The Shell Theorem

Every gravitation problem you have solved so far leaned on an enormous, unstated favour: you treated the whole Earth — six thousand trillion trillion kilograms of layered rock and iron, some of it 6,400 km away, some of it directly under your shoes — as a single point at the planet's centre. Why on earth is that legal? An apple two metres above the ground is absurdly closer to the nearest tonne of soil than to Australia. Newton asked the same question, and the answer bothered him for nearly twenty years — historians suspect it is a big part of why the Principia took until 1687. The answer is the shell theorem, and it comes in two acts. Outside a uniform spherical shell, its gravity is exactly what you'd get by squashing all its mass into the central point — not approximately, exactly, at every exterior distance. And inside the shell, the net pull is exactly zero — everywhere, not just at the centre, even with your face against the inner wall. Stack shells like an onion and you get the licence for every "centre-to-centre" calculation you have ever done, plus a gorgeous bonus: descend into the Earth and gravity dies off linearly, hitting zero at the core. NCERT states both results in a box; JEE-Advanced builds cavity and tunnel problems on them; NEET tests the statements straight. This article gives you the two theorems, the beautiful cancellation argument that algebra can fully reach, honest signposting of the one proof that needs university calculus, and the exam consequences — mine shafts, nested shells, and the famous uniform-field cavity.

Key takeaways

Theorem 1 (outside): a uniform spherical shell attracts any external mass as if the shell's entire mass sat at its centre — exactly, for every $r > R$ .
Theorem 2 (inside): a uniform spherical shell exerts zero net force anywhere in its hollow interior — even far off-centre.
The inside cancellation is pure geometry: patch area grows as $s^2$ while pull-per-kilogram falls as $1/s^2$ — near-small-strong exactly balances far-big-weak. It works only for the inverse square.
A solid sphere is an onion of shells: from outside it is a point mass; from radius $r$ inside, only the mass below you counts — $g(r) = \dfrac{GM r}{R^3} = g_0\,\dfrac{r}{R}$ for uniform density.
So gravity inside the Earth decreases going down (linearly, in the uniform model), and vanishes at the centre; the maximum is at the surface.
A shell blocks its own gravity from the interior, but does not shield the interior from outside sources — gravity cannot be screened.
Key idea: spheres gravitate as points and shells are silent inside — the two facts that make "centre-to-centre" legal and "journey to the centre" calculable.

The core insight

For an inverse-square force, a uniform spherical shell pulls external masses as if it were a point at its centre and pulls interior masses not at all — so any spherically layered body acts, from outside, as a central point mass, and at interior radius $r$ only the mass enclosed within $r$ contributes.

Where this lives today

The shell theorem's enclosed-mass logic is how astronomers weigh the invisible. A star orbiting a galaxy at radius $r$ responds only to the mass inside its orbit — equation (2) at cosmic scale — so measuring orbital speeds at many radii reads out $M(r)$ ring by ring. Run on spiral galaxies in the 1970s, that audit came back wrong: the outskirts orbit far too fast for the visible stars and gas beneath them. The fix — vast halos of dark matter, now a pillar of cosmology — entered physics through precisely the bookkeeping you used in worked example 3. Closer to home, borehole gravimetry reads $g(r)$ down drill-holes to map ore bodies and reservoirs (the linear law plus density anomalies), and physicists test the inverse square itself with modern "inside the shell" experiments, hunting for the tiny interior residue that any fifth force or extra dimension would leave. And when you meet Gauss's law in electrostatics next year, you will recognize it instantly: the shell theorem is its gravitational twin, proved by Newton with cones a century before Gauss had flux.

The physical picture

Start with the theorem that sounds false. Stand inside a huge hollow shell, well off-centre, close to one wall. The near wall looms over you — surely it wins the tug-of-war? Draw a narrow double cone through your position: it punches a small patch of shell nearby and a large patch on the far side. The near patch is close — strong pull per kilogram — but small. The far patch is weak per kilogram but big. Now count honestly. For a thin cone, the patch area grows as the square of the distance $s$ from you to the wall ( $\text{area} \propto s^2$ : double the distance, the cone's footprint is four times larger). But each kilogram's pull falls as $1/s^2$ . Multiply: patch pull $\propto s^2 \times 1/s^2 =$ independent of distance. The two patches tug with equal and opposite force, and cancel exactly. Sweep the double cone over every direction and the entire shell cancels, patch against patch, wherever you stand inside. No calculus — just two squares annihilating each other.

