← AS & A Level Mathematics

AS & A Level Mathematics · Pure Mathematics 1 · Chapter 5

Differentiation

Differentiation measures how fast one quantity changes with respect to another. Starting from the slope of a chord shrinking to a tangent, you will build fast algebraic rules for gradients, tangents, normals, turning points, optimisation and connected rates of change.

Learning objectives

  1. Understand that the gradient of a curve at a point is the gradient of the tangent at that point.
  2. Find the gradient of a curve from first principles as the limit of the gradient of a chord.
  3. Use function and Leibniz notation: f′(x), dy/dx, f′′(x) and d²y/dx².
  4. Differentiate kxⁿ (for any rational n) and sums or differences of terms, rewriting products, quotients and roots first.
  5. Find the equations of tangents and normals to a curve at a given point.
  6. Locate stationary points (dy/dx = 0), classify them using the sign of dy/dx or the second derivative d²y/dx², and test for increasing or decreasing functions.
  7. Solve practical maximum and minimum problems by forming and differentiating a function of one variable.
  8. Apply the chain rule dy/dx = (dy/du) × (du/dx) to composite functions and connected rates of change.

Book exercises: 5A, 5B, 5C, 5D, 5E, 5F, 5G and 5H, with Activities 5.1 to 5.7, the Think about it questions and the end-of-chapter Key points.

What do we mean by the gradient of a curve?

  • A curve changes steepness from point to point, unlike a straight line y = mx + c.
  • At any point of contact P, imagine a straight line touching the curve without cutting straight across it: that line is the tangent.
  • Definition: the gradient of a curve at P is the gradient m of the tangent to the curve at P.
  • Drawing a tangent by hand with a ruler (Activity 5.1) only gives a rough estimate because every hand-drawn tangent is slightly different.

Chords approaching the tangent

  • A straight line joining two points on a curve is called a chord.
  • We can calculate the gradient of a chord exactly using (y₂ − y₁)/(x₂ − x₁), with no drawing error.
  • On y = x², take P(3, 9) and join it to closer and closer points Q on the curve:
−1123369xyP(3, 9)(1, 1)(2, 4)slope 3slope 4slope 5y = x²
Chords from (0, 0), (1, 1) and (2, 4) to P(3, 9) on y = x² have gradients 3, 4 and 5. Closer points give gradients closer to 6.
Point Q on y = x²Calculation of chord QP gradientGradient
(2.7, 7.29)(9 − 7.29) / (3 − 2.7)5.7
(2.8, 7.84)(9 − 7.84) / (3 − 2.8)5.8
(2.9, 8.41)(9 − 8.41) / (3 − 2.9)5.9
(2.99, 8.9401)(9 − 8.9401) / (3 − 2.99)5.99
(3.001, 9.006001) (Activity 5.2)(9.006001 − 9) / (3.001 − 3)6.001
  • From both left and right, as Q approaches P(3, 9), the gradient of chord QP approaches 6.
  • We cannot just put Q = (3, 9) directly into (y₂ − y₁)/(x₂ − x₁) because 0/0 is undefined — instead we simplify algebraically first and then take the limit.

🎛️ Simulation: shrink the chord to the tangent

  • Choose the base point x₀ on y = x² and slide the step h towards 0 (from either side).
  • Watch the chord PQ (secondary line) rotate onto the exact tangent at P (dashed line), while the working shows why the h cancels.
Curve y = x² (primary), chord PQ (secondary) and exact tangent at P (dashed).

    Differentiation from first principles

    • To prove the gradient at P(3, 9) on y = x² is 6, let Q have x co-ordinate 3 + h, so its y co-ordinate is (3 + h)².
    • Even better, take a general point P(x, y) and Q(x + h, f(x + h)) to get the gradient function f′(x) in one go.
    xyP(x, y)Q(x + δx, y + δy)δx (or h)δy = f(x + h) − f(x)tangent at P
    A small change h (or δx) in x produces a change δy = f(x + h) − f(x) in y. As h → 0, the chord gradient δy/δx tends to the tangent gradient dy/dx.

    📘 Example · y = x² at P(3, 9) and at general point (x, x²) from first principles

    1. At P(3, 9) and Q(3 + h, (3 + h)²), gradient of PQ = = .
    2. Simplify the numerator: = = 6 + h (for h ≠ 0).
    3. Let h → 0: 6 + h → 6, so the gradient at (3, 9) is 6.
    4. For general P(x, x²) and Q(x + h, (x + h)²): gradient of PQ = = = 2x + h.
    5. Let h → 0: the gradient function of y = x² is dy/dx = 2x. Check at x = 3: 2(3) = 6. ✔

    📘 Example 5.1 · Gradient of y = x³ at the general point (x, y) from first principles

    1. Let P be (x, x³) and Q be (x + h, (x + h)³).
    2. Expand (x + h)³ = x³ + 3x²h + 3xh² + h³.
    3. Gradient of chord PQ = = .
    4. Factor out h and cancel: = 3x² + 3xh + h².
    5. Let Q → P so h → 0: the terms with h vanish, leaving f′(x) = 3x².

