Special relativity

Special relativity describes physics in Minkowski spacetime, where inertial frames assign different space and time coordinates to the same spacetime events. Its two postulates are Einstein's principle of relativity and the same vacuum speed of light $c$ in every inertial frame, independent of the source's motion. Together they replace the Galilean transformation with the Lorentz transformation, coupling space and time rather than providing a universal clock. Einstein's 1905 paper developed this account through measuring rods, clocks, and light signals. ^[ref].

A spacetime event is an occurrence at a definite place and time: a pulse reaching a detector, for example. An inertial frame supplies a Cartesian coordinate system and clocks at rest throughout it. Einstein synchronisation sets a distant clock's reflection reading to the midpoint between a light signal's departure and return readings at the sending clock. The resulting coordinate time records the local synchronized clock at each spacetime event; it differs from the time when a distant observer receives its image. This distinction makes time dilation, length contraction, and the relativity of simultaneity precise measurements. ^[ref].

The beauty of special relativity is that these different measurements fit one geometry: all inertial frames agree on the Minkowski interval. It lets us translate an experiment between a laboratory and a moving apparatus while preserving the light-signal relations between spacetime events.

Motion preserving local proper distances measured on instantaneous comoving simultaneity slices. It constrains the acceleration of different points of a body, rather than describing an infinitely fast material signal. For rectilinear positive acceleration, the front of an extended Born-rigid body accelerates less strongly than its rear.

A front worldline at fixed proper distance $L$ along a rear observer's simultaneity slice is parametrized by the displayed equations, with velocity and Lorentz factor evaluated at the rear event. The front time is locally future-oriented when $1+{\alpha }_{F}L/c^2>0$, where ${\alpha }_F$ is the rear signed proper acceleration. At a zero of that factor the rigid coordinate extension encounters an acceleration-horizon obstruction.

Two rockets maintaining fixed separation in an inertial laboratory frame do not preserve their comoving separation. A string connecting them is stretched rather than remaining unstrained; sufficiently large strain can break a real string. The contrast with Born rigidity follows from relativity of simultaneity and does not contradict Lorentz contraction.

The apparent rotation of a moving object arises when length contraction is combined with different light emission times for photons arriving at a camera together. For an edge-on plane lamina moving along x, let X,Y be its rest coordinates and put a distant camera on the negative y axis. Simultaneous reception requires $t+Y/c=\text{constant}$. Since $x=X/\gamma +vt$, its transverse image coordinate is $X/\gamma -\beta Y$. This is the projection of a stationary lamina rotated through $\theta$ with $\cos \theta =1/\gamma$, $\sin \theta =\beta$. It describes a photographic projection; simultaneous laboratory positions retain length contraction. The simple angle formula assumes the transverse distant-camera geometry.

The Minkowski inner product of two event-position four-vectors is invariant under a common Lorentz transformation fixing the origin, but changes under a common translation. It is therefore not a criterion for the causal separation of the events. For example $(cT,0)$ and $(0,R)$ are orthogonal while their separation has squared interval $c^2{T}^2-R^2$, which can have any sign. Causal classification uses their difference, not their individual positions.

The kinematic minimum energy needed to produce an allowed final state. In the centre-of-momentum frame, the threshold for freely separated massive final particles is the sum of their rest energies, with no relative kinetic energy. Conservation of total energy and momentum then determines the corresponding laboratory threshold. Charges and other selection rules must also permit the reaction; the mass bound alone does not guarantee an observable transition.

For coordinates $x^{\mu }=(ct,x,y,z)$, the four-gradient of a scalar field is its covector of coordinate derivatives. Under a constant Lorentz transformation, ${\partial }_{\mu }^{\prime }=({\Lambda }^{-1}{)}^{\nu }{}_{\mu }{\partial }_{\nu }$. In metric signature $(+,-,-,-)$, raising its index gives ${\partial }^{\mu }=(c^{-1}{\partial }_t,-\nabla )$. Contracting it with a four-vector field produces the Lorentz scalar four-divergence. In particular ${\partial }_{\mu }{J}^{\mu }=0$ is charge continuity in every inertial frame.

Angular concentration and Doppler enhancement of radiation from a relativistically moving emitter. Frequency, photon arrival rate and apparent solid angle transform together; inferred brightness depends on viewing angle and the emission geometry.

