Kepler’s Orbits and Special Relativity

in

Introductory Classical Mechanics

Tyler J. Lemmon∗ and Antonio R. Mondragon†

(Dated: April 21, 2016)

Kepler’s orbits with corrections due to Special Relativity are explored using the Lagrangian

formalism. A very simple model includes only relativistic kinetic energy by defining a Lagrangian

that is consistent with both the relativistic momentum of Special Relativity and Newtonian gravity.

The corresponding equations of motion are solved in a Keplerian limit, resulting in an approximate

relativistic orbit equation that has the same form as that derived from General Relativity in the

same limit and clearly describes three characteristics of relativistic Keplerian orbits: precession

of perihelion; reduced radius of circular orbit; and increased eccentricity. The prediction for the

rate of precession of perihelion is in agreement with established calculations using only Special

Relativity. All three characteristics are qualitatively correct, though suppressed when compared to

more accurate general-relativistic calculations. This model is improved upon by including relativistic

gravitational potential energy. The resulting approximate relativistic orbit equation has the same

form and symmetry as that derived using the very simple model, and more accurately describes

characteristics of relativistic orbits. For example, the prediction for the rate of precession of perihelion

of Mercury is one-third that derived from General Relativity. These Lagrangian formulations of the

special-relativistic Kepler problem are equivalent to the familiar vector calculus formulations. In this

Keplerian limit, these models are supposed to be physical based on the likeness of the equations

of motion to those derived using General Relativity. The resulting approximate relativistic orbit

equations are useful for a qualitative understanding of general-relativistic corrections to Keplerian

orbits. The derivation of this orbit equation is approachable by undergraduate physics majors and

nonspecialists whom have not had a course dedicated to relativity.

PACS numbers: 45.20.Jj, 03.30.+p, 45.50.Pk, 04.25.Nx

I. INTRODUCTION

The relativistic contribution to the rate of precession

of perihelion of Mercury is calculated accurately using

General Relativity [1–6]. However, the problem is

commonly discussed in undergraduate and graduate

classical mechanics textbooks, without introduction of

an entirely new, metric theory of gravity. One approach

[7–10] is to define a Lagrangian that is consistent with

both the momentum-velocity relation of Special Relativity

and Newtonian gravity. The resulting equations of motion

are solved perturbatively, and an approximate rate of

precession of perihelion of Mercury is extracted. This

approach is satisfying in that a familiar element of Special

Relativity—relativistic momentum—produces a small

modification to a familiar problem—Kepler’s orbits—and

results in a characteristic of general-relativistic orbits—

precession of perihelion. On the other hand, one must

be content with an approximate rate of precession that

is one-sixth the correct value. Another approach [11–15]

∗ Colorado College† Colorado College; antonio.mondragon@gmail.com

is that of a history lesson and mathematical exercise. A

modification to Newtonian gravity is postulated, resulting

in an equation of motion that is the same as that derived

from General Relativity. The equation of motion is

solved perturbatively, and the correct rate of precession

of perihelion of Mercury is extracted. This method

is satisfying in that the modification to Newtonian

gravity results in the observed value for the relativistic

contribution to perihelic precession. On the other hand,

one must be content with a mathematical exercise, rather

than an understanding of the metric theory of gravity from

which the modification of Newtonian gravity is derived.

Both approaches provide an opportunity for students of

introductory classical mechanics to learn that relativity is

responsible for a small contribution to perihelic precession

and to calculate that contribution.

A review of the approach using only Special Relativity

and an alternative solution of the equations of motion

in a Keplerian limit are presented—resulting in an

approximate relativistic orbit equation. This orbit

equation has the same form as that derived using

General Relativity and clearly describes three relativistic

corrections to Keplerian orbits: precession of perihelion,

reduced radius of circular orbit, and increased eccentricity.

arX

iv:1

012.

5438

v6 [

astr

o-ph

.EP]

20

Apr

201

6

2

The approximate rate of perihelic precession is in

agreement with established calculations using only Special

Relativity. The method of solution makes use of a simple

change of variables and the correspondence principle,

rather than standard perturbative techniques, and is

approachable by undergraduate physics majors.

Two models are considered. A very simple model

(Secs. II and III) consists of relativistic kinetic energy

and unmodified Newtonian gravity. An improved

model (Sec. IV) includes both relativistic kinetic energy

and relativistic modification to Newtonian gravity. A

special-relativistic force is approximated in a Keplerian

limit, resulting in a conservative approximate relativistic

force, from which a relativistic potential energy is

derived—thereby enabling the usage of the Lagrangian

formalism. For both models, the Lagrangian formalism

is demonstrated to be equivalent to the vector calculus

formalism. The keplerian limit and validity of the

approximate relativistic orbit equations is discussed

thoroughly in Sec. VI. These models are supposed to be

physical based on the likeness of the equations of motion

to those derived using General Relativity (Sec. VI B). The

physical relevance of these models is emphasized in Sec. V.

A method of construction of similar models using more

general modifications to Newtonian gravity is discussed,

including a more broadly-defined Keplerian limit that is

useful in the derivation of orbit equations for such models

(App. A and Sec. VI C).

II. RELATIVISTIC KINETIC ENERGY

Conceptually, the simplest relativistic modification to

Kepler’s orbits is to define a Lagrangian with a kinetic

energy term that is consistent with both the momentum-

velocity relation of Special Relativity and the Newtonian

gravitational potential energy [7–9, 16–21];

L = −mc2γ−1 +GMm

r, (1)

where γ−1 ≡√

1− v2/c2, and v2 = r2 + r2θ2. (G is

Newton’s universal gravitational constant, M is the mass

of the sun, and c is the speed of light in vacuum.) The

equations of motion follow from Lagrange’s equations

d

dt

∂L

∂qi− ∂L

∂qi= 0, (2)

where qi ≡ dqi/dt for each of {qi} = {θ, r}. The results

are

d

dt(γr2θ) = 0, (3)

and

γr + γr +GM

r2− γrθ2 = 0. (4)

Using Eq. (3), a relativistic analogue to the Newtonian

equation for conservation of angular momentum per unit

mass is defined

` ≡ γr2θ = constant. (5)

This is used to eliminate the explicit occurrence of θ in

the equation of motion Eq. (4)

γrθ2 =`2

γr3. (6)

Time is eliminated by successive applications of the chain

rule, together with conserved angular momentum [22, 23];

r = − `γ

d

dθ

1

r, (7)

and, therefore,

γr = −γr − `2

γr2

d2

dθ2

1

r. (8)

Substituting Eqs. (6) and (8) into the equation of motion

Eq. (4) results in

`2d2

dθ2

1

r− γGM +

`2

r= 0. (9)

Anticipate a solution of Eq. (9) that is near Keplerian and

introduce the radius of a circular orbit for a nonrelativistic

particle with the same angular momentum, rc ≡ `2/GM .

The result is

d2

dθ2

rc

r+rc

r= 1 + λ, (10)

where λ ≡ γ − 1 is a velocity-dependent correction to

Newtonian orbits due to Special Relativity. The conic

sections of Newtonian mechanics [24, 25] are recovered by

setting λ = 0 (c→∞)

d2

dθ2

rc

r+rc

r= 1, (11)

resulting in the well-known orbit equation

rc

r= 1 + e cos θ, (12)

where e is the eccentricity. Kepler’s orbits are described

by 0 < e < 1.

