Charge Conservation and the Continuity Equation

A new chapter opens here, and it opens with a confession: electrodynamics has been quietly spending conservation laws for six chapters without ever auditing them. Energy left batteries and arrived in resistors; momentum will shortly go missing between misbehaving currents; fields are about to be caught storing both. Before any of those ledgers can be trusted, we need the master template every one of them will copy — and physics keeps that template filed under its oldest, safest conservation law: electric charge. You have met the continuity equation twice already, and this track will not insult you by deriving it a third time. The E4 article built $\partial\rho/\partial t+\nabla\cdot\mathbf{J}=0$ from honest counting in a fixed volume and spent it on Kirchhoff's law and charge relaxation; the E6 twin proved it is an identity of the completed Maxwell equations, the consistency test the displacement current was born to pass. What neither article asked is the question this chapter runs on: what kind of statement is a continuity equation, and what else can be conserved that way? The answer — any additive quantity with a density, a flux, and an honest account of its transfers — is the grammar in which Poynting's theorem, the Maxwell stress tensor, and field angular momentum will all be written over the next nine articles. So this is the chapter's constitution: charge as the exemplar — globally conserved to fantastic experimental precision, locally conserved by the deepest structural demands of relativity — and the sourced continuity equation as the sentence structure that energy, momentum, and angular momentum are about to fill in with their own nouns.

Key takeaways

A conserved quantity obeys a continuity equation: $\partial_t(\text{density}) + \nabla\cdot(\text{flux}) = (\text{source})$ . For electric charge the source is exactly zero — an experimental statement, tested to extraordinary precision, not a mathematical necessity.
The equation's two readings: over any volume, "rate of change inside = − flux out through the boundary + creation inside"; at any point, the same sentence shrunk to zero size.
The source term is the chapter's key generalization: $\sigma_X=0$ means strictly conserved (charge); $\sigma_X\neq0$ means transferred to another ledger (field energy loses $-\mathbf{J}\cdot\mathbf{E}$ to matter) — transfer, not magic.
Splitting the flux $\mathbf{J}=\rho\mathbf{v}+\mathbf{J}_{\rm cond}$ gives the co-moving (convective) form: charge riding with matter versus charge leaking through it.
Even where charges are created — pair production, beta decay — creation happens in charge-neutral combinations at a single event: the vertex itself obeys continuity.
Charge is both conserved (continuity) and Lorentz-invariant (same value in every frame) — two different facts; energy manages only the first.
KEY IDEA: conservation, done locally, is always the same sentence — density rate plus flux divergence equals source — and this chapter conjugates that sentence for charge, energy, momentum, and angular momentum.

The core insight

A conserved quantity obeys a continuity equation $\partial(\text{density})/\partial t + \nabla\cdot(\text{flux}) = (\text{source})$ . For charge the source is zero, so $\partial\rho/\partial t+\nabla\cdot\mathbf{J}=0$ — the template that energy, momentum, and angular momentum conservation in electrodynamics will each instantiate.

Where this lives today

Charge bookkeeping is a live engineering discipline. Every particle detector is an exercise in it: calorimeters and time-projection chambers reconstruct events by demanding the collected charge balance, and "missing charge" in a vertex is how exotic-particle searches flag candidates. Spacecraft charging is continuity with hostile fluxes — a satellite in a substorm collects electrons faster than it sheds them, floats kilovolts negative, and arcs across its own solar panels; mitigation (plasma contactors, conductive coatings) is literally source-and-flux management. Semiconductor device simulators solve drift–diffusion continuity equations for electrons and holes billions of times per design cycle — with genuinely nonzero $\sigma_X$ : generation and recombination transfer carriers between bands, in exactly the paired-ledger pattern this article's template classified. And the template's deepest modern job is theoretical quality control. Lattice gauge theories are engineered so that a discrete continuity equation holds exactly on the grid (the E6 twin's Yee-cell observation, industrialized); machine-learned surrogate models for plasmas and electrochemistry are now routinely built with conservation laws hard-wired into the network architecture, because a model that leaks charge drifts into nonsense. When charge conservation appears to fail in the lab, it is a detector calibration problem — an epistemic status no other equation in this track quite enjoys.