For isotropic photon emission in a source rest frame, aberration maps its forward hemisphere into a laboratory cone of half-angle $\arccos \beta$. The cone contains half the emitted photons. Direction-dependent Doppler energies mean that this count statement is not an identical half-power statement.

For an unresolved, isotropically emitting rest-frame blob with power-law spectrum $F_{\nu }^{\prime }\propto {\nu }^{-\alpha }$, the observed spectral flux density at a fixed frequency is enhanced by $D^{3+\alpha }$. The factors follow from Lorentz-invariant specific intensity and the frequency shift. The bolometric or corresponding-logarithmic-bin luminosity transforms with $D^4$. A spatially steady continuous jet has a different equal-arrival-time sampling and can have exponent $2+\alpha$ instead.

The photon phase-space occupation is a Lorentz scalar. For radiation propagating through vacuum, this makes specific intensity divided by cubed frequency invariant under a change of inertial observer. With Doppler factor $D$, $I_{\nu }(\nu )=D^3{I}_{{\nu }^{\prime }}^{\prime }(\nu /D)$. This describes corresponding ray directions; it does not assert that an angularly integrated radiation field is isotropic in both frames.

An approaching jet can be enhanced and its receding counterpart suppressed by different Doppler factors. A discrete moving emitter with rest-frame power-law spectrum $F_{\nu }\propto {\nu }^{-s}$ has the usual boost ${\delta }_D^{3+s}$, whereas a steady continuous-jet convention changes the exponent. Absorption and environmental asymmetry can also affect the contrast.

For velocity $\beta c$ at angle $\theta$ toward the observer, ${\delta }_D=[\Gamma (1-\beta \cos \theta ){]}^{-1}$. Frequencies transform as $\nu ={\delta }_D{\nu }^{\prime }$ before cosmological redshift. Specific-intensity transformation follows invariance of $I_{\nu }/{\nu }^3$.

The rest energy of a body is $E_0=m{c}^2$. A body emitting energy $dE$ and no net momentum in its rest frame loses mass $dE/c^2$. For a radiating source of luminosity $L$, the corresponding radiative mass-loss rate is $L/c^2$.

The first postulate is Einstein's principle of relativity: nongravitational laws have the same form in every inertial frame. The second states that light in vacuum propagates at the same finite speed of light $c$ in every inertial frame, independently of the source's motion. Together with the usual spacetime symmetries and calibrated coordinates, these postulates yield the Lorentz transformation. The relativity principle alone also admits the Galilean transformation with its infinite invariant-speed limit.

Special relativity connects measurable clock rates, length contraction, light propagation, and four-momentum through the Lorentz transformation. Tests probe particular predictions rather than measuring a coordinate transformation in isolation. Examples include cosmic-ray muon time-dilation experiments, muon storage-ring time-dilation experiments, the Michelson-Morley experiment, the Fizeau experiment, and particle-collision energy and momentum balances.

William Bertozzi compared electron time-of-flight speeds with independently measured kinetic energies using electrostatic and linear acceleration. The 1964 report covers 0.5–15 MeV and uses calorimetry to check deposited energy. Speed approaches the light-speed limit while energy continues rising, testing relativistic energy against an unrestricted Newtonian extrapolation.^[ref]

An interferometer compares round-trip light travel times along perpendicular arms as its orientation changes. The absence of the classical stationary-aether fringe shift at the expected scale constrains orientation dependence of the round-trip speed of light. The result is consistent with special relativity, but alone does not select it uniquely from every alternative theory.^[ref]

The instrument's stone support and floating mounting allowed the experimenters to change its orientation while limiting mechanical disturbance. The measured quantity was the fringe shift produced as the apparatus rotated, rather than a direct timing of a single light pulse.^[ref]

Thomas precession is the kinematic rotation arising when the successive instantaneous rest frames of an accelerated particle are connected by noncollinear Lorentz boosts. Its contribution is needed in deriving the leading atomic spin-orbit coupling: for an electron in central potential energy $U(r)$, the corrected term is $H_{\mathrm{so}}=[2m_e^2{c}^{2}r{]}^{-1}(dU/dr)\mathbf{L}\cdot \mathbf{S}$. A rest-field argument alone gives twice this coefficient.^[ref]

For a particle instantaneously at rest in a primed frame moving at $v\hat{\mathbf{x}}$ relative to the laboratory, the ordinary three-force transforms as $F_{\parallel }=F_{\parallel }^{\prime }$ and $F_{\perp }=F_{\perp }^{\prime }/\gamma$. The spatial components of the four-force are ${\gamma }_u\mathbf{F}$, whereas its time component is ${\gamma }_u\mathbf{F}\cdot \mathbf{u}/c$. For a co-moving material segment, length contraction also changes its laboratory length, so transforming total force is not the same as transforming force per unit length.

