MATH 138A discussion Ryan Ta
University of California, Riverside Winter 2019
Solutions to assigned homework problems from Differential Geometry of Curves and Surfaces by Manfredo Perdigão do Carmo
Assignment 4 – pages 88-92: 1,2,3,4,11,15
2-4.1. Show that the equation of the tangent plane at (x0, y0, z0 ) of a regular surface given by f (x, y, z) = 0, where 0 is a regular
valueo f f , is
fx (x0, y0, z0 )(x − x0 ) + fy (x0, y0, z0 )(y − y0 ) + fz (x0, y0, z0 )(z − z0 ) = 0.
Proof. Let α(t) = (x(t), y(t), z(y)) be a curve on the surface passing through (x0, y0, z0 ) at t = 0. We can take the derivative in
t of both sides of f (x, y, z) = 0, using the multivariable chain rule in doing so, to get
fx (x, y, z)x 0(t) + fy (x, y, z)y 0(t) + fz (x, y, z)z 0(t) = 0.0, y0,
Equivalently, we have
∇ f (x, y, z) · α 0(t) = ( fx (x, y, z), fy (x, y, z), fz (x, y, z)) · (x 0(t), y 0(t), z 0(t))
= fx (x, y, z)x 0(t) + fy (x, y, z)y 0(t) + fz (x, y, z)z 0(t)
= 0.
In particular, at t = 0, we have
∇ f (x0, y0, z0 ) · α 0(0) = 0.
Since our choice of α(t) passing through p at t = 0 is arbitrary, α 0(0) is also arbitrary. In other words, all vectors tangent to
the surface at (x0, y0, z0 ) must be perpendicular to ∇ f (x0, y0, z0 ). In particular, if (x, y, z) is another point in the plane tangent
to the surface at (x0, y0, z0 ), then the vector (x − x0, y − y0, z − z0 ) lies in that tangent plane, and so we have
∇ f (x0, y0, z0 ) · (x − x0, y − y0, z − z0 ) = 0,
or equivalently,
fx (x0, y0, z0 )(x − x0 ) + fy (x0, y0, z0 )(y − y0 ) + fz (x0, y0, z0 )(z − z0 ) = 0,
as desired.
2-4.2. Determine the tangent planes of x 2 + y 2 − z2 = 1 at the points (x, y, 0) and show that they are all parallel to the z-axis.
Proof. Let f (x, y, z) = x 2 + y 2 − z 2 . Then fx (x, y, z) = 2x, fy (x, y, z) = 2y, fz (x, y, z) = −2z; in particular, at the point
(x0, y0, 0), we have fx (x0, y0, 0) = 2x0 , fy (x0, y0, 0) = 2y0 , fz (x0, y0, 0) = 0. So the tangent equation at the point (x0, y0, 0) is
fx (x0, y0, 0)(x − x0 ) + fy (x0, y0, 0)(y − y0 ) + fz (x0, y0, 0)(z − 0) = 0,
or equivalently,
∇ f (x0, y0, 0) · (x − x0, y − y0, z − z0 ) = ( fx (x0, y0, 0), fy (x0, y0, 0), fz (x0, y0, 0)) · (x − x0, y − y0, z − z0 )
= fx (x0, y0, 0)(x − x0 ) + fy (x0, y0, 0)(y − y0 ) + fz (x0, y0, 0)(z − z0 )
= 2x0 (x − x0 ) + 2y0 (y − y0 ) − 0(z − 0)
= 0,
which signifies that ∇ f (x0, y0, 0) is perpendicular to the tangent plane. Now, we consider the vectors (0, 0, ±1) in the z-axis.
Then we also have
∇ f (x0, y0, 0) · (0, 0, ±1) = ( fx (x0, y0, 0), fy (x0, y0, 0), fz (x0, y0, 0)) · (0, 0, ±1)
= fx (x0, y0, 0)(0) + fy (x0, y0, 0)(0) + fz (x0, y0, 0)(±1)
= (2x0 )(0) + (2y0 )(0) + (0)(±1)
= 0,
which signifies that ∇ f (x0, y0, 0) is also perpendicular to the z-axis. Therefore, the tangent plane and the z-axis are parallel.
Since we argued this for the point (x0, y0, 0), we can extend our argument to arbitrary points of the form (x, y, 0), as desired.
2-4.3. Show that the equation of the tangent plane of a surface which is the graph of a differentiable function z = f (x, y), at the point
p0 = (x0, y0 ), is given by
z = f (x0, y0 ) + fx (x0, y0 )(x − x0 ) + fy (x0, y0 )(y − y0 ).
Recall the definition of the differential df of a function f : R2 → R and show that the tangent plane is the graph of the
differential dfp .
