Look up on a clear night from a dark place and ask three questions. Why is that star where it is, and why will it be somewhere else in two hours? Why do some stars shine and others barely show? And how would anyone know how far away they are?
People have asked these for thousands of years with no instrument at all. The Babylonians wrote down the risings of stars on clay tablets; Chinese astronomers kept star catalogues from at least the 4th century BCE; Polynesian navigators crossed the Pacific by memorised star paths. An astronomy student relearns this old, practical skill first. After it, a graduate never looks at the night sky as a ceiling again. They see a rotating sphere, tilted for their latitude, with a brightness scale and a ruler made of angles.
Treat the stars as painted on the inside of a huge sphere around you. The Earth spins once in about 23 hours 56 minutes, so the sphere seems to turn the other way around an axis pointing at the celestial poles. In the north, a moderately bright star, Polaris, sits within about a degree of the pole. Everything else circles it.
Two coordinate systems matter. Altitude and azimuth are local: how high above the horizon, and which way round from north. They change as the sky turns and differ for every observer. Right ascension and declination are fixed to the stars, like longitude and latitude painted on the sphere. Declination is the star's "latitude" on the sphere, in degrees from −90 to +90.
Two things follow without any more algebra. A star crosses its highest point at an altitude of 90° minus the gap between your latitude and its declination. And the pole star stands at an altitude equal to your latitude: stand at 40° north and Polaris is about 40° up. A star never sets for you if it lies within your co-latitude of the pole, and never rises if it lies the same distance around the opposite pole. That is why Crux, the Southern Cross, is never seen from Europe but never sets from Australia or South Africa.
Real J2000 declinations. Precession shifts positions by about 0.4° over a few decades, so Polaris's altitude wobbles within a degree of your latitude. Refraction and the star's colour are ignored.
Without JavaScript: at latitude 40° north, Polaris is about 40° up all night and never sets; Sirius peaks at about 33° in the south; Acrux of the Southern Cross never rises (it would peak 13° below the horizon).
A solar day (noon to noon) is 24 hours. A sidereal day (a star back on the meridian) is about 23 hours 56 minutes 4 seconds. The difference is the Earth's orbit: in a day it has moved about one degree around the Sun, so it must turn about one degree extra to face the Sun again. A star therefore rises about 3 minutes 56 seconds earlier each night, roughly two hours earlier each month, and the whole sky is back in the same place after a year. That is why Orion is a winter constellation in the north and Scorpius a summer one.
The same tilt gives us seasons. The Sun's declination swings between about −23.44° and +23.44° over the year (the tilt of Earth's axis). At noon the Sun is at altitude 90° minus the gap between your latitude and that declination, and day length follows. Same formula, different object.
Approximate formulas good to about a degree and a few minutes. Day length counts sunrise and sunset at the Sun's upper edge with standard refraction. The star hint compares right ascension only.
Without JavaScript: at 40° north on 21 June the Sun is 73° high at noon and the day is about 15 hours; on 21 December it is 27° high and the day is about 9 hours 20 minutes. At 70° north in June the Sun does not set.
The classical scale comes from Ptolemy's star catalogue (about 150 CE), usually traced to Hipparchus: the brightest stars are "first magnitude", the faintest naked-eye ones "sixth". So a bigger number means a dimmer star. In 1856 Norman Pogson fixed the scale mathematically so that five magnitudes are exactly a factor of 100 in light. One magnitude is therefore a factor of 1001/5, about 2.512.
The scale has gone beyond its first range. Sirius, the brightest night star, is at about −1.46. Venus can reach about −4.9, the full Moon is about −12.7 and the Sun about −26.7. At the faint end, the dimmest sources of Hubble's deepest images are near magnitude 30 or beyond. Your eye follows the same kind of log scale, which is why the scale was natural.
The faintest star you can see depends on two things: how dark your sky is, and how much light your instrument gathers. A telescope's aperture D collects light in proportion to D². Compared with a dark-adapted pupil of about 7 mm, that is a gain of 5 · log₁₀(D / 7 mm) magnitudes: about 4.3 for 50 mm binoculars and 7.3 for a 200 mm telescope, in the ideal case. Real gains are somewhat smaller.
| Object | Mag. | vs your limit | Seen? |
|---|
Sky limits from the Bortle scale (Bortle 2001), mid-range values. Magnitudes are typical or peak values. Spread-out objects such as galaxies and nebulae are harder than their magnitude suggests, because their light is spread over an area.
Without JavaScript: from a Bortle class 3 rural sky (limit about 6.8), 50 mm binoculars reach about 11.1, enough for Neptune (7.8). From a Bortle class 9 city sky (about 3.8) the naked eye sees only stars and planets brighter than about magnitude 4.
You cannot tell from the sky alone how big or how far anything is, only how big it looks: its angle. A degree is one of 360 in a circle; an arcminute is 1/60 of a degree; an arcsecond is 1/60 of that. Your outstretched fist covers about 10°, your little finger at arm's length about 1°, and a healthy eye can just separate two points an arcminute apart. For small angles, angle in radians = size / distance. That is the whole of the astronomer's ruler.
Famously, the Sun and Moon appear almost the same size, each about half a degree. It is a coincidence of this epoch, not a law: the Sun is about 400 times wider and about 400 times farther. Both distances vary, so the match does too. When the Moon looks bigger we get a total eclipse, and when it looks smaller a ring of Sun remains (an annular eclipse).
Moon diameter 3,474.8 km; Sun diameter 1,391,400 km (IAU 2015 nominal radius 695,700 km). The ranges are the typical extremes of the Moon's and Earth's orbits.