The laws of physics have the same form in all inertial frames; uniform absolute motion cannot be identified by experiments confined to an inertial laboratory. The invariant vacuum speed of light is a separate postulate. Together they motivate the Lorentz transformations of special relativity.

Einstein synchronisation sets clocks at rest in one inertial frame using exchanged light signals. If the sending clock records departure at $t_A$ and return at $t_A^{\prime }$, the distant clock's reflection reading is set to $t_B=(t_A+t_A^{\prime })/2$. This convention assigns equal outgoing and incoming travel times and defines the frame's distant simultaneity. It does not set the arrival time of an image equal to the time of the remote spacetime event.

For ideal clocks in an inertial frame separated by fixed distance $L$, transporting a clock at small speed $u$ produces lag $\delta =(L/u)(1-\sqrt{1-u^2/c^2})\sim Lu/(2c^2)$. Hence the limit agrees with Einstein synchronisation, assuming the clock hypothesis. A relabelling $t_{\kappa }=t+\kappa x/c$ retains an endpoint offset $\kappa L/c$, so the result does not remove the choice of distant-time convention.

A relativistic massive particle is described by a timelike worldline with proper-time four-velocity $u^{\mu }$. In signature $(+---)$ its normalization is $u^{\mu }{u}_{\mu }=c^2$ and its momentum is $p^{\mu }=m{u}^{\mu }$. The relativistic charged-particle action adds electromagnetic coupling, whose variation gives the Lorentz force.

In the reparametrization-invariant massive-particle action, vary the square-root kinetic term before choosing proper time. Its velocity derivative gives $mc{u}_a/\sqrt{-u^2}$, which becomes $m{u}_a$ after imposing $u^2=-c^2$. Replacing the square root by a constant before variation would incorrectly remove the inertial momentum.

A massive particle is nonrelativistic when its speed is small compared with the speed of light. Its energy expansion is $E=m{c}^2+p^2/(2m)+O(p^4/(m^3{c}^2))$. The leading momentum-dependent term is the classical kinetic energy.

Hyperbolic motion in special relativity is rectilinear motion with constant proper acceleration. A representative positive-$x$ world line is $x=c{t}_0\cosh (\tau /t_0)$, $t=t_0\sinh (\tau /t_0)$, with proper acceleration $c/t_0$. This describes a hyperbola in spacetime, not a hyperbolic Kepler orbit.

In units $c=1$, constant signed proper acceleration $a$ givesPosition data determine $t_0,x_0$ but leave the initial rapidity ${\eta }_0$ free. Proper-acceleration magnitude alone also leaves its orientation unspecified. The familiar centred hyperbola $x^2-t^2=a^{-2}$ follows for the additional initial data $t_0=0$, $x_0=1/a$, ${\eta }_0=0$ and acceleration toward increasing $x$.

Replace instantaneous acceleration and braking by legs of constant proper-acceleration magnitude $a$, retaining the same cruise rapidity $\eta$ and fixed endpoint separation. Each leg takes proper time $c\eta /a$ and displaces $(c^2/a)(\cosh \eta -1)$ in the endpoint rest frame. Subtracting the proper cruise time displaced by both legs gives the displayed correction. A cruise segment exists only if the total endpoint separation exceeds twice this acceleration distance.

Proper length is spatial length measured using the induced metric on a specified spacelike slice. For a moving extended object, specify the simultaneity convention; local rest lengths need not coincide with a coordinate length.

Spacetime combines spatial position and time into one geometric setting for physical events. In special relativity it is modeled by Minkowski spacetime; in general relativity its geometry is dynamical.

The region of spacetime occupied by an extended object through its history. Different simultaneous slices intersect a moving tube in different endpoint events, explaining why length contraction compares different sections rather than successive physical compressions.

A spacetime event is an idealized occurrence at one point of spacetime, such as a light pulse reaching a detector. An inertial frame assigns spatial coordinates and a time coordinate to it. Different frames can assign different coordinates to the same spacetime event; the event itself is not changed by a coordinate transformation.