Proof. Define g(x, y, z) := f (x, y) − z. Then we are describing the set of all points (x, y, z) satisfying g(x, y, z) = 0 (that
is, satisfying z = f (x, y)). We obtain the partial derivatives gx (x, y, z) = fx (x, y), gy (x, y, z) = fy (x, y), gz (x, y, z) = −1.
According to Exercise 2-4.1, the equation of the tangent plane (applied to g(x, y, z)) is
gx (x0, y0, z0 )(x − x0 ) + gy (x0, y0, z0 )(y − y0 ) + gz (x0, y0, z0 )(z − z0 ) = 0,
or equivalently
fx (x0, y0 )(x − x0 ) + fy (x0, y0 )(y − y0 ) − 1(z − f (x0, y0 )) = 0.
Solving for z, we arrive at
z = f (x0, y0 )(x − x0 ) + fx (x0, y0 )(x − x0 ) + fy (x0, y0 )(y − y0 ),
as desired.
2-4.4. Show that the tangent planes of a surface given by z = x f ( yx ), x , 0, where f is a differentiable function, all pass through the
origin (0, 0, 0).
Proof. Let t = yx . Then z(x, y) = x f ( yx ) = x f (t), and so we obtain the partial derivatives
∂z ∂
zx (x, y) = = (x f (t))
∂x ∂x
∂
= f (t) + x f (t)
∂x
∂t
= f (t) + x f 0(t)
∂x
y y ∂ y
= f + xf0
xy x ∂ x y x
0 y
= f +xf − 2
xy y xy x
= f − f0
x x x
and
∂z ∂
zy (x, y) = = (x f (t))
∂y ∂y
∂
= x f (t)
∂y
∂t
= x f 0(t)
∂y
y ∂ y
= xf0
x ∂y x
∂ 1
0 y
= xf
x ∂y x
y
= f0 .
x
The tangent equation in Exercise 2-4.3 gives us
z(x, y) = f (x0, y0 ) + fx (x0, y0 )(x − x0 ) + fy (x0, y0 )(y − y0 )
= z(x0, y0 ) + zx (x0, y0 )(x − x0 ) + zy (x0, y0 )(y − y0 )
x0 y0 y0 0 y0 y0
= x0 f + f − f (x − x0 ) + f 0 (y − y0 ).
y0 x0 x0 x0 x0
In particular, our equation of the tangent plane satisfies
x0 y0 y0 0 y0 0 y0
z(0, 0) = x0 f + f − f (0 − x0 ) + f (0 − y0 )
y0 x0 x0 x0 x0
x0 y0 y0 0 y0 y0
= x0 f −f x0 + f x0 − f 0 y0
y0 x0 x0 x0 x0
=0
which implies that the tangent plane goes through the point (0, 0, 0) (that is, the origin).
2-4.11. Show that the normals to a parametrized surface given by
x(u, v) = ( f (u) cos v, f (u) sin v, g(u)),
f (u) , 0, g 0 , 0, all pass through the z-axis.
Proof. Given x(u, v) = ( f (u) cos v, f (u) sin v, g(u)), we obtain the derivatives
∂x ∂
xu (u, v) = = ( f (u) cos v, f (u) sin v, g(u))
∂u ∂u
∂ ∂ ∂
= ( f (u) cos v), ( f (u) sin v), g(u)
∂u ∂u ∂u
= ( f 0(u) cos v, f 0(u) sin v, g 0(u))
and
∂x ∂
xv (u, v) = = ( f (u) cos v, f (u) sin v, g(u))
∂v ∂v
∂ ∂ ∂
= ( f (u) cos v), ( f (u) sin v), g(u)
∂v ∂v ∂v
= (− f (u) sin v, f (u) cos v, 0).
So we obtain the cross product
i j k
xu (u, v) × xv (u, v) = f 0(u) cos v f 0(u) sin v g 0(u)
− f (u) sin v f (u) cos v 0
f 0(u) sin v g 0(u) f 0(u) cos v g 0(u) f 0(u) cos v f 0(u) cos v
= i− j+ k
f (u) cos v 0 − f (u) sin v 0 − f (u) sin v f (u) cos v
= (( f 0(u) sin v)(0) − ( f (u) cos v)(g 0(u)))i − (( f 0(u) cos v)(0) − (− f (u) sin v)(g 0(u)))j
− (( f 0(u) cos v)( f (u) cos v) − (− f (u) sin v)( f 0(u) sin v))k
= (− f (u)g 0(u) cos v)i + (− f (u)g 0(u) sin v)j + ( f (u) f 0(u))k
= (− f (u)g 0(u) cos v, − f (u)g 0(u) sin v, f (u) f 0(u))
and its associated magnitude
p
|xu (u, v) × xv (u, v)| = (− f (u)g 0(u) cos v)2 + (− f (u)g 0(u) sin v)2 + ( f (u) f 0(u))2
q
= ( f (u)g 0(u))2 (cos2 v + sin2 v) + ( f (u) f 0(u))2
p
= ( f (u)g 0(u))2 + ( f (u) f 0(u))2
p
= ( f (u))2 ((g 0(u))2 + ( f 0(u))2 )
p
= | f (u)| (g 0(u))2 + ( f 0(u))2 .
