Distinguish the input, value and point of an extremum
The location of an extremum is an input xE. Its extreme value is the height f(xE), and its point is the ordered pair (xE,f(xE)). “Local” compares nearby allowed inputs; “absolute” or “global” compares the entire specified domain. A local maximum can therefore lie below a distant endpoint.[1]
For a local maximum or minimum respectively, the comparison holds near the input; for an absolute extremum it holds throughout the allowed domain.[1]
If f is differentiable at an interior extremum, then f′(xE)=0 is necessary. Do not transfer that condition to an endpoint. A corner can also be an extremum even though no derivative exists there. Interior local extrema and endpoint extrema are named separately below.[1]
Why derivatives provide the tests
At a differentiable interior maximum, nearby differences f(xE+h)−f(xE) are nonpositive. The difference quotient is therefore nonpositive for h>0 and nonnegative for h<0. If both sides have the same finite limit, it must be zero. This justifies the necessary condition without proving its converse.[1]
For this version of the second derivative test, assume f is twice continuously differentiable near the interior point.[2]
The first derivative test needs less: f is continuous at the candidate and differentiable on the adjacent open intervals. If f′ is positive on the left and negative on the right, f rises to a local maximum and then falls. A negative-to-positive change gives a local minimum. Establish the signs on adjacent intervals; two arbitrary numerical samples alone are not a proof.[2]
For x1<x2 in the same interval of differentiability, the mean value theorem gives f(x2)−f(x1)=f′(ξ)(x2−x1) at an intermediate input ξ. Hence f′>0 throughout that interval forces strict increase and f′<0 forces strict decrease. Conversely, a strictly increasing function may have derivative zero at isolated inputs, as x3 does at zero. A gap in the domain splits the analysis into separate intervals.[3]
The second derivative test follows from the sign test: if f″ is positive near xE, then f′ increases through zero, from negative to positive. A negative second derivative reverses that pattern. If f″(xE)=0, the test is inconclusive. Both a flat minimum and a stationary inflection point are possible, so continue with another test.[2]
A cubic example with a sign check
Classify a cubic curve completely
- Given
- on ℝ; all quantities are dimensionless.
- Find
- Find the extrema and intervals of monotonicity, distinguishing local from global conclusions.
Differentiate first and obtain stationary candidates from the factored derivative.
Between their only zeros, the factors have fixed signs. The sign sequence therefore describes entire intervals, not just isolated sample points.
The derivative changes from positive to negative at −2 and from negative to positive at 1. Evaluate the second derivative as a second check.
For the point coordinates, substitute into f rather than f′.
Result: The point (−2,24) is a local maximum and (1,−3) a local minimum. f is strictly increasing on (−∞,−2) and (1,∞), and strictly decreasing on (−2,1). Over all of ℝ it has neither an absolute maximum nor an absolute minimum.
Check: Expanding around the maximum gives , which is negative for small nonzero h. At the minimum, is positive for small nonzero h. The leading term 2x3 also produces unbounded values in both directions as x→±∞.
Increasing, decreasing, then increasing again
Free with a Quanta account. Your download stays ready after registration.
Recall the method later with the formula’s flashcards.
10 flashcards · English materials
Already have an account? Sign in for flashcardsAvailable from Essential. A ready-made formula set, no AI credits required. Article exercises are not all included.
Check endpoints and absolute extrema
A continuous function attains a smallest and largest value on a closed bounded interval [a,b]. Collect the endpoints and interior inputs where f′=0 or the derivative does not exist, provided f itself is defined there. Then compare the original function values at every candidate. For open endpoints, discontinuities or unbounded domains, this comparison is not a general existence guarantee.[1]
On a bounded interval, the endpoints win
- Given
- on the closed interval [0,4].
- Find
- Find interior local extrema and the absolute minimum and maximum on the allowed domain.
As a polynomial, g is continuous. Comparing absolute extrema requires both endpoints and all interior stationary points.