The physical picture

Picture charge as a mobile but indestructible fluid painted into space. The continuity equation is nothing but the bookkeeping rule for such a fluid: the amount in any box changes only by flow through the walls. $\nabla\cdot\mathbf{J}$ measures net outflow per unit volume — where it is positive, more current leaves than arrives, and the local density must be draining at exactly that rate. The two terms of $\partial_t\rho+\nabla\cdot\mathbf{J}=0$ are mirror images of one another, a point the E4 article's counting derivation made rigorous once and for all. What earns this old equation a second home at the head of the conservation-laws chapter is the realization that nothing in the bookkeeping cared that the fluid was charge. Any additive stuff — energy, a momentum component, probability, mass, entropy — can be assigned a density $\rho_X$ and a flux $\mathbf{J}_X$ , and the same sentence can be written. The only new degree of freedom is a slot on the right-hand side: a source density $\sigma_X$ , the rate per unit volume at which the stuff is created, destroyed, or — the reading that will do all the work in this chapter — handed to a different ledger. Charge is the aristocrat of the family precisely because its slot is empty: $\sigma_q=0$ , always, everywhere, in every observed process since measurement began.

Figure

The conservation-law template — one box, three terms, and the dictionary this chapter will fill in

How secure is the empty slot? Globally: no reaction ever observed changes net charge — beta decay manufactures an electron but pairs it with a proton's worth of compensation; annihilation destroys an electron only against a positron. The cleanest single number is the electron's stability: it cannot decay at all without violating charge conservation (nothing lighter is charged), and the Borexino bound already cited by the E6 twin puts its lifetime beyond $6.6\times10^{28}$ years — five billion billion universe-ages. Locally: relativity leaves no alternative. If charge could vanish here and reappear on the Moon with only the total preserved, observers in different frames would disagree about which happened first — some would see an interval with too little charge, others too much, and "the total" would stop being frame-independent at all. That argument — global-but-not-local conservation is relativistically incoherent — is the centrepiece of the next article, Global vs. Local Conservation Laws, deferred there by name; this article takes the local law as established twice over and puts it to work as a template. One clarification the chapter needs stated early, because energy will fail half of it: conservation and invariance are different honors. Charge is conserved (its total never changes in time, in any one frame) and also invariant (every frame agrees on its value — the electron's charge does not Lorentz-contract or dilate; the four-current packaging $\partial_\mu J^\mu=0$ that makes this manifest was previewed by the E6 twin and is built properly in the relativity chapter, both recalled by name). Energy is conserved but not invariant: every frame keeps its own energy ledger perfectly, while disagreeing with every other frame about the amount. Momentum likewise. Keeping the two words separate now will save real confusion when the field's energy and momentum start transforming into each other in E10.

The template, made precise

Derivation AThe sourced continuity equation - grammar for any conserved stuff

We derive the general local budget for an arbitrary additive quantity $X$ — with the source term that charge never needs but this chapter's other laws will — and establish the dictionary between its integral and differential readings. The strategy is the fixed-volume budget of the E4 twin, generalized: where that article counted charge and set creation to zero as a law of nature, we keep the creation slot open and classify what can legally fill it. The localization step (shrinking from "every volume" to "every point") was proven honestly there; we invoke it by name rather than re-perform it.

Step 1. Let $X$ be additive stuff with volume density $\rho_X(\mathbf{r},t)$ , so a region $V$ holds $X_V=\int_V\rho_X\,\mathrm{d}\tau$ . Additivity is doing real work: charge, energy, and momentum components qualify; temperatures and phases do not.

