Length Contraction

Ask a cosmic-ray muon how tall Mount Washington is and it will tell you: 190 metres. Frisch and Smith, who counted those muons in 1963, measured the drop from the summit to their detector in Cambridge as $1907\ \mathrm{m}$ . Neither is lying. The muon arrives at $\beta = 0.995$ , so $\gamma = 10.0125$ , and the mountain — which is the thing that moves, in the muon's frame — is squashed to $1907/10.0125 = 190.5\ \mathrm{m}$ . Watch what that does to the survival count. The muon's mean lifetime is $\tau_0 = 2.197\ \mu\mathrm{s}$ , and in its own frame nothing runs slow — its clock ticks one second per second, as every clock does at rest. But it has only $190.5\ \mathrm{m}$ of contracted mountain to cross, which at $0.995c$ takes $0.6385\ \mu\mathrm{s}$ , less than a third of a lifetime, so its survival probability is $e^{-0.291} = 0.748$ . Now redo the sum in the laboratory, where the mountain is an ordinary $1907\ \mathrm{m}$ and nothing is contracted: the flight lasts $6.393\ \mu\mathrm{s}$ , but the muon's clock is dilated by $\gamma$ , so it ages only $6.393/10.0125 = 0.6385\ \mu\mathrm{s}$ . Survival $0.748$ — the same number, from two arithmetics that share no step. That is not luck. It is a theorem, and this article proves it — deriving $L = L_0/\gamma$ from the Lorentz transformation and locating the one place the physics enters: the demand that you mark both ends of a moving object at the same time. Proper Length and Consistency of Time Dilation and Length Contraction carry the story onward.

Key takeaways

An object of proper length $L_0$ — its length in its own rest frame — measures $L = L_0/\gamma = L_0\sqrt{1-\beta^2}$ where it moves at $\beta = v/c$ along its length. Transverse dimensions are untouched.
The whole effect lives in one word: simultaneously. A length is the gap between two end-markings made at one instant, and frames disagree about which markings are simultaneous.
It falls out of the transformation in two lines: put $\Delta t = 0$ into $x' = \gamma(x - vt)$ and the $vt$ terms cancel, leaving $L_0 = \gamma L$ .
It is reciprocal, harmlessly. A $100\ \mathrm{m}$ train at $\beta = 0.6$ measures $80.0\ \mathrm{m}$ from the platform, while the train finds the platform's rulers short by the same $1/\gamma$ .
The frame-consistency theorem: a muon's survival $e^{-\tau/\tau_0}$ depends only on its proper time, which is frame-independent. The lab (dilated clock, full mountain) and the muon (normal clock, contracted mountain) are therefore obliged to agree — and they do, exactly: $0.748$ .
Key idea: length contraction is not a squeezing of matter but a fact about measurement — mark the ends of a moving rod at one instant and you get $L_0/\gamma$ , because "one instant" is not the same instant for everyone. Dilation and contraction are one fact seen from two frames, and they agree to the last digit because they must.

The core insight

To measure something that is moving you must locate both of its ends at once — and "at once" is exactly what relativity has taken away. Mark the ends simultaneously in your frame and those markings are not simultaneous in the rod's frame: the leading end was marked first, and the rod ran on before the trailing end was marked. Your two marks therefore span less than the rod's own length, by exactly $1/\gamma$ . Nothing has been done to the rod: in its own frame it is as long as it ever was.

Where this lives today

Contraction is routine engineering. A $7\ \mathrm{TeV}$ LHC proton has $\gamma = E/mc^{2} = 7000/0.938272 = 7460.5$ , so in the laboratory it is flattened along the beam to $1.34\times10^{-4}$ of its rest diameter; the gold nuclei at RHIC ( $100\ \mathrm{GeV}$ per nucleon, $\gamma \approx 107$ ) arrive as thin discs, not spheres. That is an input, not a curiosity: the initial energy density of the quark–gluon plasma depends on two pancakes overlapping in a sheet, and a model built from round nuclei gets the physics wrong. And the misconception has finally been photographed. In 2025 a group at TU Wien built an optical analogue — picosecond laser pulses and a gated camera, reassembled so that the effective speed of light was about $2\ \mathrm{m/s}$ — and photographed ordinary objects "moving" at a large fraction of it. The images show what Terrell and Penrose predicted: the sphere still a sphere, the cube apparently rotated. The contraction is as real as ever; the photograph declines to show it.