The derivative has the sign sequence positive, negative, positive. The second derivative confirms the local maximum and minimum.
Now substitute the endpoints into the original function as well and compare all four heights.
Here both the smallest and largest values occur at endpoints.
Result: The interior local maximum is (1,10) and the interior local minimum is (2,9). On the specified interval, the absolute minimum is (0,5) and the absolute maximum is (4,37). Thus the local maximum is not the highest allowed point.
Check: Monotonicity on [0,1], [1,2] and [2,4] excludes additional extreme values inside those pieces. Direct substitution gives . The endpoint derivatives and are nonzero; the interior-point criterion was deliberately not applied to them.
Units and the cubic special case
The location of an extremum has the input unit; its value has the output unit. If x is time in hours and f(x) is temperature in °C, then f′ and f″ have units °C/h and °C/h2. The equation f′=0 is understood in the derivative unit. Do not interchange the units of f and f′ when calculating a point’s height.
Four real coefficients describe only this cubic special case; in a physical quantity equation, a carries the required unit.
In this special case, P′=0 is a quadratic equation. Its discriminant determines whether there are two, one or no real stationary candidates. A repeated zero of P′ in a genuine cubic gives a stationary inflection point rather than an extremum. Do not extend that special result to arbitrary functions with f″=0. A calculator limited to cubic coefficients replaces neither domain checks nor a general analysis of functions.[2]
Seven exercises on tests, limits and interpretation
Calculate · Application
Find the local extrema of on ℝ.
Show hint
Factor h′ and compute the heights after classifying the stationary points.
Learning goal: Classify stationary inputs and calculate complete point coordinates.
Show solution and calculation
Use h′=0 followed by a sign test or the second derivative test.
Here , so x=1 and x=4.
Since , the respective values are 18>0 and −18<0.
Hence the local minimum is (1,−5) and the local maximum is (4,22).
Check: h′ is negative left of 1, positive between 1 and 4, and negative right of 4, agreeing with both classifications.
Interpret · Application
Only is known for a differentiable function. Find its monotonicity and the inputs where extrema occur. Can you give the extreme values?
Show hint
The squared factor is positive away from its zero; the height of F has not been specified.
Learning goal: Infer graph behaviour from derivative signs and identify missing height information.
Show solution and calculation
For x<−2, F′ is negative; for −2<x<1 and x>1, it is positive.
At −2 the sign changes from negative to positive, giving a local minimum.
At 1 it remains positive, so there is no extremum.
F is strictly decreasing on (−∞,−2) and strictly increasing on (−2,∞), despite F′(1)=0.
Without height information, F is undetermined up to an additive constant.
Check: F and F+7 have the same derivative but minimum values differing by 7.
Error analysis · Application
For , u′(2)=u″(2)=0. A solution therefore calls (2,1) a stationary inflection point. Check this and compare with .
Show hint
Check the sign of u′; for v also check the change of concavity.
Learning goal: Resolve an inconclusive second derivative test using an appropriate method.
Show solution and calculation
The conclusion from u″(2)=0 is invalid.
changes from negative to positive, so u has a local minimum at (2,1).
Indeed, (x−2)4≥0 directly proves the absolute minimum value 1.
For v, whenever x≠2, so there is no extremum.
However, changes sign at 2, making (2,1) a stationary inflection point.
Check: for h>0, v(2−h)=1−h3 is below 1 and v(2+h)=1+h3 is above it.
Rearrange · Application
For , x=3 is required to be stationary. Find a and classify the point.
Show hint
Treat a as constant when differentiating. Set P′(3)=0 first.
Learning goal: Solve a stationarity condition for a parameter, then classify the point.
Show solution and calculation
Use .
The condition gives 27+6a−9=0, hence a=−3.
Then , so this is a local minimum.
Its value is P(3)=27−27−27+2=−25, giving (3,−25).
Check: P′=3(x−3)(x+1) changes from negative to positive at 3. The parameter also satisfies the original requirement P′(3)=0.