Step 2. Posit the transport hypothesis: $X$ moves through space with a flux density $\mathbf{J}_X$ ( $X$ per unit area per unit time crossing a surface), and may be created at a rate $\sigma_X$ per unit volume. The budget for any fixed volume $V$ with boundary $S$ :

\frac{\mathrm{d}}{\mathrm{d}t}\int_V \rho_X\,\mathrm{d}\tau \;=\; -\oint_S \mathbf{J}_X\cdot\mathrm{d}\mathbf{a} \;+\; \int_V \sigma_X\,\mathrm{d}\tau,

where the minus sign reads outward flux as loss, and every term is measurable independently in principle — that is what makes the budget falsifiable rather than definitional.

Step 3. Convert the surface term with the divergence theorem (E0, by name) and collect everything under one integral: $\int_V\left[\partial_t\rho_X+\nabla\cdot\mathbf{J}_X-\sigma_X\right]\mathrm{d}\tau=0$ for every volume $V$ . The localization lemma of the E4 twin then forces the integrand to vanish pointwise:

\frac{\partial\rho_X}{\partial t}+\nabla\cdot\mathbf{J}_X=\sigma_X.

Step 4. Classify the right-hand side — the step that turns grammar into physics. Three legal fillings. $\sigma_X=0$ : strict conservation; the stuff has nowhere to go but through walls (charge; total energy; total momentum). $\sigma_X = -(\text{transfer rate to ledger }Y)$ : partitioned conservation — $X$ alone is not conserved, but the pair $(X,Y)$ is, and the same term appears in $Y$ 's equation with opposite sign. This is the chapter's engine: field energy will obey $\partial_t u+\nabla\cdot\mathbf{S}=-\mathbf{J}\cdot\mathbf{E}$ , with $+\mathbf{J}\cdot\mathbf{E}$ reappearing as the work rate on matter (Work Done by Electromagnetic Fields on Charges, two articles ahead, by name). Genuine $\sigma_X\neq0$ with no partner ledger: the stuff simply is not conserved (kinetic energy alone, in an inelastic world).

Step 5. Read the theorem's converse as a research tool: if experiment establishes that $X$ 's total never changes, locality (next article's argument) demands there exist a flux $\mathbf{J}_X$ making the sourceless equation true. Conservation laws are promissory notes for the existence of currents — and finding the explicit current (Poynting's $\mathbf{S}$ , Maxwell's stress $\mathsf{T}$ ) is precisely the work of the coming articles. Verified: setting $X=$ charge, $\sigma=0$ recovers the E4 twin's equation exactly; integrating Step 3's equation back over a volume regenerates Step 2's budget (the dictionary is invertible); and dimensional audit gives $[\sigma_X]=[\rho_X]/\mathrm{s} = [\mathbf{J}_X]/\mathrm{m}$ consistently.

Go deeper · Derivation B · 2 minCharge riding on moving matter - the convective split

Derivation BCharge riding on moving matter - the convective split

Electrodynamics increasingly meets charge carried by moving media — plasmas, electrolytes, rotating machinery, the interior of stars — where the current has two physically distinct parts: charge dragged along with the material, and charge drifting through it. We separate them and derive the co-moving form of continuity, the form fluid dynamics and magnetohydrodynamics actually use. The one subtlety: "the charge in a moving parcel" requires differentiating an integral over a volume whose shape is itself changing — the transport theorem, handled honestly.

Step 1. Split the flux: $\mathbf{J}=\rho\,\mathbf{v}+\mathbf{J}_{\rm cond}$ , where $\mathbf{v}(\mathbf{r},t)$ is the material's velocity field, $\rho\mathbf{v}$ is convection (charge riding the medium), and $\mathbf{J}_{\rm cond}$ is conduction (drift relative to the medium — the part Ohm's law $\mathbf{J}_{\rm cond}=\sigma\mathbf{E}'$ governs, in the material's own frame). Substituting into continuity:

\frac{\partial\rho}{\partial t}+\nabla\cdot(\rho\mathbf{v})=-\nabla\cdot\mathbf{J}_{\rm cond}.