The physical picture

Measuring something at rest is so easy that nobody examines the procedure: lay a ruler alongside, read one end, walk over, read the other. The walk takes time and it does not matter, because the object is going nowhere. Now let it move. Read the front end at noon and the back end a second later and the answer is nonsense — by dawdling you can obtain any number you like, including negative ones. The only honest procedure is to record both ends at the same instant: a demand about simultaneity, which is precisely the thing that stopped being absolute.

Interactive Simulation

Length Contraction — drag β to watch a moving rod's measured length shrink as L = L₀/γ, and see why (both ends are marked at one instant in the observer's frame).

Figure

Measuring a moving rod. In the ground frame S both ends are marked at the SAME instant, 80.0 m apart. The same two events, replayed in the rod's frame S′, are 200 ns apart — front end first — and the ground slides 36.0 m in between, so the marks sit 64.0 m apart on the contracted ground: 64.0 + 36.0 = 100.0 m, the full proper length

The rod is not compressed, stressed or heated; in its own frame its atoms sit where they always sat. You simply sliced spacetime along a different "now" from the one the rod uses — and a rod is a strip in spacetime. Slice a strip at an angle and the cross-section comes out shorter. Which brings us to the number this article is really about.

Figure

PUNCHLINE — the frame-consistency theorem in numbers. LEFT, the laboratory: a full 1907 m mountain, a 6.393 μs flight, a clock dilated by γ = 10.0125, so the muon ages 0.6385 μs. RIGHT, the muon: no dilation at all, but a mountain contracted to 190.5 m and crossed in 0.6385 μs. Both roads end on 0.29 lifetimes and a survival of 0.748

The core equations

Standard configuration: $S'$ moves at $+v$ along the $x$ -axis of $S$ , axes parallel, origins coinciding at $t = t' = 0$ , with $\beta = v/c$ and $\gamma = 1/\sqrt{1-\beta^2}$ . A rod lies at rest along the $x'$ -axis; its proper length $L_0$ is the separation of its ends in $S'$ , where it is at rest and there is no hurry.

L = \frac{L_0}{\gamma} = L_0\sqrt{1-\beta^{2}}, \qquad L_\perp = L_{0\perp} \qquad (1)

where: $L$ — the length measured where the rod moves at $v$ ; $L_0$ — the proper length; $L_\perp$ — any dimension perpendicular to the motion, unchanged. Physical reading: since $\gamma \geq 1$ , the proper length is the largest any frame reports; a moving sphere is an oblate ellipsoid, flattened along its velocity and untouched across it. This is not an extra postulate. It is two lines of the Lorentz transformation, whose time line does the real damage.

x' = \gamma\,(x - v t), \qquad t' = \gamma\!\left(t - \frac{v x}{c^{2}}\right) \qquad (2)

where: $(x, t)$ — an event in $S$ ; $(x', t')$ — the same event in $S'$ ; $c = 299\,792\,458\ \mathrm{m/s}$ . Physical reading: the second equation makes $t'$ depend on where you are, so two events sharing a $t$ but not an $x$ come out with different $t'$ . That is why "mark both ends at once" is a frame-dependent instruction.

\Delta x' = \gamma\,(\Delta x - v\,\Delta t) \;\xrightarrow{\ \Delta t\, =\, 0\ }\; L_0 = \gamma L, \qquad \Delta t' = -\,\frac{\gamma v L}{c^{2}} = -\,\frac{v L_0}{c^{2}} \qquad (3)

where: $\Delta t = 0$ — the ends marked simultaneously in $S$ ; $\Delta x = L$ — the length that measurement returns; $\Delta t'$ — the time between those markings in the rod's frame. Physical reading: the measurement gives $L_0/\gamma$ because the markings are $vL_0/c^{2}$ apart in the rod's frame, leading end first. Note the $\gamma$ cancels — the one from the time transform against the $1/\gamma$ of the contracted separation. Finally, the quantity the muon cares about: it decays by an exponential law read on its own clock.

