Evidence for colour

Colour is the charge of the strong interaction, and unlike the electric charge you meet in a first physics course it comes in three kinds. Each quark flavour — up, down, strange, and the rest — exists in three otherwise identical copies, conventionally labelled red, green, and blue, and antiquarks carry the three anticolours. The label was not invented to be pretty. It was forced on us twice over: once by a bookkeeping crisis in the quark model, where a baryon like the Δ⁺⁺ = u↑u↑u↑ appeared to put three identical fermions in the same state and violate the Pauli principle, and again by hard numbers pouring out of electron–positron colliders. The threefold multiplicity is not a private convenience of theorists; it changes measurable rates by exact integer factors, and the experiments agree. The cleanest of those measurements is the ratio R, the rate for e⁺e⁻ to annihilate into hadrons divided by the rate into muon pairs. The first step, e⁺e⁻ → qq̄, is pure QED — identical to e⁺e⁻ → μ⁺μ⁻ apart from the quark's fractional charge and its three colours — and whatever the quarks do afterwards to become the hadrons a detector sees cannot change the count. The prediction is startlingly simple, R = 3∑Q², and it rises in a staircase of integer-controlled steps as each new quark threshold opens. Strip the factor of three away and every step lands three times too low. Two further witnesses corroborate it to the decimal point: the neutral pion's decay rate, which scales as the square of the number of colours, and the antisymmetric colour wavefunction that lets three quarks sit in a baryon at all.

Key takeaways

Colour is an exact SU(3) charge: every quark flavour comes in three copies (r, g, b); antiquarks carry the anticolours, and leptons are colourless.
The flagship evidence is the ratio R = σ(e⁺e⁻→hadrons)/σ(e⁺e⁻→μ⁺μ⁻) = 3∑Q², whose overall factor of 3 directly counts the colours.
R climbs a staircase: 2 below charm (u d s), 10/3 with charm added, 11/3 with bottom — each step three times the no-colour value.
The neutral-pion rate Γ(π⁰→γγ) ∝ N_c² gives 7.8 eV for N_c = 3 versus 0.9 eV for N_c = 1 — a factor of nine, decisively measured.
Three quarks form a baryon only because the totally antisymmetric colour singlet ε_{ijk} exists — which needs exactly three colours.
Key idea: each quark flavour exists in three independent colour copies, and this multiplicity is counted directly by the factor of 3 in R(e⁺e⁻→hadrons), corroborated by the π⁰ rate and baryon statistics.

The core insight

Each quark flavour exists in three independent colour copies, and this multiplicity is counted directly by the factor of 3 in R(e⁺e⁻→hadrons) — the same integer that fixes the neutral-pion decay rate and lets three quarks form an antisymmetric baryon.

Where this lives today

Colour counting is woven into every quantitative use of the strong interaction. The R ratio is still the primary spectroscopy of quark thresholds and, through its QCD correction (1 + α_s/π + …), one of the cleanest determinations of the strong coupling; the same colour factor of three appears inverted as the 1/N_c colour-averaging factor in the Drell-Yan process qq̄ → μ⁺μ⁻ at hadron colliders, where a quark and antiquark must carry matching colour to annihilate. At the Z pole, LEP measured the hadronic width of the Z and the ratio R_Z = Γ_had/Γ_lep, both of which carry the factor N_c and confirmed three colours to sub-percent precision, while also counting three light neutrino generations. Three-jet events — a quark, an antiquark, and a radiated gluon — remain the direct image of the gauge boson colour demands, and their rate measures α_s from the gluon-emission vertex of equation (6). Lattice QCD, the parton distributions used at the LHC, and every Monte-Carlo shower that dresses coloured partons into jets all take N_c = 3 as bedrock — an input so well tested that experiments now use it to calibrate rather than to question.

The physical picture

Send an electron and a positron into each other at high energy and they annihilate into a single virtual photon carrying the whole centre-of-mass energy, q² = s. That photon does not care what it makes next — it couples to electric charge and nothing else — so it converts into any charged pair light enough to produce. When it makes a muon pair the story ends there. When it makes a quark pair the quarks cannot escape as free particles; they dress themselves into jets of hadrons on a timescale far longer than the production itself. But the production step is over before any of that happens, so counting hadronic events against muon-pair events counts quark production, cleanly. The one twist is that the photon can make the qq̄ pair in any of three colours, and those three channels are distinct final states that add. The first figure is that counting drawn as a diagram.

