Math221: HW# 6 solutions

Andy Royston
November 7, 2005

9.3.16 The line element (infinitesimal distance) squared on the cylinder is ds2 = R2 dθ2 + dz 2 ,
where R is the radius of the cylinder, a constant. Hence, we wish to minimize
Z B Z B√ Z B√
I= ds = 2 2 2
R dθ + dz = R2 + z 02 dθ. (1)
A A A
I chose θ as the independent variable. By making this choice, I’ve already restricted the set of lines
I’m considering. The lines parallel to the z-axis have θ = const, so that dθ = 0. So by “dividing
by dθ,” I’ve already assumed that the curve I’m trying to solve for is not one of these. In order to
see whether these lines are geodesics, we will need to pick z as the independent
√ variable. But first,
let’s assume dθ 6= 0. Then the Euler equation following from F (θ, z, z 0 ) = R2 + z 02 is

z0
 
d ∂F ∂F
= 0 ⇒ = √ = c1 , (2)
dθ ∂z 0 ∂z 0 R2 + z 02
where c1 is a constant. Let’s solve this for z 0 :

c1 R
z 02 = c21 (R2 + z 02 ) ⇒ z 02 (1 − c21 ) = c21 R2 ⇒ z0 = p ≡ c2 , (3)
1 − c21

where I’ve just defined the constant c2 . The solution is z = c2 θ+c3 , where c3 is yet another constant.
These are helices. In the special case c2 = 0, we have that z is constant (and θ is unrestricted), so
these are circles.
Let us now return to the issue of the lines parallel to the z-axis. In order to include them in the
analysis, choose z as the independent variable:
Z B Z B√ Z B√
I= ds = 2 2
R dθ + dz = 2 R2 θ02 + 1dz. (4)
A A A
0
√
The Euler equation for F̃ (z, θ, θ ) = R2 θ02 + 1 is

R2 θ 0
 
d ∂ F̃ ∂ F̃
=0 ⇒ = √ = c4 . (5)
dz ∂θ0 ∂θ0 R2 θ02 + 1

1
These yield

R4 θ02 = c24 (R2 θ02 + 1) ⇒ θ02 (R4 − c24 R2 ) = c24 ⇒ θ 0 = c5 , (6)

ie. θ0 is a constant. The solution is θ = c5 z + c6 . These are in general the same solutions we found
before; take c2 = 1/c5 , c3 = −c6 /c5 . However, in the case c5 = 0, we can’t do this. These are
indeed the parallel lines θ = const, and we see they are geodesics as well.
The set of all geodesics can be compactly summarized by three arbitrary constants:

az + bθ = c. (7)
If a 6= 0, divide by it to get the first set of results, circles and helices. If a = 0, we recover the
θ = const. lines.

9.3.17 The line element squared in cylindrical coordinates is ds2 = dr2 + r2 dθ2 + dz 2 . (Note that
if we were on the surface of a cylinder, so that r is constant and dr zero, then we would recover
the form in the previous problem). Now we are on the surface of a cone, described by the relation
z 2 = r2 . Then z = ±r, so dz = ±dr, or dz 2 = dr2 . Hence, the line element on the cone is

ds2cone = 2dr2 + r2 dθ2 . (8)

Then, to find the geodesics, minimize
Z B Z B √ Z B √
I= dscone = 2dr2 + r2 dθ2 = 2 + r2 θ02 dr. (9)
A A A
√
I’ve chosen r as the independent variable.1 The Euler equation for F (r, θ, θ0 ) = 2 + r2 θ02 is

r2 θ0
 
d ∂F ∂F
= 0 ⇒ = √ = K, (10)
dr ∂θ0 ∂θ0 r2 θ02 + 2
for K a constant. Solve for θ0 :
√
2K
r4 θ02 = K 2 (r2 θ02 + 2) ⇒ θ02 (r4 − K 2 r2 ) = 2K 2 ⇒ θ0 = √ . (11)
r r − K2
2

Hence,
Z √ Z √
2K 2K
θ+c= dr √ = dr p . (12)
r r2 − K 2 r2 1 − K 2 /r2
Following the example in the book, let cos ξ = K/r. Then − sin ξdξ == − rK2 dr. The integral
becomes
1
Again, this means I am missing possible solutions which have dr = 0 along them. In this case, these are circles
parallel to the x-y plane. Doing the analysis with θ as the independent variable would lead one to conclude that the
circles are also geodesics, but I’m not going to do it here.