    Wallis’s rule and Leibniz notation (δx, δy and dy/dx)

    • Using the binomial theorem, (x + h)ⁿ = xⁿ + nxⁿ⁻¹h + xⁿ⁻²h² + … + hⁿ.
    • Subtracting xⁿ and dividing by h leaves nxⁿ⁻¹ + (terms with h in them), which tends to nxⁿ⁻¹ as h → 0 (Wallis’s rule).
    • The Greek letter δ (delta) stands for “a small change in”: δx is a small change in x (same as h) and δy is the corresponding small change in y.
    • In the limit as δx → 0, the chord gradient becomes the derivative (Gottfried Leibniz’s notation, read “dee y by dee x”): = limδx → 0 .

    ✏️ Try it

    1. Use first principles to find the gradient of y = x² at (−2, 4) (Activity 5.3(ii)).
    2. Expand (x + h)⁴ and hence prove from first principles that if y = x⁴ then dy/dx = 4x³ (Exercise 5A Q2).
    3. Differentiate y = x² + 5x from first principles.
    Answers

    1. Step 1: Take P(−2, 4) and Q(−2 + h, (−2 + h)²). Step 2: Chord gradient = ((4 − 4h + h²) − 4)/h = (−4h + h²)/h = −4 + h. Step 3: Let h → 0 to get −4 (matches 2x at x = −2 ✔).

    2. Step 1: (x + h)⁴ = x⁴ + 4x³h + 6x²h² + 4xh³ + h⁴. Step 2: Subtract x⁴ and divide by h: ((x + h)⁴ − x⁴)/h = 4x³ + 6x²h + 4xh² + h³. Step 3: Let h → 0: dy/dx = 4x³.

    3. Step 1: δy = (x + h)² + 5(x + h) − (x² + 5x) = 2xh + h² + 5h. Step 2: Divide by h: δy/h = 2x + 5 + h. Step 3: Let h → 0: dy/dx = 2x + 5.

    Next: differentiating powers, sums & differences →

    Standard differentiation rules

    • Power rule: if y = xⁿ, then = nxⁿ⁻¹ (works for any real rational power n: positive, negative or fractional).
    • Constant multiple rule: if y = kxⁿ, then = knxⁿ⁻¹.
    • Constant rule: if y = c (a horizontal line), its gradient is zero everywhere, so = 0.
    • Sums and differences: if y = f(x) + g(x), differentiate term by term: = f′(x) + g′(x).
    • Shifting a curve up or down by + c does not change its slope at any x (Activities 5.4 and 5.5).
    −2−112−4−224xyy = x³ + 2y = x³y = x³ − 2Same slope 3x² at every x
    Activities 5.4 & 5.5: the family y = x³ + c. Differentiating c gives 0, so every curve in the family has the same gradient 3x² at a given x.
    Original expressionRewrite as kxⁿ firstDerivative dy/dx
    y = c (constant)cx⁰0
    y = kxkx¹k
    y = x⁻¹−x⁻² = −
    y = kx⁻ⁿ−knx⁻ⁿ⁻¹ = −
    y = √xx1/2(1/2)x−1/2 =
    y = x−1/2−(1/2)x−3/2 = −

    📘 Examples 5.2 & 5.3 · Using the standard rules

    1. (i) y = x⁵ ⇒ multiply by 5, reduce power by 1: dy/dx = 5x⁴.
    2. (ii) z = 7x⁶ ⇒ dz/dx = 7 × 6x⁵ = 42x⁵.
    3. (iii) p = 11 ⇒ 11 is a constant, so dp/dx = 0.
    4. (iv) f(x) = 3/x = 3x⁻¹ ⇒ f′(x) = (−1) × 3x⁻² = −3/x².
    5. (v) y = 3x² + 4x³ ⇒ differentiate term by term: dy/dx = 6x + 12x².

    Rewriting products, quotients and roots before differentiating

    1. Expand / splitremove brackets & fractions 2. Index formwrite every term as kxⁿ 3. Differentiatekxⁿ → knxⁿ⁻¹ term by term
    Never differentiate a product or quotient term by term directly: expand brackets and split single-term denominators into kxⁿ first.

    📘 Example 5.4 · Differentiate f(x) =

    1. Expand the numerator: (x² + 1)(x − 5) = x³ − 5x² + x − 5.
    2. Divide each term by x: f(x) = − + − = x² − 5x + 1 − 5x⁻¹.
    3. Differentiate term by term: f′(x) = 2x − 5 + 0 − 5(−1)x⁻².
    4. Rewrite with positive powers: f′(x) = 2x − 5 + .

    📘 Example 5.5 · Given y = √x − , find (i) dy/dx and (ii) the gradient at (4, 1½)

    1. Rewrite in index form: y = x1/2 − 8x⁻².
    2. Differentiate: dy/dx = x−1/2 − 8(−2)x⁻³ = + .
    3. Substitute x = 4: dy/dx = + = + = + = .

