physics2u
PhysicsMathsChemistryBiology
Tier
⌕ Search ⌘K
Derivation

Length Contraction

Statement

The length of a rigid object measured along its direction of motion is shorter than its proper length \(L_0\) (the length measured in the object's rest frame) by the Lorentz factor: \(L = L_0/\gamma\), where \(\gamma = 1/\sqrt{1 - v^2/c^2}\) and \(v\) is the relative speed of the two inertial frames.

Why it matters

Length contraction is one of the two elementary kinematic consequences of the Lorentz transformation (the other being time dilation), and together they encode how space and time coordinates mix between inertial observers. Without it, special relativity is internally inconsistent: it is precisely what lets a fast muon "see" a thin atmosphere while a ground observer "sees" a slow-ticking muon, so that both agree the muon reaches the ground.

It also disciplines the intuition that "length" is an absolute property of an object. A ruler has no single length; it has a length in each frame, and the answer depends on how the two ends are located simultaneously in that frame. This makes length contraction a direct, measurable fingerprint of the relativity of simultaneity.

Assumptions
The Lorentz transformation holds between the two inertial frames.If the transformation between frames were Galilean, lengths would be invariant and no contraction would occur; the entire result is a corollary of the Lorentz coordinate mixing.
The object is rigid and its rest frame is inertial (unaccelerated during measurement).If the object accelerates, "its length" is frame-and-instant dependent in a subtler way (Born rigidity; there are no rigid bodies in the Newtonian sense), and \(L_0\) is not a single well-defined number.
Length in a frame is defined as the coordinate distance between the two ends recorded at one instant of that frame's time (\(\Delta t' = 0\)).Drop the simultaneity requirement and you can obtain any answer at all — recording the two ends at different times in the moving frame lets the object drift between readings, so the "length" becomes arbitrary.
Measurement is along the direction of relative motion.Transverse dimensions are unaffected (they satisfy \(y'=y,\ z'=z\)); applying the \(1/\gamma\) factor to them is simply wrong.
Derivation

Let frame \(S\) be the rest frame of the rod, with the rod lying along the \(x\)-axis; frame \(S'\) moves at speed \(v\) in the \(+x\) direction relative to \(S\), so in \(S'\) the rod moves at \(-v\). We compute the length the \(S'\) observer measures.

1
\[ x' = \gamma\left(x - vt\right), \qquad t' = \gamma\left(t - \frac{vx}{c^2}\right), \qquad \gamma = \frac{1}{\sqrt{1 - v^2/c^2}} \]
Prior result: the Lorentz transformation from the postulates gives the map from \(S\) to \(S'\). Everything follows from these two lines. A
2
\[ L_0 \equiv x_2 - x_1 \quad\text{(endpoints fixed for all }t\text{ in }S) \]
Definition of proper length: in the rod's rest frame the two ends sit at constant coordinates \(x_1, x_2\), so their difference is unambiguous and time-independent. A
3
\[ L \equiv x'_2 - x'_1 \quad\text{measured at a single instant } t'_1 = t'_2,\ \text{i.e. } \Delta t' = 0 \]
Operational definition of length in \(S'\): the ends must be located simultaneously in \(S'\). This step is where the relativity of simultaneity enters and is the crux of the whole result. B
4
\[ \Delta t' = \gamma\left(\Delta t - \frac{v\,\Delta x}{c^2}\right) = 0 \;\Longrightarrow\; \Delta t = \frac{v\,\Delta x}{c^2} = \frac{v L_0}{c^2} \]
Apply the time transformation to the coordinate differences of the two measurement events and impose \(\Delta t' = 0\). Simultaneous readings in \(S'\) are not simultaneous in \(S\): the \(S\) readings are separated by \(\Delta t = vL_0/c^2\), using \(\Delta x = L_0\). B
5
\[ L = \Delta x' = \gamma\left(\Delta x - v\,\Delta t\right) = \gamma\left(L_0 - v\,\Delta t\right) \]
Apply the space transformation to the same two events; \(\Delta x = L_0\) is the fixed rest-frame separation, \(\Delta t\) is the (nonzero) \(S\)-time separation found in step 4. B
6
\[ L = \gamma\left(L_0 - v \cdot \frac{v L_0}{c^2}\right) = \gamma L_0\left(1 - \frac{v^2}{c^2}\right) \]
Substitute \(\Delta t = vL_0/c^2\) from step 4 and factor out \(L_0\). Pure algebra. A
7
\[ 1 - \frac{v^2}{c^2} = \frac{1}{\gamma^2} \;\Longrightarrow\; L = \gamma L_0 \cdot \frac{1}{\gamma^2} = \frac{L_0}{\gamma} \]
Use the definition of \(\gamma\) to collapse the bracket. The two factors of \(\gamma\) leave a single \(1/\gamma\). A
Result
\[ \boxed{\,L = \frac{L_0}{\gamma} = L_0\sqrt{1 - \frac{v^2}{c^2}}\,} \]

