

 tan−1 (y/x) x >0

ϑ= tan−1 (y/x) + 2π x > 0, y < 0

Summary Notes for tan−1 (y/x) + π x <0
d V = r dr dϑ dz
Spherical Coordinates
Vector Calculus x = ρ cos(ϑ) sin(φ), 0≤φ≤π
y = ρ sin(ϑ) sin(φ), 0 ≤ ϑ ≤ 2π
z = ρ√cos(φ), 0≤ρ
Differentiability ρ = x 2 + y 2 + z 2

f : U ⊂ Rn → R is differentiable at x0 if there is a  tan−1 (y/x) x >0
vector a such that ϑ= tan−1 (y/x) + 2π x > 0, y < 0




f ( tan−1 (y/x) + π x <0
x − x0 )

x0 ) − a · (
x ) − f (  √
−1 √ x 2 +y 2

lim =0 φ = sin
x→
x0 
x − x0  x 2 +y 2 +z 2
d V = ρ 2 sin(φ) dρ dφ dϑ
If such an a exists it must be General Change of Coordinates
( ∂∂xf1 (
x0 ), ∂∂xf2 (
x0 ), . . . , ∂∂xfn (
x0 )) called D( f )(
x0 ).  ∂x ∂y 
x = x(u, v) ∂(x,y) ∂u ∂u

F : U ⊂ R → Rn is differentiable at x0 if there is
m
= det ∂ x ∂ y
y = y(u, v) ∂(u,v)
∂v ∂v
a n × m matrix A such that
  D f (x, y) d x d y =


 F(x ) − F( x0 ) − A( x − x0 )
∂(x,y)

lim =0 D  f (x(u, v), y(u, v))
∂(u,v)

du dv
x→x0 x − x0 
where D  (u, v)
→ (x, y) ∈ D
If such an A exists it must be Gradient
 ∂ F1 ∂ F1 
del or nabla: ∇ = ( ∂∂x , ∂∂y , ∂z ∂
)
∂ x1
( x0 ) · · · ∂ xm
(
x0 ))
 . .. 
grad( f ) = ∇ f = ( ∂∂ xf , ∂∂ yf , ∂∂zf )
 .. . 
∂ Fn ∂ Fn
∇ f points in the direction of greatest increase of f .
∂ x1
(x0 ) · · · ∂ xm
(x 0 ))

called D(F)( x0 ).
|∇ f | is the rate of increase in this direction.

If all the partials ∂∂ xFji are continuous on a
Du ( f ) is the rate of change of f in the direction of
the unit vector u: Du ( f ) = u · ∇ f
neighbourhood of x0 then F is differentiable at x0 .

∇ f is perpendicular to the level sets of f : if f (x, y)
Curves
is a scalar field on R2 then ∇ f is perpendicular to the

parametrization: c(t) = (x(t), y(t), z(t)), for
level curves of f ; if f (x, y, z) is a scalar field on R3
a≤t ≤b
then ∇ f is perpendicular to the level surfaces of f .

tangent vector: c  (t) √ = (x  (t), y  (t), z  (t))
 Line Integrals

speed: v = | c (t)| = x  (t)2 + y  (t)2 + z  (t)2

unit tangent: T (t) = c  (t)/| c  (t)|
C is a curve parametrized by
b 

arc length s = a | c (t)| dt c(t) = (x(t), y(t), z(t)) for a ≤ t ≤ b, F is a vector
on R . b
3

arc length parametrization: s(t) gives the length of field
the curve from a to t: s(t) = a |
t
c  (u)| du, thus
C F · ds = a F( c(t)) · c  (t) dt where the curve is
ds
= |c  (t)|, i.e. ds = | c  (t)| dt. traversed from c(a)to c(b).
dt

− C is the same curve as C geometrically, but
traversed
in the opposite
direction.