    Finding gradients at axis crossings and points with a given gradient

    −2246−30−20−1010xyA(0, 0): slope 0B(6, 0): slope 36y = x³ − 6x²
    Example 5.6: y = x²(x − 6) = x³ − 6x² touches the x axis at A(0, 0) with gradient 0 and cuts it at B(6, 0) with gradient 36.
    −6−4−2110203040xy(−3, 32)(−1, 10)Both tangents have slope −9
    Example 5.7: y = x³ + 6x² + 5 has gradient −9 at two points: (−3, 32) and (−1, 10).

    📘 Example 5.6 · Gradient of y = x²(x − 6) where it meets the x axis

    1. Set y = 0: x²(x − 6) = 0 ⇒ x = 0 (repeated root) or x = 6, giving A(0, 0) and B(6, 0).
    2. Expand y = x³ − 6x² and differentiate: dy/dx = 3x² − 12x.
    3. At A(0, 0): dy/dx = 0. At B(6, 0): dy/dx = 3(36) − 12(6) = 108 − 72 = 36.

    📘 Example 5.7 · Points on y = x³ + 6x² + 5 where the gradient is −9

    1. Differentiate: dy/dx = 3x² + 12x.
    2. Set dy/dx = −9: 3x² + 12x = −9 ⇒ 3x² + 12x + 9 = 0 ⇒ 3(x + 1)(x + 3) = 0.
    3. So x = −1 or x = −3.
    4. Find y: at x = −1, y = (−1)³ + 6(−1)² + 5 = 10; at x = −3, y = (−3)³ + 6(−3)² + 5 = 32.
    5. The points are (−1, 10) and (−3, 32).

    ✏️ Try it

    1. Differentiate g(x) = (Exercise 5B Q37).
    2. Find the gradient of y = + x at (−2, −1) and at (2, 3) (Exercise 5C Q12).
    3. Find the x co-ordinates of the points on y = x³ − 6x² + 9x + 1 where the gradient is −3.
    Answers

    1. Step 1: Split over √x = x1/2: g(x) = 3x3/2 − 2x1/2. Step 2: Differentiate term by term: g′(x) = (9/2)x1/2 − x−1/2 = (9/2)√x − 1/√x.

    2. Step 1: y = 4x⁻² + x ⇒ dy/dx = −8x⁻³ + 1 = 1 − 8/x³. Step 2: At x = −2, dy/dx = 1 − 8/(−8) = 2. Step 3: At x = 2, dy/dx = 1 − 8/8 = 0 (a stationary point).

    3. Step 1: dy/dx = 3x² − 12x + 9. Step 2: Set 3x² − 12x + 9 = −3 ⇒ 3x² − 12x + 12 = 0 ⇒ 3(x − 2)² = 0. Step 3: x = 2.

    Next: tangents & normals →

    Equations of tangents and normals

    • At a point P(x₁, y₁) on y = f(x), the gradient of the tangent is m₁ = value of when x = x₁.
    • The normal at P is the straight line perpendicular (at 90°) to the tangent at P.
    • Since perpendicular gradients satisfy m₁m₂ = −1, the gradient of the normal is m₂ = − (provided m₁ ≠ 0).
    • Use the point-gradient line equation y − y₁ = m(x − x₁) with m = m₁ for the tangent and m = m₂ for the normal.
    12341234xyP(x₁, y₁)tangent (m₁)normal (m₂ = −1/m₁)y = f(x)
    Figure 5.13: at P(x₁, y₁), the tangent touches the curve with slope m₁ and the normal is at right angles with slope m₂ = −1/m₁.
    −3−2−112351015xy(2, 12)y = 7x − 2y = x² + 3x + 2
    Example 5.8: the tangent to y = x² + 3x + 2 at (2, 12) has gradient 7 and equation y = 7x − 2.

    📘 Example 5.8 · Tangent to y = x² + 3x + 2 at (2, 12)

    1. Differentiate: dy/dx = 2x + 3.
    2. Find gradient m₁ at x = 2: m₁ = 2(2) + 3 = 7.
    3. Use y − y₁ = m₁(x − x₁) with (2, 12): y − 12 = 7(x − 2).
    4. Simplify: y = 7x − 2. Check: 7(2) − 2 = 12. ✔

    📘 Example 5.9 · Normal to y = − 4√x at (4, −4) meeting the y axis at P

    1. Rewrite in index form: y = 16x⁻¹ − 4x1/2.
    2. Differentiate: dy/dx = −16x⁻² − 2x−1/2 = − − .
    3. Substitute x = 4: m₁ = − − = −1 − 1 = −2 (tangent gradient).
    4. Normal gradient: m₂ = − = − = .
    5. Normal equation at (4, −4): y − (−4) = (x − 4) ⇒ y + 4 = x − 2 ⇒ y = x − 6.
    6. On the y axis, x = 0 ⇒ y = −6, so P is (0, −6).