According to the definition on page 87 of do Carmo, we obtain the unit normal vector
xu (u, v) × xv (u, v)
N=
|xu (u, v) × xv (u, v)|
(− f (u)g 0(u) cos v, − f (u)g 0(u) sin v, f (u) f 0(u))
= p .
| f (u)| (g 0(u))2 + ( f 0(u))2
Since the problem assumed that f (u) , 0 and g 0 , 0, it follows that |xu × xv | > 0, which means N cannot possibly be
undefined. The line that contains N is given by
α(t) = x(u, v) + tN
(− f (u)g 0(u) cos v, − f (u)g 0(u) sin v, f (u) f 0(u))
= ( f (u) cos v, f (u) sin v, g(u)) + t p
| f (u)| (g 0(u))2 + ( f 0(u))2
for all t ∈ R. However, this expression of α(t) is quickly getting complicated. So we should consider a slight workaround.
To thispend, it is important to observe that, if the line described by α(t) contains N, the same line also contains the vector
| f (u)| (g 0(u))2 + ( f 0(u))2 N = (− f (u)g 0(u) cos v, f (u)g 0(u) sin v, f (u) f 0(u)), which is a scalar multiple of N. So we can
describe the same line as
p
β(t) = x(u, v) + t(| f (u)| (g 0(u))2 + ( f 0(u))2 N)
= ( f (u) cos v, f (u) sin v, g(u)) + t(− f (u)g 0(u) cos v, − f (u)g 0(u) sin v, f (u) f 0(u))
= ( f (u) cos v − t f (u)g 0(u) cos v, f (u) sin v − t f (u)g 0(u) sin v, g(u) + t f (u) f 0(u))
= (1 − tg 0(u)) f (u) cos v, (1 − tg 0(u)) f (u) sin v, g(u) + t f (u) f 0(u)),
which is a much easier expression to work with. Now, to prove that this line passes through the z-axis, we need to find some
t0 ∈ R such that β(t0 ) = (0, 0, k) for some k ∈ R. To satisfy this condition, we must set the first and second coordinates of
β(t0 ) equal to zero; that is, we must set
(1 − t0 g 0(u)) f (u) cos v = 0
(1 − t0 g 0(u)) f (u) sin v = 0,
from which we will solve for t0 . To this end, we can multiply both sides of the first equation by cos v and both sides of the
second equation by sin v so that our system of equations becomes
(1 − t0 g 0(u)) f (u) cos2 v = 0
(1 − t0 g 0(u)) f (u) sin2 v = 0.
So we can add up the two equations to obtain (1 − t0 g 0(u)) f (u) = 0. Since f (u) , 0, we can divide both sides of our latest
equation by f (u) to obtain 1 − t0 g 0(u) = 0, and so t0 = g01(u) . Hence, we have β(t0 ) = (0, 0, g(u) + t0 f (u) f 0(u)), which means
that the line β(t) crosses the z-axis at t = t0 .
2-4.15. Show that if all normals to a connected surface pass through a fixed point, the surface is contained in a sphere.
Proof. Notice that this question is a three-dimensional analog of Exercise 1-5.4; the proof for Exercise 2-4.15 will be copied
verbatim from Exercise 1-5.4 except for minor adjustments. Let p = (x0, y0, z0 ) be a fixed point and n(s) a unit normal vector
of the parametrized surface α(s) = (x(u(s), v(s)), y(u(s), v(s)), z(u(s), v(s))); that is, α 0(s) · n(s) = 0. Since n(s) passes through
p, we have α(s) − p = λn(s) for some scalar λ ∈ R. So we have
d d
(|α(s) − p| 2 ) = ((α(s) − p) · (α(s) − p))
ds ds
d d
= (α(s) − p) · (α(s) − p) + (α(s) − p) · (α(s) − p)
ds ds
d
= 2 (α(s) − p) · (α(s) − p)
ds
= 2α 0(s) · (α(s) − p)
= 2α 0(s) · λn(s)
= 2λα 0(s) · n(s)
= 0.
Hence, |α(s) − p| 2 is constant, which means |α(s) − p| is constant. In other words, for all s ∈ I the distance between α(s) and
p is the same, which implies that the surface parametrized by α(s) is contained in a sphere of center p.
