Imagine hiking a trail with a slope meter strapped to your boot. While the reading is positive you are climbing; while it is negative you are descending; and at the exact top of a hill, or the bottom of a valley, the reading passes through zero, because for an instant the ground is level. Without ever seeing a map, the sequence of readings tells you the shape of the trail. The instantaneous rate of change is that slope meter for a function. This lesson connects the numbers you learned to estimate in the last topic with the features of a graph you already know how to describe — increasing and decreasing intervals, local maximums and minimums, the steepest part of a curve — and then looks at two families whose rates follow a striking pattern.
The sign of the rate and the direction of the graph
If the instantaneous rate of change is positive at every point of an interval, every tangent line there slopes upward, so the function is increasing on that interval. If the rate is negative throughout, the function is decreasing. At a local maximum the curve stops rising and starts falling, so the rate changes sign from positive to negative and passes through zero; the tangent there is horizontal. At a local minimum the rate changes from negative to positive. The sign CHANGE is what matters. For the rate is zero at , but it is positive on both sides, so the curve only flattens for an instant and keeps rising: a zero rate is a candidate for a turning point, not a guarantee.
| x | f(x) | Estimated rate (centred difference) | What the graph is doing |
|---|---|---|---|
| 0 | 1 | 9 | Rising steeply |
| 0.5 | 4.125 | 3.75 | Still rising, but less steeply |
| 1 | 5 | 0 | Level: rate changes + to −, so a local maximum |
| 2 | 3 | −3 | Falling as steeply as it ever falls between the turns |
| 3 | 1 | 0 | Level: rate changes − to +, so a local minimum |
| 4 | 5 | 9 | Rising again |
Where the rate is greatest
The rate of change is itself a quantity that changes. In the table, the rate drops from at , through at the maximum, down to at , then climbs back through to . So the curve is falling fastest at , halfway between its two turning points. That point, where the graph switches from bending downward to bending upward, is the point of inflection, and on many real curves it is the most interesting moment: the instant a disease outbreak is spreading fastest, the place a road is steepest, the time a car on a distance–time graph is moving fastest. To find it from data, estimate the rate at several points and look for where the rate itself is largest or smallest. With a formula, a table of centred differences on a fine grid, or graphing technology, does the same job.
Sinusoids: fastest at the axis
A Ferris wheel rider has height m. At the bottom and the top the rider is moving sideways, so the vertical rate of change is : these are the local minimum and maximum. The height changes fastest when the rider passes the level of the axle, moving straight up on one side and straight down on the other. Estimating the rate at min, a quarter turn after boarding, gives about m/min; at min it is about m/min. That pattern holds for every sinusoid: the rate is zero at maximums and minimums and largest in size where the curve crosses its axis. In fact, numerical estimates show that the greatest rate of is — the amplitude times . Here . The same reasoning tells you that tides rise fastest at mid-tide, and days lengthen fastest around the spring equinox.
Exponentials: rate proportional to value
A bacteria culture modelled by doubles every 3 hours. Estimate its instantaneous growth rate at several times and something remarkable appears: each time the population doubles, so does the rate. Dividing the rate by the population gives the same number, about per hour, at every time. That is the defining property of exponential change — the rate is proportional to the current value — and it is why exponential growth feels slow at first and then explosive. A population of 4000 is growing eight times as fast as a population of 500, simply because there is eight times as much of it to grow. Decay works the same way with a negative constant: a cooling drink or a decaying isotope changes fastest at the start and ever more slowly as it approaches its asymptote.
| t (h) | N(t) | Estimated rate (bacteria/h) | Rate ÷ N(t) |
|---|---|---|---|
| 0 | 500 | 115.5 | 0.231 |
| 3 | 1000 | 231.0 | 0.231 |
| 6 | 2000 | 462.1 | 0.231 |
| 9 | 4000 | 924.2 | 0.231 |