Figure

Annihilation e⁺e⁻ → γ* → qq̄ proceeds exactly as e⁺e⁻ → μ⁺μ⁻ but with two changes: the quark charge Q_f replaces 1, and the quark pair can be produced in any of three colours r g b, so σ(qq̄) = 3 Q_f² σ(μμ) and the ratio becomes R = 3 ∑ Q_f² = 10/3 for the u d s c flavours

Because the production step is pure electrodynamics, the theory has nowhere to hide: the hadronic rate is the muon rate scaled by ∑Q² and by however many colours there are. Plotted against the collision energy, R therefore sits flat between quark thresholds and jumps up whenever the energy climbs past 2m_f and a new flavour switches on. The heights of those plateaus are the measurement. The second figure is the staircase itself, with the colour prediction sitting three times above the colourless one that the data rule out.

Figure

The measured R staircase versus centre-of-mass energy: R holds a plateau between quark thresholds and steps up as each flavour opens — 2 below charm u d s, then 10/3 adding charm, then 11/3 adding bottom (blue). The no-colour prediction N_c = 1 (amber, dashed) lies a factor of three lower at 2/3, 10/9, 11/9; sharp resonances J/ψ and Υ spike above the continuum

The core equations

The defining relation is the ratio itself. The numerator and denominator share the identical initial state and identical photon propagator, so everything except the final-state charge and colour multiplicity cancels — which is exactly why R is a clean colour counter and not a messy dynamical quantity.

R \equiv \frac{\sigma(e^{+}e^{-}\to\text{hadrons})}{\sigma(e^{+}e^{-}\to\mu^{+}\mu^{-})} = N_c\sum_{f}Q_f^{2} \qquad (1)

where: N_c — the number of colours (= 3 in nature); Q_f — the electric charge of quark flavour f in units of e; the sum runs over every flavour with 2m_f < √s. Reading it: the muon-pair cross section is the yardstick, and R measures how many charged, coloured quark states the same photon can produce. The staircase is just equation (1) evaluated flavour by flavour as thresholds open. Below charm only u, d, s are accessible; charm adds a +2/3 quark, bottom a −1/3 quark, and the top is far too heavy to reach in e⁺e⁻ collisions.

R_{uds}=3\cdot\tfrac{2}{3}=2,\quad R_{udsc}=3\cdot\tfrac{10}{9}=\tfrac{10}{3},\quad R_{udscb}=3\cdot\tfrac{11}{9}=\tfrac{11}{3} \qquad (2)

where: ∑Q² = 4/9 + 1/9 + 1/9 = 2/3 for u d s, then +4/9 for charm and +1/9 for bottom; each measured plateau is exactly three times the value the no-colour hypothesis N_c = 1 would give. The integer heights are the fingerprint. Above tree level the strong interaction dresses the qq̄ pair with gluons, raising R by a small, calculable amount — the leading QCD correction. This is not a loophole in the colour count; it is a percent-level bump on top of the integer, and measuring it returns the strong coupling.

R = N_c\sum_f Q_f^{2}\left(1+\frac{\alpha_s}{\pi}+\mathcal{O}(\alpha_s^{2})\right) \qquad (3)

where: α_s — the strong coupling at the scale √s. At √s ≈ 34 GeV, α_s ≈ 0.14, so α_s/π ≈ 0.045 lifts R from 11/3 = 3.67 to ≈ 3.83 — the small excess above the plateau, not a change in the leading 3. A completely independent counter is the neutral pion's two-photon decay, which proceeds through a quark triangle (the chiral anomaly). Each quark in the loop contributes its charge-squared, and the three colours add, so the amplitude scales as N_c and the rate as N_c².

\Gamma(\pi^{0}\to\gamma\gamma)=\left(\frac{N_c}{3}\right)^{2}\frac{\alpha^{2}m_{\pi}^{3}}{64\pi^{3}f_{\pi}^{2}} \qquad (4)

where: m_π = 135 MeV — the π⁰ mass; f_π = 92 MeV — the pion decay constant; the factor N_c(Q_u² − Q_d²) = N_c/3 equals 1 precisely when N_c = 3. Numerically the N_c = 3 prediction is 7.8 eV; the N_c = 1 value is nine times smaller. The oldest evidence is structural: three quarks can only bind into a baryon because a totally antisymmetric colour combination — the colour singlet — exists. That combination is the Levi-Civita tensor on the colour indices, which demands exactly three colours to be non-trivial.