2
 √ √
Z
2K
Z
2 sin ξdξ √ Z √ √
dr p = p = 2 dξ = 2ξ = 2 arccos (K/r). (13)
r2 1 − K 2 /r2 1 − cos ξ 2
Hence,
√ θ+c
θ+c= 2 arccos (K/r) r cos ( √ ) = K.
or (14)
2
Incidentally, this is the equation for a plane intersecting the cone, and is known as a conic section.
Conic sections describe the general motion of planets about the sun, say, in Newtonian gravity.

9.4.3 We essentially follow the analysis of the book. Define the zero of gravitational potential
energy to be at the height of the starting point y0 = 0. If positive y corresponds to downwards,
then the potential energy is V = −mgy. The kinetic energy is as usual, K = 12 mv 2 = 12 m(ds/dt)2 .
Total energy is conserved, so the energy at any point along the wire is equal to the initial total
energy. This time, the initial total energy is E0 = 12 mv02 . Thus
 2  2
1 ds 1 2 ds
−mgy + m = mv0 ⇒ = v02 + 2gy. (15)
2 dt 2 dt
The total trip time, which is what we want to minimize, is
B B B B
√
x02 + 1dy
Z Z Z Z
ds 1 ds 1
T = dt = p = √ √ =√ √ . (16)
A A v02 + 2gy 2g A a+y 2g A a+y
Here, I’ve defined the constant a = v02 /2g. Also, I chose x as the dependent variable because it
does not appear in the functional
s
1 + x02
F (y, x, x0 ) = . (17)
a+y
The Euler equation gives

x0 √
 
d ∂F ∂F
=0 ⇒ = √ √ = c, (18)
dy ∂x0 ∂x 0
1 + x02 a + y
c a constant. Solve for x0 :

s
c(a + y)
x02 = (1 + x02 )c(a + y)2 ⇒ x02 (1 − c(a + y)) = c(a + y) ⇒ x0 = . (19)
1 − c(a + y)

This is just like the book result for the case E0 = 0, with y replaced by y + a. Therefore, try the
substitution

3
 1
c(a + y) = sin2 (θ/2) = (1 − cos θ) ⇒ cdy = sin (θ/2) cos (θ/2)dθ. (20)
2
Then,

s
sin2 (θ/2)
Z Z
1 1
x= ( sin (θ/2) cos (θ/2)dθ) = dθ sin (θ/2) cos (θ/2) tan (θ/2)
c 1 − sin2 (θ/2) c
Z Z
1 1 1
= 2
dθ sin (θ/2) = dθ(1 − cos θ) = (θ − sin θ) + c0 , (21)
c 2c 2c
where c0 is an integration constant. Thus, we again get the parametric equations for the cycloid:

1
x= (θ − sin θ) + c0 ,
2c
1 v2
y = (1 − cos θ) − 0 . (22)
2c 2g
This time, we can not set c0 = 0 if we want the particle to start at the origin (x0 , y0 ) = (0, 0).

