Rolle's Theorem and the Mean Value Theorem

Drive from one town to another averaging 60 mph, and at some instant your speedometer reads exactly 60. The MVT makes that precise: on any well-behaved stretch, the instantaneous rate somewhere equals the average rate over the whole stretch.

By the end you'll be able to state the hypotheses of Rolle's Theorem and the Mean Value Theorem, explain what each guarantees, and solve \(f'(c) = \frac{f(b)-f(a)}{b-a}\) to find the guaranteed point c.

Predict: before you touch anything — where on the arc between the two orange endpoints will the tangent line run parallel to the secant through them? Drag the endpoints (or use the sliders) and watch the marked point c track it.

The curve is \(f(x) = \tfrac{x^3}{16} - \tfrac{3x}{4} + 2\). The solid secant joins \((a, f(a))\) to \((b, f(b))\) — its slope is the average rate of change. The dashed tangent at each marked c has the exact same slope — an instantaneous rate matching the average. Turn on Rolle mode to lock \(f(b) = f(a)\): the secant goes flat and \(f'(c) = 0\).

Secant (average) slope = −0.31 · tangent slope f′(c) = −0.31 at c ≈ −1.53 and 1.53 — parallel.

f(x) = x³/16 − 3x/4 + 2 — secant through the endpoints, parallel tangent(s) at the guaranteed c
f(x) secant (average rate) tangent at c (instantaneous rate)

Rolle's Theorem and the Mean Value Theorem guarantee a point where the instantaneous rate of change equals the average rate of change over an interval, given continuity on the closed interval and differentiability on the open interval.

Rolle first: same height in, same height out

Start with the special case. Suppose f is continuous on \([a, b]\), differentiable on \((a, b)\), and \(f(a) = f(b)\) — the curve leaves and returns at the same height. Geometrically, a smooth curve that comes back to its starting height must level off somewhere in between: it either stays flat or goes up and comes back down (or down and back up), and at the turnaround the tangent is horizontal. That's Rolle's Theorem: there is at least one c in \((a, b)\) with \(f'(c) = 0\). You've already seen the engine behind this — a continuous function on a closed interval attains a maximum and minimum, and at an interior extreme point of a differentiable function the derivative vanishes.

Tilt it: the Mean Value Theorem

Now drop the requirement \(f(a) = f(b)\) and let the endpoints sit at different heights. The secant through \((a, f(a))\) and \((b, f(b))\) is no longer horizontal — it's tilted, with slope \(\frac{f(b)-f(a)}{b-a}\). The same leveling-off logic still works, just measured relative to the tilted line instead of the horizontal.

Formal

Mean Value Theorem. If f is continuous on the closed interval \([a, b]\) and differentiable on the open interval \((a, b)\), then there exists at least one c in \((a, b)\) with \[ f'(c) = \frac{f(b) - f(a)}{b - a}. \] The right side is the average rate of change across the interval; the left side is the instantaneous rate at c — the tangent at c is parallel to the secant, exactly what you watched in the demo. Rolle's Theorem is the special case \(f(a) = f(b)\), where the right side is 0.

This should feel familiar from the definition of the derivative: the quantity \(\frac{f(b)-f(a)}{b-a}\) is a difference quotient, the very expression whose limit as the two points merge defines \(f'\). The MVT runs that story in reverse: instead of shrinking the interval to reach an instantaneous rate, it keeps the interval fixed and promises the instantaneous rate already agrees with the average rate at some interior point.

Three cautions. The theorem is an existence claim — it tells you c exists, not how to find it (for that you solve \(f'(c) = \frac{f(b)-f(a)}{b-a}\) yourself). There may be several valid c's — the default demo view shows two. And the hypotheses are load-bearing: \(f(x) = |x|\) on \([-1, 1]\) has \(f(-1) = f(1)\) but no point with \(f' = 0\), because the corner at 0 breaks differentiability on the open interval.

Worked example

Find the c guaranteed by the MVT for \(f(x) = x^2\) on \([1, 3]\).
Average rate: \(\dfrac{f(3) - f(1)}{3 - 1} = \dfrac{9 - 1}{2} = 4\).
Set the instantaneous rate equal: \(f'(c) = 2c = 4\), so \(c = \) 2 — and \(2\) lies in \((1, 3)\), as the theorem promises. ✓

Your turn

Same function, a different interval: \(f(x) = x^2\) on \([2, 6]\).
Average rate: \(\dfrac{f(6) - f(2)}{6 - 2} = \dfrac{36 - 4}{4} = 8\).
Now finish it: \(f'(c) = 2c = 8\), so \(c = \) ____ — and check that it lies in the open interval.

Reveal the answer

\(2c = 8 \Rightarrow c = \) 4, which lies in \((2, 6)\). ✓ Notice c landed at the midpoint again — that's a special feature of parabolas, not a general rule. Try dragging the demo's endpoints and watch c sit off-center for the cubic.

More info — why the MVT is the bridge from f′ to f

The MVT looks like a curiosity, but it's the tool that converts local derivative facts into global statements about f. If \(f'(x) = 0\) at every point of an interval, apply the MVT to any two points a < b in it: \(f(b) - f(a) = f'(c)(b - a) = 0\), so f is constant. If \(f' > 0\) throughout, the same identity forces \(f(b) > f(a)\) — f is increasing. That second corollary is exactly what justifies the next lesson's First Derivative Test. Proof-wise, the MVT is Rolle's Theorem applied to the "gap" function \(g(x) = f(x) - \big[f(a) + \frac{f(b)-f(a)}{b-a}(x-a)\big]\), the vertical distance from the curve to the secant — g vanishes at both endpoints, so Rolle gives \(g'(c) = 0\), which rearranges to the MVT. The OpenStax link in Dive deeper walks through both the proof and the corollaries.

Check your understanding

Question 1 of 4

For \(f(x) = x^3\) on \([0, 3]\), find the value of \(c\) guaranteed by the Mean Value Theorem.

Question 2 of 4

\(f(x) = |x|\) on \([-1, 1]\) satisfies \(f(-1) = f(1) = 1\), yet no point has \(f'(c) = 0\). Why doesn't this contradict Rolle's Theorem?

Question 3 of 4

In the MVT conclusion \(f'(c) = \dfrac{f(b) - f(a)}{b - a}\), how do the two sides relate to the definition of the derivative?

Question 4 of 4

A function f is continuous on \([1, 5]\), differentiable on \((1, 5)\), and \(f(1) = f(5) = 7\). What does Rolle's Theorem guarantee?

Recap

  • Hypotheses (both theorems): f continuous on \([a, b]\) and differentiable on \((a, b)\) — check them before invoking the conclusion.
  • Rolle's Theorem: if additionally \(f(a) = f(b)\), then some c in \((a, b)\) has \(f'(c) = 0\).
  • Mean Value Theorem: some c in \((a, b)\) has \(f'(c) = \frac{f(b)-f(a)}{b-a}\) — instantaneous rate equals average rate; the tangent at c is parallel to the secant.
  • It's an existence claim: at least one c (possibly several); to find c, solve \(f'(c) = \frac{f(b)-f(a)}{b-a}\).
  • Key corollaries: \(f' = 0\) on an interval ⇒ f constant; \(f' > 0\) ⇒ f increasing — the bridge from derivative information to the function's behavior.

Dive deeper

Sources

  • Rolle's Theorem and the Mean Value Theorem