Reading. An object of proper length \(L_0\) is measured, in any frame moving at speed \(v\) along its length, to be shorter by the factor \(\sqrt{1 - v^2/c^2} = 1/\gamma \le 1\). Proper length is the longest length any observer measures; every moving observer measures less. The contraction is real in the sense of coordinate geometry, not an optical artifact, and it is reciprocal: each observer finds the other's ruler contracted.

Units check. \(\gamma\) and \(\sqrt{1 - v^2/c^2}\) are dimensionless (\(v^2/c^2\) is a ratio of squared speeds). Hence \(L\) carries the same units as \(L_0\), i.e. length \([\text{m}]\). Correct.

Limiting cases
  • \(v \to 0\): \(\gamma \to 1\), so \(L \to L_0\) — the Galilean/Newtonian limit, no contraction.
  • \(v \ll c\): expand \(L \approx L_0\left(1 - \tfrac{1}{2}v^2/c^2\right)\); the leading correction is second order in \(v/c\), which is why everyday contraction is unobservable.
  • \(v \to c\): \(\gamma \to \infty\), so \(L \to 0\) — the object contracts to a point in the direction of motion.
  • Transverse directions: \(y' = y,\ z' = z\), so dimensions perpendicular to \(\vec{v}\) are unchanged for any \(v\).
Breaks when
  • Non-inertial motion / acceleration. If the object is accelerating (or the frames are non-inertial), a single \(L_0\) and a single \(\gamma\) no longer describe it; the naive \(L_0/\gamma\) fails and one needs a Born-rigidity analysis or general relativity.
  • Transverse measurement. Applied to a dimension perpendicular to the motion, the formula gives a spurious contraction; the correct factor there is \(1\).
  • Gravitational / curved spacetime. Over regions where spacetime curvature matters, "the length of a moving rod" is not captured by a global Lorentz boost and the flat-space result does not apply.
  • Confusing measurement with visual appearance. What a single camera or eye photographs is not the contracted length: finite light-travel time adds an apparent rotation (Terrell–Penrose effect). The \(L_0/\gamma\) result is the measured length (simultaneous end positions), not the photographed shape.
Failure modes
  • Inverting the factor: writing \(L = \gamma L_0\) and concluding moving objects are longer. Proper length is always the maximum; you divide by \(\gamma\).
  • Wrong "proper" frame: treating the length measured in the lab as \(L_0\). \(L_0\) is measured in the frame where the object is at rest.
  • Skipping simultaneity: reading the two ends at different times in the moving frame, which invalidates steps 3–4 and gives a meaningless "length".
  • Contracting transverse dimensions: shrinking the height or width of an object moving horizontally.
  • Calling it an illusion: claiming the contraction is "just" due to light delay. The coordinate contraction is genuine; the optical appearance (Terrell rotation) is a separate, additional effect.
  • Double counting: applying \(1/\gamma\) to distance and \(1/\gamma\) to time within one frame's own single measurement, when each frame measures a proper quantity for only one of them.
Discussion

The physical root of length contraction is the relativity of simultaneity, not any dynamical squeezing of the object. Steps 3–4 make this explicit: the \(S'\) observer's "simultaneous" location of the two ends corresponds to two events that are not simultaneous in \(S\). Because the far end's position is recorded at a different \(S\)-time than the near end's, and the rod is at rest in \(S\), the geometry of the boost converts that simultaneity offset into a shortened spatial interval. Remove the \(\Delta t' = 0\) condition and the contraction vanishes; it is a statement about how observers slice spacetime, not about forces on the rod.