9.4.4 Use energy conservation to find an expression for the velocity v. The gravitational potential
energy inside the earth, r < R, at a distance r from the center is
mg 2
(r − 3R2 ).
V (r) = (23)
2R
You can get
R this from problem 6.8.21 as the book suggests, or you can derive it yourself by inte-
grating − F · dr from infinity into r, where F is the gravitational force. Hence, at any point along
the path the total energy is
 2
1 2 mg 2 2 1 ds mg 2
E = mv + (r − 3R ) = + (r − 3R2 ). (24)
2 2R 2 dt 2R
Energy is conserved, so this must equal the initial energy. The initial energy is entirely potential,
since the train starts from rest. It is given by setting r = R in the expression for V (r):
mg 2
E0 = (R − 3R2 ) = −mgR. (25)
2R
Now set E = E0 and solve for ds/dt:

 2
1 ds mg 2
+ (r − 3R2 ) = −mgR
2 dt 2R
 2
ds g g
= (3R2 − r2 ) − 2gR = (R2 − r2 ). (26)
dt R R

4
Hence, we obtain the following expression for the total trip time:

B B
s
B
√ s Z r
dr2 + r2 dθ2 R B 1 + r2 θ02
Z Z Z
ds R
T = dt = pg = √ = dr. (27)
A A R
(R 2 − r2 ) g A R2 − r 2 g A R2 − r 2

I chose θ as the dependent variable along the path, since it does not appear in the functional
r
1 + r2 θ02
F (r, θ, θ0 ) = . (28)
R2 − r 2
Therefore, the Euler equation is

r2 θ0
 
d ∂F
= 0 ⇒ √ √ = K. (29)
dr ∂θ0 1 + r2 θ02 R2 − r2
Now, by symmetry, when the path comes closest to the center of the earth, it’s change in radial
direction, dr, will be zero. Call this distance r0 . Then we can eliminate the integration constant
K in favor of r0 by demanding that r0 = dr/dθ = 0 when r = r0 . First, however, we need to
temporarily replace θ0 in favor of r0 . Divide top and bottom by θ0 . Then

r2 r2
p √ =√ √ = K. (30)
1/θ02 + r2 R2 − r2 r02 + r2 R2 − r2
Now set r0 = 0 and r = r0 :

r02 r0
p =K ⇒ K=p . (31)
r02 (R2 − r02 ) R2 − r02
Hence,

r2 θ0 r0
√ √ =p . (32)
2 02
1+r θ R −r 2 2 R2 − r02
Let us solve for θ0 as a function of r.

r02
 2 2

4 02 2 02 2 R −r
r θ = 2 2 2
(1 + r θ )(R − r ) = r0 (1 + r2 θ02 )
R − r02 r2 − r02
  2 2
  2 2

02 4 2 2 R −r 2 R −r
⇒ θ r − r0 r = r0
r2 − r02 r2 − r02
 2
R − r2

r02 2
r − r02 r02 (R2 − r2 )
⇒ θ02 r2 =  2 2 = . (33)
r2 (R2 − r02 ) − r02 (R2 − r2 )

2 2 R −r
r − r0
r2 − r02

5
Now, one could attempt to integrate this to find an expression for the path, θ = θ(r). However,
since we now know the quantity
R r2 θ02 along the path, we can plug it back into the expression for
the total trip time, T = ds/v, and evaluate that. Recall that the trip time is
s Z r
R B 1 + r2 θ02
T = dr. (34)
g A R2 − r 2
A and B are two points on the surface of the earth, r = R. Again, by symmetry, the time it takes
to go from A to B must be twice the time it takes to go half way, from r = r0 to r = R. Thus
s Z r
R R 1 + r2 θ02
T =2 dr. (35)
g r0 R2 − r 2
Now we need to plug in the expression for r2 θ0 and simplify considerably. Let us look at the
numerator in the square root first:

r02 (R2 − r2 ) r2 (R − r02 ) − r02 (R2 − r2 ) + r02 (R2 − r2 )

1 + r2 θ0 = 1 + =
r2 (R2 − r02 ) − r02 (R2 − r2 ) r2 (R2 − r02 ) − r02 (R2 − r2 )
r2 (R2 − r02 )
= 2 . (36)
r (R − r02 ) − r02 (R2 − r2 )