3

III. KEPLERIAN LIMIT AND ORBIT

EQUATION

The planets of our solar system are described by

near-circular orbits (e � 1) and require only small

relativistic corrections (v/c� 1). Mercury has the largest

eccentricity (e ≈ 0.2), and the next largest is that of

Mars (e ≈ 0.09). Therefore, λ [defined after Eq. (10)] is

taken to be a small relativistic correction to near-circular

orbits of Newtonian mechanics—Keplerian orbits. This

correction is approximated using the first-order series

γ ≈ 1 + 12 (v/c)2, and neglecting the radial component of

velocity v ≈ rθ

λ ≈ 12 (rθ/c)2. (13)

See Sec. VI for a thorough discussion of this approximation.

Using angular momentum Eq. (5) to eliminate θ results

in λ ≈ 12 (`/rc)2(1 + λ)−2, or

λ ≈ 12 (`/rc)2. (14)

The equation of motion Eq. (10) is now expressed

approximately as

d2

dθ2

rc

r+rc

r≈ 1 + 1

2ε(rc

r

)2. (15)

where ε ≡ (GM/`c)2. The conic sections of Newtonian

mechanics, Eqs. (11) and (12), are now recovered by

setting ε = 0 (c → ∞). The solution of Eq. (15) for

ε 6= 0 approximately describes Keplerian orbits with small

corrections due to Special Relativity.

If ε is taken to be a small relativistic correction to

Keplerian orbits, it is convenient to make the change

of variable 1/s ≡ rc/r − 1 � 1. The last term on

the right-hand-side of Eq. (15) is then approximated

as (rc/r)2 ≈ 1 + 2/s, resulting in a linear differential

equation for 1/s(θ)

d2

dθ2

2

εs+

2(1− ε)εs

≈ 1. (16)

The additional change of variable α ≡ θ√

1− ε results in

the familiar form

d2

dα2

sc

s+sc

s≈ 1, (17)

where sc ≡ 2(1− ε)/ε. The solution is similar to that of

Eq. (11)

sc

s≈ 1 +A cosα, (18)

where A is an arbitrary constant of integration. In terms

of the original coordinates

rc

r≈ 1 + e cos κθ, (19)

where

rc ≡ rc1− ε

1− 12ε

(20)

e ≡12εA

1− 12ε

(21)

κ ≡ (1− ε) 12 . (22)

According to the correspondence principle, Kepler’s orbits

[Eq. (12) with 0 < e < 1] must be recovered in the limit

ε → 0 (c → ∞), so that 12εA ≡ e is the eccentricity of

Newtonian mechanics. To first order in ε, Eqs. (20)–(22)

are

rc ≈ rc(1− 12ε) (23)

e ≈ e(1 + 12ε) (24)

κ ≈ 1− 12ε, (25)

so that relativistic orbits in this limit are described

concisely by

rc(1− 12ε)

r≈ 1 + e(1 + 1

2ε) cos (1− 12ε)θ. (26)

When compared to Kepler’s orbits [Eq. (12) with

0 < e < 1], this orbit equation clearly displays three

characteristics of near-Keplerian orbits: precession of

perihelion; reduced radius of circular orbit; and increased

eccentricity. This approximate orbit equation has the

same form as that derived from General Relativity in this

limit [26]

rc(1− 3ε)

r≈ 1 + e(1 + 3ε) cos (1− 3ε)θ. (27)

The equations of motion, Eqs. (3) and (4), are

identical to those derived using the simple force

equation (Newton’s 3rd Law) p = −GMmr/r2, where

p = pr r + pθθ, and {pr, pθ} = {γmr, γmrθ}, verifying

that the unfamiliar relativistic kinetic energy term in

the Lagrangian, T ≡ −mc2γ−1 in Eq. (1), is consistent

with the familiar definition of relativistic momentum

p = γmv. For example, with p = pr r + pθθ, and using

{dr/dt, dθ/dt} = {θθ,−rθ}, the conserved relativistic

angular momentum Eq. (3) is verified

pθ/m =1

r

d

dt(γr2θ) = 0. (28)

IV. RELATIVISTIC GRAVITATIONAL

POTENTIAL ENERGY

The effect of using the special-relativistic γ factor to

define both relativistic kinetic energy and relativistic

gravitational potential energy is explored [27–33]. A

4

relativistic gravitational potential energy is derived from

a conservative approximate gravitational force. In this

Keplerian limit, a Lagrangian formulation of this problem

is demonstrated to be equivalent to the vector calculus

formulation.

A. Vector Calculus Formalism

The central-mass problem including both relativistic

kinetic energy and relativistic gravitational force is

described by the simple force equation

p = −γGMm

r2r, (29)

where p = γmv. The equations of motion are derived

as described in the final paragraph in Sec. III and are

identical to Eqs. (3) and (4), except that the gravitational

force is multiplied by the relativistic γ factor

` ≡ γr2θ = constant, (30)

and

γr + γr + γGM

r2− γrθ2. (31)

An approximate relativistic orbit equation is derived as

described in Secs. II and III

d2

dθ2

rc

r+rc

r= 1 + λ, (32)

where λ ≡ γ2 − 1 ≈ (rθ/c)2. Notice that λ = 2λ, where

λ ≡ γ − 1 ≈ 12 (rθ/c)2 is defined after Eq. (10) and in

Eq. (13). Therefore, the orbit equation is found using the

simple replacement ε→ 2ε in Eq. (26)

rc(1− ε)r

≈ 1 + e(1 + ε) cos (1− ε)θ. (33)

This approximate relativistic orbit equation has the

same form as that derived in Secs. II and III using only

relativistic kinetic energy Eq. (26) and is more accurate—

when compared to that derived from General Relativity

in this same limit Eq. (27).

B. Lagrangian Formalism

An equivalent Lagrangian formulation of this problem

is accomplished by defining a conservative approximate

relativistic gravitational force, from which a corresponding

potential energy is derived. A relativistic gravitational

force is defined as the Newtonian force due to gravity

with the replacement m→ γm

Fg ≡ −γGMm

r2= −GMm

r2(1 + λ), (34)

where λ ≡ γ − 1 is a small correction to Newtonian

gravity due to Special Relativity. Also, relativistic

angular momentum (per unit mass) is defined to be

that of Newtonian mechanics with the replacement

m→ γm, ` ≡ γr2θ. Anticipating near-circular orbits, λ is

approximated using the first-order series γ ≈ 1 + 12 (v/c)2,

neglecting the radial component of velocity v ≈ rθ, and

using angular momentum to eliminate θ, as described in

the first paragraph in Sec. III

λ ≈ 12ε(rc

r

)2

. (35)

The result is a conservative approximate relativistic

gravitational force Eq. (34) that describes a small

correction to Kepler’s orbits due to Special Relativity

Fg ≈ −GMm

r2

[1 + 1

2ε(rc

r

)2]. (36)