\psi_{\text{colour}}=\frac{1}{\sqrt{6}}\,\varepsilon_{ijk}\,\lvert q_i\,q_j\,q_k\rangle=\frac{1}{\sqrt{6}}\big(rgb-rbg+gbr-grb+brg-bgr\big) \qquad (5)

where: ε_{ijk} — the fully antisymmetric symbol on colour indices i, j, k ∈ {r, g, b}; the sign flips under any pairwise swap, making the three-quark state antisymmetric under colour exchange. This is what lets Δ⁺⁺ = u↑u↑u↑ obey Fermi statistics. Finally, colour is a genuine gauge charge: the quark couples to the gluon field through the Gell-Mann matrices λ^a, so the electromagnetic current picks up a colour index and a colour matrix at the vertex.

j^{\mu}_{a}=\bar u(p_3)\,\Big(-\tfrac{1}{2}\,i\,g_s\,\lambda^{a}\gamma^{\mu}\Big)\,u(p_1),\qquad a=1,\dots,8 \qquad (6)

where: g_s — the strong coupling (α_s = g_s²/4π); λ^a — the eight 3×3 Gell-Mann matrices, the generators of SU(3) acting on the colour vector; the eight values of a label the eight gluons. Colour is the charge the gluon reads.

Derivation ACounting colours in the annihilation ratio R

The goal is to show that the ratio R = σ(e⁺e⁻→hadrons)/σ(e⁺e⁻→μ⁺μ⁻) equals N_c∑Q², deriving it entirely from the muon-pair cross section already computed for e⁺e⁻ → μ⁺μ⁻, so that the only new ingredients are the quark's fractional charge and its colour multiplicity. The strategy is to argue that the production step is pure QED, that hadronisation is a downstream event invisible to the counting, and that the three colours are three independent final states whose rates add.

Step 1. Start from the point cross section σ(e⁺e⁻→μ⁺μ⁻) = 4πα²/3s, valid well above threshold where masses are negligible. Replace the muon by a single quark of flavour f and one fixed colour: the diagram is identical except the final vertex carries charge Q_f instead of 1, so its amplitude scales by Q_f and the cross section by Q_f². Why: cross section goes as amplitude squared, and the only vertex that changes is the final one, whose coupling is eQ_f rather than e.

Step 2. Sum over colour. A red qq̄ pair, a green one, and a blue one are physically distinct final states — the photon can make any of them with equal amplitude, since QED is colour-blind and the coupling does not depend on colour. Distinct final states add incoherently, so the single-flavour rate is multiplied by N_c. Why: colour is conserved and unobserved, so the three channels neither interfere nor are double-counted; they are three separate ways to end up as hadrons.

Step 3. Sum over flavour. Every quark light enough to pair-produce at energy √s contributes its own N_c Q_f² σ_μμ. Adding them, σ(e⁺e⁻→qq̄) = N_c σ_μμ ∑_f Q_f², where the sum is over accessible flavours. Why: different flavours are also distinct final states, so their production rates add, each weighted by its charge-squared.

Step 4. Divide by the muon rate. The universal factor 4πα²/3s cancels top and bottom, leaving R = N_c ∑_f Q_f², independent of s except through which flavours are open. Hadronisation — the quarks turning into jets — happens with unit probability once the pair is made, so it drops out of the ratio entirely. Why: confinement guarantees every produced qq̄ pair becomes hadrons, so counting hadronic events is exactly counting quark-pair production, and the messy long-distance physics never enters. *As a check: for u d s c, R = 3(4/9+1/9+1/9+4/9) = 3·10/9 = 10/3 ≈ 3.33, exactly three times the colourless value 10/9 ≈ 1.11.*

Go deeper · Derivation B · 2 minWhy a baryon needs exactly three colours

Derivation BWhy a baryon needs exactly three colours

The aim is to resolve the statistics paradox that first demanded colour: the ground-state baryon Δ⁺⁺ = u↑u↑u↑ appears to place three identical spin-½ fermions in the same quantum state, in flat violation of the Pauli exclusion principle, and to show that the only economical cure is a new three-valued degree of freedom whose antisymmetric combination restores Fermi statistics. We build the total wavefunction piece by piece and find the missing antisymmetric factor.