Figure

Inside a shell — every double cone cancels exactly

Notice what the argument leaned on: the exponent 2, twice. Area grows as $s^2$ because we live in three-dimensional space; the pull must fall as exactly $1/s^2$ for the cancellation to close. A $1/r^{2.01}$ gravity would leave a residue pulling you toward the near wall — the theorem is a fingerprint test for the inverse square. The outside theorem is the harder act. Its honest proof slices the shell into rings and adds their pulls with an integral — genuine university calculus, and this article will not pretend otherwise (the UG deep-dive runs it in full). What you should carry instead: the statement (exact point-equivalence, at every exterior point), the plausibility (far away, any blob looks like a point; the miracle is that for spheres the "far away" approximation is exact even up close), and the authority — NCERT states both theorems in Chapter 8, and every board that touches gravitation lets you use them by name. Stack the two theorems and you can compute the strangest journey in Newtonian gravity: down a shaft toward the centre of the Earth. At interior radius $r$ , every shell of rock above you (radii $> r$ ) switches off — you are inside them all. Every shell below still acts as a central point. The planet effectively shrinks as you descend, and gravity — far from growing toward the "source" — dies away, reaching exactly zero at the centre, where all shells surround you and all cancel.

Figure

The onion audit — shells above you switch off, shells below act as a point

The math: statements, and what algebra can honestly reach

The two theorems, stated exam-ready. For a uniform (or spherically layered) shell of mass $M$ and radius $R$ , and a point mass $m$ :

F_{\text{outside}} = \frac{GMm}{r^2} \quad (r > R), \qquad\qquad F_{\text{inside}} = 0 \quad (r < R) \qquad (1)

where: $r$ — distance from the shell's centre (not its surface). Read the fine print that examiners test: the outside result is exact, not a far-field approximation; the inside result holds at every interior point, off-centre included; and both require spherical symmetry — a cube, a ring or a hemisphere obeys neither. Because gravity superposes, any body built of concentric shells inherits the results layer by layer. From outside, the whole Earth — crust, mantle, iron core, any density profile that depends only on radius — is a point mass at its centre. That is the licence stamped on every $F = GMm/r^2$ you have ever written between "centres." Inside a solid sphere, the onion audit gives the interior field. At radius $r$ : outer shells silent, inner sphere (mass $M(r)$ , enclosed within $r$ ) acts centrally:

g(r) = \frac{G\,M(r)}{r^2} \qquad (2)

where: $M(r)$ — the enclosed mass only. This equation is the theorem's practical face: gravity at radius $r$ is set by the mass beneath you, concentrated at the centre. For a uniform sphere, $M(r) = M\,(r/R)^3$ , and the competition of $r^3$ upstairs against $r^2$ downstairs nets a single power:

g(r) = \frac{GM}{R^3}\,r = g_0\,\frac{r}{R} \qquad (r \le R) \qquad (3)

where: $g_0 = GM/R^2$ — the surface value. Linear growth from zero at the centre to the maximum at the surface, then the familiar $1/r^2$ fall outside: the peak of gravity is the ground you stand on.

where you are	field	why
outside any spherical body, $r > R$	$\dfrac{GM}{r^2}$	theorem 1: point-equivalence, exact
inside a hollow shell	$0$ , everywhere	theorem 2: double-cone cancellation
inside a uniform solid sphere	$g_0\,\dfrac{r}{R}$	onion audit: enclosed mass only
between two nested shells	inner shell(s) only	outer silent, inner central
inside an off-centre cavity	uniform: $\tfrac43\pi G\rho\,d$	superposition of sphere + negative sphere

Derivation AZero inside the shell — the double-cone cancellation, in full

We prove theorem 2 — zero net gravitational force at any point $P$ inside a uniform shell — using only the inverse-square law and the geometry of cones; this is the one shell-theorem proof that lives entirely within HS algebra, and it is examinable reasoning. Strategy: pair up the shell patch by patch using narrow double cones through $P$ , show each pair's tugs cancel exactly, and let the pairing exhaust the shell. Watch how the exponent 2 is used exactly once on each side of the balance — the proof fails for any other power.

