Calculus › Section D · Applications of derivatives › Topic D.3

Increasing, decreasing, and the first derivative test

The sign of the derivative is the mood of the graph: positive slopes mean climbing, negative mean falling. Read the sign changes and you can find every peak and valley a function has.

Slopes are the story

Take any smooth curve and run your eye along it. Where the tangent tilts upward, the function is increasing; where it tilts downward, decreasing; the places where one gives way to the other are exactly the turning points, the peaks and valleys we hunted in D.1. This topic makes that eyeball test rigorous and mechanical. The derivative's sign tells you the graph's direction, the critical numbers tell you where the sign can change, and a small sign chart turns the whole up-and-down shape of a function into a few minutes of arithmetic.

The increasing-decreasing test

The core theorem: if \(f' > 0\) at every point of an interval, then \(f\) is increasing there, meaning \(x_1 < x_2\) forces \(f(x_1) < f(x_2)\); and if \(f' < 0\) throughout, \(f\) is decreasing. One line of the Mean Value Theorem is the proof: between any two points, the secant slope \(\frac{f(x_2) - f(x_1)}{x_2 - x_1}\) equals \(f'\) somewhere inside, and a positive secant slope across a positive run means the function rose. The theorem works interval by interval, so a function can rise on one stretch and fall on the next. It is a one-way guarantee: \(x^3\) is increasing everywhere even though \(f'(0) = 0\), because a single level point does not stop it rising.

Where the sign can change

Here is the observation that makes the method practical. For the functions in this course, \(f'\) is continuous wherever it exists, so it can change sign only by passing through zero or by failing to exist. That means the sign of \(f'\) can flip only at a critical number or at a point where \(f\) itself is undefined, such as \(x = 0\) for \(\frac{1}{x^2}\). Mark those points, and they chop the number line into intervals inside each of which \(f'\) keeps one sign. One test point per interval then settles the entire interval. That is the sign chart.

Worked example: mapping a cubic's moods

Map the rise and fall of \(f(x) = x^3 - 3x\). The derivative \(f'(x) = 3x^2 - 3 = 3(x - 1)(x + 1)\) is zero at \(x = \pm 1\), splitting the line into three intervals. Test a convenient point in each. At \(x = -2\): \(f' = 3(3) = 9 > 0\), rising. At \(x = 0\): \(f' = -3 < 0\), falling. At \(x = 2\): \(f' = 9 > 0\), rising again. So \(f\) increases on \((-\infty, -1)\), decreases on \((-1, 1)\), and increases on \((1, \infty)\). Three test values, and the function's entire behaviour is mapped: up to \((-1, 2)\), down to \((1, -2)\), and up forever after.

The first derivative test

Now classify the turning points. Let \(c\) be a critical number of a function \(f\) that is continuous on an open interval around \(c\). The first derivative test reads the sign change of \(f'\) across \(c\). If \(f'\) is positive to the left and negative to the right, the graph climbs into a summit: \(f(c)\) is a local maximum. Negative then positive: it falls into a valley, a local minimum. No sign change: neither, just a pause the function passes through. Why it works: rising up to \(c\) and falling after it means \(f(c)\) is at least as high as every nearby value. Continuity at \(c\) matters, because a jump at \(c\) could break this.

Worked example: classify the turning points

Classify the critical numbers of the same cubic, \(f(x) = x^3 - 3x\). The sign chart from before gives everything. At \(x = -1\): positive before, negative after, so a local maximum, with value \(f(-1) = -1 + 3 = 2\). At \(x = 1\): negative before, positive after, a local minimum, value \(f(1) = 1 - 3 = -2\). Notice what was not needed: a calculator or a plotted graph. One derivative, its two critical numbers and three test values give the full up-and-down picture, and the two extreme values come from substituting the critical numbers back into \(f\), not into \(f'\).

When the derivative does not exist

Critical numbers include the points where \(f'\) fails, and the test works there too, because it only reads signs either side. Take \(f(x) = x^{2/3}\), whose graph is a cusp-bottomed curve. Its derivative \(\frac{2}{3}x^{-1/3}\) is undefined at \(0\), negative for \(x < 0\) and positive for \(x > 0\). Falling then rising: \(0\) is a local minimum, sitting precisely at the point with no derivative at all. The absolute value function tells the same story with a corner. The machine did not need the derivative to exist at the critical number; it needed the signs around it, and the signs were unambiguous.

Turning points in the wild

Suppose a shop's daily revenue from selling \(x\) units is modelled by \(R(x) = 50x - 0.5x^2\) dollars, where the price drops as more units are sold. When does extra sales volume stop paying? \(R'(x) = 50 - x\) is positive below \(50\) and negative above, so revenue rises up to \(x = 50\) and falls after it. The sign change certifies a maximum, not just a level point: \(R(50) = 2500 - 1250 = 1250\) dollars. The same logic finds the moment a drug's concentration in the blood stops rising after a dose and starts to fall. In every case the steps are identical: differentiate, find the critical numbers, chart the signs, read the shape.

Common mistakes to avoid

The recurring errors: testing only one side of a critical number and declaring victory, when the whole test is the comparison between sides; assuming \(f'(c) = 0\) implies an extremum at \(c\), with \(x^3\) at \(0\) as the eternal counterexample; forgetting the critical numbers where the derivative is undefined, like the cusp above, and the points where \(f\) itself is undefined; choosing messy test points, when only the sign matters and the friendliest number in each interval will do; and reversing the test's conclusion: positive-to-negative is a maximum because the graph climbs then falls, not the other way round. Sketch the little hill in the margin if you ever doubt it.

Key takeaways

Where \(f' > 0\) the function increases, and where \(f' < 0\) it decreases, by the Mean Value Theorem. Critical numbers, \(f' = 0\) or undefined, are the only places the sign can flip, so one test point per interval maps the whole function. The first derivative test classifies each critical number by its sign change: positive to negative is a local maximum, negative to positive a local minimum, no change is neither. It works at corners and cusps too. This is half of the shape story; the other half, which way the graph curves between the turning points, belongs to the second derivative, and it is next.

Flashcards

Flashcard

f′ > 0 on an interval

f is increasing there

The MVT makes each secant slope positive.

Test your understanding

Question 1 of 5

For f(x) = x³ − 3x, on which interval is f decreasing?

Question 2 of 5

At x = −1, the function f(x) = x³ − 3x has what feature?

Question 3 of 5

If f′ changes from negative to positive at a critical number c, then c is…

Question 4 of 5

The function f(x) = x^(2/3), whose derivative is undefined at 0, has what at x = 0?

Question 5 of 5

Suppose f′(x) > 0 for all x in (0, 5). What must be true?

Standards: 5

← Previous topicNext topic →