The gradient is still
On a parametric curve, x and y are both given in terms of a parameter t. The gradient at a point still means what it always means: , how fast y changes compared with x.
What is different is that neither equation says y in terms of x. Instead each says how its coordinate changes with t. Differentiating gives two rates, and , and the gradient has to be built from them.
Divide the two rates
The chain rule gives . The second factor is the rate of t compared with x, which is the reciprocal of the rate of x compared with t: . So .
Picture a small step in t. The point moves a small distance across, about times the step, and a small distance up, about times the step. The gradient is the distance up divided by the distance across, and the step cancels.
The rule needs . Where and is not 0, the point is moving straight up or down for that instant, and the tangent is vertical.
An example
Take and y = 2t. Differentiate each with respect to t: and . Divide: .
At t = 2 the point is (4, 4) and the gradient is . The tangent through (4, 4) with gradient is , which is .
Check it numerically. At t = 2.001 the point is (4.004001, 4.002). From t = 2 to t = 2.001 the point rises 0.002 and moves across 0.004001, and 0.002 ÷ 0.004001 = 0.499875, very nearly .
It agrees with the Cartesian equation too. Eliminating t gave , so the upper half is , whose gradient is at x = 4.
One x, two gradients
The curve passes x = 4 twice: at t = 2, the point (4, 4), and at t = −2, the point (4, −4). The rule gives a gradient for each value of t, so the two points get different gradients: at t = 2 and at t = −2. A single formula in x could not give both.
At t = 0, while . The rule has no value there, and the curve passes through the origin moving straight up: its tangent there is the y-axis.
The curve , y = 2t. The gold tangent at t = 2 has gradient and the plain one at t = −2 has gradient . At the origin, where t = 0, the curve stands upright along the y-axis.
Two more curves
Take and . Then and , so . At t = 2 the point is (4, 8) and the gradient is 3, so the tangent is y − 8 = 3(x − 4), which is y = 3x − 4. Check: at t = 2.001, . Eliminating t agrees as well: has gradient , which is 3 at x = 4.
Take the circle x = 3 cos t, y = 3 sin t. Then and , so . At the point is and the gradient is −1: the tangent runs at right angles to the radius, as a tangent to a circle must.
At t = 0 and at , sin t = 0, so while . The tangents at (3, 0) and (−3, 0) are vertical. At , cos t = 0, so and the tangent at (0, 3) is level.
The circle x = 3 cos t, y = 3 sin t. The gold tangent at has gradient −1. At t = 0, , and the tangent through (3, 0) is the vertical line x = 3.
The usual mistakes
Dividing the wrong way up. For x = 3t, , and , so . The upside-down is , the run for each unit of rise.
Multiplying the rates. 2t × 3 = 6t is not a gradient; the two time-derivatives divide, so that dt cancels.
Taking as the gradient. says how fast y changes with time. For , y = 4t at t = 1, , but the gradient is 4 ÷ 2 = 2.
A fairground carriage
In the application below, a carriage on a curved rail is at x = 3m and , both in meters, after m seconds. Its two rates give its speed by Pythagoras' theorem and the gradient of the rail by dividing one by the other.
Worked example: A Fairground Carriage on a Curved Rail: Its Speed, and the Steepness of the Rail Beneath It
Question A fairground carriage climbs a curved rail. After m seconds it is x = 3m meters out from the start and y = 2m2 meters above it, for 0 ≤ m ≤ 2. (a) Find the speed of the carriage after 1 second. (b) Find the gradient of the rail at that point, and check it against the equation of the rail itself.
1.Differentiate each coordinate with respect to the time: dxdm = 3 meters a second out from the start, and dydm = 4m meters a second upwards.
Each coordinate has its own rate: dxdm = 3 meters a second out and dydm = 4m meters a second up. 2.After 1 second the carriage is moving 3 meters a second out and 4 meters a second up, and it is at the point x = 3, y = 2.
After 1 second the carriage is at x = 3, y = 2, moving 3 m a second out and 4 m a second up. 3.(a) The two rates are at right angles, so Pythagoras' theorem gives the speed: √32 + 42 = √25 = 5 meters a second.
(a) The two rates are at right angles, so the speed is √32 + 42 = 5 meters a second. 4.For the steepness of the rail, divide the upward rate by the rate out from the start: dydx = dydm ÷ dxdm = 4m3.
For the steepness of the rail, divide one rate by the other: dydx = 4m3. 5.(b) After 1 second that gradient is 43, or about 1.333: the rail rises 4 m for every 3 m out.
(b) After 1 second the gradient is 43: the rail rises 4 m for every 3 m out. 6.Check it against the rail itself. From x = 3m comes m = x3, so y = 2 × x29 = 2x29 and dydx = 4x9. At x = 3 that is 129 = 43, the same gradient by a second route.
Checking against the rail itself, y = 2x29 gives dydx = 4x9, which is 129 = 43 at x = 3.
Answer: (a) 5 meters a second; (b) a gradient of 43, about 1.333
Common mistakes
- Adding the two rates for the speed: 3 + 4 = 7 meters a second. The rates are at right angles, so they combine by Pythagoras' theorem rather than by addition, and the speed is 5 meters a second.
- Differentiating y = 2m2 to get 4m and calling that the gradient of the rail. It is the rate the height grows with TIME; the gradient is the height gained for each meter traveled out, so dydm must be divided by dxdm.