Length contraction is perfectly reciprocal. Two rods of equal proper length moving past each other each measure the other to be shorter, with no contradiction, because each uses its own notion of simultaneity to lay off "the length." This reciprocity is impossible to reconcile with an absolute-space picture and is one of the sharpest demonstrations that simultaneity is frame-dependent. It is the same reciprocity that resolves the ladder-in-the-barn paradox: the two barn doors that close "at the same time" in the barn frame do not close at the same time in the ladder frame.

Geometrically, length contraction and time dilation are two projections of the single invariant spacetime interval \(\Delta s^2 = c^2\Delta t^2 - \Delta x^2\). The worldlines of the rod's two ends are two parallel timelike lines; "length in a frame" is the interval cut out where that frame's simultaneity hyperplane crosses them. Boosting tilts the hyperplane, and hyperbolic (not Euclidean) geometry governs the resulting cut — which is why the moving length shrinks while the invariant interval between corresponding events is unchanged. In this language time dilation and length contraction are the same tilt seen along the timelike and spacelike directions respectively.

Common misconceptions. Length contraction is not an optical illusion and not a material compression; nothing pushes on the rod. It is also not something "happening to" the moving object as judged from a privileged frame — there is no privileged frame, and each observer's measurement is equally valid. Finally, a photograph of a relativistic object does not simply show it squashed: the Terrell–Penrose effect means it appears rotated, because light from different parts of the object left at different times.

Worked examples

Example 1 — A relativistic spacecraft. A spacecraft of proper length \(L_0 = 100\ \text{m}\) flies past a station at \(v = 0.80c\). Find its length as measured at the station.

1
\[ \gamma = \frac{1}{\sqrt{1 - v^2/c^2}} = \frac{1}{\sqrt{1 - 0.80^2}} = \frac{1}{\sqrt{1 - 0.64}} = \frac{1}{\sqrt{0.36}} = \frac{1}{0.60} \]
Compute the Lorentz factor from the given speed. A
2
\[ \gamma = 1.667 \]
Evaluate numerically. A
3
\[ L = \frac{L_0}{\gamma} = \frac{100\ \text{m}}{1.667} \]
Apply the result \(L = L_0/\gamma\); symbols first, then insert numbers with units. A
\[ L \approx 60\ \text{m} \]

Reading. The station measures the 100 m ship to be only 60 m long. In the ship's own frame it is still 100 m, and the station appears contracted to the crew.

Units check. \(\text{m}/(\text{dimensionless}) = \text{m}\). Correct.

Example 2 — Cosmic-ray muon and the atmosphere. A muon descends through the atmosphere; the vertical distance to the ground is \(d_0 = 10.0\ \text{km}\) as measured in the Earth frame. In the muon's rest frame the atmosphere rushes past at \(v = 0.98c\). Find the thickness of atmosphere the muon must traverse in its own frame.

1
\[ \gamma = \frac{1}{\sqrt{1 - v^2/c^2}} = \frac{1}{\sqrt{1 - 0.98^2}} = \frac{1}{\sqrt{1 - 0.9604}} = \frac{1}{\sqrt{0.0396}} \]
The 10.0 km is a proper length in the Earth frame (Earth and atmosphere at rest); the muon frame sees it contracted. Compute \(\gamma\). A
2
\[ \gamma = \frac{1}{0.1990} = 5.025 \]
Evaluate numerically. A
3
\[ d = \frac{d_0}{\gamma} = \frac{10.0\ \text{km}}{5.025} \]
The distance the muon travels is contracted in the muon frame; apply \(L = L_0/\gamma\) with \(L_0 = d_0\). B
\[ d \approx 1.99\ \text{km} \]

Reading. In its own frame the muon crosses only about 2.0 km, not 10 km. This is exactly why so many muons reach the ground despite their short proper lifetime: from the muon's viewpoint the trip is short (length contraction); from Earth's viewpoint the muon's clock runs slow (time dilation). The two descriptions agree on the observable outcome.

