Example 21: : when the second derivative test fails
Find all the points of local maxima and local minima of the function given by
, and . So the second derivative test fails.
For , . So does not change sign as increases through 1.
Hence is a point of inflexion, and has no point of local maxima or local minima.
Want the reason behind every step, the basics and the checks?
The second derivative test can only say "maximum" or "minimum" when is not zero. When and the test gives no answer at all, and we must go back to the first derivative test and look at the sign of on each side of . Here is positive on both sides of 1, so the graph keeps climbing through : it only flattens for an instant, with no hill and no valley.
- Given
- for every real (a polynomial, so differentiable everywhere).
- To find
- Every point of local maxima and every point of local minima of .
- Plan
- Find and the critical points. Try the second derivative test (Theorem 4). It fails, so apply the first derivative test (Theorem 3). Finally confirm the answer directly from the definition, using .
Basics you need
An interior point is a point of local maxima if, for some , for every in ; a point of local minima if there. Lesson: maxima and minima
A point where or where is not differentiable. By Theorem 2, local extrema can only occur at critical points, but a critical point need not be one (the book's example is at 0).
First derivative test: changes from to through : local maxima; from to : local minima; no change of sign: neither (a point of inflection). Second derivative test: and : local maxima; : local minima; : the test fails. Lesson: first and second derivative tests
for every , and only at . So the sign of is easy to read. Basics: sign of a product
, a perfect square. Basics: factorising and the discriminant
Solution
Differentiate twice
Both tests need derivatives: the first test needs , the second needs as well. Factorising now pays off twice: it shows at once where and what sign has elsewhere.Power rule, term by term: , constants differentiate to 0. is a perfect square (Basic 5).Find the critical points
is a polynomial, so it is differentiable at every real ; there are no critical points of the "not differentiable" kind. So is the only critical point, and .
By Theorem 2, a local maximum or minimum of a function on an open interval can only happen where or is not differentiable. So is the only candidate, and the whole question reduces to: what happens at ?Try the second derivative test: it fails
So and . This is case (iii) of Theorem 4: the test fails.
"The test fails" means the test gives no information. It does not mean "neither maximum nor minimum". Functions with can have a minimum at ( at 0), a maximum ( at 0) or neither ( at 0). So we need a different tool.Theorem 4 (Second Derivative Test), part (iii): the test fails if and ; "in this case, we go back to the first derivative test". Lesson: first and second derivative testsApply the first derivative test at
Read the sign of on each side of 1:
is (e.g. 0.9) (0.06) increasing flat for an instant (e.g. 1.1) (0.06) increasing is positive just to the left of 1 and positive just to the right: it does not change sign as increases through 1. By Theorem 3(iii), is neither a point of local maxima nor a point of local minima; it is a point of inflexion.
A local maximum needs the graph to rise and then fall ( from to ); a local minimum needs fall then rise ( to ). Here the graph rises, pauses with a horizontal tangent at , and rises again. The book reaches the same conclusion in Example 18, which is this very function done with the first derivative test from the start.Theorem 3 (First Derivative Test), part (iii): if does not change sign as increases through , then is neither a point of local maxima nor of local minima. Basics: sign of a productConfirm from the definition
Complete the cube: , so
If then and . If then and . So every interval , however small, contains points where is below and points where it is above. Neither " for all nearby " nor " for all nearby " can hold, so by Definition 4, is neither kind of point.
The tests are shortcuts; the definition is the final word. Writing around the point shows the conclusion with no theorem at all, and it also shows that is strictly increasing on all of (a cube is strictly increasing). A strictly increasing function never has a turning point.(expansion of ). Definition 4 is Basic 1.Why the book calls a point of inflexion
is negative for and positive for . So the graph bends downward (its slope is decreasing) to the left of 1 and bends upward to the right of 1: the bending changes direction exactly at , where the tangent is horizontal.
. The tangent at is horizontal (dashed), yet the curve rises on both sides of it: no hill, no valley. Theorem 3(iii) uses the name "point of inflection" for a critical point where keeps its sign. The picture shows why the word fits: the curve crosses its own tangent line there, switching from bending down to bending up.The sign of tells whether the slope is decreasing () or increasing (). Lesson: first and second derivative tests
has no point of local maxima and no point of local minima. Its only critical point, (where ), is a point of inflexion: the second derivative test fails there (), and does not change sign through 1.
, , : the value just before 1 is smaller and the value just after is larger, so is neither the biggest nor the smallest nearby value. Also , while and have opposite signs, as Step 6 says.
- Reading "the test fails" as "no extremum". When the second derivative test says nothing; you must use the first derivative test (or the definition).
- Calling every critical point a maximum or minimum. only makes a candidate (Theorem 2 has no converse).
- Checking the sign of at only one side. The first derivative test compares both sides of .
- Expanding instead of factorising . In the form the sign is not obvious; in the form it is.
- Order of attack: critical points, then the second derivative test, and if , the first derivative test.
- A derivative of the form with never changes sign: the function rises straight through .
- Rewriting a cubic about its critical point, here , turns the whole question into a one-line check.
Every fact used, and where it comes from
| Fact | Used in | Source |
|---|---|---|
| , | Step 1 | Power rule (Chapter 5); perfect square |
| Critical point; extrema only at critical points | Step 2 | Theorem 2 and the Note after it, Section 6.4 |
| Second Derivative Test, case (iii) | Step 3 | Theorem 4, Section 6.4 (Lesson: tests) |
| First Derivative Test, case (iii) | Step 4 | Theorem 3, Section 6.4 |
| Same function by the first test | Step 4 | Example 18, Section 6.4 |
| Local maxima and minima | Step 5 | Definition 4, Section 6.4 |
| Steps 5, 6 | Expansion of |
Tell us if something on this page is wrong or unclear, or if you would like a fuller explanation, another method or a new feature. Send feedback on this problem →