P_{\text{survive}} = e^{-\tau/\tau_{0}}, \qquad \tau = \frac{\Delta t}{\gamma} = \frac{\Delta x}{\gamma v} \qquad (4)

where: $\tau_0 = 2.197\ \mu\mathrm{s}$ — the mean lifetime at rest; $\tau$ — the proper time accumulated crossing the mountain; $\Delta t = \Delta x/v$ — the lab flight time over the lab height $\Delta x$ . Physical reading: the two forms of $\tau$ are the whole theorem — $\Delta t/\gamma$ is the lab's story (full mountain, dilated clock), $\Delta x/(\gamma v)$ the muon's (contracted mountain, normal clock). Same expression, so the survival cannot be frame-dependent.

Where the results come from

Derivation AMarking both ends at once — and why the answer is L₀ over gamma

The goal is to extract $L = L_0/\gamma$ from the transformation without ever asserting it, and to see where the physics enters. It enters at one point only: the instruction "mark both ends at the same time", a condition on the time coordinates of two events and therefore not frame-neutral. Everything else is algebra — and Step 3 shows that dropping that condition returns any answer you like.

Step 1. Say what a length measurement is. The rod is at rest in $S'$ with ends at fixed $x'_A$ , $x'_B$ , so $L_0 = x'_B - x'_A$ . An observer in $S$ , watching it stream past at $v$ , must define its length as $L \equiv x_B(t) - x_A(t)$ with the same $t$ in both terms. Why this is the crux and not the preamble: for a moving object, "where the ends are" means nothing until you say "when".

Step 2. Transform the two marking events — chalk on a platform, and events are what the transformation converts. Equation (2) gives $x'_A = \gamma(x_A - vt)$ , $x'_B = \gamma(x_B - vt)$ , same $t$ in both. Subtract: the identical $-\gamma vt$ terms cancel, leaving $L_0 = \gamma L$ , that is $\boxed{L = L_0/\gamma = L_0\sqrt{1-\beta^{2}}}$ . Why the $\gamma$ lands where it does: we boosted a separation from the frame that made the marks into the frame that owns the rod, and that direction carries a $\gamma$ — the rod's frame reports the LARGER number.

Step 3. Show the simultaneity condition was load-bearing. Keep $\Delta t$ free: $\Delta x = L_0/\gamma + v\,\Delta t$ . Mark the trailing end later and the "length" grows; at $\Delta t = L_0(1 - 1/\gamma)/v$ it reaches the full $L_0$ ; wait longer and it exceeds $L_0$ ; mark them in the wrong order and it goes negative. Why this settles it: without "at the same time" there is no such thing as the length of a moving object — and with it there is exactly one.

Step 4. Ask the rod's frame what it made of our marks. From (3), $\Delta t' = -\gamma vL/c^2 = -vL_0/c^2$ , and the sign says the leading end was marked first: the ground observer marked the far end, waited $vL_0/c^2$ , then marked the near end — by which time the ground had slid onward. Why the frames now agree about everything that matters: each can compute exactly what the other did, and they differ on the ordering of two events and nothing else. *Verified in verify_aq-sr-length-contraction.py: sympy boosts the two marking events, imposes $\Delta t = 0$ and returns $\Delta x' = \gamma\,\Delta x$ identically in $v$ ; the free- $\Delta t$ family is confirmed to reach $L_0$ , exceed it and change sign; the offset simplifies to exactly $-vL_0/c^2$ ; and $\gamma(0.6) = 1.25$ , $L = 80.0\ \mathrm{m}$ , $\Delta t' = 200\ \mathrm{ns}$ , slide $36.0\ \mathrm{m}$ , marks $64.0\ \mathrm{m}$ all reproduce.*

Go deeper · Derivation B · 2 minThe frame-consistency theorem — the muon must survive equally in both frames, and does

Derivation BThe frame-consistency theorem — the muon must survive equally in both frames, and does

This is the article's central claim, stronger than the usual remark that the two effects "conspire nicely". They cannot fail to agree, for a structural reason: whether a muon reaches the ground is a fact about the world — the detector clicks or it does not — so every frame must predict the same probability. That probability is $e^{-\tau/\tau_0}$ , governed by the muon's proper time; proper time is an invariant; therefore the answer is an invariant. What follows shows the two computations landing on one expression, though one uses dilation with no contraction and the other the reverse.

Step 1. State the decay law properly. A muon at rest decays with probability $e^{-t/\tau_0}$ , $\tau_0 = 2.197\ \mu\mathrm{s}$ on its own clock, and it cannot know what a laboratory is doing. So along any worldline the survival probability is $e^{-\tau/\tau_0}$ , with $\tau$ the elapsed proper time. Why this is load-bearing: an unstable particle is a clock, and its rate depends on nothing but the proper time it has accumulated — the clock hypothesis, which Proper Time develops.