Units check. \(\text{km}/(\text{dimensionless}) = \text{km}\). Correct.

Problems
  1. (Easy) A metal rod has proper length \(2.0\ \text{m}\) and moves lengthwise at \(v = 0.60c\). What length is measured in the lab?
    Solution \(\gamma = 1/\sqrt{1 - 0.60^2} = 1/\sqrt{0.64} = 1/0.80 = 1.25\). Then \(L = L_0/\gamma = 2.0\ \text{m}/1.25 = 1.6\ \text{m}\).
  2. (Medium) A spaceship of proper length \(50\ \text{m}\) is measured by a ground observer to be \(40\ \text{m}\) long. Find its speed \(v\).
    Solution \(L/L_0 = 1/\gamma = 40/50 = 0.80\), so \(\gamma = 1.25\). Then \(1 - v^2/c^2 = 1/\gamma^2 = 0.64\), giving \(v^2/c^2 = 0.36\) and \(v = \sqrt{0.36}\,c = 0.60c\).
  3. (Medium–Hard) A muon moves at \(v = 0.99c\) toward the ground. The Earth-frame distance to the surface is \(10.0\ \text{km}\). (a) Find the contracted distance in the muon frame. (b) Find the travel time in the muon frame and confirm it is consistent with the Earth-frame travel time related by time dilation.
    Solution \(\gamma = 1/\sqrt{1 - 0.99^2} = 1/\sqrt{1 - 0.9801} = 1/\sqrt{0.0199} = 1/0.1411 = 7.09\). (a) \(d = d_0/\gamma = 10.0\ \text{km}/7.09 = 1.41\ \text{km}\). (b) Muon-frame time \(t' = d/v = 1.41\times 10^3\ \text{m}/(0.99\times 3.00\times 10^8\ \text{m/s}) = 1.41\times10^3/(2.97\times10^8) = 4.75\times10^{-6}\ \text{s} = 4.75\ \mu\text{s}\). Earth-frame time \(t = d_0/v = 1.00\times 10^4/(2.97\times 10^8) = 3.37\times10^{-5}\ \text{s} = 33.7\ \mu\text{s}\). Check: \(t = \gamma t' \Rightarrow 7.09 \times 4.75\ \mu\text{s} = 33.7\ \mu\text{s}\). Consistent — the Earth measures a longer (dilated) coordinate time, the muon a shorter proper time, and \(t = \gamma t'\) holds.
  4. (Medium) A pole has proper length \(5.0\ \text{m}\); a barn has proper length \(4.0\ \text{m}\). At what speed must the pole travel (lengthwise) so that, in the barn's frame, it exactly fits inside the barn?
    Solution Require the contracted pole length to equal the barn length: \(L_0/\gamma = 4.0\ \text{m}\) with \(L_0 = 5.0\ \text{m}\), so \(\gamma = 5.0/4.0 = 1.25\). Then \(1 - v^2/c^2 = 1/1.25^2 = 0.64\), \(v = \sqrt{0.36}\,c = 0.60c\). (In the pole's frame the barn is contracted to \(4.0/1.25 = 3.2\ \text{m}\) and the pole never fits at one instant — the "paradox" is resolved by the relativity of simultaneity.)
  5. (Hard) A cube has proper edge length \(1.00\ \text{m}\) and moves at \(v = 0.80c\) parallel to one set of edges. Find (a) its measured dimensions, (b) its measured volume, and (c) the factor by which its density (proper mass over measured volume) is increased.
    Solution \(\gamma = 1/\sqrt{1 - 0.80^2} = 1/0.60 = 1.667\). (a) Only the edge along the motion contracts: \(L_\parallel = 1.00\ \text{m}/1.667 = 0.60\ \text{m}\); the two transverse edges are unchanged at \(1.00\ \text{m}\). Dimensions: \(0.60\ \text{m} \times 1.00\ \text{m} \times 1.00\ \text{m}\). (b) \(V = 0.60 \times 1.00 \times 1.00 = 0.60\ \text{m}^3\), i.e. \(V = V_0/\gamma\). (c) Since the volume falls by \(1/\gamma\) while rest mass is invariant, the mass-per-measured-volume rises by the factor \(\gamma = 1.667\).