Step 1. Through the interior point $P$ , draw a narrow double cone with a small opening angle: it intersects the shell in two patches — patch 1 at distance $s_1$ from $P$ , patch 2 on the opposite side at distance $s_2$ .

Step 2. Compare the patch areas. A cone's cross-section widens in proportion to the distance travelled from its apex — its footprint area grows as the distance squared. The two patches therefore have areas in the ratio

\frac{A_1}{A_2} = \frac{s_1^2}{s_2^2}

(The patches are slightly tilted against the cone where they meet the sphere — but by a circle's symmetry the same tilt factor applies at both intersections, so it cancels in the ratio.)

Step 3. Compare the patch masses. The shell is uniform — mass proportional to area — so $m_1/m_2 = s_1^2/s_2^2$ .

Step 4. Compare the pulls on a test mass at $P$ . Each patch pulls along the cone's axis, in opposite directions, with magnitude $Gm_{\text{patch}}m/s^2$ :

\frac{F_1}{F_2} = \frac{m_1/s_1^2}{m_2/s_2^2} = \frac{s_1^2/s_1^2}{s_2^2/s_2^2} = 1

Equal magnitudes, opposite directions: the pair cancels exactly.

Step 5. Exhaust the shell. Every direction from $P$ belongs to some double cone, every patch of shell is claimed exactly once, and every pair cancels. Net force at $P$ : zero — for any interior $P$ , however close to the wall. Why only the inverse square survives this audit: with force $\propto 1/s^n$ , Step 4's ratio becomes $s_1^{2-n}/s_2^{2-n}$ , which is 1 only for $n = 2$ . Under a $1/s^3$ force the near patch would win and the interior of a shell would pull you wall-ward — the zero is a precision test of the exponent, and laboratory versions of exactly this cancellation constrain deviations from $1/r^2$ today. *Verified: at the centre, cancellation is obvious by symmetry; the argument extends it everywhere inside. Newton's own proof (Principia, Book I, Prop. LXX) is this construction; modern texts (HRW Ch. 13, Kleppner Ch. 9) reproduce it as here. The companion exterior proof (Prop. LXXI / the ring integral) needs calculus — deep-dive material, used here by statement only.*

Go deeper · Derivation B · 1 minGravity inside the Earth — the linear law from the onion audit

Derivation BGravity inside the Earth — the linear law from the onion audit

We derive $g(r) = g_0\,r/R$ inside a uniform solid sphere — the equation behind every mine-shaft and tunnel problem. Strategy: split the planet at your radius into "shells above" (silenced by theorem 2) and "the sphere below" (centralized by theorem 1), then scale the enclosed mass by volume. Watch the honest asterisk in Step 4: uniform density is a model, and the real Earth deviates in an instructive way.

Step 1. Stand at interior radius $r$ . Decompose the planet: all shells with radii greater than $r$ — you are inside each of them — contribute zero (theorem 2). The remaining inner sphere of radius $r$ acts as a point mass at the centre (theorem 1, applied shell by shell).

Step 2. Enclosed mass for uniform density $\rho = M/(\tfrac43\pi R^3)$ :

M(r) = \rho \cdot \frac{4}{3}\pi r^3 = M\,\frac{r^3}{R^3}

Step 3. Field from the enclosed point:

g(r) = \frac{GM(r)}{r^2} = \frac{GM}{R^3}\,r = g_0\,\frac{r}{R}

Why linear: the mass beneath you shrinks as $r^3$ while the inverse square concentrates as $1/r^2$ ; three powers down, two powers up, one net power remains. At $r = R$ both branches — this linear law and the exterior $1/r^2$ — meet at exactly $g_0$ : the field is continuous at the surface.