Step 2. Compute $\tau$ the laboratory's way. There the mountain has its full height $h = 1907\ \mathrm{m}$ — measured with a mountain-at-rest ruler, so no contraction appears — and the muon crosses it at $v = \beta c$ in $\Delta t = h/v$ . Its clock is dilated, so it accumulates $\tau_{\text{lab}} = \Delta t/\gamma = h/(\gamma v)$ . Why all the relativity sits in the clock here: this is the frame at rest with respect to the mountain, and Time Dilation owns this half of the story.

Step 3. Compute $\tau$ the muon's way. The muon is at rest, so its clock is not dilated and coordinate time is proper time. What moves is the mountain, at $v$ , so its height contracts to $h' = h/\gamma$ by equation (1) and sweeps past in $\tau_{\text{muon}} = h'/v = h/(\gamma v)$ . Why there is nowhere for an error to hide: the frames use disjoint machinery — a $\gamma$ on the clock and no contraction, versus a $1/\gamma$ on the mountain and no dilation.

Step 4. Compare. $\tau_{\text{lab}} = h/(\gamma v) = \tau_{\text{muon}}$ , identically, for every $h$ , $v$ and $\gamma$ — so the two survival probabilities are equal, necessarily. Deeper: $\tau$ is the invariant interval along the worldline, $c^{2}\tau^{2} = c^{2}\Delta t^{2} - \Delta x^{2}$ ; feed it the lab's $(h/v,\ h)$ and it returns $h/(\gamma v)$ , feed it the muon's $(\tau, 0)$ and it returns $\tau$ . Why this is a theorem and not an arithmetic accident: dilation and contraction are not two effects that happen to cancel — they are two coordinate descriptions of one invariant, and that invariant is what the muon's own clock reads.

Step 5. Put in the numbers. $\beta = 0.995 \Rightarrow \gamma = 10.0125$ . Lab: $\Delta t = 1907/(0.995c) = 6.393\ \mu\mathrm{s}$ , so $\tau = 6.393/10.0125 = 0.6385\ \mu\mathrm{s}$ . Muon: $h' = 190.5\ \mathrm{m}$ , so $\tau = 190.5/(0.995c) = 0.6385\ \mu\mathrm{s}$ . Both give $\tau/\tau_0 = 0.291$ and $e^{-0.291} = \boxed{0.748}$ . Without relativity the muon would age the full $6.393\ \mu\mathrm{s}$ and survive with probability $e^{-2.910} = 0.0545$ , fourteen times smaller. Why the experiment is decisive rather than delicate: that is the difference between 31 muons an hour and 421, and Frisch and Smith could count muons. *Verified in verify_aq-sr-length-contraction.py: sympy proves $h/(\gamma v) - (h/v)/\gamma = 0$ identically and that the interval on the lab coordinates returns $h/(\gamma v)$ ; numpy reproduces $\gamma = 10.0125$ , $6.393\ \mu\mathrm{s}$ , $190.5\ \mathrm{m}$ , $\tau = 0.6385\ \mu\mathrm{s}$ by both routes, $\tau/\tau_0 = 0.2906$ , survivals $0.7478$ and $0.0545$ , and rates $563 \to 421/\mathrm{hr}$ and $563 \to 31/\mathrm{hr}$ .*

Limiting cases

Case	Result	Meaning
At rest, $\beta = 0$	$\gamma = 1$ , $L = L_0$	the proper length is the maximum any frame reports
Low speed, $\beta \ll 1$	$L \approx L_0(1 - \tfrac12\beta^{2})$	second order: at the Earth's orbital speed ( $\beta = 9.94\times10^{-5}$ ) a metre stick loses $4.9\ \mathrm{nm}$
Transverse	$L_\perp = L_{0\perp}$	unchanged — else frames could disagree about a rod passing through a ring
Half length	$L = L_0/2$ at $\beta = 0.866$	$86.6\%$ of $c$ before a metre stick reads $50\ \mathrm{cm}$
Ultra-relativistic, $\beta \to 1$	$L \to 0$	a $7\ \mathrm{TeV}$ LHC proton has $\gamma = 7460.5$ : in the lab, $1.34\times10^{-4}$ of its rest diameter
Galilean limit, $c \to \infty$	$L \to L_0$	length is absolute again — and so is "at the same time". The two facts are one

Worked examples

1 · The muon's-eye view of Mount Washington

Frisch and Smith counted $563$ muons per hour at the summit of Mount Washington in 1963, then took their detector to Cambridge, $1907\ \mathrm{m}$ lower. The muons arrive at $\beta = 0.995$ with a rest lifetime $\tau_0 = 2.197\ \mu\mathrm{s}$ . Working <strong>entirely in the muon's frame</strong>, where no clock is dilated, predict the rate in Cambridge.