A spacetime extension is an isometric embedding of a spacetime as a proper open subset of a larger connected spacetime with a nondegenerate Lorentzian metric. The allowed differentiability class must be specified: analytic, smooth and low-regularity extensions are different questions. Divergent curvature can obstruct a regular extension even when an incomplete geodesic has finite affine length.

A maximal analytic extension of an analytic spacetime is an analytic spacetime extension admitting no further proper analytic extension as a regular spacetime of the same dimension. It is different from a maximal Cauchy development, which can end at a smooth Cauchy horizon. Additional global choices or quotients can affect uniqueness. The simply connected extension of positive-mass Schwarzschild spacetime through both horizons is the Kruskal spacetime; its curvature singularities remain boundaries.

An extendible spacetime admits a spacetime extension in the specified regularity class. A coordinate horizon need not obstruct an extension: the exterior Schwarzschild chart extends through its horizon in regular null coordinates. Maximality refers to extensions of the given spacetime, not to the mere absence of a singularity in one coordinate chart.

Proper time is the time measured by a clock following a timelike worldline. With metric signature $(-,+,+,+)$ and units $c=1$,along that worldline.

Restoring the speed of light, the same metric signature gives $d\tau =\sqrt{-d{s}^2}/c=dt\sqrt{1-|\mathbf{v}{|}^2/c^2}$ for a future-directed timelike curve. Integrating along a clock's world line gives its elapsed proper time, even when its velocity varies. For uniform motion this becomes the time dilation formula; it is not generally the square root of the interval between endpoints of a curved world line.

An inertial particle passes a stationary detector, then recedes at speed $v=\beta c$ before decaying. The detector's reception time obeys $t_{\mathrm{receive}}=t_d+v{t}_d/c$. Dividing the decay time $t_d$ by its Lorentz factor gives the title formula for elapsed proper time. A reception event at the detector and a decay event at the particle are not interchangeable endpoints for a time-dilation calculation.

A nonzero spacetime vector $k$ is null when its Minkowski norm vanishes, $k^a{k}_a=0$. Null tangent vectors describe the local directions of light rays.

Special relativity describes inertial frames using Lorentz transformations and Minkowski spacetime.

In Minkowski spacetime with metric signature $(-,+,+,+)$, write two nonzero future-pointing null vectors as $A=a(1,\mathbf{n})$ and $B=b(1,\mathbf{m})$, where $a,b>0$ and $\mathbf{n},\mathbf{m}$ are unit vectors. Their Lorentzian inner product givesThus their sum is a future-pointing causal vector. It is a null vector exactly when the spatial directions coincide, and a timelike vector otherwise. The positivity of the time components is essential: adding past- and future-pointing null vectors can instead produce a spacelike vector.

The null condition for a nonzero vector $k$ in a Lorentzian metric is $g(k,k)=0$. Applied to the tangent of a curve it makes the curve a null curve. Nullity alone does not imply the geodesic equation; an additional zero or tangent-parallel covariant acceleration is required.

For a relative speed $0\le v<c$, the Lorentz factor is

At $v=0.6c$, $\gamma =1/\sqrt{1-0.36}=1/0.8=1.25$. Thus a moving clock accumulates $0.8$ seconds of proper time per laboratory second, and a parallel moving rod has $0.8$ times its rest length. The Lorentz factor approaches one at small speed and grows without bound as $v$ approaches the speed of light; a massive object's rest frame never requires a boost at $v=c$.

A rod of rest length $L_0$ moving parallel to its length has simultaneous laboratory length $L=L_0/\gamma$, where $\gamma$ is the Lorentz factor. Here $L_0$ is the rod's proper length measured in its rest frame. Transverse lengths are unchanged by the corresponding Lorentz transformation.

For a rod at rest in $S^{\prime }$ with endpoints $x^{\prime }=0$ and $x^{\prime }=10\mathrm{m}$, let $S^{\prime }$ move at $v=0.6c$ relative to the laboratory $S$. Record both endpoint positions at laboratory time $t=0$. The Lorentz transformation then gives $x^{\prime }=1.25x$, so their recorded positions are $0$ and $8\mathrm{m}$: the laboratory length is $10/1.25=8\mathrm{m}$. This is an operational definition: synchronized laboratory detectors register the endpoints at the same laboratory time. A photograph collects light emitted at different times and need not show this length. ^[ref].

For those same two endpoint spacetime events, $\Delta t^{\prime }=-1.25(0.6c)(8\mathrm{m})/c^2=-6\mathrm{m}/c$. Thus the front-end measurement occurs earlier in $S^{\prime }$ by about $20\mathrm{ns}$. Length contraction compares simultaneous endpoint measurements in each frame, which select different pairs of spacetime events; it does not mean that the rod has been compressed in its own rest frame.

