The HR Diagram

Section 2 of 6

The Observer's HR Diagram

Part 2: The Observer’s HR Diagram — Patterns from Data

A Radical Idea: Plot Everything

By the early 1900s, astronomers had measured apparent magnitudes and spectral types for thousands of stars, and parallaxes were giving absolute magnitudes. The obvious question: what happens if you plot absolute magnitude against spectral type? Ejnar Hertzsprung (1911) and Henry Norris Russell (1913) independently did exactly this. Every star has a different mass, age, and composition — so the natural expectation was noise, a random spray of dots. There was no theoretical reason in 1911 to expect anything else.

What they found instead was one of the most stunning patterns in science. The rice didn’t scatter — it fell along a narrow highway (a single diagonal band) with a few outlying clusters. This plot is the Hertzsprung-Russell diagram. Stars were not infinitely varied; something was organizing them. The pattern demanded an explanation — and that explanation would require an entirely new physics of stellar interiors.

Hertzsprung-Russell diagram

A plot of stellar luminosity (or absolute magnitude) against effective temperature (or spectral type) that organizes stars into the main sequence, giant branch, and white dwarf sequence. The single most important diagram in astrophysics — a classification chart, a mass map, and an evolution diagram at once.

Cecilia Payne-Gaposchkin: The Thesis That Decoded the HR Diagram

Figure 1These women analyzed hundreds of thousands of stellar spectra from photographic plates, discovering the patterns (OBAFGKM, period-luminosity relation, luminosity classification) that transformed astronomy into astrophysics.Harvard College Observatory / Smithsonian Institution

In Lecture 3 you met the Harvard Computers — the women who built the OBAFGKM system from $3.5 \times 10^5$ glass-plate spectra. Their work gave the HR diagram its horizontal axis. But that axis is just labels — to make the diagram mean something, someone had to decode what the labels physically represent.

That someone was Cecilia Payne, whose 1925 doctoral thesis — written at age 25 — connected the diagram’s two axes to physics. Applying the Saha ionization equation (then brand-new quantum theory) to the Harvard classifications, Payne showed the horizontal axis is a temperature sequence, not a composition sequence. Her second result was more radical still: all stars, regardless of spectral type, are overwhelmingly hydrogen and helium. The diversity of spectra is a temperature effect, not a chemistry effect.

Astronomer Otto Struve later called it “undoubtedly the most brilliant PhD thesis ever written in astronomy.” Yet her conclusion about hydrogen dominance was initially rejected by Henry Norris Russell (the “R” in HR diagram), who persuaded her to soften it — then reached the same conclusion independently four years later. Payne-Gaposchkin became the first woman to hold a full professorship at Harvard, in 1956, three decades after she was right.

Quick check

If spectral type primarily tracked composition rather than temperature, what would you expect the HR diagram to look like?

If the observer’s HR diagram still feels abstract, orient yourself with a more old-school map first. The classic version below labels familiar stars directly, so you can see the geography before translating it into the stripped-down data view.

Classic Hertzsprung-Russell diagram with luminosity on the vertical axis and surface temperature on the horizontal axis, labeled with spectral classes O through M and representative stars such as Spica, Rigel, Betelgeuse, the Sun, Sirius B, Procyon B, and Proxima Centauri in the main sequence, giant, and white dwarf regions. — Figure 2Even a classic labeled HR diagram shows the same three geographies immediately: the main sequence, the cool luminous giant/supergiant region, and the hot faint white dwarf region. Named stars make the map feel physical rather than abstract.Course illustration (A. Rosen)

Building the Diagram

The observer’s HR diagram (also a color-magnitude diagram, CMD) plots absolute magnitude $M_V$ on the vertical axis (brighter/more negative at the top) and spectral type O B A F G K M — equivalently color index $B - V$ — on the horizontal axis (hot blue stars left, cool red stars right, temperature decreasing rightward).

Color-magnitude diagram

The observer’s form of the HR diagram: absolute magnitude versus color index (or spectral type), built from photometry and distance alone — no physical theory required.

Crucially: neither axis requires theory. Spectral type is a direct classification from line patterns; absolute magnitude comes from apparent magnitude and parallax distance. This diagram is pure measurement — Observable → Model → Inference: we measure $m$ and spectral type, use the distance modulus to model $M$ , then infer structure from how stars populate the diagram.

What the Diagram Shows

1. The Main Sequence — a narrow diagonal band from upper-left (hot, bright) to lower-right (cool, faint). About 90% of all stars fall on it, the Sun roughly in the middle. As you learned in Lecture 4, mass determines luminosity ( $L \propto M^{3.5}$ ) and temperature for main-sequence stars, so the main sequence is a mass sequence — high-mass at upper left, low-mass at lower right.