Step 1. The Lorentz factor: $\gamma = 1/\sqrt{1 - 0.995^{2}} = 1/\sqrt{0.009975} = \boxed{10.0125}$ . It is the only $\gamma$ we need — and it will not be applied to a clock, because in the muon's frame its clock is the one clock in the universe that is certainly not running slow.

Step 2. Contract the mountain: here the mountain moves, at $0.995c$ , so $h' = h/\gamma = 1907/10.0125 = \boxed{190.5\ \mathrm{m}}$ . Mount Washington, to a muon, is a small hill.

Step 3. Cross it: $\tau = h'/v = 190.5/(0.995 \times 2.998\times10^{8}\ \mathrm{m/s}) = \boxed{0.6385\ \mu\mathrm{s}}$ , which is $\tau/\tau_0 = 0.6385/2.197 = 0.291$ of a lifetime.

Step 4. Survive: $P = e^{-0.291} = \boxed{0.748}$ , so of $563$ muons an hour at the summit, $563 \times 0.748 = \boxed{421\ \mathrm{per\ hour}}$ should reach Cambridge; Frisch and Smith measured about $412$ . With no relativity — the full $6.393\ \mu\mathrm{s}$ ageing an undilated clock — the prediction is $563\,e^{-2.910} = 31$ per hour, against their published classical expectation of $27$ . The small gaps from the published rates are not fudge, and the reason matters: the muons decelerate by ionization loss on the way down, so $\gamma$ falls from $\approx 10.2$ at the summit to $\approx 6.8$ at Cambridge. That is why their measured dilation factor of $8.8 \pm 0.8$ must be read against the effective factor for their detection geometry, which the paper gives as $\boxed{8.4 \pm 2}$ — not against the single-speed $\gamma = 10.01$ used in this idealized calculation. Compared correctly, the agreement is excellent. (Reading $8.8$ against $10.01$ as a leftover discrepancy is the most commonly botched fact about this experiment; inverting the counts $563 \to 412$ gives $\gamma \approx 9.3$ , between the two, exactly as a decelerating beam requires.) Notice, finally, what never appeared above: no time dilation anywhere.

2 · The 100-metre train, and who is short

A train of proper length $L_0 = 100\ \mathrm{m}$ runs through a station at $\beta = 0.6$ . What length does the platform measure — and how can the passengers simultaneously insist that the platform's rulers are the short ones?

Step 1. $\gamma = 1/\sqrt{1 - 0.36} = 1.25$ , so the platform measures $L = 100/1.25 = \boxed{80.0\ \mathrm{m}}$ : its two chalk marks, laid down at one platform instant, are $80.0\ \mathrm{m}$ apart.

Step 2. Board the train. Those marks were not simultaneous here: by equation (3), $\Delta t' = vL_0/c^{2} = (0.6)(100)/(2.998\times10^{8}) = \boxed{200\ \mathrm{ns}}$ , front of the train marked first.

Step 3. Follow the money. In those $200\ \mathrm{ns}$ the platform — moving backwards at $0.6c$ here — slides $v\,\Delta t' = \boxed{36.0\ \mathrm{m}}$ . The two chalk marks, at rest on that platform, are separated by $L/\gamma = 80.0/1.25 = \boxed{64.0\ \mathrm{m}}$ in the train's frame, the platform's own marks being contracted here. And $64.0 + 36.0 = \boxed{100.0\ \mathrm{m}}$ — the train's full proper length, exactly.

Step 4. So the reciprocity is real and harmless: each frame finds the other's rulers short by $1/\gamma = 0.8$ , with no contradiction, because the two answer different questions using different pairs of events. There is no frame-independent fact of the form "the train is really $80\ \mathrm{m}$ long". (Forcing the issue — a $100\ \mathrm{m}$ pole run into an $80\ \mathrm{m}$ barn with both doors slammed — is the business of The Pole-and-Barn Paradox.)