Step 2. Expand the convective divergence with the product rule (E0 identities, by name): $\nabla\cdot(\rho\mathbf{v})=\mathbf{v}\cdot\nabla\rho+\rho\,\nabla\cdot\mathbf{v}$ , and group the advective pair into the material derivative $\mathrm{D}\rho/\mathrm{D}t \equiv \partial_t\rho+\mathbf{v}\cdot\nabla\rho$ — the rate of change seen riding with the parcel:

\frac{\mathrm{D}\rho}{\mathrm{D}t}=-\rho\,\nabla\cdot\mathbf{v}-\nabla\cdot\mathbf{J}_{\rm cond}.

Step 3. Read each term physically. $-\rho\nabla\cdot\mathbf{v}$ : the parcel's charge density dilutes when the flow expands ( $\nabla\cdot\mathbf{v}>0$ stretches the same charge over more volume) — a density change with no charge lost. $-\nabla\cdot\mathbf{J}_{\rm cond}$ : genuine leakage through the parcel's walls by conduction. For a perfectly insulating moving medium ( $\mathbf{J}_{\rm cond}=0$ ), the total charge of a material parcel is exactly constant — integrate Step 2 over a co-moving volume and the transport theorem's boundary term absorbs the dilution piece identically.

Step 4. The payoff regime: in a highly conducting moving fluid, the E4 twin's relaxation clock $\tau_e=\varepsilon/\sigma$ is so short that parcels stay neutral, conduction current dominates convection, and continuity collapses to $\nabla\cdot\mathbf{J}\approx0$ in the moving frame — one of the two pillars (with flux freezing) on which magnetohydrodynamics stands, deferred to the advanced chapter by name. At the opposite pole — charged insulators in motion, Rowland's classic spinning-disk experiment — the convective term is everything: moving charge is a current, sourcing real magnetic fields, exactly as the magnetostatics chapter's surface-current bookkeeping assumed. Verified: summing the two terms of Step 1 under a divergence reproduces the standard continuity equation; the co-moving conservation statement matches the Reynolds transport theorem of the mechanics track (recalled by name); and the neutral-parcel limit reproduces the steady-current contract $\nabla\cdot\mathbf{J}=0$ of magnetostatics.

Figure

Creation is legal only in neutral pairs — worldlines from a vertex balance the ledger; a lone birth cannot

Go deeper · Derivation C · 2 minCreation events obey continuity - the pair-production audit

Derivation CCreation events obey continuity - the pair-production audit

Charge conservation does not say charges are never created — accelerators create them by the trillion. It says creation is locally charge-neutral. We audit the sharpest case: a photon converting to an electron–positron pair at a single spacetime event, written in the distributional language of the E0 delta-function articles (recalled by name), and show the continuity equation holds through the creation instant — then show why creating a lone charge cannot be written down at all.

Step 1. Model the pair born at position $\mathbf{r}_0$ at time $t_0$ , thereafter following worldlines $\mathbf{r}_+(t)$ and $\mathbf{r}_-(t)$ with $\mathbf{r}_\pm(t_0)=\mathbf{r}_0$ . Density and current, as point-charge distributions for $t>t_0$ :

\rho(\mathbf{r},t)=\theta(t-t_0)\,q\left[\delta^3(\mathbf{r}-\mathbf{r}_+(t))-\delta^3(\mathbf{r}-\mathbf{r}_-(t))\right],

with $\mathbf{J}$ the corresponding sum of $q_i\mathbf{v}_i\,\delta^3(\mathbf{r}-\mathbf{r}_i(t))$ terms.

Step 2. Check continuity away from the event: each moving point charge separately satisfies $\partial_t\rho_i+\nabla\cdot\mathbf{J}_i=0$ — the chain rule pulls $-\mathbf{v}_i\cdot\nabla\delta^3$ from the time derivative and $+\mathbf{v}_i\cdot\nabla\delta^3$ from the divergence; they cancel identically. Moving charge never violates continuity; only appearing or disappearing could.