At speed $v\ll c$, an endpoint separation $L$ measured simultaneously in its own rest frame becomes $L\sqrt{1-v^2/c^2}$ in the moving frame, shorter by approximately $L{v}^2/(2c^2)$. The stable evaluation $L{\beta }^2/[1+\sqrt{1-{\beta }^2}]$ avoids subtracting nearly equal floating-point numbers.

If one clock measures a proper time interval $\Delta \tau$ while moving at constant speed $v$ in an inertial frame, that frame measures $\Delta t=\gamma \Delta \tau$, where $\gamma$ is the Lorentz factor. The two spacetime events must lie on that clock's world line.

Suppose a clock moving at $0.6c$ advances by $4\mu \mathrm{s}$ between two ticks. Set $t=x=t^{\prime }=x^{\prime }=0$ at the first tick. The laboratory's synchronized clocks assign $\Delta t=1.25\times 4=5\mu \mathrm{s}$. During that interval the clock travels $\Delta x=0.6c\times 5\mu \mathrm{s}=3c\mu \mathrm{s}$, approximately $899.38\mathrm{m}$. The Minkowski interval checks the result:Both frames therefore recover the same $4\mu \mathrm{s}$ of proper time. This calculation is useful whenever a process is timed on a moving apparatus but described using laboratory clocks.

A light clock makes the mechanism visible. Two mirrors separated by $h$ perpendicular to the motion exchange a pulse; define one tick as its complete return to the lower mirror. In the clock's rest frame the tick takes $\Delta \tau =2h/c$. In the laboratory each half-tick lasts $T$ and the light traverses a diagonal, so the Pythagorean theorem givesFor $h=1\mathrm{m}$ and $v=0.6c$, each leg covers $1.25\mathrm{m}$ while the mirrors move $0.75\mathrm{m}$. The return tick lasts $2.5\mathrm{m}/c$ instead of $2\mathrm{m}/c$: the light has the same speed of light in both accounts. ^[ref].

Reciprocal time dilation is consistent: each inertial frame times the other frame's single moving clock using its own synchronized clock network. These comparisons use different simultaneous spacetime events. Comparing reunited clocks instead requires their complete world lines, as in the twin paradox.

A freely moving unstable particle of speed $v$ travels a laboratory distance $R$ in time $R/v$, accumulating proper time $R/(\gamma v)$. For rest-frame half-life $T$ and exponential decay, its survival probability is the displayed expression. At least half survive exactly when $\gamma vT\ge R$. The minimum speed is thereforeThe relevant flight length is $\gamma vT$, rather than $\gamma cT$ for a particle whose speed is appreciably less than $c$.

Accelerator-produced muons circulating in a magnetic storage ring allow a direct comparison between their measured laboratory mean lifetime and $\gamma$ times the rest lifetime. The 1977 experiment at $\gamma =29.33$ measured positive- and negative-muon lifetimes near $64.4\mu \mathrm{s}$, consistent with time dilation. This is a controlled complement to atmospheric survival measurements, with different systematic effects.^[ref]

Atmospheric cosmic rays generate muons whose mean lifetime at rest is about $2.2\mu \mathrm{s}$. For a selected speed $v$ and production-to-detection distance $h$ in the terrestrial frame, special relativity predicts survival fraction $e^{-h/(\gamma v{\tau }_0)}$, where $\gamma$ is the Lorentz factor. Without time dilation, the exponent would instead be $-h/(v{\tau }_0)$. Comparing altitude-dependent counts at controlled energies, allowing for acceptance and energy loss, tests these alternatives. In the muon rest frame the terrestrial distance is shortened to $h/\gamma$ by length contraction, so the proper flight time and survival fraction agree with the terrestrial calculation. This frame comparison does not turn the same decay measurement into two independent tests.

An unstable particle of rest lifetime ${\tau }_0$ traversing distance $H$ at constant speed $v$ has survival probability $p=\exp [-H/(\gamma v{\tau }_0)]$. Requiring a flight time no greater than one dilated mean lifetime selects $p\ge e^{-1}$. In the ultrarelativistic approximation $v\simeq c$, this gives $\gamma \gtrsim H/(c{\tau }_0)$ and $E\gtrsim m{c}^{2}H/(c{\tau }_0)$. It is a specified survival-probability criterion, not a sharp cutoff for individual random decays. Continuous energy loss requires integrating the decay probability along the path.