    🧪 Tool: tangent, normal & stationary-point explorer for y = ax³ + bx² + cx + d

    • Enter coefficients a, b, c, d and a point x₀. The tool plots the curve, tangent and normal at x₀, and shows every step of the working.
    Curve (primary), tangent at x₀ (secondary) and normal at x₀ (dashed).

      ✏️ Try it

      1. Find the equation of the tangent to y = x³ − 4x² at (2, −8), and find where this tangent meets the curve again (Exercise 5D Q3).
      2. Find the equations of the tangent and normal to y = √x at (1, 1), and find the distance AB between their x-axis intercepts (Exercise 5D Q10).
      3. Show that the normal to y = 5 − at P(2, 1) is 2y + x = 4, and find where it meets the curve again at Q (Exercise 5D Q20).
      Answers

      1. Step 1: dy/dx = 3x² − 8x; at x = 2, m₁ = 12 − 16 = −4. Step 2: Tangent is y − (−8) = −4(x − 2) ⇒ y = −4x. Step 3: Intersect with curve: x³ − 4x² = −4x ⇒ x(x − 2)² = 0 ⇒ x = 0, so it meets the curve again at (0, 0).

      2. Step 1: dy/dx = 1/(2√x); at (1, 1), m₁ = 1/2 and m₂ = −2. Step 2: Tangent: y = (1/2)x + 1/2 (cuts y = 0 at A(−1, 0)); Normal: y = −2x + 3 (cuts y = 0 at B(1.5, 0)). Step 3: Length AB = 1.5 − (−1) = 2.5.

      3. Step 1: dy/dx = 8/x²; at x = 2, m₁ = 2 so normal has m₂ = −1/2. Step 2: y − 1 = −(1/2)(x − 2) ⇒ 2y + x = 4 ✔. Step 3: Substitute y = 5 − 8/x: 2(5 − 8/x) + x = 4 ⇒ x² + 6x − 16 = 0 ⇒ (x − 2)(x + 8) = 0 ⇒ x = −8, y = 6, so Q is (−8, 6).

      Next: stationary points & increasing/decreasing functions →

      Maximum and minimum (stationary) points

      • At any maximum or minimum turning point on a smooth curve, the tangent is horizontal, so = 0.
      • Points where = 0 are called stationary points (or turning points when the gradient changes sign).
      • Local maximum: as x increases through the point, goes from + to 0 to − (rising → flat → falling).
      • Local minimum: as x increases through the point, goes from − to 0 to + (falling → flat → rising).
      • Note (Activity 5.6): a local maximum need not be the highest point on the whole graph — a curve can rise higher elsewhere.
      −2−112−224xyMax (−1, 3)Min (1, −1)slope +slope −slope +
      Example 5.10: y = x³ − 3x + 1 has a local maximum at (−1, 3) (+ → 0 → −) and a local minimum at (1, −1) (− → 0 → +).
      −1−0.50.510.811.2xyMax (0, 1)Min (−0.5, 0.875)Min (0.5, 0.875)
      Example 5.11: y = 2t⁴ − t² + 1 has three stationary points close together: minima at (±0.5, 0.875) and a maximum at (0, 1).

      📘 Example 5.10 · Find the stationary points on y = x³ − 3x + 1

      1. Differentiate: dy/dx = 3x² − 3 = 3(x + 1)(x − 1).
      2. Solve dy/dx = 0: x = −1 or x = 1.
      3. Test the sign of dy/dx either side of x = −1: at x = −2, dy/dx = +9 (+); at x = 0, dy/dx = −3 (−). Since + → 0 → −, x = −1 gives a maximum.
      4. Test either side of x = 1: at x = 0, dy/dx = −3 (−); at x = 2, dy/dx = +9 (+). Since − → 0 → +, x = 1 gives a minimum.
      5. Substitute into y = x³ − 3x + 1: at x = −1, y = 3; at x = 1, y = −1. Answer: maximum at (−1, 3), minimum at (1, −1).

      📘 Example 5.11 · Stationary points of y = 2t⁴ − t² + 1 using a sign table

      1. Differentiate: dy/dt = 8t³ − 2t = 2t(4t² − 1) = 2t(2t − 1)(2t + 1).
      2. Solve dy/dt = 0: t = −0.5, t = 0 or t = 0.5.
      3. Test points in each interval (e.g. t = −1, −0.25, 0.25, 1):
      tt < −0.5−0.5−0.5 < t < 000 < t < 0.50.5t > 0.5
      Sign of dy/dt−0+0−0+
      Nature↘min↗max↘min↗
      1. Find y: at t = 0, y = 1; at t = ±0.5, y = 2(1/16) − 1/4 + 1 = 0.875 (7/8). Answer: maximum (0, 1), minima (−0.5, 0.875) and (0.5, 0.875).

      Increasing and decreasing functions

      • A function is increasing on an interval if its gradient is positive: > 0.
      • A function is decreasing on an interval if its gradient is negative: < 0.
      • To prove a function is increasing for all real x, complete the square or use an even power (like x² ≥ 0) to show dy/dx > 0 everywhere.
      −1123456−9−6−336xyDecreasing (x < 3)Increasing (x > 3)(3, −9): dy/dx = 0
      Example 5.13: y = x² − 6x has dy/dx = 2x − 6, so it is decreasing for x < 3 and increasing for x > 3.