Step 1. Write the total baryon wavefunction as a product over its spaces: ψ = ψ_space × ψ_spin × ψ_flavour × ψ_colour. Fermi statistics require ψ to be totally antisymmetric under exchange of any two quarks. Why: quarks are spin-½ fermions, so the full state must change sign under any pairwise swap — this is the constraint the naive quark model appeared to break.

Step 2. Evaluate the three known factors for the Δ⁺⁺. In the L = 0 ground state ψ_space is symmetric; the spin-3/2, J_z = +3/2 state |↑↑↑⟩ is symmetric; and flavour uuu is symmetric. Their product is symmetric — the opposite of what statistics demands. Why: three symmetric factors multiply to a symmetric state, so without a fourth factor the Δ⁺⁺ would be a forbidden configuration, yet it plainly exists.

Step 3. Introduce a colour factor and require it to be totally antisymmetric, so that ψ_symmetric × ψ_colour is antisymmetric overall. For three quarks the only totally antisymmetric combination of a three-valued index is the Levi-Civita contraction ε_{ijk}|q_i q_j q_k⟩, equation (5). Why: an antisymmetric tensor on three indices exists only if the index takes at least three values — two colours give no antisymmetric singlet, four give more structure than nature shows — so exactly three colours is selected.

Step 4. Check that this factor is a colour singlet, invariant under SU(3) rotations, so the baryon carries no net colour and can exist as a free, colourless particle. The same ε_{ijk} is the unique singlet in 3 ⊗ 3 ⊗ 3, and confinement then permits only such singlets. Why: observed hadrons are colourless, and the antisymmetric ε combination is precisely the colour-neutral state, tying the statistics fix to confinement. *As a check: ε_{ijk} is non-zero only for {i,j,k} a permutation of {r,g,b}; the six terms of equation (5) carry alternating signs, giving a state antisymmetric under every pairwise swap, as required.*

Go deeper · Derivation C · 2 minThe neutral pion lifetime as a colour counter

Derivation CThe neutral pion lifetime as a colour counter

The purpose is to show that the decay π⁰ → γγ measures N_c independently of any collider, because its amplitude runs through a triangle loop of quarks whose colours add coherently, making the rate proportional to N_c². We sketch how the anomaly amplitude assembles the quark charges and colours, then turn the resulting rate into a number that separates N_c = 3 from N_c = 1 by a full factor of nine.

Step 1. Identify the mechanism. The π⁰ is the u ū − d d̄ combination; it decays to two photons through a triangle diagram in which a quark loop emits both photons (the Adler–Bell–Jackiw chiral anomaly). Each quark flavour in the loop contributes its charge-squared with the pion's isospin weight, giving the factor Q_u² − Q_d² = 4/9 − 1/9 = 1/3. Why: the two photons each couple with a factor Q, so the loop brings in Q² per flavour, and the π⁰ wavefunction weights up and down with opposite sign.

Step 2. Sum over colour. The loop runs independently over the three colours of the quark, and because colour is conserved around the loop the three contributions add in the amplitude, supplying a factor N_c. The full amplitude is therefore proportional to N_c(Q_u² − Q_d²) = N_c/3. Why: the same coherent colour sum that boosts R by a factor of 3 boosts the anomaly amplitude by N_c, but here it sits inside an amplitude, so the rate will carry N_c².

Step 3. Square to a rate. The anomaly fixes the amplitude normalisation, and the two-body decay rate is Γ = (N_c/3)² α²m_π³/(64π³f_π²), equation (4). The N_c = 3 case makes the prefactor (N_c/3)² = 1. Why: the decay rate is amplitude-squared times phase space, so the N_c in the amplitude becomes N_c² in the rate — the observable that distinguishes colour counts quadratically.

Step 4. Put in numbers. With m_π = 135 MeV and f_π = 92 MeV, N_c = 3 gives Γ = 7.8 eV (lifetime 8.4×10⁻¹⁷ s), while N_c = 1 gives 0.87 eV. The measured width is 7.8 eV. Why: the factor-of-nine gap between the two hypotheses is far larger than any uncertainty, so the neutral-pion lifetime alone establishes three colours. *As a check: (α²m_π³)/(64π³f_π²) = 7.80 eV for N_c = 3 and 0.867 eV for N_c = 1; the ratio is exactly 9 = (3/1)², matching the PDG value Γ(π⁰) ≈ 7.8 eV.*