Step 4. The honest asterisk. The real Earth is not uniform: its iron core is about five times denser than the crust. Descending through light mantle rock removes little mass while you close in on the heavy core, so measured $g$ actually rises slightly (peaking near $10.7\ \mathrm{m/s^2}$ at the core–mantle boundary) before falling to zero at the centre. Theorems 1 and 2 survive untouched — only the $r^3$ scaling of Step 2 assumed uniformity. Boards examine the uniform model; saying "for a uniform Earth" costs nothing and is exactly right. Verified: $r = R$ : $g = g_0$ ✓ (continuous with outside); $r = 0$ : $g = 0$ ✓ (all shells surround you); $r = R/2$ : $g = 4.9\ \mathrm{m/s^2}$ — the halfway-down value used in the worked examples. This linear law is also what makes the tunnel-through-the-Earth oscillate harmonically with the 84-minute period — the tunnel archetype cashes this cheque.

Limiting cases

case	result	meaning
$r \gg R$ outside	$F \to GMm/r^2$	far away, everything is a point — for spheres it is exact even up close
just outside vs just inside a shell	$GMm/R^2$ vs $0$	the force jumps across the shell wall; the shell's mass sheet carries the jump
centre of any spherical body	$g = 0$	total surrounding cancellation
uniform sphere, $r = R/2$	$g = g_0/2$	the linear interior law
nested shells, between them	only inner ones act	the onion audit, shell by shell
force law $\ne 1/r^2$	both theorems fail	the zero inside is an exponent test
non-spherical body (cube, ring)	neither theorem applies	symmetry is a load-bearing hypothesis

Common trap“Inside the shell, the nearer wall pulls you toward it”

The intuition is reasonable — closer means stronger — and it is exactly wrong, because it forgets that how much shell sits in each direction also depends on where you stand. Off-centre, less shell is near you and more is far, in the precise proportion (area $\propto s^2$ ) that cancels the strength advantage ( $\propto 1/s^2$ ). Zero net force everywhere inside — not just at the centre. Two siblings of this error patrol exam papers. First: "so a shell shields gravity" — no; the theorem silences the shell's own pull on interior objects, but an external source (the Moon, say) reaches through the shell untouched: gravity cannot be screened, unlike electric fields in a conductor. Second: "gravity is zero inside the Earth" — the Earth is not hollow; only the shells above your radius switch off, and at the bottom of the deepest mine nearly the whole planet still pulls from below.

Worked examples

1 · Board level — reading the theorems (NEET/CBSE pattern)

A uniform hollow shell has mass $6.0 \times 10^{22}\ \mathrm{kg}$ and radius $2000\ \mathrm{km}$ . Find the gravitational force on a $10\ \mathrm{kg}$ probe (a) at $1000\ \mathrm{km}$ from the centre, (b) at $1500\ \mathrm{km}$ from the centre but close to the inner wall in direction, (c) at $4000\ \mathrm{km}$ from the centre.

Step 1. (a) $r = 1000$ km $< R$ : inside. Theorem 2: $F = \boxed{0}$ .

Step 2. (b) Still $r < R$ — and position inside does not matter; the cancellation is exact everywhere in the hollow: $F = \boxed{0}$ . (The "near the wall" phrasing is the planted misconception.)

Step 3. (c) $r = 4000$ km $> R$ : outside. Theorem 1 — the shell is a point at its centre: $F = \dfrac{GMm}{r^2} = \dfrac{6.67\times10^{-11} \times 6.0\times10^{22} \times 10}{(4.0\times10^6)^2}$ .

Step 4. Numerator $= 4.0\times10^{13}$ ; denominator $= 1.6\times10^{13}$ : $F = \boxed{2.5\ \mathrm{N}}$ . Note $r$ is centre-to-centre — the shell's radius enters only the inside/outside verdict, never the outside formula.

2 · Board level — weight at the bottom of a deep shaft

In the uniform-Earth model ( $R = 6371\ \mathrm{km}$ , surface $g_0 = 9.8\ \mathrm{m/s^2}$ ), what does a $60\ \mathrm{kg}$ student weigh at the bottom of a (fictional) 3000 km shaft — nearly halfway to the centre?