      📘 Examples 5.12 & 5.13 · Increasing and decreasing functions

      1. Example 5.12: Show that y = x³ + x is an increasing function for all x.
      2. Differentiate: dy/dx = 3x² + 1. Since x² ≥ 0 for all real x, dy/dx ≥ 1 > 0 for all x, so y = x³ + x is always increasing. ✔
      3. Example 5.13: Find the range of values of x for which y = x² − 6x is decreasing.
      4. Differentiate: dy/dx = 2x − 6. Set dy/dx < 0: 2x − 6 < 0 ⇒ x < 3.

      ✏️ Try it

      1. Find the stationary points of y = x³ + 3x² − 9x + 6 and determine their nature (Exercise 5E Q7).
      2. Find the stationary points of f(x) = 4x + and the set of values of x for which f(x) is increasing (Exercise 5E Q13).
      3. Find the set of values of x for which y = (2x − 3)³ − 4x is an increasing function (Exercise 5E Q14).
      Answers

      1. Step 1: dy/dx = 3x² + 6x − 9 = 3(x + 3)(x − 1) = 0 ⇒ x = −3 or x = 1. Step 2: Sign of dy/dx goes + → 0 → − at x = −3 (max) and − → 0 → + at x = 1 (min). Step 3: Maximum at (−3, 33) and minimum at (1, 1).

      2. Step 1: f′(x) = 4 − 1/x² = 0 ⇒ x² = 1/4 ⇒ x = ±1/2. Step 2: Maximum at (−1/2, −4) and minimum at (1/2, 4). Step 3: Increasing when 4 − 1/x² > 0 ⇒ x² > 1/4 ⇒ x < −1/2 or x > 1/2.

      3. Step 1: dy/dx = (2x − 3)² − 4. Step 2: Increasing when (2x − 3)² − 4 > 0 ⇒ (2x − 3)² > 4 ⇒ 2x − 3 < −2 or 2x − 3 > 2. Step 3: x < 1/2 or x > 5/2.

      Next: points of inflection & the second derivative →

      Points of inflection

      • A point of inflection is a point where the curve changes from bending one way (concave) to bending the other way (convex) — so the tangent actually crosses the curve.
      • At a stationary point of inflection (such as (0, 0) on y = x³), = 0, but the gradient has the same sign on both sides (+ → 0 → + or − → 0 → −).
      • A non-stationary point of inflection can also occur where ≠ 0 (Figure 5.24: the steering wheel switches from left lock to right lock).

      The second derivative d²y/dx²

      • Differentiating a second time gives the second derivative, written or f′′(x): = ().
      • Warning: is the rate of change of the gradient — it is not the same as ()².
      • Example 5.14: if y = x⁵ + 2x, then = 5x⁴ + 2 and = 20x³.
      xyP (max)Q (min)xy−11xy< 0 at P> 0 at Qy = f(x)dy/dxd²y/dx²
      Figure 5.25 & Activity 5.7: lining up y = f(x), dy/dx and d²y/dx². At a maximum P, dy/dx = 0 and is decreasing so d²y/dx² < 0. At a minimum Q, dy/dx = 0 and is increasing so d²y/dx² > 0.

      Second-derivative test for stationary points

      At a stationary point (dy/dx = 0)What dy/dx is doingConclusion
      < 0dy/dx is decreasing (+ → 0 → −)Local maximum
      > 0dy/dx is increasing (− → 0 → +)Local minimum
      = 0Inconclusive! Could be max, min or inflectionTest the sign of dy/dx on both sides
      xyy = x³ (inflection)slope: + → 0 → +xyy = x⁴ (minimum)slope: − → 0 → +xyy = −x⁴ (maximum)slope: + → 0 → −
      Figures 5.29–5.31: why d²y/dx² = 0 is inconclusive. All three curves have dy/dx = 0 and d²y/dx² = 0 at (0, 0), yet they are an inflection, a minimum and a maximum.

      📘 Example 5.15 · Stationary points of y = 2x³ + 3x² − 12x using d²y/dx²

      1. Differentiate: dy/dx = 6x² + 6x − 12 = 6(x² + x − 2) = 6(x + 2)(x − 1).
      2. Solve dy/dx = 0: x = −2 or x = 1.
      3. Differentiate again: d²y/dx² = 12x + 6.
      4. At x = −2: d²y/dx² = 12(−2) + 6 = −18 < 0 ⇒ maximum. With y = 2(−8) + 3(4) − 12(−2) = 20, the maximum is (−2, 20).
      5. At x = 1: d²y/dx² = 12(1) + 6 = +18 > 0 ⇒ minimum. With y = 2 + 3 − 12 = −7, the minimum is (1, −7).