This force is integrated to derive an approximate

relativistic gravitational potential energy

U(r) = −GMm

r

[1 + 1

6ε(rc

r

)2]. (37)

The Lagrangian

L = −mc2γ−1 − U(r) (38)

results in the equations of motion

d

dt(γr2θ) = 0, (39)

and

γr + γr +GM

r2

[1 + 1

2ε(rc

r

)2]− γrθ2. (40)

The first equation Eq. (39) verifies the definition of

relativistic angular momentum, and the second equation

Eq. (40) includes a small relativistic correction to Kepler’s

orbits due to Special Relativity. Compare these equations

of motion to those derived using only relativistic kinetic

energy in Sec. II, Eqs. (3) and (4). A relativistic orbit

equation is derived as described in Secs. II and III

d2

dθ2

rc

r+rc

r= 1 + λ, (41)

where λ ≡ γ[1 + 12ε(rc/r)

2] − 1 is a small correction to

Keplerian orbits due to Special Relativity. Using the

first-order series γ ≈ 1 + 12 (v/c)2, neglecting the radial

component of velocity v ≈ rθ, and keeping terms first

order in ε results in λ ≈ 2λ, where λ ≡ γ−1 ≈ 12 (rθ/c)2 is

defined after Eq. (10) and in Eq. (13). Therefore, the orbit

equation is found using the simple replacement ε→ 2ε in

5

Eq. (26), thereby reproducing the orbit equation Eq. (33)

derived using the vector calculus formalism in Sec. IV A

rc(1− ε)r

≈ 1 + e(1 + ε) cos (1− ε)θ. (42)

Characteristics of near-Keplerian orbits are most easily

understood by comparing the approximate relativistic

orbit equations to that derived from the nonrelativistic

Kepler problem Eq. (12). For this purpose, it is convenient

to express the orbit equation as

rc

r≈ 1 + e cos κθ, (43)

where, using the simple replacement ε→ 2ε in Eqs. (20)–

(25),

rc ≡ rc1− 2ε

1− ε ≈ rc(1− ε) (44)

e ≡ e 1

1− ε ≈ e(1 + ε) (45)

κ ≡ (1− 2ε)12 ≈ 1− ε. (46)

V. CHARACTERISTICS OF

NEAR-KEPLERIAN ORBITS

The approximate relativistic orbit equation Eq. (42)

predicts a shift in perihelion through an angle

∆θ ≡ 2π(κ−1 − 1) ≈ 2πε (47)

per revolution. This prediction is twice that derived

using the standard approach [7–9, 34] to incorporating

Special Relativity into the Kepler problem, and is

compared to observations assuming that relativistic and

Keplerian angular momenta are approximately equal. For

a Keplerian orbit [24, 25] `2 = GMa(1 − e2), where

G = 6.670× 10−11 m3/kg·s2, M = 1.989× 1030 kg is the

mass of the Sun, and a and e are the semimajor axis

and eccentricity of the orbit, respectively. Therefore, the

relativistic correction defined after Eq. (15),

ε ≈ GM

c2a(1− e2), (48)

is largest for planets closest to the Sun and for

planets with very eccentric orbits. For Mercury

[35, 36] a = 5.79× 1010 m and e = 0.2056, so that

ε ≈ 2.66 × 10−8. (The speed of light is taken to

be c2 = 8.987554× 1016 m2/s2.) According to Eq. (47),

Mercury precesses through an angle

∆θ ≈ 2πGM

c2a(1− e2)= 1.67× 10−7 rad (49)

per revolution. This angle is very small and is usually

expressed cumulatively in arc seconds per century. The

rsinθ/r c

r cos θ/rc

KeplerEinstein

−1.5

−1

−0.5

0

0.5

1

1.5

−1.5 −1 −0.5 0 0.5 1 1.5

FIG. 1. A relativistic orbit in a Keplerian limit (solid)

Eq. (42) is compared to a Keplerian orbit (dotted) Eq. (12)

with the same angular momentum. Precession of perihelion

is one characteristic of relativistic orbits and is illustrated

here for 0 ≤ θ ≤ 6π. Precession of perihelion is also predicted

by General Relativity Eq. (27) in greater magnitude. This

characteristic of relativistic orbits is exaggerated by both the

choice of eccentricity (e = 0.25) and relativistic correction

parameter (ε = 0.1) for purposes of illustration. Precession is

present for smaller (non-zero) reasonably chosen values of e

and ε as well. (The same value of e is chosen for both orbits.)

orbital period of Mercury is 0.24085 terrestrial years, so

that

∆Θ ≡ 100 yr

0.24085 yr× 360× 60× 60

2π×∆θ (50)

≈ 14.3 arcsec/century. (51)

Precession, as predicted by Special Relativity is illustrated

in Fig. 1.

The general-relativistic (GR) treatment of this problem

results in a prediction of 43.0 arcsec/century [26, 34–55],

and agrees with the observed precession of perihelia

of the inner planets [36–45, 56–59]. Historically, this

contribution to the precession of perihelion of Mercury’s

orbit precisely accounted for the observed discrepancy,

serving as the first triumph of the general theory of

relativity [1–4]. The present approach, using only Special

Relativity, accounts for approximately one-third of the

observed discrepancy Eq. (51).

The approximate relativistic orbit equation, Eq. (42)

with e = 0, predicts a reduced radius of circular orbit—

when compared to that of Newtonian mechanics, Eq. (12)

with e = 0. This characteristic is not discussed in the

6

rc

Rc

(rc/`)

2Veff

r/rc

Newtonian Mechanics

General Relativity

−0.75

−0.5

−0.25

0

0.25

0 0.5 1 1.5 2 2.5 3

FIG. 2. The effective potential commonly defined in the

Newtonian limit to General Relativity (dashed) Eq. (52) is

compared to that derived from Newtonian mechanics (solid)

with the same angular momentum. The vertical (dotted)

lines identify the radii of circular orbits, Rc and rc, as

calculated using General Relativity and Newtonian mechanics,

respectively. General Relativity predicts a smaller radius of

circular orbit, when compared to that predicted by Newtonian

mechanics. This reduction in radius of a circular orbit Eq. (53),

Rc − rc ≈ −3εrc, is three times that predicted by the present

treatment using only Special Relativity Eq. (44), rc−rc ≈ −εrc.A curve representing an effective potential including small

corrections predicted by Special Relativity is expected to be

nearly identical to that for Newtonian mechanics (solid), with a

slightly smaller radius of circular orbit, rc. The value ε = 0.06

is chosen for purposes of illustration. Reduction in radius

of circular orbit is present for smaller (non-zero) reasonably

chosen values of ε as well.