Step 1. Radius: $r = 6371 - 3000 = 3371$ km, so $r/R = 0.529$ .

Step 2. Onion audit: shells above 3371 km silent; inner sphere acts centrally. Linear law: $g = g_0 \times 0.529 = 5.19\ \mathrm{m/s^2}$ .

Step 3. Weight: $W = 60 \times 5.19 = \boxed{311\ \mathrm{N}}$ — down from $588$ N at the surface, despite being 3000 km closer to the centre.

Step 4. Contrast with altitude to feel both theorems: at 3000 km up ( $r = 9371$ km), the inverse square gives $g = 9.8 \times (6371/9371)^2 = 4.5\ \mathrm{m/s^2}$ . Above the surface distance dilutes the pull; below it, disappearing shells do.

3 · Competitive level — nested shells (JEE-Main pattern, ~2.5 minutes)

Two concentric uniform shells have masses $M_1 = 2.0\times10^{22}\ \mathrm{kg}$ (radius $1000\ \mathrm{km}$ ) and $M_2 = 6.0\times10^{22}\ \mathrm{kg}$ (radius $2000\ \mathrm{km}$ ). Find the field at distances (a) $500\ \mathrm{km}$ , (b) $1500\ \mathrm{km}$ , (c) $2500\ \mathrm{km}$ from the common centre.

Step 1. (a) $r = 500$ km: inside both shells — both silent. $g = \boxed{0}$ .

Step 2. (b) $r = 1500$ km: outside shell 1 (acts as central point), inside shell 2 (silent): $g = \dfrac{GM_1}{r^2} = \dfrac{6.67\times10^{-11} \times 2.0\times10^{22}}{(1.5\times10^6)^2} = \dfrac{1.33\times10^{12}}{2.25\times10^{12}} = \boxed{0.59\ \mathrm{N/kg}}$ .

Step 3. (c) $r = 2500$ km: outside both — total mass $8.0\times10^{22}$ kg acts centrally: $g = \dfrac{6.67\times10^{-11} \times 8.0\times10^{22}}{(2.5\times10^6)^2} = \dfrac{5.34\times10^{12}}{6.25\times10^{12}} = \boxed{0.85\ \mathrm{N/kg}}$ .

Step 4. Sanity: the field is discontinuous at each shell wall (jumps as a shell's mass switches on) but uses only enclosed mass everywhere — the onion audit in pure form. Timing note: one sentence per region — "enclosed mass, centralized" — then arithmetic. Two and a half minutes. The distractor answers include ALL the mass at every radius; the audit ("what am I inside of?") is the whole question.

4 · Competitive level — JEE-Advanced stretch: the uniform-field cavity (~4 minutes)

A uniform sphere of density $\rho = 5500\ \mathrm{kg/m^3}$ contains a spherical cavity whose centre lies $d = 1000\ \mathrm{km}$ from the sphere's centre. Show the field inside the cavity is the same everywhere, and compute it.

Step 1. Superposition move: the actual body $\equiv$ (complete uniform sphere of density $\rho$ ) $+$ (sphere of negative density $-\rho$ filling the cavity). Gravity adds linearly, so add the two fields.

Step 2. The key tool — the interior field of a uniform sphere in vector form. From the linear law (3), at displacement $\vec r$ from a sphere's centre: $\vec g = -\dfrac{4}{3}\pi G\rho\,\vec r$ (inward, proportional to position).

Step 3. At any point $P$ in the cavity, with $\vec r$ from the big centre and $\vec r\,'$ from the cavity centre: $\vec g = -\tfrac43\pi G\rho\,\vec r + \tfrac43\pi G\rho\,\vec r\,' = -\tfrac43\pi G\rho\,(\vec r - \vec r\,')$ .