Step 3. Check the event itself: differentiating the step function contributes $q\,\delta(t-t_0)\left[\delta^3(\mathbf{r}-\mathbf{r}_+(t_0))-\delta^3(\mathbf{r}-\mathbf{r}_-(t_0))\right]$ — and because both particles are born at the same point $\mathbf{r}_0$ , the bracket vanishes exactly. The creation burst cancels within $\rho$ itself. Continuity holds distributionally through the vertex: nothing needs to flow anywhere at $t_0$ , because zero net charge appeared.

Step 4. Now attempt a lone creation: $\rho=\theta(t-t_0)\,q\,\delta^3(\mathbf{r}-\mathbf{r}_0)$ , motionless for simplicity. Then $\partial_t\rho = q\,\delta(t-t_0)\,\delta^3(\mathbf{r}-\mathbf{r}_0)$ , while $\nabla\cdot\mathbf{J}=0$ (nothing moved). Continuity fails by an irreducible point burst — and integrating over a small ball gives $\mathrm{d}Q/\mathrm{d}t=q\,\delta(t-t_0)$ with zero boundary flux: the budget is simply violated. No choice of $\mathbf{J}$ localized at the event can repair it, since the created charge would have to arrive from somewhere. This is the current-level shadow of the E6 twin's field-level verdict — Maxwell's equations refuse to be sourced by a non-conserved current — recalled by name, not re-derived.

Step 5. The physical moral: relativistic quantum theory, which is where creation and annihilation live, builds charge conservation into every interaction vertex — each QED vertex conserves charge exactly, event by event, which is why beta decay ( $n\to p+e^-+\bar\nu$ : $0\to+1-1+0$ ) and annihilation balance to the last unit. The empirical scoreboard: no charge-nonconserving decay has ever been seen; the electron's lower lifetime bound stands at $6.6\times10^{28}$ years. Verified: Step 2's cancellation is the standard point-charge continuity check (Jackson §6.1 problems); Step 3–4's contrast reproduces the textbook statement "charge is created only in neutral pairs" in distributional form; the Step 4 obstruction matches the Gauss-law argument — a lone new charge would demand its Coulomb flux appear at infinity instantaneously.

The chapter map: one sentence, four nouns

The rest of the E7 chapter is this table, row by row — each conservation law an instantiation of the template, each flux a physical object with its own article:

conserved stuff	density	flux	transfer term (to/from matter)	built in
charge	$\rho$	$\mathbf{J}$	none — $\sigma=0$ exactly	this article (E4/E6 twins recalled)
energy	$u=\tfrac{\varepsilon_0}{2}E^2+\tfrac{1}{2\mu_0}B^2$	Poynting $\mathbf{S}$	$-\mathbf{J}\cdot\mathbf{E}$	Work Done…, Energy Density…, The Poynting Vector, Poynting's Theorem
momentum	$\mathbf{g}=\varepsilon_0\,\mathbf{E}\times\mathbf{B}$	$-\mathsf{T}$ (Maxwell stress)	$-(\rho\mathbf{E}+\mathbf{J}\times\mathbf{B})$	Momentum Density…, The Maxwell Stress Tensor, Conservation of Linear Momentum…
angular momentum	$\boldsymbol{\ell}=\mathbf{r}\times\mathbf{g}$	$-\mathbf{r}\times\mathsf{T}$	torque density on matter	Angular Momentum of the EM Field, Feynman Disk Paradox

Two features of the table repay a pause. First, only charge's transfer slot is empty: fields and matter trade energy, momentum, and angular momentum incessantly, so each of those rows is really half of a field-plus-matter pair whose sum is strictly conserved. Second, the fluxes climb in tensor rank — scalar charge is moved by a vector, vector momentum by a rank-two tensor — which is why the stress tensor is unavoidable three articles from now, not an optional sophistication.