standard approach to incorporating Special Relativity

into the Kepler problem, but is consistent with the GR

description. An effective potential naturally arises in the

GR treatment of the central-mass problem [26, 35, 39, 40,

42, 45, 52, 53, 60, 61],

Veff ≡ −GM

r+

`2

2r2− GM`2

c2r3, (52)

that reduces to the Newtonian effective potential in

the limit c → ∞. In the Keplerian limit, the GR

angular momentum per unit mass ` is also taken to be

approximately equal to that for a Keplerian orbit [26, 35–

42, 45, 53, 61, 62]. Minimizing Veff with respect to r

results in the radius of a stable circular orbit,

Rc =1

2rc +

1

2rc

√1− 12ε ≈ rc(1− 3ε), (53)

rsinθ/r c

r cos θ/rc

KeplerEinstein

−1.5

−1

−0.5

0

0.5

1

1.5

−2 −1.5 −1 −0.5 0 0.5 1

FIG. 3. A relativistic orbit in a Keplerian limit (solid)

Eq. (42) is compared to a Keplerian orbit (dotted) Eq. (12)

with the same angular momentum. Precession of perihelion

has been removed from the relativistic orbit equation, Eq. (42)

with (1 − ε)θ → θ, to emphasize two other characteristics

of relativistic orbits—reduced orbital radii and increased

eccentricity. These two characteristics are also predicted

by General Relativity Eq. (27) in greater magnitude. These

characteristics of relativistic orbits are exaggerated by both the

choice of eccentricity (e = 0.45) and the relativistic correction

parameter (ε = 0.15) for purposes of illustration. Reduced

orbital radii and increased eccentricity are present for smaller

(non-zero) reasonably chosen values of e and ε as well. (The

same value of e is chosen for both orbits.)

so that the radius of circular orbit is predicted to be

reduced, Rc − rc ≈ −3εrc. (There is also an unstable

circular orbit, as illustrated in Fig. 2.) This reduction

in radius of a circular orbit is three times that predicted

by the present treatment using only Special Relativity

Eq. (44), for which rc − rc ≈ −εrc. Reduced size of an

orbit, as predicted by Special Relativity, is illustrated in

Fig. 3.

Many discussions of the GR effective potential Eq. (52)

emphasize relativistic capture. The 1/r3 term in Eq. (52)

contributes negatively to the effective potential, resulting

in a finite—rather than infinite—centrifugal barrier and

affecting orbits very near the central mass (large-velocity

orbits), as illustrated in Fig. 2. This purely GR effect is

not expected to be described by the approximate orbit

equation Eq. (42), which is derived using only Special

Relativity and implicitly assumes orbits very far from the

central mass (small-velocity orbits).

An additional characteristic of relativistic orbits is that

7

of increased eccentricity. The relativistic orbit equation

Eq. (42) predicts increased eccentricity, when compared

to a Keplerian orbit Eq. (12) with the same angular

momentum Eq. (45), e − e ≈ εe. This characteristic

of relativistic orbits is not discussed in the standard

approach to incorporating Special Relativity into the

Kepler problem, but is consistent with the GR description.

The GR orbit equation in this Keplerian limit Eq. (27)

predicts an increase in eccentricity e− e ≈ 3εe, which is

three times that predicted by the present treatment using

only Special Relativity. Increased eccentricity of an orbit,

as predicted by Special Relativity, is illustrated in Fig. 3.

VI. DISCUSSION

The debate concerning the usage of relativistic inertial

mass and relativistic gravitational mass [63–70] is

irrelevant in the present context due to the fundamental

incompatibility of Special Relativity and gravitation

[71]. Accordingly, language describing the replacement

m → γm is intentionally omitted. Instructors may

feel more comfortable using the equivalent replacements

v → γv to introduce relativistic momentum, and G→ γG

to introduce relativistic gravitational force. Rather,

Lagrangians are constructed using familiar elements of

Special Relativity for the purpose of simulating general-

relativistic orbital effects using only Special Relativity

in a well-defined Keplerian limit. One purpose of

this presentation is to provide students with interesting

and tractable problems that arise from small special-

relativistic modifications to a familiar problem—Kepler’s

orbits, the solutions of which provide a qualitative

understanding of corrections to Kepler’s orbits due to

General Relativity. In this Keplerian limit, these models

are supposed to be physical based on the likeness of

the equations of motion to those derived using General

Relativity [26]. Another purpose is to present methods

by which similar models may be constructed and solved.

A. Domain of Validity

The following discussion concerning the derivation,

validity, and scope of the approximate special-relativistic

orbit equations follows the presentation in Secs. II and III,

in which only relativistic kinetic energy is included. The

arguments and conclusions also apply to the presentation

in Sec. IV, in which both relativistic kinetic energy and

relativistic gravitational potential energy are included.

The approximate relativistic orbit equation Eq. (26)

provides small corrections to Kepler’s orbits Eq. (12) due

to Special Relativity. A systematic verification may be

carried out by substituting Eq. (26) into Eq. (15), and

only keeping terms of orders e, ε, and eε. The domain of

validity is expressed by subjecting the solution Eq. (26)

to the condition

rc

r− 1� 1 (54)

for the smallest value of r. Evaluating the orbit equation

Eq. (42) at perihelion rp results in

rc

rp=

1 + e(1 + 12ε)

1− 12ε

. (55)

Substituting this into Eq. (54) results in the domain of

validity

e(1 + 12ε) + ε� 1. (56)

Therefore, the relativistic eccentricity e = e(1 + 12ε)� 1,

and Eq. (26) is limited to describing relativistic corrections

to near-circular (Keplerian) orbits. Also, the relativistic

correction ε� 1, and thus the orbit equation Eq. (26) is

valid only for small relativistic corrections.

The correction to Keplerian orbits due to Special

Relativity λ ≡ γ − 1 [defined after Eq. (10) and in

Eq. (13)] is approximated using the first-order series

γ ≈ 1 + 12 (v/c)2, and neglecting the radial component of

the velocity v ≈ rθ. Neglecting the radial component of

velocity in the relativistic correction λ is consistent with

the assumption of approximately Keplerian (near-circular)

orbits, and is supported by the condition rc/r − 1 � 1

preceding Eq. (16). It is emphasized that the radial

component of velocity is neglected only in the relativistic

correction λ; it is not neglected in the derivation of the

relativistic equation of motion Eq. (10). That there is

no explicit appearance of r in the relativistic equation

of motion, other than in the definition of γ, is due to a

fortunate cancellation after Eq. (8).

The presentation in Sec. IV includes relativistic

gravitational potential energy using the replacement

m→ γm in the Newtonian gravitational force. Although

the use of relativistic gravitational mass in the special-

relativistic Kepler problem is discouraged by some

authors [68], there are several useful results in the

Keplerian limit. This relativistic gravitational force is

approximated, resulting in a conservative force, from

which an approximate relativistic potential energy is

derived—thereby enabling the use of the Lagrangian

formalism. The resulting approximate relativistic orbit

equation Eq. (42) is more accurate, when compared

to that derived from General Relativity in the same

limit Eq. (27), than that derived using only relativistic

kinetic energy Eq. (26). Specifically, this more accurate

orbit equation demonstrates that—in this Keplerian

8

limit—the only consequence of neglecting relativistic

gravitational potential energy is that corrections due to

Special Relativity are decreased by a factor of two.