Now divide top and bottom by (R2 − r02 ) and something nice happens:

r2 (R2 − r02 ) r2 r2
= r02
= r02 r 2 R2
r2 (R − r02 ) − r02 (R2 − r2 ) r2 − (R2 − r2 )
r2 (1 + ) − R20−r2
R −r02
2 R −r02
2
0
2 2 2 2 2
 
r r R − r0 r
= R2 R2
= R2 = . (37)
2 2
r ( R2 −r2 ) − r0 ( R2 −r2 ) R2 −r 2
2 2
(r − r0 ) R 2 r − r02
2
0 0 0

The term is parentheses is just a constant which we will be able to pull out of the integral. Putting
everything together,

s r s 1/2 s
R R 2
r2 θ02 R − r02 r2
Z Z
R 1+ R
T =2 dr = 2 dr
g r0 R2 − r 2 g r0 R2 (R2 − r2 )(r2 − r02 )
s
R2 − r02 R
Z
rdr
=2 √ p . (38)
gR r0 R2 − r2 r2 − r02
The remaining integral to do is not easy, but it is of a standard form that you can look up in a
table. The result is simply the numerical factor π/2; in particular, it does not depend on R or r0
as one might naively expect. Hence, the total trip time is

6
 s
R2 − r02
T =π . (39)
gR
For r0 = 0, so that the path goes through the center of the earth, we have
s
R
T =π ≈ 42.2min. (40)
g
This is precisely half of the total period we found for a particle executing simple harmonic motion
along a tunnel through the center of the earth (problem 8.5.35). For r0 = 0.99R, we find
s
R(1 − (0.99)2 )
T =π ≈ 6.0min. (41)
g

9.5.3 The line element squared in spherical coordinates is ds2 = dr2 + r2 dθ2 + r2 sin2 θdφ2 . Hence,
the kinetic energy of a particle is
 2
1 ds 1
K= m = m(ṙ2 + r2 θ̇2 + r2 sin2 θφ̇2 ), (42)
2 dt 2
where a dot denotes time derivative. Then the action2
Z Z
1
S = dtL = dt( m(ṙ2 + r2 θ̇2 + r2 sin2 θφ̇2 ) − V (r, θ, φ)), (43)
2
where V (r, θ, φ) is the potential energy. The independent variable is time, and the dependent
variables are the coordinates r, θ, φ; ie. L = L(t, r, ṙ, θ, θ̇, φ, φ̇). The Euler-Lagrange equations are

 
d ∂L ∂L
− = 0,
dt ∂ ṙ ∂r
 
d ∂L ∂L
− = 0,
dt ∂ θ̇ ∂θ
 
d ∂L ∂L
− = 0. (44)
dt ∂ φ̇ ∂φ

These become
2
The action is always denoted S in physics. It is the same object as the I in the text.

7
 d ∂V
(mṙ) − (mrθ̇2 + mr sin2 θφ̇2 − ) = 0,
dt ∂r
d ∂V
(mr2 θ̇) − (mr2 φ̇2 sin θ cos θ − ) = 0,
dt ∂θ
d ∂V
(mr2 sin2 θφ̇) − (− ) = 0. (45)
dt ∂φ

Now let the d/dt’s act on whatever is in the parentheses. Be careful to make appropriate use of the
chain rule and product rule:

 
2 2 2 ∂V
m r̈ − r(θ̇ + sin θφ̇ ) + = 0,
∂r
 
2 2 2 ∂V
m 2rṙθ̇ + r θ̈ − r φ̇ sin θ cos θ + = 0,
∂θ
 
2 2 2 2 ∂V
m 2rṙ sin θφ̇ + 2r φ̇θ̇ sin θ cos θ + r sin θφ̈ + = 0. (46)
∂φ

9.5.4 The kinetic energy of the pendulum is K = 12 mvtan.

2
, where vtan. is the tangential velocity,
given by vtan. = lθ̇. Hence
1
K = ml2 θ̇2 . (47)
2
Let us define the zero of gravitational potential energy to be at the lowest height of the pendulum
bob, where θ = 0. Then, V = mgh, where h is the height in the y-direction. In terms of θ,
h = l − l cos θ. Thus,

