When the Input Is Not Time
A derivative often describes how an output changes as time passes. But time is only one possible input. A temperature can depend on altitude, a cost can depend on the number of items produced, and the area of a circle can depend on its radius.
In each case, the derivative describes how the output changes as the input changes. Its units name both quantities: output units per input unit. Reading those units carefully helps you explain what the derivative means without accidentally treating every rate as a rate per second.
Identify the Input and Output
Suppose a function \(T(h)\) gives temperature \(T\) at altitude \(h\). The input is altitude, and the output is temperature. The derivative \(T'(h)\) therefore describes temperature change per unit of altitude, not temperature change per unit of time.
Similarly, if \(C(q)\) gives the cost of producing \(q\) items, then \(C'(q)\) describes cost change per item produced. The input is the production quantity \(q\); it is not elapsed time. When a production quantity is a whole-number count, calculus models typically treat it as a continuous variable to estimate how cost changes near a particular quantity.
The sign has the usual contextual meaning, but it must be described relative to the input. A positive derivative means the output is increasing as the input increases near the specified value. A negative derivative means the output is decreasing as the input increases. Neither sign, by itself, says that time is passing or that an output is changing over time.
Why the Units Are Output per Input
The derivative is defined through a limit of difference quotients. For a small nonzero input change \(h\), the quotient compares the output change to the input change. Each quotient has units of output divided by input; taking its limit does not reverse or change those units.
Here, \(f(a+h)-f(a)\) is an output change, while \(h\) is an input change. Thus the quotient has units \(U_y/U_x\). When the limit exists, it is the instantaneous rate at \(a\), with those same units.
The derivative also supports a local estimate. If the input changes by a small amount \(\Delta x\), then the corresponding output change is approximately \(f'(a)\Delta x\). This estimate follows from the derivative definition: for small nonzero \(\Delta x\), the quotient \(\frac{f(a+\Delta x)-f(a)}{\Delta x}\) is close to \(f'(a)\). Multiplying by \(\Delta x\) gives an output change close to \(f'(a)\Delta x\). The estimate need not be exact unless the function is linear over the interval.
Worked Example: Temperature Change per Unit of Altitude
Worked Example: Temperature Change per Unit of Altitude
Suppose a model for temperature near a location is \(T(h)=22-0.006h\), where \(h\) is altitude in meters and \(T(h)\) is temperature in degrees Celsius. Find and interpret \(T'(1500)\).
Differentiate the model:
Therefore, \(T'(1500)=-0.006\) degrees Celsius per meter. The negative sign means that, near an altitude of \(1500\) meters, temperature decreases as altitude increases. For each additional meter of altitude, the model's local temperature change is \(-0.006\) degrees Celsius.
For a small altitude increase of \(100\) meters, the estimated temperature change is:
Check against the model: \(T(1500)=22-0.006(1500)=22-9=13\) °C, and \(T(1600)=22-0.006(1600)=22-9.6=12.4\) °C. The actual change is \(12.4-13=-0.6\) °C, matching the estimate. It is exact here because the model is linear and its derivative is constant.
Worked Example: Cost Change per Item Produced
Worked Example: Cost Change per Item Produced
A cost model is \(C(q)=80+3q+0.01q^2\) dollars, where \(q\) is the number of items produced. Find and interpret \(C'(40)\).
Differentiate each term:
Substitute \(q=40\):
The units are dollars per item. Near a production level of \(40\) items, the cost is increasing at about \(3.80\) dollars for each additional item. This is an instantaneous rate in the model, not a claim that cost changes by exactly \(3.80\) dollars for every whole item.
Check the derivative using the difference quotient. First, \(C(40)=80+3(40)+0.01(40^2)=80+120+16=216\). For an input change \(h\),
Therefore, for \(h\ne 0\),
As \(h\) approaches \(0\), this quotient approaches \(3.8\), confirming \(C'(40)=3.8\). As a separate check on the local estimate, producing one more item changes the model cost by \(C(41)-C(40)\). Since \(C(41)=80+123+0.01(1681)=80+123+16.81=219.81\), the actual change is \(219.81-216=3.81\) dollars. This is close to, but not exactly, the derivative estimate of \(3.80\) dollars.
Worked Example: Area Change per Unit of Radius
Worked Example: Area Change per Unit of Radius
The area of a circle with radius \(r\) meters is \(A(r)=\pi r^2\) square meters. Find and interpret \(A'(3)\), then estimate the area change for a radius increase of \(0.1\) meter.
Differentiate the area function with respect to its input \(r\):
At \(r=3\) meters,
The derivative is approximately \(18.85\ \text{m}^2\) per meter of radius, to two decimal places. Near a radius of \(3\) meters, the area increases by about \(6\pi\) square meters for each additional meter of radius. Using \(\Delta r=0.1\) meter gives the estimate:
Check against the exact area change: \(A(3)=9\pi\ \text{m}^2\) and \(A(3.1)=\pi(3.1)^2=9.61\pi\ \text{m}^2\). Thus \(A(3.1)-A(3)=0.61\pi\ \text{m}^2\), which is close to the estimate \(0.6\pi\ \text{m}^2\). The estimate is not exact because the area function is curved, not linear.
Common Pitfall: Treating Every Rate as “Per Time”
Suppose you find a temperature derivative of \(-0.006\) degrees Celsius per meter of altitude. A wrong interpretation would be: “Temperature decreases by \(0.006\) degrees Celsius per second.” That changes the denominator from meters to seconds without any information supporting the change.
The correction is to identify the input in the model first. If the function is \(T(h)\), then its input is altitude \(h\), so \(T'(h)\) is measured in degrees Celsius per meter. A time-based rate would require temperature to be modeled as a function of time, such as \(T(t)\), with a derivative measured in degrees Celsius per unit of time. These are different derivatives and answer different questions.
A related units mistake is to reverse the ratio. If output is measured in dollars and input in items, the derivative is dollars per item, not items per dollar. The difference quotient makes the order explicit: output change divided by input change.
How to State a Contextual Derivative
A clear interpretation names the rate, the direction of change, and the input value. In practice, identify the variables and units before explaining the derivative.
For example, altitude is the input and temperature is the output.
Divide output units by input units, such as degrees Celsius per meter.
Explain whether the output is increasing or decreasing as the input increases near that value.
For a small input change, multiply the derivative by that change to estimate the output change.
Derivatives with non-time inputs appear whenever one quantity depends on another. They describe local sensitivity: how responsive the output is to a small change in the input. The input might be a length, a production quantity, or another measured quantity. In every case, the derivative's units and interpretation follow from the same rule: output change per input change.
Check Your Understanding
Use the input and output named in each question, and include units in your interpretations.
- A model \(S(d)\) gives salinity in grams per liter as a function of distance \(d\) in kilometers. What are the units of \(S'(d)\)?
- If \(S'(5)=-0.2\) grams per liter per kilometer, what does the negative sign mean near \(d=5\)?
- A revenue model \(R(n)\) depends on the number of tickets sold. What quantity does \(R'(n)\) describe, and what are its units if revenue is measured in dollars?
- A derivative with units of square meters per meter describes which output-per-input relationship? Explain why its denominator is not automatically seconds.
- If \(f'(a)=4\) output units per input unit and the input increases by \(0.2\) units, what is the estimated output change?