B. Structure of the Models

These models are appealing because they produce

equations of motion that are similar to those derived

using General Relativity. Compare the special-relativistic

(SR) equation of motion, Eq. (9) in Sec. II [using

γGM = GM +GM(γ − 1)], to the general-relativistic

(GR) equation of motion, Eq. (10) in Ref. 26,

(SR) `2d2

dθ2

1

r−GM +

`2

r−GM(γ − 1) = 0 (57)

(GR) ¯2 d2

dϕ2

1

r−GM +

¯2

r−GM

( 3¯2

c2r2

)= 0. (58)

The terms in parentheses describe corrections to

Newtonian orbits due to Special Relativity Eq. (57)

and General Relativity Eq. (58); these terms are zero

in the Newtonian limit c → ∞. Note that both of

these equations are exact in the sense that they are

derived directly from Lagrangians without making any

approximations. The SR equation of motion Eq. (57)

has a correction to Newtonian orbits GM(γ − 1). Using

the first-order expansion γ ≈ 1 + 12 (v/c)2, neglecting the

radial component of velocity v ≈ rθ, and using angular

momentum to eliminate θ results in

GM(γ − 1) ≈ GM( `2

2c2r2

). (59)

Aside from a constant, this term is identical to the

term that naturally arises in the GR derivation Eq. (58),

and is directly responsible for the 12ε corrections that

appear in the approximate relativistic orbit equation

Eq. (26). The factor of γ that is responsible for

this term is a direct result of the γ that appears

in the definition of angular momentum (per unit

mass) Eq. (5) ` ≡ γr2θ. Fundamentally, this angular

momentum has a γ factor due to the definition

of relativistic kinetic energy T ≡ −mc2γ−1. The GR

angular momentum is found to be simply ¯ ≡ r2ϕ

[Eq. (4) in Ref. 26] because that problem is solved using

proper time τ rather than coordinate time t, so that

ϕ ≡ dϕ/dτ . A transformation to coordinate time results

in ¯≡ γr2ϕ, where γ ≡ dt/dτ = [1 + 2V (r)/c2]−1/2, and

V (r) ≡ −GM/r is the Newtonian gravitational potential.

The vector calculus formalism in Sec. IV A includes a γ

factor in the definition of the relativistic gravitational

force. The γ factor from the relativistic angular

momentum propagates through the derivation of the

orbit equation as described in the preceding paragraph,

resulting in a correction to Newtonian orbits GM(γ2− 1).

Using the first-order expansion γ2 ≈ 1+(v/c)2, neglecting

the radial component of velocity v ≈ rθ, and using angular

momentum to eliminate θ results in

GM(γ2 − 1) ≈ GM( `2

c2r2

). (60)

This is exactly twice the correction term that results

from using only relativistic kinetic energy Eq. (59), and is

directly responsible for the ε corrections that appear in the

approximate relativistic orbit equation, Eqs. (33) and (42).

Using the Lagrangian formalism in Sec. IV B, there is an

additional requirement that the relativistic correction

to the Newtonian gravitational force does not depend

explicitly on θ or θ, so that the simple γ factor appears

in the definition of angular momentum.

More generally, any function f(γ) may be used in the

definition of relativistic gravitational force, provided a

first-order expansion of the form γf(γ) − 1 ≈ α(v/c)2

exists, where α > 0 is a constant. For example, choosing

a simple power-law dependence f(γ) = γn, so that

Fg ≡ γnGMm/r2, results in an approximate relativistic

orbit equation identical to Eq. (26) with the replacement

ε → (n + 1)ε. Notice the physical condition n ≥ 0 that

is necessary to insure the proper direction of precession,

reduced orbital radii, and increased eccentricity.

Using these types of models, the Keplerian limit is

defined precisely by only approximating the correction

term that represents a modification of the Keplerian

equation of motion. The approximation of this correction

term consists of: a series expansion to first order in (v/c)2;

neglecting the higher-order r2 term; and making a change

of variable subject to the condition rc/r − 1 ≡ 1/s� 1,

allowing the linearization (rc/r)n ≈ 1 + n/s.

With appropriate conditions and approximations,

modifications to Newtonian gravity that depend, more

generally, on velocity and radial coordinate may be used.

This is most easily described by example. A toy model is

presented in App. A and discussed in the following section

that describes methods—including a more broadly-defined

Keplerian limit—that are useful for solving problems with

more general modifications to Newtonian gravity.

C. A Toy Model

There are many attempts to motivate a physically

appropriate modification to Newtonian gravity in the

literature [27–33]. The possibility of simply replacing

m→ γm in the Newtonian gravitational potential energy

and using the Lagrangian formalism is explored in App. A

and discussed in this section. This toy model is useful

for describing methods for solving problems with more

9

general modifications to Newtonian gravity, for which

a more broadly-defined Keplerian limit is needed. The

Lagrangian for this model is

L = −mc2γ−1 + γGMm

r. (61)

Lagrange’s equations result in an abstruse set of

differential equations, Eqs. (A2) and (A3), that—using

an appropriate approximation—reduce to the equations

of motion derived in Sec. IV, Eqs. (30) and (31), thereby

reproducing the orbit equation, Eqs. (33) and (42)

rc(1− ε)r

≈ 1 + e(1 + ε) cos (1− ε)θ. (62)

Another solution to the differential equations,

Eqs. (A2) and (A3), is derived using a more broadly-

defined Keplerian limit. For near-circular approximately

Newtonian orbits, r and γ are very slowly varying

functions, so that the higher-order terms γr and

(γ/c)2V (r)γr2/r are taken to be negligible. The result

is the Newtonian equation of motion with a relativistic

correction term Eq. (A10)

˜2 d2

dθ2

1

r−GM +

˜2

r−GMλ = 0, (63)

where

λ ≡ γ2[1− (γ/c)2V (r)

]− 1, (64)

˜≡ γr2θ[1− (γ/c)2V (r)

]is a constant of motion, and

V (r) ≡ −GM/r is the Newtonian gravitational potential.

Compare this equation of motion to Eqs. (57) and (58).

This equation of motion has a small relativistic correction

to Keplerian orbits Eq. (A12)

GMλ ≈ GM[ ˜2

c2r2− V (r)

c2

]. (65)

Compare this to the correction terms used in the two

simple models, Eqs. (59) and (60). The corresponding

orbit equation in the Keplerian limit Eq. (A16),

rc(1− 2ε)

r≈ 1 + e(1 + ε) cos (1− 3

2 ε)θ, (66)

does not have the symmetry of those derived using the

two simple models, Eqs. (26) and (42). The methods and

approximations used to derive this orbit equation may

be applied to other, more physical, models that include

relativistic corrections to Kepler’s orbits.

VII. CONCLUSION

The Lagrangian formalism is useful for describing small

relativistic corrections to Kepler’s orbits. The simplest

model includes only relativistic kinetic energy using the

replacement m→ γm in the Newtonian linear momentum,

p = γmv. A solution to the corresponding equations of

motion in a Keplerian limit results in an approximate

relativistic orbit equation Eq. (26) that has the same

form as that derived from General Relativity in this limit

Eq. (27) and is easily compared to that describing Kepler’s

orbits Eq. (12). This form is that of elliptical orbits

of Newtonian mechanics with corrections to radius and

eccentricity, and exhibiting precession. Specifically, the

approximate relativistic orbit equation clearly describes

three characteristics of relativistic orbits: precession of

perihelion; reduced radius of circular orbit; and increased

eccentricity. The predicted rate of precession of perihelion

of Mercury is in agreement with that of established

calculations using only Special Relativity. Each of these

characteristics of relativistic Keplerian orbits is exactly

one-sixth of the corresponding correction described by

General Relativity in this limit—providing a qualitative

description of corrections to Keplerian orbits due to

General Relativity.

