Calculus · Mean value theorem

Mean value theorem

If a function f is continuous on [a, b] and differentiable on (a, b), then at some point c between a and b its tangent is parallel to the secant through the endpoints: f′(c) = (f(b) − f(a))/(b − a).

01

Mean value theorem in this visualization

The mean value theorem, also called Lagrange’s mean value theorem, says: if a function ff is continuous on the closed interval [a,b][a, b] and differentiable on the open interval (a,b)(a, b), then there is at least one point cc with a<c<ba < c < b where

Mean value theorem
f′(c)=f(b)−f(a)b−a.f'(c) = \frac{f(b) - f(a)}{b - a}.

The left side is the slope of the tangent at cc; the right side is the slope of the secant through (a,f(a))(a, f(a)) and (b,f(b))(b, f(b)), the average rate of change of ff over the interval. The point is also written ξ\xi. Rolle’s theorem is the case f(a)=f(b)f(a) = f(b), where the secant is flat and f′(c)=0f'(c) = 0. The page shows the theorem on four smooth functions and on two whose hypotheses can fail: ∣x∣|x| is not differentiable at its corner, and the jump function, on an interval with the jump inside, is neither continuous nor differentiable there.

The blue points A and B sit on the curve above aa and bb: drag them, or slide the blue segment on the x-axis to move the whole interval. The yellow line through them is the secant; call its slope mm. Each blue point marks a point cc: its tangent is parallel to the secant, and a dashed vertical joins it to the secant: where ff is smooth, the height of the curve above the secant has derivative zero at every cc (the two have the same slope there). Under the graph, the slope strip draws f′(x)f'(x) in blue and the secant slope mm as a yellow horizontal line: every cc is a crossing. With the answers hidden, a tangent point appears that you can move along the curve to find cc yourself.

02

What not to read into it

  • The theorem promises at least one cc, not exactly one: x3−3xx^3 - 3x on [−2,2][-2, 2] has two, sin⁡x\sin x on [−6,6][-6, 6] has four. It does not say where cc is: for a non-degenerate quadratic the unique cc is always the midpoint, but for other functions it can lie anywhere inside.
  • Differentiability is needed only inside the interval. x\sqrt{x} on [0,4][0, 4] has a vertical tangent at 0, and still c=1c = 1 works.
  • When a hypothesis fails, the theorem is silent. ∣x∣|x| on [−1,2][-1, 2] has no cc, nor does the jump function on [0.5,1.5][0.5, 1.5]; yet the same jump function on [0,2][0, 2] happens to have c=1.5c = 1.5.
  • The points cc on this page are solved exactly from f′(c)=f'(c) = secant slope for each function, not found by a numerical search.
03

Where it is used

  • Average speed: a car covers 10 km in 5 minutes, an average of 120 km/h. If the distance it has travelled changes continuously and has a derivative, the speed, at every moment inside those 5 minutes, then at some instant its speed was exactly 120 km/h. So a driver whose average over a camera-timed stretch is above the limit was over the limit at some moment.
  • If f′=0f' = 0 everywhere on an interval, ff is constant there; two functions with the same derivative differ by a constant, which is why an antiderivative carries + C+\,C.
  • Estimates: if ∣f′∣≤M|f'| \le M on the interval, then ∣f(b)−f(a)∣≤M ∣b−a∣|f(b) - f(a)| \le M\,|b - a|. The theorem is also a step in the proof of the fundamental theorem of calculus.
04

Related