From Geometric Regions to Integrals
A graph can enclose a region that is easy to measure even when an integral has not been evaluated by an antiderivative. If a function is linear on an interval, the region between its graph and the horizontal axis is triangular, trapezoidal, or a combination of the two. Its geometric area gives the integral, provided we account for whether the graph is above or below the axis.
There are two details to keep separate. Ordinary geometric area is never negative, while a definite integral records the graph’s position relative to the axis. Also, reversing the endpoints reverses the sign of the integral. These distinctions turn simple area calculations into reliable integral calculations.
Triangle and Trapezoid Formulas
Let \(a<b\), and write \(L=b-a\) for the interval’s positive width. A triangle with base \(L\) and perpendicular height \(H\) has area \(\frac{1}{2}LH\). A trapezoid with parallel sides of lengths \(H_1\) and \(H_2\), separated by distance \(L\), has area \(\frac{1}{2}L(H_1+H_2)\). For a graph, the endpoint heights are the absolute values of its ordinates when calculating ordinary area.
For a linear graph, it is useful to keep the endpoint ordinates signed. Let the graph have value \(u=f(a)\) at the left endpoint and \(v=f(b)\) at the right endpoint. Its signed integral is the interval width times the average of those endpoint values. This formula also works when one endpoint value is negative; in that case, the positive and negative regions cancel algebraically.
If both endpoint values are positive, this is the area of a trapezoid above the axis. If one endpoint is zero, it gives the area of a triangle. If both values are negative, the integral is the negative of the corresponding geometric area.
Why the Endpoint-Average Formula Works
Here is a derivation using equal-width Riemann sums. Suppose \(f\) is linear, with \(f(a)=u\) and slope \(m\), so that \(f(x)=u+m(x-a)\). Set \(L=b-a\), and divide \([a,b]\) into \(n\) equal subintervals of width \(\Delta x=L/n\). Using the right endpoint of each subinterval gives the Riemann sum
Separate the constant and slope contributions, and use \(\sum_{i=1}^{n} i=n(n+1)/2\):
As \(n\) tends to infinity, these Riemann sums converge to the integral. Since \(v=u+mL\), their limit can be written in terms of the endpoint values:
This proves the formula for a linear function. Geometrically, it is the signed trapezoid rule: average the signed endpoint heights, then multiply by the width. When the graph crosses the axis, the same formula gives the net area, because the part below the axis contributes negatively.
Worked Example: A Positive Triangular Region
Worked Example: A Positive Triangular Region
Suppose \(f\) is linear on \([0,4]\), with \(f(0)=0\) and \(f(4)=3\). Find \(\int_0^4 f(x)\,dx\).
The graph starts on the horizontal axis and ends at height \(3\). It stays above the axis, so the region is a triangle with base \(4-0=4\) and height \(3\). Its area is
The region is above the axis, so its signed integral is positive: \(\int_0^4 f(x)\,dx=6\).
Check: The endpoint-average formula gives \(\frac{4-0}{2}(0+3)=2\cdot3=6\), matching the triangular area.
Worked Example: A Trapezoid and Reversed Endpoints
Worked Example: A Trapezoid and Reversed Endpoints
Suppose \(g\) is linear on \([-1,3]\), with \(g(-1)=2\) and \(g(3)=6\). Find \(\int_{-1}^{3}g(x)\,dx\), and then find \(\int_3^{-1}g(x)\,dx\).
Both endpoint heights are positive. The interval width is \(3-(-1)=4\), and the trapezoid’s parallel sides have lengths \(2\) and \(6\). Therefore,
Reversing the endpoints changes the sign of a definite integral: \(\int_3^{-1}g(x)\,dx=-\int_{-1}^{3}g(x)\,dx\). Thus \(\int_3^{-1}g(x)\,dx=-16\).
Check: The endpoint-average formula for the forward interval gives \(\frac{4}{2}(2+6)=16\). Reversing the endpoints gives \(-16\), not \(16\), as required by the orientation rule.
When a Linear Graph Crosses the Axis
If the endpoint ordinates have opposite signs, the graph crosses the horizontal axis somewhere between them. The signed endpoint-average formula still gives the integral, but splitting at the crossing makes the cancellation visible. Each part is a triangle: count the part above the axis positively and the part below it negatively.
Worked Example: Two Triangles with Opposite Signs
Let \(h\) be linear on \([0,5]\), with \(h(0)=3\) and \(h(5)=-2\). Find \(\int_0^5 h(x)\,dx\).
The change in height is \(-2-3=-5\) over a width of \(5\), so the slope is \(-5/5=-1\). Thus \(h(x)=3-x\), and the graph crosses the axis where \(3-x=0\), namely at \(x=3\).
