The Chain Rule
Most of the functions we use are composites of simpler functions. For example, (2x
2
+ 1)3
is calculated by evaluating 2x
2
+ 1 and then cubing that number. If
f(x)
= 2x
2
+ 1 and
g(x) = x
3,
then
g(f(x»
=
f(x)3
= (2x
2
+ 1)3.
Similarly, we can write
,J3x
+ 1 as the composition
of
..jX
and 3x +1: if
f(x)
=3x +1 and
g(x)
=
..jX,
then
g(f(x»
= J
f(x)
=
,J3x
+ 1.
Here are some other examples:
1 . 1
(i)
-2-
=
g(f(x»
with g(x) =
-,
f(x)
= x
2
+4;
x
+4
x
(ii) (3x - 1)4 =
g(f(x»
with
g(x)
= x
4
,
f(x)
= 3x - 1;
1 1
(iii)
.JX2+X
=
h(g(f(x)))
with
h(x)
=
-,
g(x)
=
,JX,
f(x)
= x
2
+x.
x
2
+x
x
Using dependent variable notation
(y,
z,
W,
etc.) instead of functional notation
(f(x),
g(x),
etc.), the above examples would look like this:
(1
') .
f 1 d 2 4 h 1
1 Z = - an y = x +
,t
en z =
-2--;
y x
+4
(ii) if z = y4 and y = 3x - 1, then z = (3x - 1)4;
(iii) if z = 1 and W = If and y = x
2
+ x, then z =
b.
w
VJ
~r~
Situations like these are quite common in physical applications. For example, if the force F
on an object depends on its position x, and the position x depends on the time t, then F is a
(composite) function of
t. Because most functions that occur in practice are compositions
of
simpler functions, we proceed to find the rule for differentiating such functions.
Consider the general situation in which z is a function of
y and y is a function of x, so
that z also becomes a function of
x. Let
t!,.y
be the change in y which results from changing x
2S