Limiting cases

case	statement	meaning
steady state	$\nabla\cdot\mathbf{J}=0$	closed current loops; the magnetostatic contract (E4 twin's Kirchhoff/taper corollaries, by name)
accumulation	$\mathrm{d}Q/\mathrm{d}t=I_{\rm in}$	capacitor plates, beam dumps — Worked Example 2
insulating moving medium	$\mathrm{D}q_{\rm parcel}/\mathrm{D}t=0$	charge rides matter; convection current only (Rowland disk)
perfectly conducting medium	parcels neutral, $\nabla\cdot\mathbf{J}\approx0$ co-moving	the MHD pillar; relaxation clock wins every race
relativistic packaging	$\partial_\mu J^\mu=0$	one sentence, every frame — E6 twin's preview, built in E10 (by name)
global limit	$\mathrm{d}Q_{\rm universe}/\mathrm{d}t=0$	local law integrated over everything; flux honesty at infinity handled next article

Worked examples

1 · The Earth's leaky battery — a global budget with competing fluxes (confidence builder)

Fair-weather atmosphere carries a downward current density $J\approx2$ pA/m² everywhere; the Earth's surface holds charge producing the fair-weather field $E\approx130$ V/m. Audit the planet's charge budget with the integral continuity equation.

Step 1. The stock: surface charge from Gauss's law, $Q=\varepsilon_0 E\times 4\pi R_\oplus^2 = 8.854\times10^{-12}\times130\times5.10\times10^{14} = \mathbf{-5.9\times10^{5}\ C}$ (negative — the field points down).

Step 2. The drain: total fair-weather current $I = J\times4\pi R_\oplus^2 = 2\times10^{-12}\times5.10\times10^{14} = \mathbf{1.0\ kA}$ of positive charge arriving — a planet-sized leakage neutralizing the surface.

Step 3. The clock: with no resupply, $\tau = Q/I = 5.9\times10^5/1.0\times10^3 \approx 575\ \text{s} \approx \mathbf{10\ minutes}$ . The continuity equation predicts the fair-weather field of the entire planet should collapse before you finish this article.

Step 4. It doesn't — so the budget demands a competing flux, and continuity tells you its exact size: some process must deliver $\sim1$ kA of negative charge to the ground continuously. The resolution is the global thunderstorm circuit: at any moment $\sim$ 1000–2000 active storms drive negative charge downward (lightning, point discharge) and positive charge up to the conductive ionosphere, recharging the planetary capacitor at precisely the rate fair-weather regions drain it. A century of atmospheric electricity — the Carnegie curve's daily rhythm of the global field — is this one budget equation, solved for the missing term.

2 · Charging a beam dump — accumulation you can put a voltmeter on (exam level)

A 1.0 μA proton beam runs into an electrically isolated metal dump of capacitance $C=200$ pF to ground. Find the charging dynamics and the engineering consequence.

Step 1. Continuity in integral form, with the dump as the volume: every proton that enters and stops adds its charge — $\mathrm{d}Q/\mathrm{d}t = I = 1.0\times10^{-6}$ C/s. No flux leaves (it's isolated); the source term is zero (protons carry their charge in; none is created).

Step 2. The measurable consequence: $V=Q/C$ rises at $\mathrm{d}V/\mathrm{d}t = I/C = 10^{-6}/2\times10^{-10} = \mathbf{5.0\ kV/s}$ .

Step 3. Failure clock: air-gap breakdown around $\sim$ 30 kV (centimetre-scale gaps) arrives in $t = 30\,000/5000 = \mathbf{6.0\ s}$ , at which point the accumulated $Q=CV=6.0$ μC exits as a spark — continuity enforced by arc. Real dumps are grounded through a current monitor for exactly this reason, and that monitor's reading is the beam current: charge bookkeeping as diagnostic.

Step 4. The instructive contrast with Example 1: same equation, opposite regimes — there, two huge fluxes balancing to hold a stock steady; here, one small flux integrating relentlessly into a dangerous stock. The continuity equation does not care which; it only refuses to lose count.