      ✏️ Try it

      1. For y = x⁴ − 8x², find dy/dx, d²y/dx² and the co-ordinates and nature of all three stationary points (Exercise 5F Q3).
      2. The curve y = px³ + qx² has a stationary point at (1, −1). Find p and q (Exercise 5F Q6).
      3. For y = 6√x − x√x, find the x co-ordinate of the stationary point and show it is a maximum (Exercise 5F Q9).
      Answers

      1. Step 1: dy/dx = 4x³ − 16x = 4x(x − 2)(x + 2) = 0 ⇒ x = 0, ±2. Step 2: d²y/dx² = 12x² − 16. Step 3: At x = 0, d²y/dx² = −16 < 0 ⇒ maximum at (0, 0); at x = ±2, d²y/dx² = 32 > 0 ⇒ minima at (−2, −16) and (2, −16).

      2. Step 1: Passes through (1, −1) ⇒ p + q = −1. Step 2: dy/dx = 3px² + 2qx = 0 at x = 1 ⇒ 3p + 2q = 0. Step 3: Solve simultaneously: p = 2, q = −3.

      3. Step 1: Rewrite y = 6x1/2 − x3/2 ⇒ dy/dx = 3x−1/2 − (3/2)x1/2 = 0 ⇒ 3/√x = (3/2)√x ⇒ x = 2. Step 2: d²y/dx² = −(3/2)x−3/2 − (3/4)x−1/2, which is negative at x = 2 ⇒ maximum ✔.

      Next: practical maximum & minimum problems →

      Solving practical optimisation problems

      1. Draw & labelchoose variables x, y 2. Eliminate yuse given constraint 3. Solve dA/dx = 0find stationary x 4. Check d²A/dx²& state answer
      Every practical max/min problem follows the same four steps: express the target quantity in one variable, differentiate, set to zero and verify with the second derivative.
      0.51.2522.5123xyMax (1.25, 3.125)A = 5x − 2x²old garden wallx mx my = (5 − 2x) mArea A = x(5 − 2x)Rope: 2x + y = 5 m
      Example 5.16: a 5 m rope encloses three sides of a rectangular plot against a wall (left). The area A = 5x − 2x² peaks at x = 1.25 m (right).

      📘 Example 5.16 · Kelly’s vegetable plot against a wall (5 m rope)

      1. Let the two sides perpendicular to the wall have length x m and the side parallel to the wall have length y m.
      2. Use the 5 m rope constraint to eliminate y: 2x + y = 5 ⇒ y = 5 − 2x.
      3. Write the area A in terms of x alone: A = xy = x(5 − 2x) = 5x − 2x².
      4. Differentiate and set to zero: dA/dx = 5 − 4x = 0 ⇒ x = 5/4 = 1.25 m.
      5. Check nature: d²A/dx² = −4 < 0 ⇒ maximum. With y = 5 − 2(1.25) = 2.5 m, the largest area is 1.25 m by 2.5 m (area 3.125 m²).
      123456153045xyMax height 45 m at t = 3 s (dh/dt = 0)h = 30t − 5t²
      Example 5.17: height h = 30t − 5t² of a stone thrown upwards at 30 m s⁻¹. The gradient dh/dt is velocity and d²h/dt² = −10 m s⁻² is acceleration.

      📘 Example 5.17 · Stone projected upwards: h = 30t − 5t²

      1. Differentiate with respect to t: velocity dh/dt = 30 − 10t, and acceleration d²h/dt² = −10.
      2. At maximum height, dh/dt = 0 ⇒ 30 − 10t = 0 ⇒ t = 3 s (and d²h/dt² = −10 < 0 confirms a maximum).
      3. Substitute t = 3: h = 30(3) − 5(3)² = 90 − 45 = 45 m.

      🎛️ Simulation: fold the open box and watch the volume peak

      • Exercise 5G Q2: a 12 cm × 12 cm card has squares of side x cm cut from each corner and folded up into an open box.
      • Drag x from 0.2 cm to 5.8 cm to see the cut card, the 3D box and the point moving along V(x) = x(12 − 2x)².
      Left: 12 cm × 12 cm card with x cm corners cut out. Right: volume V = 4x³ − 48x² + 144x peaking at x = 2 cm (V = 128 cm³).

        ✏️ Try it

        1. A farmer has 120 m of fencing for a complete rectangular enclosure of length x m and width y m. Find the maximum area (Exercise 5G Q1).
        2. An open-top glass fish tank with a square base of side x m and height y m has volume 0.5 m³. Find x and y to minimise the glass surface area (Exercise 5G Q10).
        3. A closed cylinder of radius r cm and height h cm has total surface area 250π cm². Express V in terms of r and find the maximum volume to 3 s.f. (Exercise 5G Q8).
        Answers

        1. Step 1: 2x + 2y = 120 ⇒ y = 60 − x. Step 2: A = x(60 − x) = 60x − x² ⇒ dA/dx = 60 − 2x = 0 at x = 30 m (d²A/dx² = −2 < 0). Step 3: 30 m by 30 m, maximum area = 900 m².