This simple model is improved upon by including

relativistic gravitational force using the replacement

m → γm in Newtonian gravity, Fg = γGMm/r2. An

approximation consistent with the Keplerian limit results

in a conservative force, from which a relativistic potential

energy is derived that is useful in a Lagrangian formulation

of the special-relativistic Kepler problem. A solution of

the corresponding equations of motion in a Keplerian

limit results in an approximate relativistic orbit equation

Eq. (42) that has the same form as that derived using

the simplest model Eq. (26), thereby describing the same

three characteristics of relativistic orbits. Each of these

characteristics of relativistic Keplerian orbits is exactly

one-third of the corresponding correction described by

General Relativity in this limit Eq. (27).

The Lagrangian formalism applied to the special-

relativistic Kepler problem is instructive, providing

several challenges appropriate for an introductory classical

mechanics course, including: solve Newton’s force

equation using vector calculus to verify the unfamiliar

relativistic kinetic energy term in the Lagrangian—as

outlined in the last paragraph of Sec. III and in Sec. IV A;

derive a potential energy function from a conservative

force; apply Lagrange’s equations to derive the conserved

relativistic angular momentum and equation of motion;

and transform and solve a differential equation to derive

an approximate relativistic orbit equation in terms

of planar coordinates. This approach also provides

an opportunity to use less familiar problem solving

strategies, including: variable transformations to cast the

differential equation into familiar form; approximation

10

methods that simplify the differential equation; and

usage of the correspondence principle to identify a

constant of integration. Most importantly, students

are rewarded with a clear understanding that a small

relativistic modification to a familiar problem results in

an approximate relativistic orbit equation that clearly

demonstrates that relativity is responsible for a small

contribution to perihelic precession, and the satisfaction

of calculating that contribution.

These models are appealing because they are easy to

motivate and, most importantly, they produce equations

of motion that are similar to those derived using General

Relativity, as discussed in Sec. VI B. A larger class

of models may be solved using similar methods and

approximations. This is demonstrated using a toy model

in App. A and in Sec. VI C, wherein a more broadly-

defined Keplerian limit is described. Exact solutions of

the special-relativistic Kepler problem require a thorough

understanding of special relativistic mechanics [72, 73] and

are, therefore, inaccessible to many undergraduate physics

majors. The present approach and method of solution is

understandable to nonspecialists, including undergraduate

physics majors whom have not had a course dedicated to

relativity.

[1] A. Einstein, “Erklarung der Perihelbewegung des Merkuraus der allgemeinen Relativitatstheorie,” Sitzungsber.Preuss. Akad. Wiss. Berlin (Math. Phys.) 1915, 831–839(1915).

[2] A. Einstein, “Explanation of the perihelion motion ofmercury from the general theory of relativity,” in TheCollected Papers of Albert Einstein, translated by A. Engel(Princeton University Press, Princeton, 1997), Vol. 6,pp. 112–116. This article is the English translation ofRef. 1.

[3] A. Einstein, “Die Grundlage der allgemeinenRelativitatstheorie; von A. Einstein,” Ann. d. Phys. 354(7), 769–822 (1916).

[4] Reference 2, pp. 146–200. This article is the Englishtranslation of Ref. 3.

[5] K. Schwarzschild, “Uber das Gravitationsfeld einesMassenpunktes nach der Einsteinschen Theorie,”Sitzungsber. Preuss. Akad. Wiss., Phys.-Math. Kl. 1916,189–196 (1916). Reprinted in translation as “On theGravitational Field of a Mass Point according to Einstein’sTheory,” arXiv: physics/9905030v1 [physics.hist-ph].

[6] J. Droste, “Het veld van een enkel centrum in Einstein’stheorie der zwaartekracht, en de beweging van eenstoffelijk punt in dat veld,” Versl. Akad. Amst. 25, 163–180 (1916-1917). Reprinted in translation as “The fieldof a single centre in Einstein’s theory of gravitation,and the motion of a particle in that field,” Proc. K.Ned. Akad. Wetensch. 19 (1), 197–215 (1917), <adsabs.harvard.edu/abs/1917KNAB...19..197D>.

[7] H. Goldstein, C. Poole, and J. Safko, Classical Mechanics(Addison Wesley, San Francisco, 2002), 3rd ed., pp. 312–314 and p. 332 (Exercise 26).

[8] J. V. Jose and E. J. Saletan, Classical Dynamics:A Contemporary Approach (Cambridge UniversityPress, Cambridge, 1998), pp. 209–212 and pp. 276–277(Problem 11).

[9] P. C. Peters, “Comment on ‘Mercury’s precessionaccording to special relativity’ [Am. J. Phys. 54, 245(1986)],” Am. J. Phys. 55 (8), 757 (1987).T. Phipps, Jr., “Response to ‘Comment on “Mercury’sprecession according to special relativity,”’[Am. J. Phys.55, 757 (1987)],” Am. J. Phys. 55 (8), 758 (1987).T. Phipps, Jr., “Mercury’s precession according to specialrelativity,” Am. J. Phys. 54 (3), 245 (1986).

[10] L. Jia, “Approximate Kepler’s Elliptic Orbits with theRelativistic Effects,” Int. J. Astron. Astrophys. 3, 29–33(2013).

[11] Reference 7, pp. 536–539.[12] S. T. Thornton and J. B. Marion, Classical Dynamics of

Particles and Systems (Thomson Brooks/Cole, Belmont,CA, 2004), pp. 292–293 and pp. 312–316.

[13] V. D. Barger and M. G. Olsson, Classical Mechanics: AModern Perspective (McGraw-Hill, Inc., New York, 1995),pp. 306–309.

[14] G. R. Fowles, Analytical Mechanics (Holt, Rinehart andWinston, New York, 1977), pp. 161–164.

[15] L. N. Hand and J. D. Finch, Analytical Mechanics(Cambridge University Press, Cambridge, 1998), pp. 420–422 (Problems 18 and 19).

[16] Reference 12, pp. 578–579.Reference 13, p. 366 (Problem 10-12).

[17] J. M. Potgieter, “Derivation of the equations of Lagrangefor a relativistic particle,” Am. J. Phys. 51 (1), 77 (1983).

[18] E. A. Desloge and E. Eriksen, “Lagrange’s equations ofmotion for a relativistic particle,” Am. J. Phys. 53 (1),83–84 (1985).

[19] Y.-S. Huang and C.-L. Lin, “A systematic method todetermine the Lagrangian directly from the equations ofmotion,” Am. J. Phys. 70 (7), 741–743 (2002).

[20] S. Sonego and M. Pin, “Deriving relativistic momentumand energy,” Eur. J. Phys. 26, 33–45 (2005).