Origins · 1 minMichael Faraday and how this was discovered

Michael Faraday

1791–1867 · Royal Institution, London — who put conservation of charge on the bench

Franklin had asserted in 1747 that electricity is never created, only transferred — bookkeeping born with the naming of positive and negative. But it was Faraday who made conservation measurable. His 1843 ice-pail experiment lowered a charged ball into a metal pail on an electrometer and showed the induced charge equaled the inducing charge exactly, through every configuration, contact, and removal — charge accounted for, coulomb by coulomb, before the coulomb existed. The same year he charged and discharged in every way chemistry and friction allowed, always recovering equal and opposite amounts. For Faraday this was one strand of a larger conviction — his 1840s "conservation of force" intuition that nature's agencies are transformed, never spent — a conviction that fed directly into the energy-conservation revolution of Joule (next articles) and Helmholtz. That the bench-top bookkeeping of a bookbinder's apprentice would one day be enforced at every vertex of quantum electrodynamics is the kind of promotion physics rarely grants twice.

History

1747

— Franklin: electricity is redistributed, never created — charge conservation stated with the naming of + and −

1843

— Faraday's ice-pail experiment makes conservation quantitative: induced equals inducing, exactly

1861

— Maxwell's displacement current makes local conservation an identity of field theory

1897

— Thomson's electron gives the conserved fluid its atom

1918

— Noether ties conservation laws to symmetries: charge to gauge invariance

1959

— Rowland's convection current (1876) reconfirmed in the modern era: moving charge is current, as continuity demands

2015

— Borexino pushes the electron lifetime beyond 6.6×10²⁸ years: the emptiest source term in physics

Going further

References. Griffiths, Introduction to Electrodynamics, 4th ed., §5.1.3 and §8.1.1 · Jackson, Classical Electrodynamics, 3rd ed., §6.1 · Purcell & Morin, Electricity and Magnetism, 3rd ed., §4.2 · Zangwill, Modern Electrodynamics, §2.1.

Provenance. Equations and derivations follow Griffiths (Introduction to Electrodynamics, 4th ed.) §5.1.3, §7.1, §8.1.1; Jackson (Classical Electrodynamics, 3rd ed.) §6.1; Purcell & Morin §4.2; Zangwill §2.1. A verifier should confirm: (1) the divergence derivation of continuity from the Ampère–Maxwell and Gauss laws; (2) the charge-relaxation time constant $\tau=\epsilon/\sigma$ ; (3) standard constants $\epsilon_0=8.854\times10^{-12}$ F/m, $\mu_0=4\pi\times10^{-7}$ T·m/A, $c=2.998\times10^8$ m/s; (4) the electron lifetime lower bound $>6.6\times10^{28}$ yr (Particle Data Group). All numbers are standard textbook/PDG values. Addendum for this edition: THIS IS THE THIRD CONTINUITY ARTICLE — items (1) and (2) are owned by the E6 twin (Maxwell-consistency identity) and the E4 twin (counting derivation + relaxation) respectively and are recalled BY NAME only; this E7 edition owns the sourced template with ledger taxonomy, the convective/material-derivative split, the pair-creation distributional audit, and the chapter dictionary (energy/momentum/angular-momentum rows). Worked numbers recomputed: planetary budget $Q=\varepsilon_0E\cdot4\pi R^2=5.9\times10^5$ C, $I=1.0$ kA, $\tau\approx575$ s; beam dump $\mathrm{d}V/\mathrm{d}t=I/C=5.0$ kV/s, breakdown at ~6 s, $Q=6$ μC.

Builds on: Divergence and the Divergence Theorem · Electric Current and Current Density · Maxwell's Equations in Vacuum.

Leads to: Global vs. Local Conservation Laws · Poynting's Theorem · Conservation of Linear Momentum in Electrodynamics · The Poynting Vector.

Three articles on the continuity equation — what exactly does each one own?