A light clock measures time by counting round trips of a light pulse between mirrors. In its rest frame, mirrors separated by $h$ give a tick duration $2h/c$, where $c$ is the speed of light. Comparing this round trip in another inertial frame provides the geometric time dilation argument above.

The twin paradox compares proper time along the different world lines of a travelling observer and a stay-at-home observer who reunite. Reciprocal inertial time dilation produces an apparent contradiction only if the traveller's change of inertial frame is ignored.

The decisive experiment is local: put two clocks together at departure, separate them, then compare their readings together at reunion. Everyone agrees on those two spacetime events and on the final difference. The twin paradox asks why two clocks can accumulate different proper times between the same spacetime events, despite reciprocal time dilation during uniform motion. Its answer is the geometry of their complete world lines, together with the relativity of simultaneity when comparing distant clocks.

Consider a change of distant-time convention $s=t+\kappa x/c$, where $\kappa$ is constant and $|\kappa |<1$. This changes time labels and simultaneity slices while leaving the world lines and the metric tensor as geometric objects unchanged. In one spatial dimension,Substitute $dt=ds-\kappa dx/c$ into the Minkowski interval to obtain the formula. For a timelike curve, the proper time calculated from this transformed metric tensor equals the old value. At shared departure and reunion events with $x=0$, the labels also obey $s=t$. Thus a distant synchronization convention cannot alter the twin paradox's reunited clock readings. It would be an error to change time labels while retaining the old diagonal metric components.

Work in Minkowski spacetime, neglect gravity, and choose the stay-at-home clock's inertial frame. Let departure be $O=(t,x)=(0,0)$ and reunion be $R=(T,0)$. The home clock follows $x_H(t)=0$. A traveller moves outward with velocity $+v$ and returns with velocity $-v$, where $0<v<c$ and $c$ is the speed of light. The ideal world line has a sharp turnaround at $P=(T/2,vT/2)$:Each straight leg lies in an inertial frame, but no single traveller inertial frame covers both legs.

For any differentiable timelike curve in this home inertial frame, its clock accumulates proper time through the integralApply it separately to the traveller's two legs. With the Lorentz factor $\gamma =(1-v^2/c^2{)}^{-1/2}$,For $v=0.6c$ and $T=10\mathrm{yr}$, the Lorentz factor is $\gamma =1.25$. The turnaround is $3$ light-years away at home time $5\mathrm{yr}$. The traveller records $4\mathrm{yr}$ outward and $4\mathrm{yr}$ returning: the reunited clocks read $10\mathrm{yr}$ and $8\mathrm{yr}$. These are illustrative model values, not experimental measurements.

The same integral gives a short geometric explanation: $\sqrt{1-v(t{)}^2/c^2}\le 1$, so every such competing timelike curve between these fixed spacetime events has $\Delta \tau \le T$. Equality requires zero velocity throughout, apart from sets of zero duration. Thus the straight home world line maximizes proper time in this Minkowski spacetime problem. This statement concerns fixed endpoints and an inertial home clock; in a gravitational field, compare paths using the appropriate metric tensor.

During either inertial leg, the traveller can correctly say that the moving home clock runs slowly. The missing step is deciding which distant home-clock reading is simultaneous with turnaround. By the Lorentz transformation, the outward inertial frame's constant-time line through $P$ obeys $t-vx/c^2=\text{constant}$. At $x=0$ it meets the home world line atFor the return inertial frame, use $t+vx/c^2=\text{constant}$ through the same spacetime event $P$; its home-clock assignment isThe dashed lines in the spacetime diagram illustrate this relativity of simultaneity. For $v=0.6c$ and $T=10\mathrm{yr}$, they assign $3.2\mathrm{yr}$ and $6.8\mathrm{yr}$ to the home clock. The difference of $3.6\mathrm{yr}$ is a change in the traveller's distant-time assignment, not a sudden physical advance of the home clock.

On each leg, the traveller's $4\mathrm{yr}$ of proper time corresponds to $4/\gamma =3.2\mathrm{yr}$ of home-clock advance according to that leg's inertial frame. Including the twin paradox simultaneity change gives $3.2+3.6+3.2=10\mathrm{yr}$, exactly the reading at reunion. Actual received light signals remain continuous across a physical turnaround; their arrival times are distinct from the distant spacetime events assigned as simultaneous by Einstein synchronisation.

