Gradient of the curve = rate of change — the limit of the gradient
Rules
Power rule first; chain (outside×inside′), product, quotient as needed
Stationary points
dy/dx = 0, then second derivative: positive = min, negative = max
Integration
Reverse differentiation +c; definite integral = area under the curve
Optimisation
Model → differentiate → set to zero → justify max/min
1. Limits — the derivative's foundation
A limit asks what value a function APPROACHES as x gets close to a point — not necessarily what it equals there. The derivative is born from a limit: take the gradient of a chord between two points, then slide the points together (h → 0) — the chord becomes the tangent and the limit becomes the derivative. Some limits do not exist: 1/x blows up as x → 0 — saying so earns the mark.
Example
Worked: gradient of y = x² at x = 2: chord to (2 + h, (2+h)²): ((2+h)² − 4)/h = (4h + h²)/h = 4 + h. As h → 0 the gradient → 4. The h in the denominator is why we MUST take the limit — and why the algebra always factors the h away first.
2. The derivative — rate of change
Key idea
Power rule
dy/dx is the gradient of the curve at x — the instantaneous rate of change. Differentiate term by term: d/dx(x^n) = nx^(n−1) — MULTIPLY by the power, then REDUCE it by one. Constants differentiate to zero (their gradient is 0 — flat lines).
Example
Worked: y = 3x⁴ − 2x² + 5 → dy/dx = 12x³ − 4x (the 5 vanishes). At x = 1 the gradient is 12 − 4 = 8; at x = 0 it is 0 — flat there. Reverse: the gradient 12x³ − 4x is itself a function — differentiating AGAIN gives the second derivative, section 8.
3. Tangents and normals — two lines, one point
The TANGENT at x = a has gradient f′(a) — it touches the curve. The NORMAL is perpendicular to the tangent, so its gradient is −1/f′(a) (flip and invert — geometry chapter 2, recycled). Both are straight-line equations: find the point y = f(a), find the gradient, then y − y₁ = m(x − x₁).
Example
Worked: y = x² at x = 3 → point (3, 9), gradient 6 → tangent: y = 6x − 9. Normal: gradient −1/6 → y − 9 = −(x − 3)/6. The normal question is a tangent question with one extra flip — and the flip is the mark.
4. Stationary points — find, then classify
Key idea
Stationary points
Where dy/dx = 0 the tangent is horizontal: the curve is momentarily flat. Classify with the SECOND derivative: f″ > 0 → MINIMUM (concave up), f″ < 0 → MAXIMUM (concave down), f″ = 0 → inconclusive (test the gradient either side instead).
Example
Worked: y = x³ − 3x → dy/dx = 3x² − 3 = 0 → x = ±1. Second derivative: 6x. At x = 1: +6 > 0 → minimum (y = −2). At x = −1: −6 < 0 → maximum (y = 2). Two stationary points, two classifications — the exam wants both, with the second-derivative test shown.
5. Optimization — the exam's favourite application
The recipe is always the same four beats: 1) translate the words into ONE variable (use the constraint to eliminate the other); 2) differentiate; 3) set dy/dx = 0 and solve; 4) classify (second derivative) AND answer the actual question asked — with units. Skipping the classification or answering the wrong quantity are the two ways candidates lose the final mark.
Example
Worked: 40 m of fence against a wall (no fence needed there): width x, length 40 − 2x → area A = x(40 − 2x) = 40x − 2x². dA/dx = 40 − 4x = 0 → x = 10. d²A/dx² = −4 < 0 → maximum ✓. Answer the QUESTION: dimensions 10 × 20, maximum area 200 m² — the area, not the width, is what was asked.
6. The chain rule — outside × inside
Key idea
Chain rule
For a function of a function: dy/dx = dy/du × du/dx — differentiate the outside (leaving the inside alone), then MULTIPLY by the derivative of the inside. In practice: d/dx[(ax + b)^n] = n(ax + b)^(n−1) × a. The ×a is the most-forgotten factor in the entire course.
Example
Worked: y = (2x + 1)⁵ → dy/dx = 5(2x + 1)⁴ × 2 = 10(2x + 1)⁴. Check against the expansion... or trust the rule. y = sin(3x) → 3cos(3x). Forgetting the ×2 (or ×3) turns a correct answer into a near-miss — examiners see it constantly.
7. Product and quotient rules
Key idea
Product rule
For y = uv: dy/dx = u′v + uv′ — first times the derivative of the second, plus second times the derivative of the first. For y = u/v: dy/dx = (u′v − uv′)/v² — the quotient rule, order matters in the numerator (the minus sign punishes a swap).
