G5 - Implicit and parametric differentiation
Find first derivatives when a curve is given by a relation between its coordinates or by a parameter. Use them for tangents, normals and horizontal tangent points, checking the original relation and allowed parameters before interpreting a gradient.
Differentiating an implicit relation
How can we find a tangent when a curve has not been rearranged into ? For , points above and below the horizontal axis have different -values at the same . We can still follow a local branch and differentiate its equation.
You already know
The chain rule differentiates a function of another function. The product rule differentiates both changing factors. Both rules are recalled; the quotient rule is supplied in the formula booklet.
An implicit relation links and without necessarily isolating . Locally, where a branch gives a differentiable , every occurrence of changes with . For example,
These implicit derivative results follow from the recalled chain and product rules; they are not supplied as separate formulae. For , the derivative is , with , also a recall consequence. For , where , the supplied quotient rule gives .
Differentiate every term, collect the terms containing , then divide only when its coefficient is nonzero.
Worked example 1: an implicit quadratic
Question. For , find and its value at .
Plan. Use product and chain rules, collect , then verify and substitute the point.
Why does give two terms? Both and vary, so it needs the product rule.
Answer. , where , with gradient at .
Check. , so the point is on the curve. Substitution in the differentiated equation gives .
Change only the power of the final -term from a square to a cube; choose the constant so that the same point remains on the curve. The chain derivative now includes .
Worked example 2: a different power
Question. For , find and its value at .
Plan. Differentiate the cubic -term by the chain rule and collect the gradient terms.
Why retain after differentiating ? The derivative is with respect to , rather than with respect to .
Answer. , with gradient at .
Check. , and the differentiated equation gives .
A simpler relation follows the same rule: gives , hence where . No rearrangement into cube roots is needed.
Watch out
Wrong: when differentiating with respect to . This omits the changing inner function . Right: .
Change the final power once more. Verify the supplied point before interpreting the derivative.
Valid points, tangents and zero denominators
The original equation decides which points exist; the local geometry decides whether a finite gradient is available.
For a valid point with finite gradient , the tangent is . If , the normal has gradient . A horizontal tangent has a vertical normal; a vertical tangent has a horizontal normal. These line and perpendicular-gradient rules are not supplied in the formula booklet: learn them.
For the circle , differentiation gives
This first-derivative formula is derived using recall rules; it is not supplied. It cannot be evaluated with .
Worked example 3: a finite gradient
Question. Find the tangent and normal to at .
Plan. Verify the point, evaluate the implicit gradient and use the two line gradients.
Why take the negative reciprocal for the normal? It is perpendicular to the tangent, and the tangent gradient here is finite and nonzero.
Answer. Tangent ; normal .
Check. Both lines contain , and their gradients multiply to .
Change only the point on the same circle to . The finite-gradient formula now divides by zero, so use the circle's geometry or regard as a function of .
Worked example 4: an axis-aligned tangent
Question. Find the tangent and normal to at .
Plan. Check the point and use near it.
Why reverse the dependent coordinate? has a finite derivative with respect to here, while has no finite derivative with respect to .
Answer. Tangent ; normal .
Check. The radius at is horizontal, so its perpendicular tangent is vertical.
More generally, differentiation may give . Where , division gives a finite candidate slope. If and , no finite slope can satisfy that equation; at a regular smooth curve point this indicates a vertical tangent. If both coefficients vanish, this equation alone says nothing about the slope. For example, gives zero coefficients at the origin, yet consists of the two lines and . There is no unique tangent there. This reasoning is not a supplied formula.
Watch out
Wrong: “Every zero denominator means a vertical tangent.” If numerator and denominator both vanish, the quotient is indeterminate. Right: check the original curve and another local representation; only conclude a vertical tangent when the geometry supports it.
Move to one more point on the same circle. This time the tangent is horizontal, so the normal must be written without a gradient.
Using a parameter to find the gradient
A parameter determines both coordinates, so the curve gradient compares the two coordinate rates at the same parameter value.
For , , increasing changes both and . The chain rule gives . Dividing is valid only when .
Continue this lesson.
You’re previewing NovaLearn without an account. Create a free account to read the rest.
Create free account