Step 4. But $\vec r - \vec r\,' = \vec d$ , the fixed vector between the two centres — independent of $P$ . The field is uniform across the whole cavity, pointing from the cavity's centre toward the sphere's centre, with magnitude $g = \tfrac43\pi G\rho\,d = \tfrac43\pi \times 6.67\times10^{-11} \times 5500 \times 10^6 = \boxed{1.5\ \mathrm{m/s^2}}$ . Timing note: two ideas carry everything — negative-mass superposition, and the vector form of the linear interior law. With both pre-loaded this is four minutes; discovering either in the hall is fifteen. The result itself ("uniform field in an off-centre cavity") recurs in electrostatics with the identical proof — learn it once, bill it twice.

Exam traps

Hollow shell versus solid sphere. Inside a hollow shell: $g = 0$ everywhere. Inside a solid sphere: $g = g_0 r/R \ne 0$ except at the centre. Options pairing these two are the topic's bread and butter.

Zero force ≠ nothing there. Inside the shell the force vanishes but the potential is a non-zero constant ( $-GM/R$ , next article) — it still costs energy to haul a mass out to infinity. "Field zero, potential constant" is the exam-grade phrasing.

No gravitational shielding. The shell silences its own interior pull; it does not block external gravity. A probe inside a shell still feels the Sun. (Contrast with electrostatic shielding — a conductor does screen; gravity has no negative mass to rearrange.)

Enclosed mass only — and centre-to-centre. At interior radius $r$ , use $M(r)$ , never total $M$ ; outside, measure $r$ from the centre, never the surface. Both slips are printed among every question's options.

Uniformity fine print. The linear interior law assumes uniform density (say so). The theorems themselves need only spherical symmetry — layered density is fine; cubes and rings are not.

The exponent licence. Both theorems are exclusive properties of $1/r^2$ . A question about a hypothetical $1/r^3$ gravity is testing exactly this: the shell would no longer be silent inside, and spheres would stop acting as points.

Which $g$ ? Ratios dominate here (mercifully $g$ -free); when absolute values appear, JEE/NEET declare $g = 10$ or supply $G$ — use what is given.

Origins · 1 minIsaac Newton and how this was discovered

Isaac Newton

1643–1727 · Woolsthorpe & Cambridge

The shell theorem is the proof Newton owed himself. His 1666 Moon test treated the Earth as a point at its centre — defensible at the Moon's distance, but the apple hangs metres from the nearest soil and 12,700 km from the farthest, and that needed proof. Historians read Newton's twenty-year publication delay partly as the wait for this theorem, which finally appears in the Principia (1687) as Propositions LXX and LXXI of Book I: the interior cancellation, argued with exactly the double-cone construction you met above, and the exterior point-equivalence. With them, "centre-to-centre" stopped being a hopeful approximation and became mathematics. The irony: Newton possessed the calculus that makes the exterior proof a one-page exercise, but dressed everything in Euclidean geometry so his readers would trust it.

History

1666

— the Moon test works, but treating Earth as a point remains an unproven convenience; Newton shelves publication.

1685

— the breakthrough proofs: a sphere gravitates exactly as a point, and a shell is silent inside.

1687

— Principia, Book I, Propositions LXX–LXXI: both theorems in Euclidean dress; the apple-and-Moon argument closes honestly.

1798

— Cavendish uses point-equivalence to turn a bench-top

G

into the mass and density of the Earth.

1911

— laboratory "inside the shell" tests begin — the interior zero becomes a precision probe of the inverse-square exponent.

1970

s — rotation curves of galaxies, read through enclosed-mass logic

g = GM(r)/r^2

, reveal that visible matter is not enough: dark matter enters physics through this article's equation (2).

Going further

Texts. NCERT Physics Class 11, Ch. 8 (the boxed shell statements) · HC Verma Vol. 1, Ch. 11 · Halliday–Resnick–Walker Ch. 13 · Kleppner & Kolenkow Ch. 9 for the honest ring integral.

Provenance. All results are standard canon. A verifier should recompute: the double-cone ratio audit ( $s^2$ against $1/s^2$ ); $g(r) = g_0 r/R$ from $M(r) = M(r/R)^3$ ; example 1's $2.5$ N; example 2's $311$ N (and the $4.5\ \mathrm{m/s^2}$ altitude contrast); example 3's $0.59$ and $0.85$ N/kg; example 4's $\frac43\pi G\rho d = 1.5\ \mathrm{m/s^2}$ .