Cylindrical Coordinates
x = r cos(ϑ) 
−C F · ds = − C F · ds
0 ≤ θ ≤ 2π
C ∇ f · ds = f (end) − f (beginning) =
y = r sin(ϑ)
0≤r f (
c(b)) − f (c(a))
z=√z 2
r = x +y 2
for a scalar field we set:

-i-
 b
f (x, y) ds = a f ( c  (t)| dt
c(t)) | area.
C b
f (x, y) d x = f (
c (t)) x 
(t) dt
area(S) = S d S = D |Tu × Tv | du dv.
C ba

for the parametrization of the sphere of radius a
C f (x, y) dy = a f ( c(t)) y  (t) dt
given above |Tφ × Tϑ | = a 2 sin(φ)
Conservative Fields
a parametrization r(u, v) of a surface S orients a


a vector field F is conservative if the line integral surface via the normal vector Tu × Tv
C F · ds = 0
for any closed loop C, equivalently for
a simple closed curve C which bounds an oriented
any path C, C F · ds depends only on the endpoints surface S inherits an orientation from S: an
of C. upstanding person will traverse C so that S is to her

∇ f is conservative for any scalar field f . left.

if F is conservative on an open connected region
(x,y) Flux through a Surface
D,then F = ∇ f where f (x, y) = (a,b) F · ds and

if F a vector field on a surface S parametrized by
the line integral is taken along any path in D joining r(u, v) = x(u, v), y(u, v), z(u, v) (u, v) ∈ D, the
(a, b) to (x, y). flux of F through S is

a region D is simply connected if it has no ‘holes’. F · d S = F r(u, v) · T × T  du dv
v

if F = P(x, y)i + Q(x, y) j and ∂∂Qx = ∂∂Py then F is S D u
Divergence
conservative provided that F is continuous on an
  + R k is a vector
open simply connected region D and the first partials
if F = P i + Q j   field, then
of P and Q are continuous on D. div(F) = ∇ · F = ∂∂ Px , ∂∂Qy , ∂∂zR .
Green’s Theorem the flux of F through S
∂ Q
div(F)(x, y, z) = lim where
∂P

C P d x + Q dy = D ∂ x − ∂ y d A where C is a vol→0 volume enclosed by S
the limit is taken over the family of closed surfaces
simple closed curve in the plane, D is the interior of
which contain (x, y, z).
C, and C is traversed with the interior to the left.
Vector Identities
Circulation and Curl    



i j 


scalar grad vector  curl  vector  div scalar


k

fields
→ fields
→ fields
→ fields

curl(F) = ∇ × F =

∂∂x ∂∂y ∂z ∂


=
curl(grad( f )) = 0 for any scalar field f

P Q R

div(curl(F)) = 0 for any vector field F
 
∂R ∂Q ∂P ∂R ∂Q
− ∂z , ∂z − ∂ x , ∂ x − ∂ y for∂P
if curl(F) = 0 then F = ∇ f for a scalar field f
∂y
  provided F is defined on all of R3 and the first
F = P(x, y, z), Q(x, y, z), R(x, y, z) .
partials of F are continuous

if F is a vector field and C is a closed curve, the

if div(F) = 0 then F = curl(G) for some vector
circulation of F around C is C F · d r.
field G provided F is defined on all of R3 and the first

if d is a unit vector the component of partials of F are continuous.
curl(F)(x, y, z) in the direction of d is
circulation around C Stokes’s Theorem
lim where the limit is taken
area→0 area in C
S curl(F) · d S = ∂ S F · ds where F is a vector
over the family of curves in the plane perpendicular field on S with continuous first partials, S is an
to d which shrink to the point (x, y, z). oriented surface bounded by a simple closed curve ∂ S
Surfaces oriented by S.

a surface S is parametrized by  Divergence Theorem

r(u, v) = x(u, v), y(u, v), z(u, v) for (u, v) ∈ D if
∂ V F · d S = V div(F) d V where F is a vector
r(u, v) runs over S as (u, v) runs over D. field on the simple solid region V with continuous

the sphere  of radius a is parametrized by first partials, and the boundary surface ∂ V is oriented
r(φ, ϑ) =  a cos(ϑ) sin(φ), a sin(ϑ) sin(φ), with an outward pointing normal.
a cos(φ) , 0 ≤ φ ≤ π, 0 ≤ ϑ ≤ 2π

d S = |Tu × Tv | du dv gives the element of surface November 29, 2002

-ii-