The Mean Value Theorem and the Extended

Mean Value Theorem

Willard Miller
September 21, 2006

0.1 The MVT

Recall the Extreme Value Theorem (EVT) from class: If the function f is
defined and continuous on a closed bounded interval [a, b] then there is some
point c ∈ [a, b] where it takes on its maximum value M = f (c) and some
point d ∈ [a, b] where it takes on its minimum value m = f (d). Thus
M = f (c) ≥ f (x) ≥ f (d) = m
for all x ∈ [a, b].
Theorem 1 Fermat’s Theorem. (Not his last one but a very useful observa-
tion that is easy to prove.) Suppose the function f is defined and continuous
on a closed bounded interval [a, b] and takes on an extreme value (either its
maximum M or its minimum m) at an interior point c of the interval, so
a < c < b. If the derivative f ′ (c) exists, then f ′ (c) = 0.

PROOF: To be definite, we assume that f (c) = M. (The proof for the case
f (c) = m is virtually the same.) By definition
f (c + ∆x) − f (c)
f ′ (c) = lim . (1)
∆x→0 ∆x
Since f (c) = M is a maximum, it follows that for ∆x > 0 we have
f (c + ∆x) − f (c)
≤ 0,
∆x

1
since the numerator is nonpositive and the denominator is positive. On the
other hand, for ∆x < 0 we have
f (c + ∆x) − f (c)
≥ 0,
∆x
since the numerator is nonpositive and the denominator is negative. However
by assumption the limit (1) exists and the same value is obtained as ∆x → 0
through postitive or negative values. Thus we must have f ′ (c) = 0. Q.E.D.
Let a < b be finite numbers.
Theorem 2 Mean Value Theorem (MVT). Suppose the function f is defined
and continuous on a closed bounded interval [a, b] and differentiable on the
open interval (a, b). Then there is a point c, a < c < b, such that
f (b) − f (a)
= f ′ (c).
b−a
PROOF: Consider the secant line ℓ that connects the endpoints (a, f (a))
and (b, f (b) on the graph of the function y = f (x). Using the point-slope
equation for a line, we see that the equation for the secant line is
f (b) − f (a)
ℓ: y − f (a) = (x − a).
b−a
For each x ∈ [a, b] let g(x) be the directed distance between the curve f (x)
and the line y(x):
" #
f (b) − f (a) f (b) − f (a)
g(x) = f (x) − f (a) + x− a .
b−a b−a
Note that g(a) = g(b) = 0 and
f (b) − f (a)
g ′ (x) = f ′ (x) − .
b−a
Since g vanishes at the endpoints and g is continuous on [a, b] by the EVT
there must be some interior point c ∈ (a, b) such that g(c) is an extreme
value of g. By Fermat’s theorem g ′ (c) = 0, which means
f (b) − f (a)
= f ′ (c).
b−a
Q.E.D.

2
0.2 The EMVT
Theorem 3 Extended Mean Value Theorem (EMVT). Suppose f (x), g(x)
are functions such that
1. f, g are defined and continuous on the closed bounded interval [a, b],
a < b.
2. f, g are differentiable on the open interval (a, b).
3. g ′(x) 6= 0 for all x ∈ (a, b).
Then there exists a c ∈ (a, b) such that
f ′ (c) f (b) − f (a)
= .
g (c)
′ g(b) − g(a)

NOTE: If g(x) = x, then this is just the statement of the Mean Value The-
orem (MVT).
PROOF: For the MVT we considered the function
!
f (b) − f (a)
F (x) = f (x) − L(x) = f (x) − (x − a) − f (a).
b−a

We observed that F (a) = F (b) = 0 and that there must be a relative ex-
tremum of F at some c ∈ (a, b). Then by the Fermat Theorem we must have
F ′ (c) = 0. But !
′ ′ f (b) − f (a)
F (x) = f (x) − .
b−a
For the EMVT we apply the same procedure to the function
!
f (b) − f (a)
F (x) = f (x) − (g(x) − g(a)) − f (a).
g(b) − g(a)
Again F (a) = F (b) = 0 and there must be a relative extremum of F at
some c ∈ (a, b). By the Fermat Theorem and assumption 1. we must have
F ′ (c) = 0. Thus
!
′ ′ f (b) − f (a) ′
F (c) = f (c) − g (c) = 0.
g(b) − g(a)

3
or
f ′ (c) f (b) − f (a)
= ,
g ′ (c) g(b) − g(a)
since g ′ (c) 6= 0. Q.E.D.