Without JavaScript: at its nearest the Moon spans 0.558°, at its farthest 0.490°. The Sun spans 0.542° at its nearest and 0.524° at its farthest. The Moon covers the Sun only when it is the larger.
The same trick gives distances to nearby stars. Look at a star in January and again in July: Earth has moved 300 million km, and a near star shifts a little against the faraway background. Half the shift is the parallax p. A star with a parallax of one arcsecond is at a distance of one parsec, 3.26 light-years.
The old argument against a moving Earth was that no parallax could be seen. It was right that none was visible; nobody could know how small it is. The nearest star, Proxima Centauri, has a parallax of 0.768 arcsecond, about 1/80 of what a good eye can resolve. Friedrich Bessel announced the first accepted stellar parallax, for 61 Cygni, in 1838: about a third of an arcsecond, a distance of around 10 light-years. The ESA satellite Gaia (launched 2013, observing until January 2025) measured parallaxes for about 1.5 billion stars, with an uncertainty near 0.02 to 0.03 milliarcseconds for stars brighter than magnitude 15.
The slider is logarithmic, from 4 to 100,000 light-years. The error uses 0.03 milliarcseconds, a round figure for bright stars (Gaia DR3 gives 0.02 to 0.03 for stars brighter than magnitude 15).
Without JavaScript: Proxima Centauri, 4.25 light-years, has a parallax of 768 milliarcseconds. At 1,000 light-years the parallax is 3.3 milliarcseconds, measured to about 1% by Gaia; at 10,000 light-years it is 0.33, measured to about 10%.
John Bortle's 2001 scale grades sky darkness from class 1, an excellent dark site, to class 9, an inner-city sky. A class 1 sky shows stars to about magnitude 7.6 to 8.0 and many Messier objects can be seen without a telescope; a class 9 sky hides all but magnitude 4.0 or brighter stars. In 2016 Fabio Falchi and colleagues combined satellite data with ground measurements into a world atlas and reported that more than 80% of the world's people, and more than 99% of people in the US and Europe, live under light-polluted skies. They estimated that the Milky Way cannot be seen from where about a third of humanity lives.
The loss is not only for astronomers. Artificial light at night affects migrating birds, insects and human sleep, and it is wasted energy. Cheap fixes exist: shielded lamps that point down, warmer colour temperature, and lights that dim or switch off when unneeded. Protected places include the Aoraki Mackenzie Dark Sky Reserve in New Zealand and the NamibRand Nature Reserve in Namibia (both designated in 2012 by what is now DarkSky International). Big observatories sit where the sky is dry, high and dark: Chile's Atacama and Andes (ESO's Paranal and the Vera C. Rubin Observatory), La Palma and Tenerife in the Canary Islands, Hawaii's Mauna Kea, South Africa's Sutherland (SALT), and the radio-quiet Karoo and Western Australia (the SKA).
A constellation is a pattern, not an object: the stars in one are usually at wildly different distances. The 88 constellations in use were fixed by the International Astronomical Union in 1922, with boundaries drawn in 1930, so that every point on the sphere belongs to exactly one. Most of the names come from Greek and Roman tradition, passed through Ptolemy's Almagest (about 150 CE) and Islamic astronomers who translated, corrected and renamed. That is why so many bright stars have Arabic names: Betelgeuse, Altair, Aldebaran, Vega. The Persian astronomer al-Sufi's Book of the Fixed Stars (964 CE) described what we now call the Andromeda Galaxy as a "little cloud".
Other peoples drew other skies. Chinese astronomy grouped stars into asterisms inside a different system of the sky, with court astronomers recording comets and "guest stars" (novae and supernovae) for two thousand years. The Maya tracked Venus's 584-day cycle in the Dresden Codex. The Inca read dark lanes in the Milky Way as animals, including a llama. Many Aboriginal Australian peoples see an emu in the same dark lanes. Polynesian and Micronesian navigators memorised where stars rise and set around the horizon, so a "star compass" could hold a course across open ocean; in 1976 the canoe Hōkūle'a sailed from Hawaii to Tahiti guided by the Micronesian navigator Mau Piailug.
None of these is "the real sky". They are different ways to organise the same stars, and any of them worked well enough to navigate, farm or keep time. The Western catalogue won for science mostly because it was standardised, not because the others were wrong.
The sky is a turning sphere whose view depends on where you stand. Brightness runs on a backwards logarithmic scale, size and distance come from angles, and the darkness of your sky decides how much of it you get. All the physics of stars begins with these plain measurements.
Try this: tonight, find Polaris (or, in the south, the pole's empty patch near the Southern Cross) and hold a fist at arm's length below it. Note what the sky looks like again two hours later, and write down which stars moved and which did not.
Sources. Pogson, "Magnitudes of thirty-six of the minor planets," Monthly Notices of the RAS 17 (1856); Ptolemy, Almagest (about 150 CE); al-Sufi, Book of the Fixed Stars (964 CE); Bessel's 1838 parallax of 61 Cygni; IAU constellation boundaries (Delporte 1930); IAU 2015 Resolution B3 on nominal solar values; Bortle, "Introducing the Bortle dark-sky scale," Sky & Telescope, February 2001; Falchi et al., "The new world atlas of artificial night sky brightness," Science Advances 2 (2016); ESA, Gaia Data Release 3 (13 June 2022) documentation; DarkSky International reserve list; Meeus, Astronomical Algorithms, for solar position approximations. Star coordinates are J2000 values from standard catalogues. All outputs of the interactives were computed in node. Cultural astronomy statements follow standard surveys of the field (for example Ruggles, ed., Handbook of Archaeoastronomy and Ethnoastronomy, 2015).