        2. Step 1: V = x²y = 0.5 ⇒ y = 1/(2x²). Step 2: Surface area A = x² + 4xy = x² + 2/x ⇒ dA/dx = 2x − 2/x² = 0 ⇒ x³ = 1 ⇒ x = 1 m, y = 0.5 m (d²A/dx² = 2 + 4/x³ = 6 > 0, min area 3 m²).

        3. Step 1: 2πr² + 2πrh = 250π ⇒ h = (125 − r²)/r. Step 2: V = πr²h = 125πr − πr³ ⇒ dV/dr = 125π − 3πr² = 0 ⇒ r = √(125/3) ≈ 6.45 cm (d²V/dr² = −6πr < 0). Step 3: V = (250π/3)√(125/3) ≈ 1690 cm³ (3 s.f.).

        Next: the chain rule & connected rates of change →

        The chain rule for composite functions

        • How do we differentiate a composite “function of a function” like y = √(x² + 1) = (x² + 1)1/2 or y = (x² − 2)⁴ without expanding?
        • Call the inside expression u: let u = x² + 1 so that y = u1/2.
        • A small change δx in x produces a small change δu in u and a small change δy in y, and by algebra = × .
        • Taking the limit as δx → 0 gives the chain rule: = × .
        Input xsmall change δx du/dx Inside u = g(x)small change δu dy/du Output y = f(u)small change δy Overall rate: dy/dx = (dy/du) × (du/dx)
        The chain rule multiplies the rate of change of the outside function (dy/du) by the rate of change of the inside function (du/dx).

        📘 Example 5.18 · Differentiate y = (x² + 1)1/2

        1. Let u = x² + 1, so y = u1/2.
        2. Differentiate each part: du/dx = 2x and dy/du = u−1/2 = .
        3. Multiply using the chain rule: dy/dx = × = × 2x.
        4. Cancel the 2 (and make sure no u remains): = .

        📘 Example 5.19 · Differentiate y = (x² − 2)⁴

        1. Let u = x² − 2, so y = u⁴.
        2. Then du/dx = 2x and dy/du = 4u³ = 4(x² − 2)³.
        3. Multiply: dy/dx = 4(x² − 2)³ × 2x = 8x(x² − 2)³.
        4. Quick mental pattern for [g(x)]ⁿ: bring down n, reduce the bracket’s power to n − 1, and multiply by g′(x).

        Connected rates of change (differentiating with respect to time t)

        • “Increasing at a rate of …” means differentiating with respect to time t (e.g. dr/dt, dA/dt, dV/dt).
        • If you know how area A or volume V depends on radius r, link the time rates using the chain rule: = × and = × .
        r = 25 cm dr/dt = 8 cm s⁻¹ Example 5.20: circular ripple A = πr² ⇒ dA/dr = 2πr dA/dt = (dA/dr) × (dr/dt) = 2πr × 8 At r = 25: dA/dt = 400π ≈ 1260 cm² s⁻¹
        Example 5.20: as the circular ripple expands at dr/dt = 8 cm s⁻¹, its area grows at dA/dt = (dA/dr) × (dr/dt).

        📘 Example 5.20 · Circular ripple with radius increasing at 8 cm s⁻¹ when r = 25 cm

        1. We are given dr/dt = 8 cm s⁻¹ and r = 25 cm, and we want dA/dt.
        2. Area of a circle: A = πr² ⇒ dA/dr = 2πr.
        3. Chain rule: = × = 2πr × .
        4. Substitute r = 25 and dr/dt = 8: dA/dt = 2π(25)(8) = 400π ≈ 1260 cm² s⁻¹ (3 s.f.).

        ✏️ Try it

        1. Differentiate (a) y = (2x + 3)⁴ and (b) y = (Exercise 5H Q1(ii), (vi)).
        2. For y = (3x − 5)³, find the equation of the tangent at (2, 1) and the normal at (1, −8) (Exercise 5H Q2).
        3. The force between two magnetic poles is F = . Find dF/dt when r = 0.2 m and dr/dt = 0.03 m s⁻¹ (Exercise 5H Q6).
        Answers

        1. Step 1: (a) Let u = 2x + 3: dy/dx = 4(2x + 3)³ × 2 = 8(2x + 3)³. Step 2: (b) Rewrite y = (x² − 3)⁻³: dy/dx = −3(x² − 3)⁻⁴ × 2x = −6x/(x² − 3)⁴.

        2. Step 1: dy/dx = 3(3x − 5)² × 3 = 9(3x − 5)². Step 2: At (2, 1), m₁ = 9(1)² = 9 ⇒ tangent is y − 1 = 9(x − 2) ⇒ y = 9x − 17. Step 3: At (1, −8), m₁ = 9(−2)² = 36 ⇒ normal has m₂ = −1/36 ⇒ y + 8 = −(1/36)(x − 1) ⇒ 36y + x + 287 = 0 ✔.