[21] T. J. Lemmon and A. R. Mondragon, “First-OrderSpecial Relativistic Corrections to Kepler’s Orbits” (2010)unpublished, arXiv: 1012.5438v1 [astro-ph.EP].

[22] Reference 7, p. 87.Reference 8, p. 84.Reference 12, p. 292.Reference 14, p. 148.

[23] P. Hamill, Intermediate Dynamics (Jones and BartlettPublishers, LLC, Sudbury, MA, 2010), p. 330.

[24] Reference 7, pp. 92–96.Reference 8, pp. 84–86.Reference 12, pp. 300–304.Reference 14, pp. 137–157.Reference 23, pp. 307–343.

[25] P. Smith and R. C. Smith, Mechanics (John Wiley &Sons, Chichester, 1990), pp. 195–202.

11

[26] T. J. Lemmon and A. R. Mondragon, “Alternativederivation of the relativistic contribution to perihelicprecession,” Am. J. Phys. 77 (10), 890–893 (2009);arXiv: 0906.1221v2 [astro-ph.EP].

[27] A. Singh and B. K. Patra, “Relativistic corrections to thecentral force problem in a generalized potential approach,”(2014), arXiv: 1404.2940v2 [physics.class-ph].

[28] F. Bunchaft and S. Carneiro, “Weber-like interactions andenergy conservation,” (1997), arXiv: gr-qc/9708047v1.

[29] R. A. V. Hidalgo-Gato, “Towards an extension of 1905relativistic dynamics with a variable rest mass measuringpotential energy,” (2012), arXiv: 1210.4157v1 [physics.gen-ph].

[30] T. E. Phipps, Jr., “On Gerber’s Velocity-dependentGravitational Potential,” (2004) unpublished,http://studylib.net/doc/5885051/on-gerber-s-

velocity-dependent-gravitational-potential.[31] R. L. Kurucz, “The Precession of Mercury and

the Deflection of Starlight from Special RelativityAlone,” (2009) unpublished, kurucz.harvard.edu/

papers/deflection/deflection.pdf.[32] C. G. Vayenas, A. Fokas, and D. Grigoriou, “Gravitational

mass and Newton’s universal gravitational law underrelativistic conditions,” J. Phys: Conf. Series 633 (2015)012033.

[33] M. A. Abramowicz, A. M. Beloborodov, X. Chen, andI. V. Igumenshchev, “Special Relativity and pseudo-Newtonian potential,” (1996) unpublished, arXiv: astro-ph/9601115v1.

[34] D. H. Frisch, “Simple aspects of post-Newtoniangravitation,” Am. J. Phys. 58 (4), 332–337 (1990).

[35] S. M. Carroll, An Introduction to General Relativity:Spacetime and Geometry (Addison-Wesley, San Francisco,2004), pp. 205–216.

[36] J. Plebanski and A. Krasinski, An Introduction to GeneralRelativity and Cosmology (Cambridge University Press,Cambridge, 2006), pp. 176–182.

[37] C. W. Misner, K. S. Thorne, and J. A. Wheeler,Gravitation (Freeman, San Francisco, 1973), pp. 1110–1115.

[38] S. Weinberg, Gravitation and Cosmology (John Wiley &Sons, New York, 1972), pp. 194–201.

[39] W. Rindler, Relativity: Special, General, andCosmological (Oxford University Press, New York, 2001),pp. 238–245.

[40] H. Ohanian and R. Ruffini, Gravitation and Spacetime (W.W. Norton & Company, New York, 1994), pp. 401–408.

[41] R. D’Inverno, Introducing Einstein’s Relativity (OxfordUniversity Press, New York, 1995), pp. 195–198.

[42] M. P. Hobson, G. P. Efstathiou, and A. N. Lasenby,General Relativity: An Introduction for Physicists(Cambridge University Press, Cambridge, 2006), pp. 205–216 and pp. 230–233.

[43] K. Doggett, “Comment on ‘Precession of the perihelionof Mercury,’ by Daniel R. Stump [Am. J. Phys. 56, 1097–1098 (1988)],” Am. J. Phys. 59 (9), 851–852 (1991).

[44] D. R. Stump, “Precession of the perihelion of Mercury,”Am. J. Phys. 56 (12), 1097–1098 (1988).

[45] G. Ovanesyan, “Derivation of relativistic corrections tobounded and unbounded motion in a weak gravitationalfield by integrating the equations of motion,” Am. J. Phys.71 (9), 912–916 (2003).

[46] A. M. Nobili and I. W. Roxburgh, “Simulation of generalrelativistic corrections in long term numerical integrationsof planetary orbits,” in Relativity in Celestial Mechanicsand Astrometry: High Precision Dynamical Thoeriesand Observational Verifications, J. Kovalevsky andV. A. Brumberg (IAU, Dordrecht, 1986), pp. 105–110,<adsabs.harvard.edu/abs/1986IAUS..114..105N>.

[47] C. Magnan, “Complete calculations of the perihelionprecession of Mercury and the deflection of light bythe Sun in General Relativity,” (2007) unpublished,arXiv: 0712.3709v1 [gr-qc].

[48] M. M. D’Eliseo, “The first-order orbital equation,” Am.J. Phys. 75 (4), 352–355 (2007).

[49] D. R. Brill and D. Goel, “Light bending and perihelionprecession: A unified approach,” Am. J. Phys. 67 (4),316–319 (1999).

[50] B. Dean, “Phase-plane analysis of perihelion precessionand Schwarzschild orbital dynamics,” Am. J. Phys. 67(1), 78–86 (1999).

[51] N. Ashby, “Planetary perturbation equations based onrelativistic Keplerian motion,” in Relativity in CelestialMechanics and Astrometry: High Precision DynamicalThoeries and Observational Verifications, edited byJ. Kovalevsky and V. A. Brumberg (IAU, Dordrecht,1986), pp. 41–52, <adsabs.harvard.edu/abs/1986IAUS.

.114...41A>.[52] R. M. Wald, General Relativity (University of Chicago

Press, Chicago, 1984), pp. 136–143.[53] J. B. Hartle, Gravity: An Introduction to Einstein’s

General Relativity (Addison-Wesley, San Francisco, 2003),pp. 191–204.

[54] A. Schild, “Equivalence principle and red-shiftmeasurements,” Am. J. Phys. 28 (9), 778–780 (1960).

[55] A. Larranaga and L. Cabarique, “Advance of PlanetaryPerihelion in Post-Newtonian Gravity,” Bulg. J. Phys. 30,1–7 (2005), arXiv: 1202.4951v1 [gr-qc].

[56] Reference 53, pp. 230–232.[57] M. G. Stewart, “Precession of the perihelion of Mercury’s

orbit,” Am. J. Phys. 73 (8), 730–734 (2005).[58] D. Brouwer and G. M. Clemence, “Orbits and masses of

planets and satellites,” in The Solar System: Planets andSatellites, edited by G. P. Kuiper and B. M. Middlehurst(University of Chicago Press, Chicago, 1961), Vol. III,pp. 31–94.

[59] C. Sigismondi, “Astrometry and relativity,” Nuovo Cim.B 120, 1169–1180 (2005).