Prerequisites. Newton's Law of Universal Gravitation · The Gravitational Field & Variation of g. Leads to. Gravitational Potential Energy & Potential · The Tunnel Through the Earth (archetype) · The Shell Theorem (UG deep dive).

Why doesn't the nearer wall of the shell win? It's so much closer.

Because "closer" buys strength but costs quantity, at exactly compensating rates. Look along any direction from your off-centre position: the nearby wall patch subtends the same cone as the far patch behind you, but the far patch's area is larger by $(s_2/s_1)^2$ while its per-kilogram pull is weaker by the same $(s_1/s_2)^2$ . Product: dead equal. The inverse-square law is the unique power law with this property — the theorem is less a fact about shells than a fingerprint of the exponent 2.

Is the outside theorem an approximation that gets better with distance?

No — that is what makes it remarkable. Any compact mass looks pointlike from far away; the theorem says a uniform sphere is exactly pointlike from any exterior distance, even one metre above its surface. The proof (the ring-slicing integral — university calculus, flagged honestly as the deep dive) shows the near-side excess and far-side deficit cancel perfectly at every $r > R$ . For exams: quote it as exact, cite the shell theorem by name, and never add "approximately."

Does the shell theorem mean gravity is zero inside the Earth?

No — the Earth is solid, not hollow. At interior radius $r$ , only the shells above you (radii $> r$ ) switch off; the entire inner sphere still pulls, acting as a central point of mass $M(r)$ . Result for uniform density: $g = g_0\,r/R$ — reduced, not zero, until the exact centre. The genuinely zero-force region exists only in a hollow body's cavity. Exams pair these two situations deliberately; audit "hollow or solid?" before anything else.

Can a shell shield me from outside gravity, like a Faraday cage shields electric fields?

No. The theorem is about the shell's own gravity; external fields pass through unmodified, and superposition guarantees it — the shell's contribution inside is zero, so the total is just the external source's field, untouched. Electrostatic shielding works because conductors carry mobile charges of both signs that rearrange to cancel interior fields; gravity has only one sign of "charge" and nothing to rearrange. No gravity shield exists — a fact worth one conceptual mark and several science-fiction plots.

What if the density isn't uniform — does the real, layered Earth break the theorems?

The theorems survive anything spherically symmetric: density may vary with radius (iron core, rocky mantle) so long as it doesn't vary with direction. Each thin shell is still uniform, so from outside, layered Earth is still a perfect point mass — Cavendish's calculation never cared about the core. What layering does break is the linear interior law, which assumed $M(r) \propto r^3$ : in the real Earth, $g$ actually creeps up slightly as you descend through light mantle toward the dense core, peaking around $10.7\ \mathrm{m/s^2}$ before dying to zero at the centre. State "uniform model" and both truths coexist.

Why is the field inside an off-centre cavity uniform? That seems too neat.

It falls out of two facts stacking cleanly: the interior field of a uniform sphere is linear in position ( $\vec g = -\frac43\pi G\rho\,\vec r$ ), and gravity superposes. Cavity = big sphere + negative-density sphere; at any cavity point the two linear fields subtract, and position-dependence cancels term by term, leaving $-\frac43\pi G\rho\,\vec d$ — set by the fixed centre-to-centre offset alone. Move within the cavity and nothing changes: a manufactured pocket of perfectly parallel gravity. The identical construction with charge density answers a famous electrostatics problem — one proof, two chapters.

Why do physicists still test the "zero inside" today?

Because it is one of the sharpest available probes of the inverse-square law. Any deviation — a fifth force, large extra dimensions leaking gravity at short range, a massive graviton — would spoil the double-cone cancellation and leave a tiny residual force inside a shell. Torsion-balance experiments (the Eöt-Wash program and successors) look for exactly that residue at sub-millimetre scales and keep finding zero, pushing the exponent ever closer to exactly 2. Newton's 1685 proposition is a live experimental frontier.