        3. Step 1: F = (1/500)r⁻² ⇒ dF/dr = −2/(500r³) = −1/(250r³). Step 2: At r = 0.2, dF/dr = −1/(250 × 0.008) = −0.5. Step 3: dF/dt = (dF/dr)(dr/dt) = (−0.5)(0.03) = −0.015 N s⁻¹.

        Next: summary and quiz →

        Chapter summary

        • The gradient of a curve at P is the gradient of its tangent at P, found from first principles as limh → 0 = .
        • Power & sum rules: y = kxⁿ ⇒ dy/dx = knxⁿ⁻¹, y = c ⇒ dy/dx = 0, and y = f(x) + g(x) ⇒ dy/dx = f′(x) + g′(x).
        • Always expand brackets and rewrite roots (√x = x1/2) and reciprocals (1/xⁿ = x⁻ⁿ) before differentiating.
        • At (x₁, y₁), tangent gradient is m₁ = dy/dx and normal gradient is m₂ = −1/m₁; use y − y₁ = m(x − x₁).
        • Function is increasing where dy/dx > 0, decreasing where dy/dx < 0, and stationary where dy/dx = 0.
        • Classify stationary points using a sign table for dy/dx or the second derivative d²y/dx² (< 0 max, > 0 min, = 0 check dy/dx either side).
        • Chain rule: = × , which also links connected rates of change such as = × .
        Key pointFormula / rule
        Power & constant rulesy = kxⁿ ⇒ dy/dx = knxⁿ⁻¹; y = c ⇒ dy/dx = 0
        Sum / difference ruley = f(x) + g(x) ⇒ dy/dx = f′(x) + g′(x)
        Tangent at (x₁, y₁)y − y₁ = m₁(x − x₁), where m₁ = dy/dx at x = x₁
        Normal at (x₁, y₁)y − y₁ = m₂(x − x₁), where m₂ = −1/m₁
        Stationary pointsdy/dx = 0; d²y/dx² < 0 ⇒ maximum, d²y/dx² > 0 ⇒ minimum
        When d²y/dx² = 0Check the sign of dy/dx just either side of the point
        Chain ruledy/dx = (dy/du) × (du/dx)

        Which differentiation tool do I need?

        Prove gradient from first principles?→Simplify (f(x + h) − f(x))/h, cancel h, then let h → 0
        Brackets, roots or fractions in x?→Expand and rewrite every term as kxⁿ before differentiating
        Equation of tangent or normal at x₁?→Find y₁ and m₁ = f′(x₁); use m₁ (tangent) or −1/m₁ (normal) in y − y₁ = m(x − x₁)
        Stationary points or max/min problem?→Express in 1 variable, solve dy/dx = 0, test with d²y/dx² or sign of dy/dx
        Power of a bracket or rate of change in time t?→Use the chain rule: dy/dx = (dy/du)(du/dx) or dV/dt = (dV/dr)(dr/dt)
        Common slipFix
        Differentiating a product or quotient term by termMultiply out brackets or split over a single-term denominator first
        Adding 1 to a negative power (−2 → −1 instead of −3)Subtract 1 carefully: d/dx(x⁻²) = −2x⁻³ and d/dx(x1/2) = (1/2)x−1/2
        Using dy/dx as the gradient of the normalFlip and change sign: mnormal = −1 / mtangent
        Forgetting the inside derivative in the chain ruleAlways multiply by du/dx: d/dx[(2x² − 3)⁵] = 5(2x² − 3)⁴ × 4x
        Assuming d²y/dx² = 0 means a point of inflectionTest the sign of dy/dx on both sides (think of y = x⁴ vs y = x³)

        📄 Practice questions

        The book's Exercises 5A to 5H, its activities and the Think about it questions are in the printable assignment for this chapter, along with a short Check yourself set.

        Mini-quiz

        1. What is the gradient of the chord joining (2, 4) to (2 + h, (2 + h)²) on y = x²?
        2. Differentiate y = 3x⁴ − 5x² + 7x − 9 with respect to x.
        3. Find dy/dx when y = 4√x − 6/x².
        4. What is the gradient of the curve y = x³ − 4x at the point (2, 0)?
        5. Find the equation of the tangent to y = x² − 3x + 4 at the point (2, 2).
        6. A curve has gradient 3 at (1, 4). What is the equation of the normal at (1, 4)?
        7. For what values of x is f(x) = x³ − 6x² + 5 a decreasing function?
        8. Find the stationary points of y = x³ − 3x + 2 and their nature.
        9. Given y = 2x⁴ − 3/x, find the second derivative d²y/dx².
        10. Two positive numbers have sum 10. What is the maximum value of their product?
        11. Differentiate y = (2x² − 3)⁵ with respect to x.
        12. A circle’s radius r cm increases at 3 cm s⁻¹. How fast is its area increasing when r = 4?

        Source: Pure Mathematics 1 (Hodder Education, Sophie Goldie), Chapter 5 (Differentiation). Explanations, the extra worked examples, the Try it answers, the tool, the simulations and the quiz are Guru Jazzy's own. This page is not made or endorsed by Hodder Education or Cambridge.