[60] Reference 37, pp. 660–662, Box 25.6.[61] F. Y.-H. Wang, “Relativistic orbits with computer

algebra,” Am. J. Phys. 72 (8), 1040–1044 (2004).[62] A. Harvey, “Newtonian limit for a geodesic in a

Schwarzschild field,” Am. J. Phys. 46 (9), 928–929 (1978).[63] L. Okun, “The Concept of Mass,” Physics Today (June,

1989), 31–36.[64] E. Hecht, “Einstein on mass and energy,” Am. J. Phys.

77 (9), 799–806 (2009).[65] E. P. Manor, “Gravity, Not Mass Increases with Velocity,”

J. Mod. Phys. 6, 1407–1411 (2015), http://dx.doi.org/10.4236/jmp.2015.610145.

[66] R. I. Khrapco, “Rest mass or inertial mass?,” (2001)unpublished, arXiv: physics/0103008v2 [physics.gen-ph].

[67] P. M. Brown, “On the concept of mass in relativity,” (2007)unpublished, arXiv: 0709.0687v2 [physics.gen-ph].

12

[68] G. Oas, “On the abuse and use of relativistic mass,”(2005), arXiv: physics/0504110v2 [physics.ed-ph].G. Oas, “On the Use of Relativistic Mass in VariousPublished Works,” (2005), arXiv: physics/0504111v1[physics.ed-ph].

[69] C. G. Adler, “Does mass really depend on velocity, dad?,”Am. J. Phys. 55 (8), 739–743 (1987).

[70] S. A. Vasiliev, “On the Notion of the Measure of Inertiain the Special Relativity Theory,” App. Phys. Res. 4 (2),136–143 (2012).

[71] Reference 37, ch. 7.[72] W. Rindler, Special Relativity (Oliver and Boyd, London,

1960), pp. 79–103.[73] J. L. Synge, Relativity: The Special Theory (North-

Holland Publishing Company, Amsterdam, 1958), pp. 394–399.

Appendix A: A Toy Model

The possibility of including relativistic gravitationalpotential by simply replacing m→ γm in the Newtoniangravitational potential energy and using the Lagrangianformalism is explored. This toy model is useful fordescribing methods for solving problems with moregeneral modifications to Newtonian gravity, for whicha more broadly-defined Keplerian limit is needed. TheLagrangian for this model is

L = −mc2γ−1 + γGMm

r. (A1)

Lagrange’s equations result in an abstruse set ofdifferential equations

d

dt

{γr2θ

[1− (γ/c)2V (r)

]}= 0, (A2)

and

γr[1− (γ/c)2V (r)

]+ γr

[1− 3(γ/c)2V (r)

]+ γ

GM

r2− γrθ2

[1− (γ/c)2V (r)

]+ (γ/c)2V (r)γ

r2

r= 0,

(A3)

where V (r) ≡ −GM/r is the Newtonian gravitationalpotential. Mercury’s orbit is approximately circularwith radius of the same order of magnitude as itssemimajor axis r ∼ a ≈ 5.79 × 1010 m, and its velocityis very small when compared to the speed of light, sothat (γ/c)2V (r) ∼ V (a)/c2 ∼ −10−8. Ignoring termsproportional to (γ/c)2V (r) results in the equations ofmotion derived in Sec. IV, Eqs. (30) and (31), therebyreproducing the orbit equation, Eqs. (33) and (42)

rc(1− ε)r

≈ 1 + e(1 + ε) cos (1− ε)θ. (A4)

Another solution to the equations of motion,Eqs. (A2) and (A3), is derived using a more broadly-defined Keplerian limit. For near-circular approximately

Newtonian orbits, r and γ are very slowly varyingfunctions, so that the higher-order terms γr and(γ/c)2V (r)γr2/r are taken to be negligible. [ForMercury’s orbit (γ/c)2V (a)γ/a ∼ −10−18 m−1.]

˜≡ γr2θ[1− (γ/c)2V (r)

]= constant, (A5)

and

γr[1− (γ/c)2V (r)

]+ γ

GM

r2

− γrθ2[1− (γ/c)2V (r)

]≈ 0.

(A6)

Conservation of angular momentum Eq. (A5) is used toeliminate the explicit occurrence of θ in the equation ofmotion Eq. (A6)

γrθ2[1− (γ/c)2V (r)

]=

˜2

γr3 [1− (γ/c)2V (r)]. (A7)

Time is eliminated by successive applications of the chainrule, together with the conserved angular momentum;

r = −˜

γ[1− (γ/c)2V (r)]

d

dθ

1

r, (A8)

and, therefore, (again taking γr to be negligible)

γr[1−(γ/c)2V (r)] ≈ −˜2

γ[1− (γ/c)2V (r)]r2

d2

dθ2

1

r. (A9)

Substituting Eqs. (A7) and (A9) into the equation ofmotion Eq. (A6) results in

˜2 d2

dθ2

1

r− γ2[1− (γ/c)2V (r)]GM +

˜2

r= 0. (A10)

Anticipate a solution of Eq. (A10) that is near Keplerianand introduce the radius of a circular orbit for anonrelativistic particle with the same angular momentum,rc ≡ ˜2/GM . The result is

d2

dθ2

rc

r+rc

r= 1 + λ, (A11)

where λ ≡ γ2[1 − (γ/c)2V (r)] − 1 is a correction toNewtonian orbits due to Special Relativity. [Tildenotation is used to emphasize that quantities dependon the exact conserved quantity ˜ defined in Eq. (A5),rather than the angular momentum ` defined in Eq. (30).]

The orbit equation is derived following the methoddescribed in Sec. III. The correction term λ isapproximated by expanding to first-order in 1/c2,neglecting the radial component of velocity, and usingangular momentum to eliminate θ

λ ≈ (˜/rc)2 − V (r)/c2. (A12)

The equation of motion Eq. (A11) is now expressedapproximately as

d2

dθ2

rc

r+rc

r≈ 1 + ε

rc

r+ ε( rc

r

)2, (A13)

13

where ε ≡ (GM/˜c)2. An orbit equation is derived, asdescribed in Sec. III. The equation of motion Eq. (A13)is linearized using rc/r = 1 + 1/s, and (rc/r)

2 ≈ 1 + 2/s.The additional change of variable α ≡ θ

√1− 3ε results

in the familiar differential equation

d2

dα2

sc

s+sc

s≈ 1, (A14)

where sc ≡ (1− 3ε)/(2ε). The solution is similar to thatof Eq. (11)

sc

s≈ 1 +A cosα, (A15)

where A is an arbitrary constant of integration. In termsof the original coordinates, and defining e ≡ 2εA, an orbitequation in the Keplerian limit is described concisely by

rc(1− 2ε)

r≈ 1 + e(1 + ε) cos (1− 3

2 ε)θ. (A16)

This orbit equation does not have the symmetry of thosederived using the two simple models, Eqs. (26) and (42).The methods and approximations used to derive this orbitequation may be applied to other, more physical, modelsthat include relativistic corrections to Kepler’s orbits. Seethe discussions in Sec. VI B and Sec. VI C.