A fair audit. The E4 article (magnetostatics chapter) derives the law from counting charge in a fixed volume, and spends it where steady currents live: Kirchhoff's node law, the taper theorem, charge relaxation τ=ε/σ. The E6 twin proves the law is an identity of the completed Maxwell equations — the divergence of Ampère–Maxwell — and runs the logic backward to show the displacement current is forced by conservation; its jurisdiction is consistency of the field theory. This article owns the law as a template: the sourced generalization, the convective split, creation-event bookkeeping, and the dictionary that energy, momentum, and angular momentum fill in across E7. Same equation; three different jobs; zero repeated derivations.

Is there any experimental wiggle room — could charge conservation be violated at some tiny level?

It is among the most over-determined facts in physics, attacked from three independent directions. Decay searches: the electron cannot decay without violating it, and Borexino's bound exceeds 6.6×10²⁸ yr. Neutrality checks: atoms are neutral to ~10⁻²¹ of an elementary charge, so protons and electrons balance absurdly precisely. Structural arguments: in quantum field theory, charge conservation follows from gauge invariance (Noether's second-theorem reading, recalled from the E6 twin), and small violations are hard to even formulate without breaking the consistency of Maxwell's equations — a lone appearing charge would need its Coulomb field to appear acausally everywhere. Any confirmed violation would demolish far more than one conservation law, which is why claimed anomalies are always hunted first in the apparatus.

If charge in an expanding plasma parcel dilutes, hasn't the parcel lost charge?

No — density and amount are different bookkeeping lines, and the convective derivation separates them cleanly. Dρ/Dt = −ρ∇·v − ∇·J_cond: the first term says the density falls as the parcel's volume grows, exactly compensated by the growth, so the parcel's total ∫ρ dτ is untouched by expansion. Only the conduction term moves charge across the parcel's material boundary. It is the same distinction as dough rising: the raisin density drops; no raisins are lost.

Why do we bother with the source-term version if charge's source is exactly zero?

Because the chapter ahead is built on quantities whose source terms are not zero — and the discipline of naming the receiving ledger is what separates physics from hand-waving. Field energy alone is not conserved (−J·E drains it into matter); mechanical momentum alone is not conserved (fields absorb the imbalance — the resolution of the third-law crisis, deferred by name to Newton's Third Law in Electrodynamics). The sourced template makes each partial law honest by displaying exactly what is owed to whom. Charge, with its empty slot, is the calibration standard: the one ledger that balances by itself.

Does the continuity equation assume charge is quantized?

No — it is a statement about the continuum bookkeeping of whatever charge exists, quantized or not. Historically it predates the electron by decades; formally, ρ and J may be smooth fluids or sums of delta functions (the derivations here used both). Quantization is a separate, additional fact — charge comes in units of e (or e/3 in confined quarks) — with its own candidate explanations (Dirac's monopole argument, recalled from Divergence of B, grand unification). Conservation says the ledger balances; quantization says the ledger's entries are integers. Millikan established the second; Faraday's ice pail the first.

What symmetry is behind charge conservation, in the Noether sense?

Global phase (gauge) symmetry: in quantum theory, shifting the phase of every charged field by the same constant leaves physics unchanged, and Noether's first theorem converts that invariance into a conserved current — exactly ∂μJ^μ = 0. Promote the symmetry to a local one (phase shifts varying point to point) and you are forced to introduce the electromagnetic potential itself — gauge symmetry doesn't just protect charge conservation, it generates electromagnetism. The identity-based reading (Noether's second theorem) was the E6 twin's closing theme; the symmetry-based one anchors the next article's doctrine that conservation laws are never accidents. Full machinery lives in the field-theory track, by name.

How literally should I take the "indestructible fluid" picture?

Take the bookkeeping literally and the substance loosely. The fluid picture nails the mathematics — density, flux, budget — and safely covers even exotic cases (two carrier species, holes in semiconductors, sign-mixed plasmas) because additivity is all it uses. It misleads only if you let it promise that charge is a thing with independent existence: charge is a property carried by particles, and the particles are creatable and destructible — in charge-neutral combinations, as the pair-production audit showed. "The fluid is conserved" is exactly true; "the fluid exists apart from its carriers" is not. Physics keeps the ledger, not the metaphysics.