Parametric Differentiation
- Sometimes and are both written in terms of a third letter (or ), e.g. .
- To get the gradient you go through : the chain rule one more time.
is the hidden inside. . So differentiate with respect to , differentiate with respect to , then divide. The answer comes out in , not .
The one-line method
(Worked Example 4.14)
Worked example
: find , then the tangent at
- Differentiate each with respect to : , .
- Divide ( on top, like ): .
- At : gradient , point .
- Line : .
Worked example
Real question: , show
9709/32 F/M 2023 Q5(a). The product rule on leads to a clean cancellation, and the whole question rests on it.
- is a product (inside derivative ): .
- : .
- Divide, and the cancels top and bottom: .
Spotting the shared factor turns a messy answer into one line. That cancellation is the whole question. (9709/32 Feb/March 2023 Q5)
Common mistake
- Inverting the fraction — writing . Memory hook: on top, just like .
- Leaving the answer in when a numeric gradient or point is asked — substitute the given .
- Axis confusion (see below) — mixing up which derivative is zero.
'Show that' and tangent parts
On a “show that ”, the tidy answer is already given. So you must show every algebra step and every trig-identity step. The target form tells you which identity to use.
Worked example
Real question: , tangent at
9709/31 M/J 2025 Q4. Find the gradient through , then write an ordinary P1 tangent line with exact and .
- Differentiate (both need the chain rule, inside with derivative ): , .
- Divide; the cancels: .
- At , gives , so .
- Line through : .
Reading off the given point is the step that gives you an exact gradient. Never reach for a decimal. (9709/31 May/June 2025 Q4)
- If the divided fraction comes out as , a target of is your hint to write it as a single cotangent.
- These parts often mix in logs and exponentials. For example, gives . (9709/31 Oct/Nov 2023 Q6)
Tangents parallel to an axis
Parallel to the -axis ⇒ gradient ⇒ set the top . Parallel to the -axis ⇒ straight up, infinite gradient ⇒ set the bottom .
- The why: a flat tangent needs the up-down rate to stop, a vertical tangent needs the side-to-side rate to stop. (Worked Example 4.15)
Your turn— tap to reveal the worked answer (9709-style: x = 3t, y = t³ + 4t² − 3t, gradient 0)
, , so . Gradient when the top is zero: .
- Maximum-gradient and second-derivative follow-ups sit in the advanced layer.