An instantaneous reversal is an idealization. A real traveller changes velocity over a finite time, follows a smooth world line, and has finite proper acceleration. Special relativity still calculates the elapsed proper time using the home inertial frame and the same integral; general relativity is not needed merely because the traveller accelerates in otherwise flat spacetime. The clock hypothesis identifies an ideal clock's reading with this geometric proper time.

One explicit smooth trip between the same $O$ and $R$ starts and ends at rest in the home inertial frame. ChooseIts velocity reverses continuously at $T/2$, and its proper acceleration is bounded because $v_{\max }<c$ and $dv/dt$ is finite. With $v_{\max }=0.6c$ and $T=10\mathrm{yr}$, numerical integration givesThis comparison fixes the reunion spacetime events and maximum speed, not the turnaround distance: the smooth trip reaches $v_{\max }T/\pi \simeq 1.910$ light-years, whereas the ideal trip reaches $3$ light-years. The smooth traveller spends less time at high speed and therefore accumulates more proper time than the ideal traveller, while still accumulating less than the home clock.

For a sufficiently short smooth reversal replacing the corner, the proper time approaches the piecewise inertial result. A reversal interval of home duration $\delta t$ contributes between zero and $\delta t$ to the clock reading; shrinking that interval cannot create a finite extra loss concentrated at the corner. The finite difference survives in the long coasting legs. Keeping the same reversal manoeuvre while lengthening those legs changes the age difference. Conversely, two travellers following mirrored paths with the same speed history accumulate equal proper times, even though both accelerate. Thus acceleration distinguishes paths and enables reunion here, but an amount of acceleration alone does not determine the age difference.

The changing simultaneity assignment also becomes continuous. At a traveller spacetime event $(t,x(t))$ with instantaneous velocity $v(t)$, the momentarily comoving inertial frame assigns home time $t_H=t-v(t)x(t)/c^2$. Differentiating givesDuring the reversal at positive $x$, the last term advances the assigned home time rapidly. It describes the traveller's changing relativity of simultaneity, not a locally measured alteration of the home clock's ticking.

The clock hypothesis says that an ideal clock following a timelike curve measures its proper time: its instantaneous rate depends on velocity through $d\tau =dt\sqrt{1-v^2/c^2}$, with no separate acceleration term. Acceleration still changes the world line and hence the accumulated proper time. A physical clock may suffer disturbances under forces or environmental changes, so experiments must distinguish its mechanism's errors from the geometric prediction.

The 1972 paper Around-the-World Atomic Clocks: Observed Relativistic Time Gains by J. C. Hafele and others reports four cesium-beam atomic clocks flown around the world in October 1971, eastward and westward. Their readings were compared with the U.S. Naval Observatory reference time scale before and after travel. An electronic time-interval counter resolved differences to the nearest nanosecond; intercomparisons among the flying clocks continued during travel. The final ensemble results were $-59\pm 10\mathrm{ns}$ eastward and $+273\pm 7\mathrm{ns}$ westward; the quoted spreads are standard deviations. These are corrected clock offsets, not raw readings or aircraft travel times. Four clocks allowed identification of individual rate changes through patterns in their mutual differences. The authors also report magnetic shielding and laboratory orientation checks. Their analysis could not detect a disturbance common to every flying clock, and the small ensemble limits statistical conclusions. This is a concrete reunited-clock experiment with apparatus, readout and drift controls.^[ref]

The companion paper Around-the-World Atomic Clocks: Predicted Relativistic Time Gains by J. C. Hafele and others computes $-40\pm 23\mathrm{ns}$ eastward and $+275\pm 21\mathrm{ns}$ westward from the recorded flight paths. These are predictions, with flight-data uncertainty estimates. Both time dilation from motion and gravitational time dilation matter: aircraft clocks at altitude experience a different gravitational potential, while the rotating Earth means a ground clock is not the inertial home clock of our simple model. Eastward and westward motion change the aircraft's velocity relative to an approximately nonrotating inertial frame, not merely its speed relative to the ground. The measured agreement therefore tests combined special relativity and general relativity; the westward gain does not contradict the maximum-proper time result for fixed endpoints in Minkowski spacetime.^[ref]