Example
Worked: y = x² sin x → dy/dx = 2x sin x + x² cos x. Worked: y = x/(x + 1) → (1·(x+1) − x·1)/(x + 1)² = 1/(x + 1)². Spot the structure FIRST: a product → product rule; a quotient → quotient rule; a function of a function → chain rule. Choosing the rule IS the first mark.
8. The second derivative — concavity and motion
f″(x) is the rate of change of the gradient: f″ > 0 means the gradient is INCREASING (concave up — a smile), f″ < 0 decreasing (concave down — a frown). The kinematics bridge: displacement s, velocity v = ds/dt, ACCELERATION a = dv/dt = d²s/dt² — the second derivative is acceleration. One idea, two subjects, same marks.
Example
Worked: s = t³ − 6t² → v = 3t² − 12t, a = 6t − 12. At rest: v = 0 → t = 0 or 4. Acceleration at t = 4: a = 12 m/s². The chain differentiate-then-substitute is the whole method — units carried (m, m/s, m/s²) are the sanity check.
9. Integration — differentiation in reverse
Key idea
The reverse power rule
∫x^n dx = x^(n+1)/(n+1) + c — RAISE the power by one, divide by the new power. The +c is NOT optional: differentiation destroys constants, so integration must restore an unknown one. Check every integral by differentiating your answer back — the two operations undo each other.
Example
Worked: ∫(3x² + 2x) dx = x³ + x² + c. Definite integral: ∫₁³(3x² + 2x) dx = [x³ + x²]₁³ = (27 + 9) − (1 + 1) = 34 — the c drops out when you subtract. ∫1/x dx = ln|x| + c; ∫e^x dx = e^x + c — the two special cases worth memorising.
10. Areas under curves — signed areas
The definite integral measures SIGNED area: regions BELOW the x-axis count negative. Area questions: split the interval at the roots and integrate each piece separately, then make everything positive. Area BETWEEN two curves: ∫(top − bottom) dx — one integral, no subtraction of separate areas needed.
Example
Worked: area under y = x² from 0 to 3: [x³/3]₀³ = 9. Worked — the trap: y = sin x from 0 to 2π integrates to 0 (positive half cancels negative half) — but the AREA is 4 (2 + 2). 'Evaluate the integral' and 'find the area' are different questions — read which one is asked.
11. Integration in kinematics — the bridge both ways
Differentiation runs displacement → velocity → acceleration; integration runs back: v = ∫a dt, s = ∫v dt. Each integration introduces a +c — fixed by a boundary condition (initial position, initial velocity). The full kinematics chapter of physics is this one rule wearing different letters.
Example
Worked: v = 6t − 4 with s = 0 at t = 0 → s = 3t² − 4t + c, c = 0. Displacement after 2 s: 12 − 8 = 4 m. The exam may ask for DISTANCE travelled, not displacement: if v changes sign mid-journey, integrate |v| piecewise — split where v = 0.
12. Differential equations — the exponential shortcut
Key idea
dy/dx = ky
The rate of change proportional to the amount: the equation of growth and decay. Its solution is y = Ae^(kx) — exponentials are their own derivatives (up to a constant). Separate variables when it is more general: collect the y terms with dy, the x terms with dx, integrate both sides.
Example
Worked: dy/dx = 2y → y = Ae^(2x); boundary y(0) = 5 fixes A = 5 → y = 5e^(2x). Worked: dy/dx = x/y → y dy = x dx → y²/2 = x²/2 + c → y² = x² + C. Population growth, radioactive decay, Newton's cooling — one equation family, infinite applications.
13. Summary — the calculus spine
Question type
Method beat
Trap
Derivatives
Power rule term by term
Constants vanish; coefficient × power
Tangent/normal
f′(a) then flip for the normal
The flip is the mark
Stationary
f′ = 0, then f″ classifies
f″ = 0 needs the sign test
Optimization
Constraint → one variable → f′ = 0
Answer the question asked, with units
Chain rule
Outside × derivative of inside
The forgotten ×a
Integration
Raise the power, divide; + c always
The +c — never omit
Areas
Signed! Split at roots
Integral ≠ area when the curve dips
Kinematics
Differentiate down, integrate up
Distance vs displacement
Note
Command discipline: 'differentiate' = the rule named, every term shown; 'find the equation of' = point + gradient + line form; 'show that' = full working to the target; 'hence' = use the previous result — starting over forfeits the method marks. Calculus papers are chains: every link shown, every link marked.
Spotted a problem with this content? Sign in to send it to our editors.