Transfer · Application
A temperature model is for 0≤t≤4 h. When does the model attain its highest and lowest temperatures?
Show hint
T′ has units °C/h; include the endpoints when comparing absolute extrema.
Learning goal: Combine derivative and coordinate units with a model restricted to a stated time interval.
Show solution and calculation
First solve , obtaining t=3 h.
Since T″=−2 °C/h2 is negative, this is an interior local maximum.
Comparing the candidates gives T(0)=20 °C, T(3 h)=29 °C and T(4 h)=28 °C.
Thus the maximum is 29 °C after 3 h and the minimum is 20 °C at the start.
Check: T rises until 3 h and then falls. The model makes no claim about measured temperatures outside the specified time range.
Error analysis · Challenge
For , x≠0, r′(x)=1/x2>0. Someone therefore calls r strictly increasing on its entire domain. Check this reasoning.
Show hint
Distinguish an interval from two intervals separated by a gap.
Learning goal: Restrict a monotonicity claim to an interval where its assumptions hold.
Show solution and calculation
The monotonicity theorem applies on an interval.
Here r is strictly increasing on (−∞,0) and separately on (0,∞).
Across the gap the claim fails: −1<1, but r(−1)=1>−1=r(1), a concrete counterexample to the asserted ordering.
At 0 there is neither a function value nor a stationary point.
Check: both individual branches still have a positive derivative throughout; the local calculation was correct but its extension across the gap was not.
Interpret · Foundation
Consider on [0,5]. Does the minimum occur at a zero of w′?
Show hint
Use the non-negativity of absolute value and its two linear branches.
Learning goal: Identify an extremum at a non-differentiable point and compare endpoint values.
Show solution and calculation
Since |x−2|≥0 and w(2)=0, (2,0) is the absolute minimum.
To the left of 2, w′=−1; to the right, w′=+1; at 2 the ordinary derivative does not exist.
The minimum therefore is not a zero of w′.
For the maximum, also compare the endpoints: w(0)=2 and w(5)=3, giving the maximum value 3 at x=5.
Check: the piecewise linear branches decrease to 2 and then increase, so no additional candidate is missing.
Check your understanding with the formula’s reference test.
8 test questions · English materials
Already have an account? Sign in for the testAvailable from Essential. A ready-made formula set, no AI credits required. Article exercises are not all included.
Relevant prerequisites and next steps
For polynomials, the power rule supplies the first and second derivatives. For quadratic functions, vertex form is often the shorter route to an extremum, while the general derivative test remains a useful check.
A point with derivative zero has a horizontal tangent. Finding its tangent equation is a next step after the point has been classified. Derivatives of sine and cosine provide periodic examples of the same tests.
Sources for this explanation
Quanta · Updated November 5, 2026
- OpenStax, Calculus Volume 1, 4.3 Maxima and Minima
- OpenStax, Calculus Volume 1, 4.5 Derivatives and the Shape of a Graph
- OpenStax, Calculus Volume 1, 4.4 The Mean Value Theorem
- North Rhine-Westphalia Ministry of School Education, upper-secondary Mathematics curriculum, 2023, introductory phase: Functions and Analysis, competencies 12 and 15–17 (German-language curriculum)
Editorial responsibility: Quanta Study editorial team
Review scope
This explanation, 2 worked examples and 7 exercises are available in German and English. Structure, source references and mathematical notation undergo automated checks. Calculations, model assumptions and both language versions receive a separate AI review.
An independent, named human subject review has not yet been documented. Our exercise numbers are worked examples, rather than measurements published by the cited sources.
Editorial responsibility, sources and methodsChange history
- Authored an independent DE/EN pilot article on extrema and monotonicity with explanations of the tests, two original examples, seven solved exercises and a native curve diagram.
Cite this reference page
Quanta. Extrema and monotonicity. Updated November 5, 2026. https://quanta-study.de/en/formula-sheet/extrempunkte-monotonie
When citing a specific scientific statement, also cite the specialist source linked to that statement.