Limiting cases

Case	Result	Meaning
Below charm (u d s)	R = 2	lowest plateau; ∑Q² = 2/3, times 3 colours
Charm open (u d s c)	R = 10/3 ≈ 3.33	+4/9 from the +2/3 charm quark
Bottom open (u d s c b)	R = 11/3 ≈ 3.67	+1/9 from the −1/3 bottom quark
With QCD correction	R → R(1 + α_s/π + …)	percent-level rise above the integer plateau; measures α_s
At a resonance (J/ψ, Υ)	R spikes far above the plateau	qq̄ bound state, Breit–Wigner peak, not the continuum count
No colour (N_c = 1)	R = 2/3, 10/9, 11/9	every step three times too low — excluded by data

Common trapTreating colour as a metaphor with no observable consequences

The name "colour" is a label — quarks are not literally red or blue — and it is tempting to conclude the whole construction is just bookkeeping with no empirical bite. That is exactly backwards. The threefold multiplicity is one of the most rigorously tested integers in physics: it multiplies the e⁺e⁻ → hadrons rate by precisely 3, so every R plateau lands at three times the colourless value; it multiplies the π⁰ → γγ rate by 3² = 9, matching the measured 7.8 eV against a colourless 0.9 eV; and it is what allows a baryon to exist at all, since without an antisymmetric colour factor the Δ⁺⁺ would violate the Pauli principle. A "mere label" cannot change a decay rate by a factor of nine or a cross-section ratio by a factor of three. Colour is as observable as electric charge — you just read it off rates rather than off a deflection in a field. The one genuinely subtle point is that you never see a single colour in isolation, because confinement permits only colourless combinations; but the number of hidden colours is fixed by counting, unambiguously, at 3.

Worked examples

1 · The R staircase across three energy windows (fully worked)

Predict R at √s = 2.5 GeV, 6 GeV, and 20 GeV, with and without colour, and compare to the measured plateaus (≈ 2.1, ≈ 3.6, ≈ 3.9). Which flavours are open at each energy?

Step 1. Identify open flavours from 2m_f < √s: at 2.5 GeV, u (2 MeV), d (5 MeV), s (95 MeV) are open but charm (2m_c ≈ 2.6 GeV) is not; at 6 GeV, charm is open but bottom (2m_b ≈ 9 GeV) is not; at 20 GeV, bottom is open, top (2m_t ≈ 346 GeV) is hopelessly out of reach. The three windows are u d s, u d s c, and u d s c b.

Step 2. Sum the charge-squares. u d s: 4/9 + 1/9 + 1/9 = 2/3. Add charm (+4/9): 10/9. Add bottom (+1/9): 11/9.

Step 3. Apply colour, R = 3∑Q². This gives R = 3·(2/3) = 2, then 3·(10/9) = 10/3 ≈ 3.33, then 3·(11/9) = 11/3 ≈ 3.67. Without colour (N_c = 1) the same sums give 0.67, 1.11, 1.22 — a factor of three lower at every step.

Step 4. Compare to data. The measured 2.1, 3.6, 3.9 sit just above the colour predictions 2, 3.33, 3.67 by a few percent — the leading QCD correction (1 + α_s/π): at 34 GeV it lifts 3.67 to ≈ 3.83. The colourless 0.67, 1.11, 1.22 miss by threefold and are ruled out. Colour reproduces every plateau; no colour fails at all three.

2 · Neutral-pion decay: nine to one against no colour

Use equation (4) to compute Γ(π⁰→γγ) for N_c = 3 and N_c = 1, and compare to the measured 7.8 eV. What lifetime does N_c = 3 predict?

Step 1. Evaluate the colour-independent prefactor with m_π = 135 MeV, f_π = 92 MeV, α = 1/137: α²m_π³/(64π³f_π²) = 7.8 eV. This is the N_c = 3 value, since (N_c/3)² = 1 there.

Step 2. For N_c = 1 multiply by (1/3)² = 1/9: Γ = 7.8/9 = 0.87 eV. The two hypotheses differ by a clean factor of nine because the colour factor enters the amplitude linearly and the rate quadratically.

Step 3. Compare to experiment. The measured width is 7.8 eV (PrimEx), landing squarely on N_c = 3 and nine standard steps away from N_c = 1. The neutral pion is thus a stand-alone colour counter, independent of any collider.

Step 4. Lifetime: τ = ħ/Γ = (6.58×10⁻¹⁶ eV·s)/(7.8 eV) = 8.4×10⁻¹⁷ s, in agreement with the PDG value. N_c = 3 gives 7.8 eV and 8.4×10⁻¹⁷ s; N_c = 1 gives 0.87 eV — excluded.

