Math 322 Homework 10

December 5, 2011
1 Part 1
3. One direction: assume that the arc length of any parametrized curve in S is equal to the arc length
of the image curve by . Let : (, ) S be a curve in S. Then () is a curve in S with the
same arc length. That is,

b
a
[

(t)[ dt =

b
a
[( )

(t)[ dt for any choice of a and b between and

. But in order for this to be true for any choice of a and b, we must have [

(t)[ = [( )

(t)[ =

(t),

(t) = ( )

(t), ( )

(t); that is, the rst fundamental form is preserved by . But we

already know that this implies isometry.
For the other direction, assume is an isometry. Then, using the same denitions as above, we want
to show that

(t)[ dt =

[( )

(t)[ dt. But this is trivial; since is an isometry, we already

know that [

(t)[ = [( )

(t)[ for all t (, ), and so were done.

4. Let f : R
2
S
2
be given by
f(u, v) =

4u
u
2
+ v
2
+ 4
,
4v
u
2
+ v
2
+ 4
,
2(u
2
+ v
2
)
u
2
+ v
2
+ 4

Then we have
f
u
=

4(u
2
+ v
2
+ 4)
(u
2
+ v
2
+ 4)
2
,
8uv
(u
2
+ v
2
+ 4)
2
,
16u
(u
2
+ v
2
+ 4)
2

f
v
=

8uv
(u
2
+ v
2
+ 4)
2
,
4(u
2
v
2
+ 4)
(u
2
+ v
2
+ 4)
2
,
16v
(u
2
+ v
2
+ 4)
2

Note that f
u
f
u
= f
v
f
v
and f
u
f
v
= 0 (shown in a previous problem set). So let

f
u
f
u
= (p),
where p = f(u, v). Consider two curves
1
(t) = (u
1
(t), v
1
(t)) and
2
(t) = (u
2
(t), v
2
(t)) with
1
(0) =

2
(0) = p. Then

1
(0) and

2
(0) are both in T
p
(R
2
), and we have
f
u
u

1
(0) + f
v
v

1
(0), f
u
u

2
(0) + f
v
v

2
(0) = f
u
f
u
u

1
(0)u

2
(0) + f
u
f
v
(u

1
(0)v

2
(0) + u

2
(0)v

1
(0)) + f
v
f
v
v

1
(0)v

2
(0)
=
2
(p)

1
(0),

2
(0)
which shows that f is a local conformal map at p. But since p was arbitrarily chosen, we conclude
then that the sphere is locally conformal to the plane.
5. Let p be a point in x
2
(V ) and w be a vector in T
p
(x
2
(V )). Consider a coordinate curve (t) = (s(t), v(t))
with x
2
((0)) = p and
d
dt
x
2
((t))

t=0
= w. First, note that x
1
x
1
2
(p) = x
1
(s(0), t(0)). Call this point
q. Since we know that to prove isometry we only need to show equivalence of the rst fundamental
form, it is enough then to show that
I
p
(w) = I
q

d
dt
x
1
((t))

t=0

1
We have, rst of all, that w is the derivative of
2
(s(t))+v(t)

2
(s(t)) evaluated at 0, or

2
(s(0))s

(0)+
v

(0)

2
(s(0)) + v(0)

2
(s(0))s

(0). Since

(0),

(0) = 0, we can write

I
p
(w) = (s

(0) + v

(0))
2

2
(s(0)),

2
(s(0)) + (v(0)s

(0))
2

(s(0)),

(s(0))
= (s

(0) + v

(0))
2
+ k
2
(s(0))
2
v(0)
2
s

(0)
2
Similarly, the right hand side becomes (s

(0) + v

(0))
2
+ k
1
(0)
2
v(0)
2
s

(0)
2
, and we are done, since
k
1
(s) = k
2
(s).
14. Suppose is locally conformal. Then p S
1
, v
1
, v
2
T
p
(S
1
), there exists
2
> 0 s.t. d
p
(v
1
), d
p
(v
2
) =