Physics of the portrait

The portrait here is an idea, not a person, and deliberately so: no single physicist owns this one. Across seventy years the formula never changed. What it meant changed three times.

Origins · 2 minLength contraction and how this was discovered

Length contraction

1889–1959 · FitzGerald · Lorentz · Einstein · Terrell and Penrose

The idea arrived as a rescue. In 1889 George Francis FitzGerald (1851–1901) sent the journal Science a note of a few hundred words, "The Ether and the Earth's Atmosphere", proposing that bodies moving through the ether change length by an amount depending on the square of the ratio of their speed to that of light — just enough to hide the ether wind from Michelson and Morley. Hendrik Antoon Lorentz (1853–1928) reached the same idea independently in 1892 and gave it a reason: if the forces holding matter together are electromagnetic, and the fields of moving charges flatten, matter should flatten too. His 1895 Versuch introduced the "local time" $t' = t - vx/c^{2}$ ; by 1904 he had the full transformation. What lifts this above a fudge is a fact the Lorentz–FitzGerald Contraction Hypothesis proves: setting the two interferometer arm times equal gives a linear equation in the arm length, so it has exactly one root — $\sqrt{1-\beta^{2}}$ is the only factor that nulls the experiment. Then in 1905 Albert Einstein obtained the same factor from the relativity principle and the constancy of $c$ , with no ether — and made it reciprocal, which no ether theory could. Finally, in 1959, James Terrell and Roger Penrose independently showed that no camera ever records it: light-travel time from the far side of a fast object rotates its image instead of squashing it. The contraction is real, and it is invisible.

History

1889

— FitzGerald, in a note to Science titled "The Ether and the Earth's Atmosphere", suggests that moving bodies change length by an amount depending on the square of their speed divided by that of light.

1892

— Lorentz, not knowing of the letter, proposes an almost identical contraction to save a stationary ether from the Michelson–Morley null result.

1895

— Lorentz's Versuch introduces the local time t′ = t − vx/c², correct to first order in β.

1904

— Lorentz publishes the transformation valid to all orders; the contraction becomes a consequence, not an assumption.

1905

— Einstein derives L = L₀/γ from the two postulates alone, with no ether; the effect is now reciprocal between frames.

1959

— Terrell ("Invisibility of the Lorentz Contraction", Physical Review 116, 1041) and Penrose, independently, show that a photograph of a fast object records a rotation, not a contraction.

1963

— Frisch and Smith count cosmic-ray muons at the Mount Washington summit (1917 m above sea level) and 1907 m below it in Cambridge; the contracted mountain and the dilated clock give the same survival, and the data agree.

2025

— Schattschneider and co-workers photograph the Terrell–Penrose rotation, using picosecond laser pulses to bring the effective speed of light down to about 2 m/s.

Going further

Texts. Taylor and Wheeler, Spacetime Physics, 2nd ed., Ch. 3 · Rindler, Introduction to Special Relativity, 2nd ed., §§2.6–2.7 · Griffiths, Introduction to Electrodynamics, 4th ed., §12.1.2 · D. H. Frisch and J. H. Smith, American Journal of Physics 31 (1963) 342 · J. Terrell, Physical Review 116 (1959) 1041.

Provenance. The contraction $L = L_0/\gamma$ , its simultaneity-based derivation and the reciprocity discussion are standard (Taylor and Wheeler Ch. 3; Rindler §2.7; Griffiths §12.1.2). Every number above is recomputed from scratch in `verify_aq-sr-length-contraction.py` from the anchors $c = 299\,792\,458\ \mathrm{m/s}$ , $\tau_0 = 2.197\ \mu\mathrm{s}$ , $h = 1907\ \mathrm{m}$ , $\beta = 0.995$ — symbolically ( $\Delta x' = \gamma\,\Delta x$ at $\Delta t = 0$ ; the offset $-vL_0/c^2$ ; the frame-consistency identity $h/(\gamma v) = (h/v)/\gamma$ and its interval form) and numerically (every $\gamma$ , time, survival and rate, the train's $80.0/200/36.0/64.0$ , and the LHC and RHIC factors). The published Frisch–Smith figures ( $\approx 412$ per hour observed, $27$ expected classically, dilation factor $8.8 \pm 0.8$ ) must be compared with the effective factor for their detection geometry, which the paper itself gives as $8.4 \pm 2$ — not with the single-speed $\gamma = 10.01$ used above. The muons decelerate by ionization loss on the way down ( $\gamma$ falls from $\approx 10.2$ at the summit to $\approx 6.8$ at Cambridge), so the descent samples a range of $\gamma$ ; inverting the published counts ( $563 \to 412$ ) returns $\gamma \approx 9.3$ , sitting between the two exactly as a decelerating beam requires. Compared correctly there is no discrepancy at all. Base correction: the dry base carried a full pole-and-barn derivation, which belongs to The Pole-and-Barn Paradox; it is removed and replaced by the frame-consistency theorem, which this article owns.