A complementary microscopic test is Measurements of relativistic time dilatation for positive and negative muons in a circular orbit by J. Bailey and others. The authors measured muon decay lifetimes in a storage ring at Lorentz factor $\gamma =29.33$: $64.419\pm 0.058\mu \mathrm{s}$ for positive muons and $64.368\pm 0.029\mu \mathrm{s}$ for negative muons. The reported positive-muon lifetime agrees with the independently measured lifetime at rest and special relativity to a fractional error of $2\times 10^{-3}$ at 95% confidence. This is a decay-lifetime readout under continuous centripetal acceleration, rather than a reunion of portable clocks. It supports the clock hypothesis for this physical process and regime; the inspected abstract does not establish universal immunity of every clock mechanism to acceleration. ^[ref]

An atomic clock uses an atomic transition as a frequency reference for an oscillator; counting its calibrated cycles realizes a time interval. In a passive beam implementation, the chemical atoms are repeatedly prepared, interrogated and detected, while feedback control corrects the continuously running oscillator. The Ramsey separated oscillatory fields experiment explains how internal quantum phase becomes a measurable population signal.

The SI second fixes the unperturbed ground state hyperfine structure transition frequency of caesium-133 to exactly $9192631770\mathrm{Hz}$. This is a defining value for the ideal transition, rather than a guarantee that an operating apparatus has no shifts.^[ref]

An oven supplies a collimated caesium beam. State-selecting magnets precede and follow two coherent microwave interaction regions; a weak static magnetic field separates sublevels. The clock transition connects $F=3,m_F=0$ and $F=4,m_F=0$. Transferred atoms reach a heated ionizing collector. A detector scan identifies the resonance before locking.^[ref]

Here $F$ labels the total angular momentum and $m_F$ its projection. The selected pair is a two-level system inside a larger set of atomic energy levels; treating every oven atom as a prepared qubit would overstate the preparation. A measured detector signal may be modeled as $S(\nu )=S_{\mathrm{bg}}+A\overline{P}(2\pi ({\nu }_0-\nu ))$, with background $S_{\mathrm{bg}}$ and gain $A>0$. Background subtraction and stable transmission are necessary before interpreting the signal as a transition probability. The velocity-averaged Ramsey line provides one explicit ideal example; its chosen dimensions are not historical specifications.

A peak has zero slope at its center, so a single peak-height reading does not tell feedback control which way to tune. Instead sample the two sides at ordinary frequencies ${\nu }_c\pm d$, where $d>0$ is a chosen modulation offset, and subtract:For an even signal around the true center ${\nu }_0$, write $ϵ={\nu }_c-{\nu }_0$. A Taylor expansion gives $E({\nu }_0+ϵ)=2S^{\prime }({\nu }_0+d)ϵ+O({ϵ}^3)$. On the descending side of a positive central peak $S^{\prime }({\nu }_0+d)<0$, so the update ${\nu }_c\mapsto {\nu }_c+gE$ has restoring sign for $g>0$ and is locally stable in this ideal discrete update when $0<-2g{S}^{\prime }({\nu }_0+d)<2$. A detector dip requires the opposite sign. These conditions follow from linearizing the update, not from assuming any particular electronics.

For the ideal signal $S=A{\cos }^2[\pi (\nu -{\nu }_0)T]+S_{\mathrm{bg}}$, choose $d=1/(4T)$. Then $E\simeq -2\pi ATϵ$. With $T=5\mathrm{ms}$ and $A=1$, a center error of $+1\mathrm{Hz}$ gives exact $E=-\sin (0.01\pi )\simeq -0.03141$. The sign directs the update toward lower frequency. Constant background cancels, but background drift between samples or asymmetric neighboring spectral lines can bias the zero. An initial scan matters because other fringes can also supply zeros.

A narrow fringe makes the center easier to resolve statistically; it does not certify an unbiased frequency. For example, with ideal $\pi /2$ pulses and a second-pulse oscillation phase offset ${\phi }_c$, the finite-pulse Ramsey transition probability approachesThis sign uses exactly the Rabi detuning convention and pulse Hamiltonian operator above, with the second pulse replacing ${\sigma }_x$ by $\cos {\phi }_c{\sigma }_x+\sin {\phi }_c{\sigma }_y$. The factor $2\pi$ converts angular frequency to ordinary frequency. For the original example, ${\phi }_c=0.010$ rad and $T=5\mathrm{ms}$ give a shift of about $-0.318\mathrm{Hz}$. For finite pulses or distributed flight times, use the full signal and its atomic clock frequency lock rather than inserting an unweighted mean $T$.