Differentiation and the gradient of a curve
A straight line has one gradient everywhere, but a curve gets steeper and shallower as you move along it. Differentiation is the tool that gives the gradient of a curve at every point at once, as a formula called the derivative. Everything else in the Paper 1 calculus questions (tangents and normals, stationary points, rates of change, optimisation, and integration as the reverse process) is built on the rules in this note, so they have to be fast and error-free.
The gradient of a curve
Gradient at a point means gradient of the tangent
The tangent to a curve at a point is the straight line that touches the curve at and has the same direction as the curve there. The gradient of the curve at is defined to be the gradient of that tangent.
The difficulty is that a tangent passes through only one known point, and you need two points to calculate a gradient. The way round this is to approximate the tangent by chords.
Chords that close in on the tangent
Take the curve and the point . Pick a second point on the curve with -coordinate , so is . The gradient of the chord is
Now slide towards by making smaller:
| Gradient of chord | |
|---|---|
As gets closer to , the chord gradients get closer and closer to . The chord turns into the tangent, so the gradient of the curve at is . In the language of limits, as .
The graph shows , the chord from to (the case , gradient ) and the tangent at , gradient . Smaller values of give chords that lie closer and closer to the tangent.
The gradient of a curve at a point is the gradient of the tangent at that point. It is the limit of the gradients of chords as moves along the curve towards .
Only an informal idea of a limit is expected: you should be able to set up the chord gradient for a specific curve, simplify it, and say what it tends to as . You are not asked to differentiate general functions from first principles.
The derivative and its notation
Repeat the chord argument at a general point on , with at :
So the gradient of at any point is . At the gradient is ; at it is . A formula like this, which gives the gradient at every point, is the gradient function or derivative.
The derivative of with respect to is the function that gives the gradient of the curve at each value of . It is written
Finding the derivative is called differentiating. Differentiating again gives the second derivative, written or .
is a single symbol, read "dee by dee ". It is not a fraction you can cancel, although in the chain rule it often behaves like one. The notation means "the derivative of with respect to ", so .
The second derivative measures how fast the gradient itself is changing. You will use it to decide whether a stationary point is a maximum or a minimum.
The power rule
The same chord argument for , , and so on reveals a pattern:
The power comes down in front, and the power drops by one. This works for every rational power, including negative powers and fractions.
- Constant multiples stay: .
- Sums and differences differentiate term by term: .
- Constants differentiate to zero: , because is a horizontal line.
- Linear terms give their coefficient: , because is a line of gradient .
For example, if then , and .
Rewrite as powers first
The power rule only works on terms of the form . Roots, reciprocals, products and quotients must be rewritten before you differentiate.
| Written as | Rewrite as | Derivative |
|---|---|---|
Two of these are traps worth naming. is , not : only the is in the denominator's power. And a fraction with several terms on top is split into separate terms, each divided by the denominator. There is no product rule or quotient rule in Paper 1; you never need one, because every P1 product or quotient can be expanded or split.
- Expand any brackets (unless it is a bracket raised to a power, which needs the chain rule).
- Split any fraction with more than one term in the numerator.
- Write every term as , using and .
- Differentiate each term: multiply by the power, then reduce the power by one.
- If a value is asked for, substitute after differentiating. Convert back to roots and fractions only if it helps.
Using the derivative
Once you have there are three things you can do with it.
- Find the gradient at a given point: substitute the -coordinate into .
- Find where the gradient has a given value: set equal to that value and solve for , then substitute into the curve's equation (not the derivative) for .
- Find unknown constants: if the curve contains letters such as and , each piece of information (a point on the curve, a gradient at a point) gives one equation.
The curve has gradient at two points, and , so it has two parallel tangents of gradient . A derivative that is a quadratic can take the same value twice, which is why "find the points where the gradient is " usually has two answers. This is Example 4 below.
Given that , find and , and find the gradient of the curve at the point where .
Solution
Differentiate term by term:
At :
The gradient at is .
Find when .
Solution
Rewrite every term as a power of . Note that .
Differentiate:
The middle term is : two negatives make a positive.
A curve has equation . Find the gradient of the curve at the point where .
Solution
Split the fraction, dividing each term of the numerator by :
Differentiate:
At , , so the gradient is
Find the coordinates of the points on the curve at which the gradient is .
Solution
Set the gradient equal to :
So or . Substitute into the curve for the -coordinates:
The points are and .
The curve , where and are constants, passes through the point , and its gradient at that point is .
(a) Find the values of and .
(b) Find the value of at the point .
Solution
(a) The point lies on the curve:
Write , so
The gradient at is :
Subtract (2) from (1): , so and . Then from (1), , so .
(b) With , , , so
At : .
The points and lie on the curve .
(a) Show that the gradient of the chord is .
(b) Hence state the gradient of the curve at , and check your answer by differentiation.
Solution
(a) The gradient of is
(b) As approaches , , so the chord gradient . The gradient of the curve at is .
Check: , which is at .
Differentiating before rewriting. is not . Write first, then differentiate to get .
Getting the new power wrong for negative powers. Reducing by one gives , not . Reducing by one gives .
Misplacing a coefficient. means , but is just . And , not .
Differentiating a product term by term. is not . Expand to first, giving .
Using the derivative to find . After solving for , substitute into the equation of the curve to find , not into .
- Show the rewritten form. Writing before differentiating makes your method visible. If you then slip on one term, the method mark is still available.
- "Find " questions are usually marked term by term: one mark for each correct term, or a method mark for reducing powers correctly plus accuracy marks.
- Simplified form. Answers like and are both acceptable unless a form is asked for. Do not waste time converting back unless you are substituting a value.
- Exact values. If a question asks for an exact gradient, give a fraction like , not rounded.
- Chord questions ask you to "show" a simplified chord gradient: expand fully, cancel the , and state the limit as in words or symbols.
- The gradient of a curve at a point is the gradient of the tangent there, the limit of chord gradients as the second point approaches the first.
- and are the first derivative; and are the second derivative.
- Power rule: for any rational ; constants give , gives .
- Multiply constants through and differentiate sums term by term.
- Rewrite roots as fractional powers, reciprocals as negative powers, expand products, and split fractions before differentiating.
- To find where the gradient is , solve , then use the curve for .
- Unknown constants need one equation per piece of information.
Practice questions
- Find when .
- Find when .
- Find when .
- Given that for , find and evaluate .
- Given that , find and the value of .
- The points and lie on . Find, in terms of , the gradient of the chord , simplifying your answer. Hence find the gradient of the curve at .
- Find the coordinates of the points on the curve at which the gradient is .
- The curve has gradient at the point where and gradient at the point where . Find and , and the value of when .
- Find the set of values of the constant for which the gradient of the curve is positive for every value of .
Answers
-
.
-
, so .
-
Expand: , so .
-
, so . Then .
-
, , . So .
-
Gradient of . As this tends to , so the gradient at is (and at agrees).
-
, so , , . At , ; at , . Points and .
-
. At : . At : . Subtracting, , so and . Then , which is at .
-
. This quadratic has positive coefficient, so it is positive for all exactly when it has no real roots: discriminant , so , , giving .