3 · The Δ⁺⁺ statistics paradox, counted out

Show explicitly that the Δ⁺⁺ = u↑u↑u↑ ground state requires an antisymmetric colour factor, and that two colours would not suffice.

Step 1. List the symmetry of each factor. Space (L = 0): symmetric. Spin (J_z = +3/2, all up): symmetric. Flavour (uuu): symmetric. Their product is symmetric under any quark swap. Three symmetric factors give a symmetric state.

Step 2. Impose Fermi statistics: the full wavefunction must be antisymmetric. The colour factor must therefore carry all the antisymmetry, ψ_colour antisymmetric under every pairwise exchange.

Step 3. Count colours needed. A totally antisymmetric state of three indices is ε_{ijk}, which is non-zero only when i, j, k are all different. With two colours you cannot make three distinct indices, so ε has no non-zero component and no antisymmetric singlet exists. Three colours are the minimum; three quarks fix it at exactly three.

Step 4. Confirm the result is colourless. The ε_{ijk} combination is the unique SU(3) singlet in 3 ⊗ 3 ⊗ 3, so the baryon carries no net colour and can propagate freely, consistent with confinement. Colour resolves the paradox and selects N_c = 3.

Physics of the portrait

Colour was born not from data but from a paradox, and the person who first named the missing degree of freedom was a young theorist puzzling over why three identical quarks could sit together at all.

Origins · 1 minOscar W. Greenberg and how this was discovered

Oscar W. Greenberg

born 1932 · introduced the threefold colour degree of freedom

Oscar Wallace Greenberg, an American theorist at the University of Maryland, gave colour its first physical form in 1964 — the same year Gell-Mann and Zweig proposed quarks. The new quark model had a glaring flaw: baryons like the Δ⁺⁺ and the Ω⁻ required three identical spin-½ quarks in a symmetric ground state, in apparent violation of the Pauli principle that had organised all of atomic and nuclear physics. Greenberg's answer was that quarks obey "parastatistics of order three" — equivalently, that each quark carries a hidden three-valued charge, so the three quarks are not truly identical and their overall wavefunction can be antisymmetrised in that new index. He had, in effect, discovered colour, though the modern language came later. In 1965 Moo-Young Han and Yoichiro Nambu recast the idea as an exact SU(3) gauge symmetry with three quark triplets, promoting colour from a statistical device to a dynamical charge coupled to gauge bosons — the seed of the gluon. The name "colour" and the theory "quantum chromodynamics" were fixed by Fritzsch, Gell-Mann, and Leutwyler in 1973, the same moment asymptotic freedom made the theory calculable. What began as Greenberg's bookkeeping trick to save the Pauli principle became, within a decade, the exact gauge theory of the strong force — and the R ratio, the pion lifetime, and the three-jet event turned his hidden index into a measured integer.

History

1964

— Greenberg proposes a threefold degree of freedom (parastatistics) to fix baryon statistics

1965

— Han and Nambu recast colour as an exact SU(3) gauge symmetry with three triplets

1973

— Fritzsch, Gell-Mann and Leutwyler name colour and quantum chromodynamics

1974

— the J/ψ is discovered, opening the charm threshold seen in the R staircase

1975

— SPEAR measures R and confirms the factor of three from colour

1979

— three-jet events at PETRA give direct evidence for the gluon

Going further

Texts. Thomson, Modern Particle Physics, §10.2 (colour and QCD), §9.6 (baryon wavefunctions and colour), Fig. 10.2 · Griffiths, Introduction to Elementary Particles, 2nd ed., §1.10 and §8.1 (hadron production in e⁺e⁻) · Halzen & Martin, Quarks and Leptons, §2.11 (colour) and §11.6 (R ratio) · Peskin & Schroeder, An Introduction to QFT, §5.1 and §19.3 (the anomaly and π⁰ → γγ) · Schwartz, QFT and the Standard Model, §30 (colour and QCD). The key numerical benchmarks are: R = 2, 10/3, 11/3 for u d s, u d s c, u d s c b — three times the N_c = 1 values 2/3, 10/9, 11/9; the QCD-corrected R(34 GeV) ≈ 3.83; Γ(π⁰→γγ) = 7.8 eV for N_c = 3 and 0.87 eV for N_c = 1, ratio 9, lifetime 8.4×10⁻¹⁷ s; and ε_{ijk} is the unique totally antisymmetric three-colour singlet. Quark charges are u, c = +2/3; d, s, b = −1/3.