Builds on. The Two Postulates of Special Relativity · The Lorentz Factor Gamma · Relativity of Simultaneity · The Lorentz Transformation Equations · Lorentz–FitzGerald Contraction Hypothesis. Leads to. Proper Length · Time Dilation · Consistency of Time Dilation and Length Contraction · The Invariant Spacetime Interval · The Pole-and-Barn Paradox.

Is the rod really shorter, or does it just look shorter?

Neither phrasing survives contact with the physics, and the second is worse: it does not look shorter — a photograph shows it rotated. And "really" smuggles in a frame-independent fact about the rod's length. There is none: length is a relation between an object and a frame, exactly as velocity is. In its own frame the rod has its proper length, atoms undisturbed; in yours, a measurement marking both ends at one of your instants returns $L_0/\gamma$ . Both are correct — they measure different pairs of events.

Which γ do I use? I keep getting L₀γ instead of L₀ over γ.

Anchor on one sentence: the proper length is the biggest. $L_0$ is measured where the object is at rest, so every other frame gets something smaller — $L_0/\gamma$ , never $L_0\gamma$ . If your algebra hands you $L_0\gamma$ you have boosted the wrong way, transforming a separation that is simultaneous in the rod's frame, which is not a length measurement in the other frame. The companion rule: proper time is the smallest, so $\Delta t = \gamma\,\Delta\tau$ is bigger. Proper length maximal, proper time minimal — that asymmetry is the minus sign in the interval.

Why does only the length along the motion contract? Why not sideways too?

Because a transverse change would let two frames disagree about an absolute fact. Fire a rod lengthwise through a ring whose proper radius slightly exceeds the rod's half-length: either it passes cleanly or it scrapes, and the scrape mark is a fact both frames must accept. If transverse dimensions changed with speed, relativity would make the situation symmetric — each frame saying the other's object shrank — so one would predict a clean pass and the other a scrape. The only change consistent with the symmetry is no change at all, and the light-clock derivation of $\gamma$ silently assumes it.

If everything contracts, including my ruler, how could contraction ever be measured?

The ruler you measure with is not moving with the object. A co-moving ruler contracts along with the rod and duly returns $L_0$ — correctly, because in that frame the rod really is $L_0$ long. The contraction appears only when the apparatus is at rest in the frame where the object moves: chalk marks on a platform, synchronized detectors, a radar echo. Contrast the ether hypothesis, where the contraction was absolute and every ruler shrank together, making it unobservable in principle. In relativity the effect is relational — which is why it is measurable.

Does the contraction happen gradually while the rod accelerates? Does that not stress it?

Take an object at rest in $S$ and accelerate it to $v$ . In $S$ its measured length does change over time, from $L_0$ to $L_0/\gamma$ , and here real physics can enter, because how the material responds depends on how you push it. Push every atom in a coordinated way (Born-rigid acceleration) and it arrives unstressed; grab one end and drag it and you will stress it. But that is ordinary mechanics, and the final inertial state is the same either way: length contraction involves no forces, no strain and no history.

Where does the contraction come from physically? Saying it follows from the transformation feels like a dodge.

Then follow the chalk. In the ground frame you mark both ends of the train at once and the marks land $80\ \mathrm{m}$ apart. Board the train and replay those two events: the front mark went down first; $200\ \mathrm{ns}$ passed; the rear mark went down — and in that $200\ \mathrm{ns}$ the platform slid $36\ \mathrm{m}$ backwards. So the marks, at rest on the platform, are $64\ \mathrm{m}$ apart in the train's frame, and they landed on the ends of a $100\ \mathrm{m}$ train because $64 + 36 = 100$ . The contraction is the relativity of simultaneity, seen with a ruler in your hand.