a 22.0 sky | Vs. a 20.0 sky |
|---|---|---|---|---|---|
| M31 Andromeda Galaxy | 3.4 | 3° x 1° | 22.1 | 0.90x | 0.14x |
| M33 Triangulum Galaxy | 5.8 | 71' x 42' | 23.1 | 0.36x | 0.06x |
| M101 Pinwheel Galaxy | 7.9 | 28.8' x 27' | 23.8 | 0.20x | 0.03x |
| M81 Bode's Galaxy | 6.9 | 27' x 14' | 22.0 | 1.0x | 0.16x |
| M51 Whirlpool Galaxy | 8.4 | 11' x 7' | 21.7 | 1.3x | 0.20x |
| M42 Orion Nebula | 4.0 | 65' x 60' | 21.6 | 1.4x | 0.23x |
| M27 Dumbbell Nebula | 7.4 | 8' x 5.6' | 20.2 | 5.5x | 0.87x |
| M1 Crab Nebula | 8.4 | 6' x 4' | 20.5 | 4.1x | 0.64x |
| M57 Ring Nebula | 8.8 | 1.5' x 1' | 17.9 | 45x | 7.1x |
Published surface brightnesses for the same galaxy often differ by a magnitude or more, because catalogs measure a galaxy's size to different faint limits; these are computed from the sizes on this site's Messier pages.
The large galaxies are no brighter per square arcsecond than the natural sky itself. M33 averages about 23.1 and M101 about 23.8, so even under a pristine sky they show only as a faint brightening of the background, and in a class 5 suburban sky M33 adds just 6% to the sky glow behind it. That is why galaxies are the first things light pollution takes away: every magnitude the sky brightens cuts their contrast against it by 2.5 times. A small, bright planetary nebula like the Ring Nebula, M57, is the opposite case: at about 18 magnitudes per square arcsecond it is about 45 times brighter than a natural sky and still several times brighter than a suburban one, which is why it shows from a city.
The rule of thumb that a diffuse object behaves about one magnitude fainter than its catalog value is a fair guide for large galaxies seen by eye. The extreme naked-eye limit is about magnitude 8 for stars but about 6.9 for galaxies. Andrew Crumey's 2014 model of human contrast vision puts M33, catalog magnitude 5.8, at an effective visual magnitude of about 6.6: you need to see stars to about 6.6 before you can see the galaxy. The correction is not a fixed number, though. It grows with size and faintness, so the biggest, faintest galaxies, such as M33 and M101, are harder than one magnitude suggests, and compact objects such as planetary nebulae, small bright galaxies like M82 and the cores of galaxies behave much like stars.
A telescope cannot make a surface brighter than it looks to the eye; it gathers more light but spreads it over a larger image. It does make the object larger, and the eye detects a large faint patch more easily than a small one, which is why a galaxy often shows better at moderate to high power than at the lowest. For emission nebulae, a nebula filter such as a UHC or OIII filter passes the nebula's light and blocks much of the sky glow, raising the contrast; it does nothing for galaxies, which shine by starlight.
A full Moon brightens even a class 1 site to about the level of a class 7 suburb
Moonlight scattered in the air swamps everything else. At a dark site, the zenith reads about 18.0 to 18.5 under a full Moon, close to the class 7 range, though the moonlit sky is blue rather than gray or orange. Bortle's classes assume a moonless sky, fully dark, with the Sun 18° below the horizon. The natural sky also varies on its own:
- The solar cycle: airglow brightens the natural sky by 0.4 to 0.6 magnitudes near the peak of the solar cycle, as in 2024 and 2025.
- The Milky Way: with the bright part of the band overhead, a pristine sky reads about 20 rather than 22.
- The zodiacal light: it brightens the sky near the ecliptic, most of all in the first and last hours of the night.
- Altitude in the sky: the sky is brighter toward the horizon, where you look through more air; readings are taken at the zenith.
A Sky Quality Meter gives a reading in seconds; a star count gives a class
- Sky Quality Meter: Unihedron's handheld meter, the SQM, measures the zenith sky and reports magnitudes per square arcsecond. Hold it overhead, away from any light, take several readings and keep the darkest. Readings from different models, and from phone apps that imitate them, are not exactly comparable.
- Naked-eye limiting magnitude: count the stars you can see inside a known pattern, such as the Little Dipper or the Square of Pegasus, and compare with a chart of star magnitudes. Let your eyes adapt for 20 to 30 minutes first.
- Globe at Night: the citizen-science program run by the National Science Foundation's NOIRLab asks observers to match the sky to a set of star charts of a constellation; the results go into a worldwide map.
- Light pollution maps: maps built from satellite images of lights at night, such as the World Atlas of Artificial Night Sky Brightness, model the sky everywhere. They are a good first guide to where to go, but a map class is a model; a measurement at the site settles it.
The darkest skies on Earth all read close to 22.0, and what sets them apart is distance from cities, dry air and protection
No site can be darker than the natural sky, so the best places on Earth all measure within a few hundredths of 22.0 on a moonless night. What separates them is how far the nearest city's light dome is, how dry and clear the air is, how high the site sits, and whether the land around it is protected from new lighting.
- United States: measurements by the National Park Service's night sky team show Big Bend National Park in West Texas has the darkest skies of any national park in the lower 48 states. It sits inside the Greater Big Bend International Dark Sky Reserve, which spans the Texas-Mexico border. Great Basin National Park in Nevada, Death Valley in California and Nevada, the Colorado Plateau parks of Utah and Arizona, and central Idaho are close behind. In the East, Cherry Springs State Park in Pennsylvania and the North Woods of Maine are the darkest.
- Chile: the Atacama Desert, with the world's driest air, is home to the large telescopes of the European Southern Observatory and the Rubin Observatory.
- Namibia: the NamibRand Nature Reserve, one of the darkest places on Earth that visitors can reach.
- New Zealand: the Aoraki Mackenzie International Dark Sky Reserve on the South Island.
- China and India: the Tibetan Plateau, with the Lenghu observatory site in Qinghai and the Ngari dark sky reserve in Tibet, and the Hanle dark sky reserve in Ladakh, India.
- Ocean islands: La Palma in the Canary Islands and Mauna Kea in Hawaii, where observatories sit above the clouds.
For places to go, see the dark sky places list and the Bortle class 1 sites.
Night skies are brightening by about 10% a year, doubling in less than a decade
The World Atlas of Artificial Night Sky Brightness, published in 2016, found that more than 80% of the world's people, and more than 99% of people in the United States and Europe, live under light-polluted skies, and that the Milky Way is hidden from more than a third of humanity, including nearly 80% of North Americans. A 2023 study in Science, by Christopher Kyba and colleagues, used more than 50,000 Globe at Night star counts from 2011 to 2022 to measure the change from the ground. Skies brightened by 9.6% a year on average, 10.4% in North America and 6.5% in Europe, far faster than the roughly 2% a year seen by satellites. Part of the gap is the switch to white LED lighting, whose blue light the dark-adapted eye is sensitive to but satellite sensors largely miss. At that rate, a site with 250 visible stars would see only 100 after 18 years.
Each range of the scale has its own targets
- Classes 1 to 3: faint galaxies, dark nebulae, the zodiacal light and gegenschein, and the Milky Way in full detail.
- Classes 4 and 5: the Milky Way overhead, the brighter galaxies and nebulae, and the Messier objects.
- Classes 6 and 7: star clusters, small bright planetary nebulae and double stars, the targets of the Astronomical League's Urban observing list.
- Classes 8 and 9: the Moon, the planets, double stars and the brightest clusters, which light pollution barely affects.