where
μ 0 =
∫ 2 π
0
F 0 ·dξ, F 0 =F(x 0 , y 0 , z 0 )
μ 1 =
∫ 2 π
0
dF
d ·d
ξ
μ 2 =
∫ 2 π
0
1
2!
d^2 F
d^2
·dξ
···
where all the derivatives are evaluated at = 0 and dξ= (−sin θˆe 1 + cos θˆe 2 )dθ. The
vector F 0 is a constant and the integral μ 0 is easily shown to be zero. The vector
dF
d
evaluated at = 0 , when expanded is given by
dF
d =
∂F
∂x cos θ+
∂F
∂y sin θ=
(
∂F 1
∂x cos θ+
∂F 1
∂y sin θ
)
ˆe 1
+
(
∂F 2
∂x cos θ+
∂F 2
∂y sin θ
)
ˆe 2
+
(
∂F 3
∂x
cos θ+∂F^3
∂y
sin θ
)
ˆe 3 ,
where the partial derivatives are all evaluated at = 0.It is readily verified that the
integral μ 1 reduces to
μ 1 =π
(
∂F 2
∂x −
∂F 1
∂y
)
.
The area of the circle surrounding P 0 is π^2 ,and consequently the ratio of the circu-
lation divided by the area in the limit as tends toward zero produces
( curl F)·ˆe 3 =∂F^2
∂x
−∂F^1
∂y
. (8 .35)
Similarly, by considering other planes through the point P 0 which are parallel to the
xz and yz planes, arguments similar to those above produce the relations
( curl F)·ˆe 2 =
∂F 1
∂z −
∂F 3
∂x and ( curl
F)·ˆe 1 =∂F^3
∂y −
∂F 2
∂z. (8 .36)
Adding these components gives the mathematical expression for curl F. One finds
the curl F can be written as
curl F =
(
∂F 3
∂y
−∂F^2
∂z
)
ˆe 1 +
(
∂F 1
∂z
−∂F^3
∂x
)
ˆe 2 +
(
∂F 2
∂x
−∂F^1
∂y
)
ˆe 3. (8 .37)