2
v
1
, v
2
. So
cos(d
p
(v
1
), d
p
(v
2
)) =
d
p
(v
1
), d
p
(v
2
)
[d
p
(v
1
)[[d
p
(v
2
)[
=

2
v
1
, v
2

[v
1
[[v
2
[
= cos(v
1
, v
2
)
So preserves angles.
Now suppose preserves angles and is dierentiable. Then p S, v
1
, v
2
T
p
(S
1
) we have
cos(d
p
(v
1
), d
p
(v
2
)) = cos(v
1
, v
2
). In particular, let e
1
, e
2
be orthonormal basis of T
p
(S
1
), we
have cos(d
p
(e
i
), d
p
(e
j
)) = cos(e
i
, e
j
) =
ij
. So d
p
(e
i
), d
p
(e
j
is an orthogonal basis of T

p
(S).
That is d is of full rank. By inverse function theorem, is a local dieomorphism.
Now cos(d
p
(e
1
+e
2
), d
p
(e
1
e
2
) = cos(e
1
+e
2
, e
1
e
2
) = 0. So d
p
(e
1
+e
2
), d
p
(e
1
e
2
) = 0. So
|d
p
(e
1
)|
2
= |d
p
(e
2
)|
2
=
2
(p) for some dierentiable function depending only on p.
d
p
(e
1
+ e
2
, d
p
(e
1
+ e
2
))
= |d
p
(e
1
)|
2
+ |d
p
(e
2
)|
2
= ( + )
2
(p)
=
2
(p)e
1
+ e
2
, e
1
+ e
2

So is locally conformal.
15. If p = (x, y), then let (p) = u
2
x
+ u
2
y
.
Let
1
(t) = (x
1
(t), y
1
(t)) and
2
(t) = (x
2
(t), y
2
(t)) be two curves in R
2
with

1
(0) = v
1
and

2
(0) = v
2
.
Then d
p
(v
1
) =
d
dt
[u(x
1
(t), y
1
(t)), v(x
1
(t), y
1
(t))]

t=0
= (u
x
x

1
(0) + u
y
y

1
(0), v
x
x

1
(0) + v
y
y

1
(0)), and
similarly for d
p
(v
2
). So
d
p
(v
1
), d
p
(v
2
) = (u
x
x

1
(0) + u
y
y

1
(0)) (u
x
x

2
(0) + u
y
y

2
(0)) + (v
x
x

1
(0) + v
y
y

1
(0)) (v
x
x

2
(0) + v
y
y

2
(0))
= u
2
x
(x

1
(0)x

2
(0) + y

1
(0)y

2
(0)) + u
2
y
(x

1
(0)x

2
(0) + y

1
(0)y

2
(0))
where we have used the fact that u
2
x
= v
2
y
, u
2
y
= v
2
x
, and u
x
u
y
= v
x
v
y
. The last expression, of course,
equal to (p)v
1
, v
2
, meaning that is a local conformal map wherever (p) ,= 0. But (p) is never
zero on R
2
Q by denition, and so were done.
16. If log tan
1
2
= u, then tan
1
2
= e
u
= e
2u
=
1cos
1+cos
, which rearranges to cos =
e
2u
1
e
2u
+1
= tanh u.
The identity tanh
2
u + sech
2
u = 1 then gives sin = sech u. So (sech ucos v, sech usin v, tanh u) is
an equivalent parametrization of the unit sphere, and we can ip our sphere across the xy plane to get
the parametrization
y(u, v) = (sech ucos v, sech usin v, tanh u)
2
as desired. We then have
y
u
= (sech utanh ucos v, sech utanh usin v, sech
2
u)
y
v
= (sech usin v, sech ucos v, 0)
E = sech
2
utanh
2
u + sech
4
u = sech
2
u
F = 0
G = sech
2
u(sin
2
v + cos
2
v) = sech
2
u
as desired.
2. We conclude directly from eqn. (2) on p. 232 that