Main sequence

The diagonal band on the HR diagram where hydrogen-core-burning stars spend ~90% of their lives, ordered by mass: high-mass stars are hot and luminous (upper left), low-mass stars cool and faint (lower right). It is the mass-luminosity relation made visible.

2. The Giant and Supergiant Region — a cluster in the upper-right: cool ( $T \sim 3{,}000\text{–}5{,}000~\text{K}$ ) but very luminous ( $100\text{–}10^4\,L_\odot$ ). These are red giants and supergiants. A main-sequence star at $3{,}500~\text{K}$ would be a dim M dwarf — so to be both cool and luminous, recall Stefan-Boltzmann ( $L = 4\pi R^2\sigma T^4$ ): they must have enormous radii. A red giant is typically $10\text{–}100\,R_\odot$ ; a supergiant can exceed $1{,}000\,R_\odot$ .

Red giant

An evolved star that has exhausted core hydrogen and expanded enormously (typically $10\text{–}100\,R_\odot$ ), becoming cool but very luminous — the upper-right region of the HR diagram.

Supergiant

The most luminous evolved stars, with radii reaching $100\text{–}1{,}000\,R_\odot$ (luminosity class I) — e.g. Betelgeuse. They occupy the very top of the HR diagram across a wide temperature range.

3. The White Dwarf Sequence — a scattering in the lower-left: hot ( $T \sim 10^4\text{–}3 \times 10^4~\text{K}$ ) but very faint ( ${\sim}0.01\,L_\odot$ ). By the same logic — hot but faint means small — these are white dwarfs, Earth-sized ( $R \sim 0.01\,R_\odot$ ): the remnant cores of dead stars.

White dwarf

The Earth-sized ( $R \sim 0.01\,R_\odot$ ) remnant core of a low- or intermediate-mass star — hot but faint, in the lower-left of the HR diagram, cooling slowly with no ongoing fusion.

2-Minute Exercise: Place Stars on the Diagram

Sketch a blank HR diagram (luminosity vertical, temperature horizontal — decreasing to the right) and place each star: (1) Sun — G2 V, $1\,L_\odot$ , $5{,}800~\text{K}$ ; (2) Sirius — A1 V, $25\,L_\odot$ , $9{,}900~\text{K}$ ; (3) Betelgeuse — M1 I, ${\sim}10^5\,L_\odot$ , ${\sim}3{,}600~\text{K}$ ; (4) Sirius B — white dwarf, ${\sim}0.03\,L_\odot$ , ${\sim}2.5 \times 10^4~\text{K}$ ; (5) Proxima Centauri — M5 V, $0.0017\,L_\odot$ , ${\sim}3{,}000~\text{K}$ . Which two are not on the main sequence? (Betelgeuse — supergiant, upper right; Sirius B — white dwarf, lower left.)

Luminosity Classes: Vertical Structure

Even at the same spectral type, stars differ enormously in luminosity. A K2 star could be a dwarf ( $L \sim 0.4\,L_\odot$ ) or a giant ( $L \sim 200\,L_\odot$ ) — same temperature, 500 times more luminous. Observers tell them apart through spectral line widths: higher surface gravity (compact dwarfs) raises atmospheric pressure and broadens lines; lower gravity (extended giants) gives narrower lines. This is the luminosity classification:

Luminosity class

The Roman-numeral part of a stellar classification (I supergiant, III giant, V dwarf, …), read from spectral line widths (surface gravity). It distinguishes stars of equal temperature but very different size and luminosity.

Luminosity Class	Name	Example
I	Supergiant	Betelgeuse (α Ori)
II	Bright giant	—
III	Giant	Arcturus (α Boo)
IV	Subgiant	Procyon (α CMi)
V	Main-sequence dwarf	Sun, Sirius A

A complete classification includes both: the Sun is G2 V (G2 temperature ${\sim}5{,}800~\text{K}$ , class V dwarf); Betelgeuse is M1 I ( ${\sim}3{,}600~\text{K}$ , class I supergiant).

Quick check

Which corner of the observer’s HR diagram holds the hottest, most luminous stars? The coolest, faintest?
A star is spectral type K5, luminosity class III. Dwarf, giant, or supergiant? Hotter or cooler than the Sun?
Two stars are both G2. Star A is class V; Star B is class III. Which is more luminous, and why?
Why does spectral line width distinguish giants from dwarfs?