Prerequisites. SU(3) Flavour Symmetry · e⁺e⁻ → μ⁺μ⁻ · Baryon Wavefunctions. Leads to. Colour and QCD · Gluons · Colour Factors.

Ask the tutor: Ask me to walk the R staircase past the charm and bottom thresholds step by step, or to trace how the three colours add coherently in the π⁰ triangle to give the factor-of-nine that rules out a single colour.

If colour is unobservable because of confinement, how can we say there are three?

Confinement hides individual colours — you never catch a lone red quark — but it does not hide how many there are. The number of colours enters observable rates as an integer multiplier that survives confinement completely. The e⁺e⁻ → hadrons rate is three times what one colour would give; the π⁰ → γγ rate is nine times larger; the Z boson's hadronic width scales with N_c. Each of these is a measurement of an inclusive rate — you sum over all the colourless hadrons in the final state — so the colour sum is done for you by nature. Counting hidden colours by their effect on rates is exactly as legitimate as counting atoms by weighing a gas: you never see one, but the total is unambiguous.

Why does the factor of three appear linearly in R but as nine in the pion rate?

Because R is a ratio of cross sections and the pion width is a rate, and colour enters them at different stages. In R the three colour channels are three distinct final states that add incoherently — you sum three cross sections — so the rate gains one factor of N_c and R = 3∑Q². In π⁰ → γγ the three colours run coherently around the same triangle loop, so they add in the amplitude, giving one factor of N_c there; squaring the amplitude to get the rate then produces N_c². The rule is simple: colour sums that happen inside an amplitude come out squared in the rate, while sums over genuinely different final states come out linearly. Both are counting the same three colours.

Doesn't the QCD correction to R spoil it as a colour counter?

No — it strengthens it. The correction is a small multiplicative factor (1 + α_s/π + …) sitting on top of the integer 3∑Q². At √s = 34 GeV it lifts R from 3.67 to about 3.83, a few percent, nowhere near the factor of three that separates colour from no colour. Far from being a nuisance, this bump is a bonus: measuring exactly how far R rises above its plateau returns the strong coupling α_s at that energy, which is one of the standard determinations. So R tells you two things at once — the leading integer fixes the number of colours, and the small excess measures the strength of the force between them.

Why can't the R staircase be explained without colour, by adding more flavours?

Because the flavours are fixed by other measurements — their masses set exactly where each step occurs, and their charges set each step's height in units of ∑Q². You cannot invent extra quarks to make up the missing factor of three without producing new thresholds (new steps) that are simply not seen, and without violating the charge assignments confirmed in deep inelastic scattering and hadron spectroscopy. The staircase has precisely the risers and heights the known six flavours predict, but every plateau sits a uniform factor of three too high for one colour. A single overall factor of three, independent of energy and flavour, is the signature of a multiplicity that every quark shares — colour — not of extra flavours.

What actually happens to the quarks after e⁺e⁻ → qq̄, and why doesn't it affect R?

The qq̄ pair flies apart, and because the strong force does not weaken with distance the energy in the colour field grows until it is cheaper to create new qq̄ pairs from the vacuum. The original quarks thus fragment into a spray of colourless hadrons — a jet along each quark's direction. This all happens long after the production vertex, on hadronic timescales, and with total probability one: every produced pair becomes hadrons. Since R is a ratio and the hadronisation probability is unity, that entire complicated process cancels. You are counting how often a quark pair was made, and every quark pair announces itself as hadrons. That separation of a clean short-distance production from an inevitable long-distance dressing is what makes R calculable.

How does the Z boson at LEP confirm three colours if R already did?

The Z gives an independent, higher-precision check. When the Z decays it can go to lepton pairs or to quark pairs, and the quark channels carry the same colour factor N_c because each flavour again comes in three copies. LEP measured the hadronic-to-leptonic width ratio R_Z = Γ(Z→hadrons)/Γ(Z→leptons), which is proportional to N_c times a sum of quark couplings, and found it consistent with three colours to better than a percent. Because the Z couples through both vector and axial currents, this also tests colour in the electroweak sector, not just in QED-mediated annihilation. The agreement of the R staircase, the π⁰ lifetime, and the Z width — three very different processes — makes N_c = 3 one of the most secure numbers in the Standard Model.