1
11
=
1
2

2
11
=
1
2

1
12
=
1
2

2
12
=
1
2

1
22
=
1
2

2
22
=
1
2

v
So the Gauss formula gives
K =

uu

2
u
2
2
+

vv

2
v
2
2


2
v
4
2
+

2
u
4
2
+

2
v
4
2


2
u
4
2
Simplifying and rearranging gives
K =
1
2

uu

2
u

2
+

vv

2
v

,
which is equivalent to what we wanted to show. So if = (u
2
+ v
2
+ c)
2
, then

1
2

uu

2
u

2
=
1
2(u
2
+ v
2
+ c)
2

(u
2
+ v
2
+ c)
2
(4(u
2
+ v
2
+ c)
3
+ 24u
2
(u
2
+ v
2
+ c)
4
) 16u
2
(u
2
+ v
2
+ c)
6
(u
2
+ v
2
+ c)
4
which simplies to 2u
2
+ 2v
2
+ 2c. The analaglous equation with v simplies to 2u
2
2v
2
+ 2c.
Adding these expressions gives K = 4c, as desired.
3. For x, we have E = 1+
1
u
2
, F = 0, G = u
2
, e =
1
u

1+u
2
, f = 0, and g =
u

1+u
2
, giving K =
1
(1+u
2
)
2
.
For x, we have E = 1, F = 0, G = 1 + u
2
, e = 0, f =
1
1+u
2
, and g = 0, also giving K =
1
(1+u
2
)
2
.
However, consider the coordinate curve (t) = (t, 0). In x, we have x((1)) = (1, 0, 0), and the tangent
line is
d
dt
x((t))

t=1
= (1, 0, 1), which has magnitude

2. But
d
dt
x((t))

t=1
= (1, 0, 0), which has
magnitude 1 ,=

2. So x x
1
is not an isometry.
3
6. We know that K =
egf
2
EGF
2
= 1. However, all the derivatives of E, F, G, e, f, and g are 0, so all
the Christoel coecients are 0. The Gauss formula then EK = 0 = K = 0. But this is a
contradiction, since we know that K = 1. So there is no such surface.
8. (a) We have x(u, v) = (u, v, 0), so x
u
= (1, 0, 0) and x
v
= (0, 1, 0). We then get E = 1, F = 0, and
G = 1. Since all the derivatives of E, F, and G are 0, all the Christoel symbols are also zero,
giving us K = 0.
(b) We have x(u, v) = (ucos v, usin v, 0), so x
u
= (cos v, sin v, 0) and x
v
= (usin v, ucos v, 0). So
E = 1, F = 0, and G = u
2
. The Christoel symbols are then
1
11
=
2
11
=
1
12
=
2
22
= 0,
2
12
=
1
u
,
and
1
22
= u. The Gauss formula then becomes EK =
1
u
2
0 +0 +
1
u
2
0 0 = K = 0,
as expected.
2 Part 2
1. (a) Let (s) be a parametrization of C by arc length. Since C is a geodesic, we know from Def. 8a
(p. 246) that the unit normal n =
1
k

(s) is equal to the normal to the surface N((s)). So if

n = b

= t n

, then we can substitue dN

p
(

(s)) for n

where p = (s). But since is a line of

curvature, we know that dN
p
(

(s)) = k
1

(s) = k
1
t for some k
1
. So n = t k
1
t = 0 = = 0.
Since a plane curve is by denition a curve with zero torsion, we are done.
3. By decree of Prof. Stoiciu, we can ignore the without using Prop. 5 part of the problem. Prop. 5
guarantees the uniqueness of a geodesic at a point in a particular direction. So we only need to show
that straight lines are geodesics in the plane. This is simple: if (s) parametrizes a straight line, then
by Denition 8a, we need only show that

(s) is normal to the tangent plane (which, in this case,

is simply the plane that the line lies in). But

(s) = 0, which is (trivially) normal to the plane.

Therefore straight lines are the only geodesics of a plane.
4