Stars are classified using two primary variables: surface temperature and luminosity. These two properties form the backbone of the Morgan-Keenan (MK) system, which assigns every star a spectral type (a letter-number code reflecting its temperature) and a luminosity class (a Roman numeral reflecting its size and brightness). The Sun, for example, is classified as G2V, where G2 indicates its temperature and V means it sits on the main sequence.
Surface Temperature and Spectral Type
The most fundamental variable in stellar classification is surface temperature, expressed through a letter sequence: O, B, A, F, G, K, and M. O-type stars are the hottest, with surface temperatures reaching 100,000 Kelvin, while M-type stars are the coolest at around 3,000 K. The Sun falls in the G category, with a surface temperature between 5,000 and 6,000 K. Each letter is further divided into ten numerical subdivisions (0 through 9) for finer detail, so a B3 star is hotter than a B7.
This sequence was developed in the early 1900s by Annie Jump Cannon at Harvard Observatory, who reorganized an earlier alphabetical system by rearranging the letter groups to reflect a smooth temperature progression. Before Cannon’s work, colleagues at the observatory fiercely debated whether spectral class even correlated with temperature. The International Astronomical Union officially adopted her system in 1922, and it remains the standard today.
What astronomers actually measure to determine spectral type is a star’s spectrum: the specific wavelengths of light the star absorbs and emits. Different temperatures cause different chemical elements to show up as dark absorption lines. Hot O and B stars show strong lines from ionized helium. A-type stars display prominent hydrogen lines. Cooler K and M stars show absorption from metals and molecules like titanium oxide and vanadium oxide. These spectral fingerprints are what place a star into its temperature class.
Luminosity Class
Temperature alone doesn’t tell the full story. Two stars can have the same surface temperature but differ enormously in size and total energy output. A red giant and a red dwarf might both be 3,000 K, yet the giant can be thousands of times more luminous simply because it has a much larger surface area radiating light. This is why the MK system adds a second variable: luminosity class, labeled with Roman numerals from I to V.
- Class Ia and Ib: Supergiants (bright and less-bright). Deneb, a blue-white supergiant, is classified A2Ia.
- Class II: Bright giants. Canopus is an example at F0II.
- Class III: Ordinary giants. Capella is classified G5III.
- Class IV: Subgiants, stars transitioning off the main sequence.
- Class V: Main sequence stars (also called dwarfs). Vega is A0V, and the Sun is G2V.
Astronomers determine luminosity class by looking at subtle differences in spectral line widths. Stars with lower surface gravity, like bloated supergiants, produce narrower, sharper absorption lines than compact main sequence stars of the same temperature. This makes it possible to distinguish a supergiant from a dwarf using only the star’s light, without needing to measure its physical size directly.
The Hertzsprung-Russell Diagram
These two classification variables map directly onto the Hertzsprung-Russell (HR) diagram, the most important visualization in stellar astronomy. The horizontal axis plots temperature (or a related measurement called color index), with hotter stars on the left. The vertical axis plots luminosity or absolute magnitude, with brighter stars toward the top. When thousands of stars are plotted this way, they don’t scatter randomly. Instead, they cluster into distinct bands that correspond to the luminosity classes: a dense diagonal stripe of main sequence stars, a clump of giants in the upper right, and supergiants stretched across the top.
The HR diagram makes it clear that temperature and luminosity aren’t independent. For main sequence stars, hotter always means more luminous, and both properties are driven by mass. A star’s luminosity scales roughly as mass raised to the 3.5 power, meaning a star twice the Sun’s mass is about 11 times more luminous. Mass is the single variable that determines where a main sequence star lands on the diagram, which is why mass is sometimes called the most fundamental stellar property, even though it isn’t directly part of the classification label.
Color as a Temperature Proxy
You don’t need a full spectrum to estimate a star’s temperature. Color works as a shortcut. Astronomers measure a star’s brightness through different filters and compare the results. The difference between blue-filter brightness and visual-filter brightness, called the B-V color index, correlates tightly with surface temperature. A negative B-V value means the star is blue-white and very hot. A large positive value means it’s red and cool. Diagrams that plot color index against magnitude are called color-magnitude diagrams, and they’re the observational workhorse for classifying stars in large surveys where taking individual spectra would be impractical.
Chemical Composition
While temperature and luminosity are the two main classification axes, chemical composition adds an important third dimension. Most stars have broadly similar compositions, dominated by hydrogen and helium, but the proportion of heavier elements (collectively called “metals” in astronomy) varies. Stars with unusually low or high metal content show noticeably different spectral lines compared to chemically normal stars at the same temperature. Metal-weak A-type stars, for instance, have weaker absorption lines from iron and other metals than typical A stars.
Some stars are chemically peculiar enough to warrant special designations. These include stars with abnormal concentrations of elements like mercury, manganese, or silicon in their atmospheres. Carbon stars, which show strong carbon-molecule absorption, also fall outside the normal classification sequence. Automated classification systems now flag these chemical oddities alongside the standard temperature and luminosity labels.
Beyond M: The Cool End of the Sequence
The original OBAFGKM sequence was designed for true stars, objects hot enough to sustain hydrogen fusion. In the late 1990s, astronomers discovered objects too cool even for M classification, and three new spectral types were added: L, T, and Y.
L dwarfs have surface temperatures between roughly 1,400 and 2,200 K. At these temperatures, metal oxides like titanium oxide disappear from the spectrum and are replaced by metal hydrides and dust. T dwarfs are cooler still, around 950 K, and their spectra are dominated by methane absorption, which gives them a distinctly blue color in infrared images. Y dwarfs are the coolest known, with temperatures that can dip below the boiling point of water. These objects blur the line between the coolest stars and the largest planets, and classifying them relies on infrared spectroscopy rather than the visible-light techniques used for hotter stars.
The variables remain the same in principle: temperature drives the spectral features, and luminosity (or in this case, the object’s intrinsic faintness) provides the second dimension. But the specific molecular markers astronomers look for shift dramatically at these low temperatures, making the cool end of the sequence a distinct classification challenge.