From \(0\) to \(3\), the graph forms a triangle above the axis, with base \(3\) and height \(3\). Its signed contribution is
From \(3\) to \(5\), the graph forms a triangle below the axis, with base \(5-3=2\) and height \(2\). Its signed contribution is \(-\frac{1}{2}\cdot2\cdot2=-2\). Adding the contributions gives
Check: The signed endpoint-average formula gives \(\frac{5-0}{2}(3+(-2))=\frac{5}{2}\cdot1=\frac{5}{2}\). This agrees with the difference between the two triangle areas.
Using Trapezoids to Bound an Unknown Integral
A function does not have to be linear for geometric areas to help. If an integrable function is bounded above and below by linear functions, the regions under those linear functions provide integral bounds. In precise terms, if \(g(x)\leq f(x)\leq h(x)\) for every \(x\in[a,b]\), then integral comparison gives \(\int_a^b g(x)\,dx\leq\int_a^b f(x)\,dx\leq\int_a^b h(x)\,dx\). The bounding integrals can be calculated from trapezoid areas when the bounds are linear and stay on one side of the axis.
Worked Example: Bounding an Integral Between Two Trapezoids
Suppose \(f\) is integrable on \([0,4]\) and satisfies \(1+\frac{x}{2}\leq f(x)\leq3+\frac{x}{2}\) throughout the interval. Find a lower and an upper bound for \(\int_0^4 f(x)\,dx\).
The lower linear bound has endpoint heights \(1+\frac{0}{2}=1\) and \(1+\frac{4}{2}=3\). Its trapezoid area is
The upper linear bound has endpoint heights \(3+\frac{0}{2}=3\) and \(3+\frac{4}{2}=5\). Its trapezoid area is
Both bounds are positive, and the assumed pointwise inequalities hold throughout the interval. Integral comparison therefore gives \(8\leq\int_0^4 f(x)\,dx\leq16\).
Check: The lower bound is the integral of a function everywhere no larger than \(f\), and the upper bound is the integral of a function everywhere no smaller than \(f\). The resulting order \(8\leq\int_0^4 f(x)\,dx\leq16\) is consistent with those pointwise bounds.
Common Pitfall: Treating Area as Always Positive
A common mistake is to report the ordinary area when the graph is below the horizontal axis, or to ignore the order of the endpoints. For instance, the trapezoid in the second example has geometric area \(16\). Writing \(\int_3^{-1}g(x)\,dx=16\) because “area is positive” is incorrect: the forward integral is \(16\), so reversing the interval gives
For a graph below the axis, the correction is different: keep the interval in its given direction, but assign the region a negative contribution. For a graph that crosses the axis, split at the crossing or use signed endpoint heights; do not add the two unsigned triangle areas unless the question asks for total geometric area.
A Reliable Geometric Procedure
For \(a<b\), use the positive width \(b-a\). If the integral’s endpoints are reversed, calculate in the forward direction and then change the sign.
For a linear graph, determine whether the region is a triangle, a trapezoid, or two regions separated by an axis crossing.
Use \(\frac{1}{2}(\text{base})(\text{height})\) for a triangle and \(\frac{1}{2}(\text{width})(\text{sum of parallel sides})\) for a trapezoid. Assign a positive sign above the axis and a negative sign below.
Check that the proposed lower and upper functions really enclose the function throughout the interval, then apply integral comparison to their signed integrals.
Why These Estimates Matter
Geometric estimates connect the shape of a graph with the value of accumulated change. A triangle or trapezoid can give an exact integral when the graph is linear, or a useful bound when a curve is enclosed by linear edges. In either case, the method is dependable only when signs and interval orientation are handled explicitly. The essential habit is to distinguish ordinary area from signed integral before doing the arithmetic.
Check Your Understanding
For each question, keep track of the endpoint order and the graph’s position relative to the horizontal axis.
- A linear function on \([1,5]\) has endpoint values \(0\) and \(4\). What is its integral on this interval, and what geometric region gives the result?
- A linear function on \([0,3]\) has endpoint values \(-2\) and \(-6\). What is its integral, and why is its sign negative?
- A linear function on \([0,4]\) has endpoint values \(2\) and \(-2\). Where does it cross the axis, and what is its signed integral?
- If \(\int_2^7 f(x)\,dx=11\), what is \(\int_7^2 f(x)\,dx\)?
- If linear functions \(g\) and \(h\) satisfy \(g(x)\leq f(x)\leq h(x)\) on an interval, how do their integrals bound the integral of \(f\)?