0
0.3 A 0 form of the L’Hospital Rule
Theorem 4 Let f (x) and g(x) be differentiable in an open neighborhood N
containing x = a, (but not necessarily at x = a), f ′ (x), g ′ (x) continuous in
the same neighborhood, and suppose g ′ (x) 6= 0 in N. Suppose
lim f (x) = 0, lim g(x) = 0.
x→a x→a
Then
f (x) f ′ (x)
lim = lim ′ ,
h→a g(x) h→a g (x)

where the equality is meant in any one of the senses

1. Both limits exist and are equal.
2. Both limits diverge to +∞.
3. Both limits diverge to −∞.
If the right hand limit fails to exist, the rule is inconclusive.
Corollary 1 The Rule is also true if the limits are right hand (x → a+) or
left-hand (x → a−).
PROOF OF THE RULE: Since limx→a f (x) = 0, limx→a g(x) = 0, we can
extend the domains of f, g to x = a, if necessary, by defining f (a) = g(a) = 0.
Then f and g are continuous at a. Now let x be in the neighborhood N, with
x > a. Then the EMVT applies to the interval [a, x] and there is a y ∈ (a, x)
such that
f ′ (y) f (x) − f (a) f (x)
= = .
g (y)
′ g(x) − g(a) g(x)
Thus
f (x) f ′ (y) f ′ (x)
lim = lim ′ = lim ′ .
x→a+ g(x) x→a+ g (y) x→a+ g (x)

if the last limit exists or diverged to ±∞. Similarly, if x < a we apply the
EMVT to the interval [x, a] and get the same result for the limit as x → a−.
Q.E.D.

4
 ∞
0.4 A ∞ form of the L’Hospital Rule
∞
The proofs of the ∞ forms of the L’Hospital Rule are a little trickier. We
will prove a right-hand limit version; the proof of the left-hand limit case
is virtually identical. If both left and right-hand limits exist and are equal,
then the two-sided limit exists.

Theorem 5 Suppose

1. f (x) and g(x) are differentiable in an open interval (a, b), and f ′ (x), g ′ (x)
are continuous in the same neighborhood

2. g ′(x) 6= 0 on (a, b)

3. limx→a+ f (x) = +∞, limx→a+ g(x) = +∞

Then if
f ′ (x)
lim = L,
h→a+ g ′ (x)

we also have
f (x)
lim = L,
h→a+ g(x)
where the equality is meant in any one of the senses

1. Both limits exist and are equal.

2. Both limits diverge to +∞.

3. Both limits diverge to −∞.

The test is inconclusive if the right hand limit fails to exist.

PROOF: Let a < x < y < b and apply the EMVT to the interval [x, y]. Then
there is a ξ ∈ (x, y) such that
f (y)
f ′ (ξ) f (y) − f (x) f (x) 1 − f (x)
= = g(y)
. (2)
g ′(ξ) g(y) − g(x) g(x) 1 − g(x)

5
Now we let both x and y approach a, but with x making the approach more
rapidly than y, so that

f (y) g(y)
lim = lim = 0.
f (x) g(x)

This can be done because of assumption 3. Thus

g(y)
1− g(x)
lim f (y)
=1
1− f (x)

f ′ (x)
and, since g ′ (x)
is continuous,

f ′ (ξ) f ′ (x)
lim = lim = L.
g ′(ξ) h→a+ g ′ (x)

It follows from (2) that

g(y)
f (x) f ′ (ξ) 1 − g(x)
lim = lim ′ lim (y)
= L · 1 = L.
x→a+ g(x) g (ξ) 1 − ff (x)

Q.E.D.