V = mgl(1 − cos θ). (48)

Therefore the Lagrangian is
1 1
L(t, θ, θ̇) = K − V = ml2 θ̇2 − mgl(1 − cos θ) = ml2 θ̇2 + mgl cos θ − mgl. (49)
2 2
The Euler-Lagrange equation is

 
d ∂L ∂L d
− =0 ⇒ (ml2 θ̇) − (−mgl sin θ) = 0
dt ∂ θ̇ ∂θ dt
g
⇒ θ̈ + sin θ = 0. (50)
l
This is the equation we have derived previously from Newton’s law.

8
9.5.6 The line element squared, on the surface of a sphere of radius a is ds2 = a2 dθ2 +a2 sin2 θdφ2 .
(Compare with problem 9.5.3, letting r = a, dr = 0). Note that in spherical coordinates, one
typically defines the 0 of the polar angle θ at the north pole. However, for us it more convenient
to define it from the south pole; clearly the line element doesn’t care since the sphere is totally
symmetric. Think of a mass hanging from a thread of length a. The angle it makes with the
vertical is θ, and the angle it rotates about the vertical in φ.
Then its kinetic energy is
1
K = m(a2 θ̇2 + a2 sin2 θφ̇2 ). (51)
2
Again, define the zero of potential energy to be at the base of the sphere. Then V = mgh, with h
the height in the z-direction above the base. The expression is the same as before: h = a − a cos θ.
Hence

V = mga(1 − cos θ), (52)

and therefore,

1
L(t, θ, θ̇, φ, φ̇) = m(a2 θ̇2 + a2 sin2 θφ̇2 ) − mga(1 − cos θ)
2
1
= m(a2 θ̇2 + a2 sin2 θφ̇2 ) + mga cos θ − mga. (53)
2
The Euler-Lagrange equation for θ is3

 
d ∂L ∂L
− =0
dt ∂ θ̇ ∂θ
d
⇒ (ma2 θ̇) − (ma2 sin θ cos θφ̇2 − mga cos θ = 0
dt
g
⇒ θ̈ − sin θ cos θφ̇2 + sin θ = 0. (54)
a
The Euler-Lagrange equation for φ is

 
d ∂L ∂L
− =0
dt ∂ φ̇ ∂φ
 
d 2 2
⇒ ma sin θφ̇ = 0. (55)
dt
3
This differs from the back of the book precisely because they are using θ as defined from the north pole of the
sphere. From a physical standpoint, this is a poor choice of coordinates. Denoting that angle θn and my angle θs , the
relation is θs = π − θn . This means θ¨s = −θ¨n , sin θs = sin θn , and cos θs = − cos θn . Plugging in these substitutions
will give the book result.

9
One could expand the time derivative, but it is much more instructive not to. This equation is
telling us that the quantity ma2 sin2 θφ̇ is a constant in time. In fact, it is the angular momentum
of the system about the z-axis in disguise, and this is the statement that angular momentum is
conserved.

9.5.9 We have the same cone as in problem 9.3.17. The line element squared is ds2 = 2dr2 +r2 dθ2 .
Therefore, the kinetic energy is
1
K = m(2ṙ2 + r2 θ̇2 ). (56)
2
Defining the zero of potential energy at the origin, we have V = mgz. We want to eliminate z in
favor of r. On the cone z = r, so V = mgr. Thus, the Lagrangian is
1
L(t, r, ṙ, θ, θ̇) = m(2ṙ2 + r2 θ̇2 ) − mgr. (57)
2
The Euler-Lagrange equation for r is
d 1 g
(2mṙ) − (mrθ̇2 − mg) = 0 ⇒ r̈ − rθ̇2 + = 0. (58)
dt 2 2
The Euler-Lagrange equation for θ is
d
(mr2 θ̇) = 0. (59)
dt

10